06-Aug-19 01:03 AM This channel is empty, so I'll just start off with a nice diagram of a transistor 06-Aug-19 01:03 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/mosfet2-F4847.png 06-Aug-19 01:03 AM #semiconductor now has semiconductor talk. 06-Aug-19 01:21 AM yay! 20-Aug-19 07:10 AM I ask experts : Does such micro-fabricated process seems doable at home with DIY style and refurbished equipment found on ebay? (goal = Micro-fabricated Ion trap). If the complete process is NOT doable at home , could you tell me, according to your experience, which steps is doable and which is not? source : https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5614346/#__sec2title 20-Aug-19 10:09 AM In theory, it's possible. You'd need to pretty much buy and refurbish an entire obsolete/scrap fab though. 14-Sep-19 09:28 PM @yann35 it would probably be less work/time and less money to just learn the processes needed (by the books) then go pay to use a university's equipment 14-Sep-19 09:28 PM Anyone in here know about Ruthenium? 14-Sep-19 11:14 PM Shiny, heavy, nonreactive, melts hot, and not used very much besides electroplating sometimes 15-Sep-19 12:48 PM It's all the new rage 15-Sep-19 12:48 PM Was at TechCon conference last week, 163 students, mostly PhD, presenting on their industry+some-govt jointly funded project progress 15-Sep-19 12:48 PM It was recommended to me for really low resistance conductors at atomic scale thicknesses 15-Sep-19 12:48 PM (Ruthenium) 15-Sep-19 12:48 PM Just found the precursor used in a thesis I read a few days ago https://www.strem.com/catalog/v/44-0056/59/ruthenium_949113-49-9 15-Sep-19 12:48 PM $370/gram 15-Sep-19 02:53 PM Yeah, it's proposed for new applications, but not widely used in industry 16-Sep-19 12:42 AM Seems like it's going to be gaining steam as a diffusion barrier 18-Sep-19 01:46 AM Found a local school that can deposit Ru with ALD and etch too... Guess I'll need to make a road trip down there soon to meet the prof who has the equipment 18-Sep-19 01:46 AM Might need to figure out how to build/fab a preamp too 18-Sep-19 01:46 AM @AdamMcCombs do you remember what the bandwidth of your STM system amplifier and ADC is/was? Do you think it had to do averaging of the signal for a given XY location, or was one measurement enough? 18-Sep-19 02:19 PM @nmz787 the probe would constantly scan and sample at some high rate. It would bin the horizontal scan into pixels, thus averaging the samples taken over that pixel. I don't remember the bandwidth exactly but it's not anything crazy so I didn't really look into it 18-Sep-19 02:19 PM The probe needed to be in constant motion instead of a move per pixel due to rigidity reasons 19-Sep-19 10:24 AM Ah, I was just trying to relate it to my application with the probe fixed in space across a nanopore, where the sample would be moving due to an electric field, but also where stopping it to get averaging probably requires some tricky magic (i e. Extra work on the design and fab) 19-Sep-19 04:22 PM Anyone know a good protocol for a safe/safer BOSCH DRIE process with normal silicon? 19-Sep-19 04:22 PM I.e no fluoro chemistry 20-Sep-19 05:34 AM how? thought you use SF6 as a process gas in DRIE of Si? 20-Sep-19 09:57 AM Well I'm sure there are alternative chemistries, that's all I'm hoping anyway 20-Sep-19 10:41 AM ah ok 20-Sep-19 10:41 AM Only thought that especially Bosch DRIE uses F 20-Sep-19 10:41 AM but I'm interested too 20-Sep-19 07:01 PM Ah maybe 20-Sep-19 07:01 PM I guess I shouldn't have said Bosch 20-Sep-19 07:01 PM I just meant DRIE 21-Sep-19 12:26 AM ah ok 23-Sep-19 10:59 AM AFAIK there is no way around F gas or acids for etching silicon EXCEPT KOH but that only etches silicon in specific directions based on crystal orientation and masking. 23-Sep-19 10:59 AM This is from my lacking independent research anyway. 23-Sep-19 10:59 AM You CAN try to sputter the silicon away but the problem is your selectivity is bad and you make a mess, you end up re-depositing silicon as-well as removing it so it just becomes useless. 23-Sep-19 10:59 AM Also I think it gets REALLY hot trying to do that. 23-Sep-19 05:27 PM Yeah sputtering is pretty obvious, but aspect ratios are limited to about that of a toilet-paper roll 26-Sep-19 02:52 AM https://giphy.com/gifs/bear-hello-waving-IThjAlJnD9WNO 26-Sep-19 03:03 AM hi 27-Sep-19 02:42 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/IMG_20190722_102530_005-DE3A1.jpg 27-Sep-19 02:42 AM So, my next plan for the e-beam drawing machine repurposing a vidicpn tube is making an adapter that has the X-Y plates and closes the glass tube with an O-rkng. 27-Sep-19 02:42 AM That way, when the filament burns out, I'll just change the whole tube. 27-Sep-19 02:44 AM https://www.ebay.de/itm/TE-Connectivity-AMP-Series-Hermetic-Seal/113256971519?epid=1768069664&hash=item1a5ea43cff:g:Z80AAOSwaf1boOpD 27-Sep-19 03:02 AM https://eu.mouser.com/ProductDetail/Johnson-Cinch-Connectivity-Solutions/142-1000-004?qs=sGAEpiMZZMuHTYi1bHPYEG2T8T2p4fct 27-Sep-19 03:02 AM this one @Nixie 27-Sep-19 03:02 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/IMG_-wf34ce-E07A3.jpg 27-Sep-19 03:02 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/IMG_-29f02c-28791.jpg 27-Sep-19 03:02 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/IMG_20190723_234025-12FCE.jpg 27-Sep-19 03:06 AM do I see some hexagons 27-Sep-19 03:10 AM Hexagons, where? 27-Sep-19 03:10 AM Ah, those PENTAGONS 27-Sep-19 03:10 AM Yep, thats for the future. 27-Sep-19 03:15 AM 3d printed ? i see glass leaking 27-Sep-19 03:21 AM Thats a model 27-Sep-19 03:21 AM Maybe you never heard of that purpose? 27-Sep-19 03:25 AM uhehueuhe 27-Sep-19 03:25 AM okay have a look in the seal i sent, will make your life easier 27-Sep-19 03:25 AM you can solder with tin 27-Sep-19 03:48 AM NAVY seals rushing in to evacuate the tube 27-Sep-19 09:16 AM @Nixie is that a vidicon pointed into a turbopump? 27-Sep-19 01:01 PM Yes, it is. 27-Sep-19 01:01 PM I want to turn it into an e-beam litography machine for making 30um device chips. 27-Sep-19 01:02 PM but first it's time to e-beam weld the tmp it would seem 27-Sep-19 01:04 PM It can't as the silicon holder will block the beam going directly into the blades. 27-Sep-19 01:04 PM That was discussed with adam long ago. 13-Oct-19 11:28 AM https://www.sciencedirect.com/science/article/pii/S0040609014002971 13-Oct-19 11:28 AM R-134a refrigerant 13-Oct-19 11:28 AM "Deposition of fluorocarbon film with 1,1,1,2-tetrafluoroethane pulsed plasma polymerization" 13-Oct-19 11:28 AM Now I wonder how 60Hz would do... 13-Oct-19 01:42 PM Oh, hmm, it says that DC and "low DC" work for some processses 13-Oct-19 09:16 PM I mean you can buy SF6 from certain suppliers. They don't seem expensive or hard to get. They have actual prices listed so I mean I would just get the gas you want vs trying to futz with a refrigerant if you have the money. Also then you have a more predictable output gas to deal with and know how to handle waste output. 13-Oct-19 10:30 PM SF6 is an etchant, that r134a is for a passivation step 13-Oct-19 10:52 PM ah 13-Oct-19 11:13 PM But yeah I was just thinking r134a is available down the street at the auto store, purity would be a concern though for sure 13-Oct-19 11:13 PM I don't see prices online for one of the common similar gasses, OCTAFLUOROCYCLOBUTANE 13-Oct-19 11:39 PM "The high-density plasma is generated by microwave surfatron or inductive radio frequency excitation. During each passivation cycle, a thin Teflon®-like film is deposited onto the walls of etched structures from C4F8-precursor species. Some scavenging of oxides from the silicon etch floor may also happen during or after the Teflon® deposition step. During the subsequent etching cycle, part of this film is removed from the coated sidewall by off-vertical ion impact and driven deeper into the trench (similar to erosion of photomasks and their redeposition to the sidewalls in some classical RIE etches). At the same time, the trench bottom is cleared of fluorocarbon polymer and etched by fluorine radicals released from SF6 in the plasma. Switching times between steps normally range from a few seconds up to one minute, depending on tolerated sidewall roughness. Since the impact of low energy ions is enough to remove the passivating polymer film, mask selectivity reaches very high values, for example, > 150:1 for photoresist and ≫300:1 for SiO2 hard masks. If harder passivating polymers were used, more aggressive ion impact would be required, and selectivity toward the masking material would be reduced as a consequence." 13-Oct-19 11:48 PM https://www.sciencedirect.com/topics/chemistry/octafluorocyclobutane 14-Oct-19 12:09 AM Has this been discussed? https://libresilicon.com/ 14-Oct-19 12:09 AM Because it looks interesting 14-Oct-19 12:53 AM I don't remember any in depth talk. 14-Oct-19 12:53 AM It is interesting, but somewhat limited if you only want a small part, as the comms to the outside must be serial because they don't have enough pads for everyone to connect to the outside. 14-Oct-19 01:10 AM That libresilicon has some good resources... I was not aware of it. 14-Oct-19 01:11 AM Their process description and the materials surrounding seem to be a neat read. 14-Oct-19 01:11 AM And it's not theoretical, its practical work. 14-Oct-19 01:11 AM Yea they even have multiple options for certain steps. 14-Oct-19 01:12 AM Yep 14-Oct-19 01:12 AM I will probably try and follow these. 14-Oct-19 01:12 AM The process design rules is especially nice 14-Oct-19 01:12 AM Even if you don't use their exact values they are nice rules to follow for a rough idea of boundries and such. 14-Oct-19 01:24 AM yep 14-Oct-19 01:24 AM And they have the math for how they ended up with the parameters they are using. 18-Oct-19 09:33 PM http://doeplasma.eecs.umich.edu/files/PSC_Godyak6.pdf 18-Oct-19 09:33 PM ICP vs CCP (inductively vs capacitively coupled) 18-Oct-19 09:33 PM Mentions lower frequencies in the kHz can actually be more efficient power wise, the only thing I'm noticing otherwise is the electron temperature increases with lower frequency.... But i don't know what electron temperature affects process wise. 18-Oct-19 09:33 PM Second to last page's title "Twelve reasons why the future belongs to the distributed ICPs with ferromagnetic core" 18-Oct-19 09:33 PM "Low driving frequency (0.4 - 2 MHz) reduces cost of rf equipment" 18-Oct-19 11:33 PM Interesting, I knew there were some complex designed for rf plasma gen, with magnetics involved... I am just planning on having a plate above my die/wafer and just yea... making plasma between the two. I don't need it to be fancy. This is already fancy compared to HF acid. 18-Oct-19 11:33 PM I am also too ignorant at the moment to discuss the complexities of this topic haha. 18-Oct-19 11:33 PM Much reading/learning to do :) 18-Oct-19 11:40 PM With anti-adhesive coating like C4F8 18-Oct-19 11:40 PM I'm looking to make the one nixie made, but also interested in removing/switching-off the diodes to get it to be an AC system. While at the same time thinking I might need to build a 4ft x 4ft plasma chamber for coating a similarly sized nanoimprint metal "shim" 20-Oct-19 06:02 AM you simply have to drive the tank, right? 20-Oct-19 09:17 AM @noon 20-Oct-19 09:17 AM @N00N tank? 21-Oct-19 02:07 AM tank as LC resonator 26-Oct-19 02:31 PM So, I might have found a small hope in the wirebonder experiment. All ultrasonic cutters sold by that description, are utterly expensive at 250+€ so an investment in that is a bit on the heavy side. However, I just found out that there are ultrasonic teeth de-scalers. Someone featured at hackaday turned one into a precision cutter, so I think I might as well try one of those, as the price is around 90€ from china. 26-Oct-19 02:31 PM https://www.ebay.com/itm/Dental-Ultraschall-zahnsteinentferner-Ultrasonic-Piezo-Scaler-Ultraschallgerat/274057253981?hash=item3fcf15d45d:g:kiMAAOSw2uxdZ508 26-Oct-19 02:31 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/HTB1RtlBKkvoK1RjSZPfq6xPKFXa0-342E3.png 26-Oct-19 02:31 PM That's what it seems to be inside. I might be able tomodify a tip to act as anvil. 26-Oct-19 03:06 PM So I think that's way overkill for wire/ball bonding... the ball bonding machine I took apart didn't even have an ultrasonic device in it! It was a simple coil drive device, I will get a photo for you in a bit. 26-Oct-19 03:36 PM there are non-ultrasonic bonders 26-Oct-19 03:36 PM that doesn't mean ultrasonic bonders don't need ultrasonic 26-Oct-19 03:50 PM Well sure ultrasonic bonders imply ultrasonic... I said wire/ball bonders not ultrasonic :) 26-Oct-19 04:30 PM you definitely need ultrasonic for DIY ball/thin stuff 26-Oct-19 04:30 PM because you need absolutely immaculate conditions for it to work otherwise 26-Oct-19 04:30 PM it can 26-Oct-19 04:30 PM (see: microwave-oriented wedge bonders for thick strip) 26-Oct-19 04:30 PM those dental scalers though, they look super interesting 26-Oct-19 04:30 PM very large, multilayer actuator and quite a nice horn already made 26-Oct-19 04:30 PM wonder if it autotunes as well 26-Oct-19 04:30 PM judging by the video in this article, it's quite violent 26-Oct-19 04:30 PM https://hackaday.com/2019/01/07/making-an-ultrasonic-cutter-for-post-processing-tiny-3d-prints/ 26-Oct-19 04:30 PM @Nixie thanks for letting me know these things exist 26-Oct-19 04:30 PM I'll try using it in my bonder project if I ever get around to it 27-Oct-19 12:43 AM I will buy one on December to try it out. 27-Oct-19 01:26 AM I have seen the power supplies for those on DiY units (for sale on aliexpress at around 120€) and they look fairly complex, unlike an ultrasonic cleaner board, so it might be monitoring the output freq and adjusting to the load. 27-Oct-19 04:27 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/HTB1cWYuKVXXXXbjXpXXq6xXFXXXr-41078.png 27-Oct-19 04:27 AM Uses 24Vdc at 1.5A 27-Oct-19 07:45 AM Oh...ffs...Why should I bother finishing my hot plate? 27-Oct-19 07:45 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/HTB1tUlRNmzqK1RjSZPxq6A4tVXaj-2438E.png 27-Oct-19 07:45 AM This just costs 40€ shipped. 27-Oct-19 07:45 AM 50/350ºC 27-Oct-19 08:16 AM No stirring at that price point? 27-Oct-19 08:56 AM No need for that, I just need to dry the resist on top of the silicon. 27-Oct-19 08:56 AM (that is actually to disssasemble phone screens) 27-Oct-19 09:17 AM guess its a good hotplate for that 27-Oct-19 12:50 PM I am not sure the surface material but I can always use an undoped wafer as top and put the other wafers on top of it. 27-Oct-19 12:50 PM Or use aluminium sheet 27-Oct-19 02:50 PM @Nixie i have played with one of these already, they are quite powerful, you would have to turn it down quite a lot 27-Oct-19 02:50 PM i didnt use the original power sppl but using a fgen, power amp and transformer i could drive it hard enough at 41kHz to get water to bubble 27-Oct-19 02:50 PM sadly i dnt have the scaper ends at hand since my dad uses different methods these days but when some urn up again ill try some more with it 27-Oct-19 02:50 PM maybe not for bondng bu im pretty sure they can do other interesting stuff too 27-Oct-19 02:54 PM Yay! will buy one as soon as possible. I'll be having vacations on December and would like to try it out. 27-Oct-19 02:57 PM they also have a hollow center for watercooling so i could imagine it would be possible to feed a wire through 27-Oct-19 02:57 PM not sure if the channel is straight all the way frm the grip though, its some time since i took it apart 27-Oct-19 02:57 PM and make sure to get tips for it, i could not identify the threads so far, no metric or imperial screw i own fit in there 27-Oct-19 03:18 PM also i have a question: 27-Oct-19 03:18 PM Does anyone know if/where you can buy empty TO-5 housings? 27-Oct-19 03:18 PM i see them from time to time as experimental carriers, but are they just NOS leftovers or are they still made? 27-Oct-19 03:22 PM @george https://www.alibaba.com/product-detail/Custom-semiconductor-packaging-for-TO-type_60482987738.html 27-Oct-19 03:22 PM disclaimer: not associated, never bought from them, first thing I found 27-Oct-19 03:53 PM No, the water channel does not run through the tips. Just on the base to get splashed out. My plan is to microdrill the end of one standard tip and try to use it as anvil, at least for tests. 27-Oct-19 03:53 PM How the hell am I going to pull that off, dunno. XD 27-Oct-19 04:08 PM Send it out and get it done with EDM...? 27-Oct-19 09:34 PM Yeah...but that would probably cost an arm, and a leg, and a few kidneys. Joking apart, I don't even know where could I send that. I have 0,1mm drill bits, but I doubt any of my equipment has small enough runout to use them (I use them to clean printer nozzles) 27-Oct-19 10:11 PM the very few times I have had EDM jobs priced out it was far more reasonable than i expected 27-Oct-19 10:11 PM just never actually needed to go that way in the end...thanks alum! 27-Oct-19 11:02 PM Hmm there's an EDM shop about 5 minutes drive away, I was just wondering if they could make an aperture for my SEM since it is relatively large (6mm OD, 1mm ID) 27-Oct-19 11:02 PM I've toured through once, seemed like nice people 27-Oct-19 11:20 PM can't hurt to ask 27-Oct-19 11:20 PM you'd need sinker EDM for this 27-Oct-19 11:20 PM and electropolishing maybe 28-Oct-19 02:49 AM joup, just ask them. I once asked for having a custom punsh and die cut from 20mm material and they wanted 50€ for it 28-Oct-19 02:49 AM never had anything drilled that fine but i doubt it will be much more expensive 28-Oct-19 02:22 PM I toyed with the idea of making a sinker edm to do that. Got my hands on a roller bearing for minimum play, and would try to use the actual gold wire itself as drill. That could be more of a problem than a solution, though. 28-Oct-19 02:22 PM (i mean one of those roller bearings wich has each roller 90° apart. 28-Oct-19 02:25 PM Applied Science(Ben Krasnow) showed a guide on building a sinker edm system 28-Oct-19 02:26 PM Yeah, but he never goes deep enough to make it easily reproducible. There are guides outthere on how to build the ower supply etc, anyways. 28-Oct-19 02:29 PM Ussually he doesn't in his videos but I think he has writeups on his website that go much more indepth, usually the video is just a quick showing the thing. 28-Oct-19 02:29 PM most people don't realize he has a website 28-Oct-19 02:36 PM Yeah, but if I recall correctly he bought the power supply for the edm from ebay, so no scratchbuild, and I can't get my hands easily on that kind of equipment. Will try to look if there is any edm factory in my area. 20-Nov-19 03:18 AM Btw, @GigaSquirrel convinced me to make a patreon for the semicobductor thing. Not for anyone to pitch in, but rather as the new source of info on the status of my research. Hackaday.io is just messy and little helpful, has file limit size, etc. I might move the whole thing to github or similar. https://www.patreon.com/user?u=14363159 20-Nov-19 04:26 AM sure, just blame me :P 20-Nov-19 07:13 AM Of course. 20-Nov-19 11:43 AM Wait so you're using patreon as a place just to upload stuff? 20-Nov-19 11:43 AM That is... Odd to me 20-Nov-19 11:42 PM No, no. I mean, if I get help from Patreon that's good, but I didn't post it here to get more backers. What I meant is that I'll be moving the files from Hackaday.io to somewhere else more convenient, as people has started to find difficult to find any particular info on the two huge PDF's I have right now. 21-Nov-19 02:15 AM Ooo gotcha 21-Nov-19 03:46 AM I have you backed Nixie xD 21-Nov-19 04:27 AM Thanks! 21-Nov-19 04:27 AM I should have sent you a thankyou message already through patreon too. 03-Dec-19 07:35 PM Curious to know if anyone here has any experience with affordable maskless lithography. Are there any non-DIY machines that have less than 5 or 6 figure prices? Several years ago I built one using a blue ray laser head and servo-control optical stage. I could make 8um lines with it but it was painfully slow. 03-Dec-19 07:38 PM Sam Zeloof made one using a projector in reverse through a microscope 03-Dec-19 07:41 PM Yeah I know about that one. 03-Dec-19 08:37 PM I hacked a projector myself but never hooked it up to my microscope 03-Dec-19 08:42 PM Yea just need a UV led suitable in the projector... 03-Dec-19 08:42 PM I am curious about e-beam litho since I have a second SEM that I could dedicate to it 03-Dec-19 08:42 PM but I also have a projector I got specifically to modify 03-Dec-19 08:42 PM problem with projector is dealing with inconsistencies in the lighting source as Sam dealt with... 03-Dec-19 08:42 PM problem with ebeam is spot size control and flood filling larger areas of stuff you want to mask off 03-Dec-19 08:42 PM positive photoresist would probably be best for e-beam stuff 03-Dec-19 08:45 PM And having a SEM to modify in the first place 03-Dec-19 08:45 PM right... I mean SEMs are cheap secondhand the problem is getting them and maintaining them mostly 03-Dec-19 08:45 PM don't really have to do much to modify them... adding x/y inputs on an SEM is generally pretty trivial 03-Dec-19 08:45 PM spot size control/beam blanking is another issue I suppose 03-Dec-19 08:46 PM That's what I meant 03-Dec-19 08:46 PM but nothing that makes it unusable as an SEM after figuring out how to add them 03-Dec-19 09:06 PM I guess I've technically done direct write lithography with a little $100 laser etcher, was playing with circuit board etching though and assumed the linewidths would be terribly wide 03-Dec-19 09:06 PM Was planning to upgrade it with a bluray laser sled, potentially achieving submicron lines 03-Dec-19 09:06 PM But linewidth has a lot more than just spot size 03-Dec-19 09:06 PM Stuff like diffusion of reactants in the resist layer, diffusion of the radiation, etc 03-Dec-19 09:06 PM I'd guess inconsistency in the light field could just mean you need to expose for longer 03-Dec-19 09:58 PM Yea I know Sam was playing with blueray lasers at somepoint 03-Dec-19 10:52 PM If only we could like tag @szeloof 03-Dec-19 10:59 PM IMO the ideal solution is a combination UV coherent and ultra-short pulse (femtosecond) NIR coherent source with a DMD backed by an optically addressed LCD. The UV source gives you your broad patterning and the NIR source gives super-fine detail beyond the standard resin limitation due to two-photon adsorption being so non-linear; there's also a method of using a second linear pulse to deactivate the reaction happening around your first pulse. 03-Dec-19 10:59 PM could do a brief pre-patterning of the optically addressed LCD via the DMD at super high resolution and get full control over amplitude, polarization and phase at sub-pixel scale 03-Dec-19 10:59 PM with the DMD alone you can do standard projection, binary masking of amplitude and super-pixel/group discretized control of phase, with the binary masking you can present fresnel zone plates that dynamically split the beam into several spots and control them individually over a small FOV 03-Dec-19 10:59 PM then with the optically addressed (no electrode patterning, just an additional 2 NIR photoelectric layers each sandwiching the LC) LCD at a refresh rate of, say, 50ms, you can load it up with DMD patterned pulses via the NIR laser and prior to the 50ms refresh trigger send a UV pulse 03-Dec-19 10:59 PM demonstration of optical addressing of LC btw, as it's pretty unheard of outside photonics (you wouldn't keep/add the polarizing films for darkening upon LC polarization) https://www.youtube.com/watch?v=R9UnGUCfJDw 04-Dec-19 03:47 AM The thing is, how small do you want the features to be? because a SEM by itself is going to be capable of detail that normal wet chemical etching is not going to be capable of doing (below 3um, textbooks say), so you will need other expensive equipment to keep working on that scale. Of course, having a SEM rocks, not arguing that. 04-Dec-19 03:47 AM I'm trying to transform a vidicon tube into an electron gun etcher to do 10/20um features. 04-Dec-19 08:43 AM imho dlp is probably best for diy people, whole system can be made <$500 and exposure is fast bc Hg lamp inside projector can give 100 04-Dec-19 08:43 AM 100+ mJ/cm^2 dose at the substrate 04-Dec-19 08:43 AM of course it has issues but the resolution is very good and comparable to anything a diy person would likely make via ebeam without real ebeam software for proximity correction etc etc and better resists than PMMA 04-Dec-19 08:43 AM * should be mW not mJ 04-Dec-19 08:43 AM some more dlp modification pics here that aren't on my website, but its mostly uncommented and undocumented sorry i will get on that later https://www.flickr.com/photos/zeloof/ 04-Dec-19 09:41 AM Broad spectrum lamp is going to have more aberration in the final image than a monochrome source though 04-Dec-19 09:41 AM SEM/FIB have relatively low bandwidth and the field of view is so tiny you have a lot more emphasis on stitching and working around that being difficult without an OMG-awesome stage 04-Dec-19 09:41 AM I also have been thinking along the lines of @Metanoic regarding a multiple source/scale system 04-Dec-19 10:04 AM All you have to do is find a free femtosecond laser in a dumpster somewhere 04-Dec-19 10:40 AM Thanks for the photos Sam! Don't worry, most of us here are used to partial documentation and or photos, we are crafty :) Also we generally just share stuff here complete or not for fun, also its great to get feedback on the direction we are headed because some people here, much smarter than myself can tell you where your about to walk off a cliff with the direction your going hehe... 04-Dec-19 10:40 AM Also @nmz787 I am thinking with my DLP I will remove the LCD stack, and replace the lamp with a high power UV LED source. I thought Sam did the same with his but I guess he may have only removed the LCD stack and just used the stock bulb? 04-Dec-19 10:45 AM Oh, hmm, I've not really considered using LCD too much 04-Dec-19 10:45 AM Only DLP really 04-Dec-19 10:45 AM My projector didnt have an LCD stack as far as I remember, it had a colorwheel though 04-Dec-19 10:56 AM oh color wheel then 04-Dec-19 10:56 AM I didn't know if they were using an LCD for the color selection or color wheel 04-Dec-19 10:56 AM I assumed LCD but yea just remove whatever adds the color is my point 04-Dec-19 10:56 AM I have a DLP projector 06-Dec-19 05:53 AM For the lulz: 06-Dec-19 05:53 AM https://www.youtube.com/watch?v=6LfFjmWCWPw&feature=youtu.be&t=107 06-Dec-19 10:00 AM Looks pretty simple. 06-Dec-19 10:34 AM XDDD 12-Dec-19 06:20 AM Anyone has had luck finding aluminium wirebonding wire? 12-Dec-19 11:43 AM Yes, I've purchased some old bonding wire from ebay. It rarely appears there. I tried to do wire bonding using medical scaler too. I currently have ~2-3% success rate. Power is too high, or on low power it drops from resonance. But I was doing it with pedal. It looks like arduino with controller pulse length is absolutely required. Custom driver might be required (PWM->amplifier->head), so that power is not that depend on mechanical resonance. 12-Dec-19 11:43 AM I still see multiple lots on ebay, around 30$: https://www.ebay.com/sch/i.html?_from=R40&_trksid=m570.l1313&_nkw=wire+bonding+aluminum&_sacat=12576&LH_TitleDesc=0&_osacat=12576&_odkw=wire+bonding 12-Dec-19 12:01 PM WOOOOOOOOO 12-Dec-19 12:01 PM Mine! 12-Dec-19 12:01 PM (if my american contact can get it for me, I mean) 12-Dec-19 12:02 PM Gold wire is pretty much a must for anything DIY afaik 12-Dec-19 12:02 PM Aluminium oxidizes a lot 12-Dec-19 12:04 PM What preheat temp did you use in the substrate? 12-Dec-19 12:05 PM I don't have any experience with bonding personally 12-Dec-19 12:05 PM Sorry, I was asking to zeptobars. 12-Dec-19 12:05 PM I'd say 60-90C is good 12-Dec-19 12:06 PM If you can do gold, you can do aluminium, the ultrasonic action ckeans both surfaces. 12-Dec-19 12:06 PM Nowhere near as good 12-Dec-19 12:06 PM I have heard of 150° substrate temps. 12-Dec-19 12:06 PM The price difference isn't that big either 12-Dec-19 12:06 PM You can buy a 300m spool of gold wire from china for ~$150-250 12-Dec-19 12:06 PM 300m will last all of us a lifetime, collectively 12-Dec-19 12:11 PM @Nixie I was trying without preheat, and it worked terrible ) Should probably try gold... I assume purple death is less of a concern for amateurs 12-Dec-19 12:11 PM *purple plague 12-Dec-19 12:13 PM Oxidation? 12-Dec-19 12:13 PM The aluminium spool just costed me 20$ 12-Dec-19 12:13 PM Good enough for tests. 12-Dec-19 12:13 PM No need to spend 200$ 12-Dec-19 12:13 PM (wich I would spend in a metallographic microscope, if I had them) 12-Dec-19 12:22 PM I've been told by John McMaster(@digshadow, but he hasn't joined on the right account yet), who succeeded at proper DIY wire bonding, that random wire off eBay, especially not gold can make bad results 12-Dec-19 12:22 PM I've also been told that gold wire(!!!) expires once it's unsealed. 12-Dec-19 12:22 PM So just set your expectations with that in mind 12-Dec-19 12:22 PM You might be doing nothing wrong and the wire could be failing you 12-Dec-19 12:35 PM Absolutely 12-Dec-19 12:35 PM But what I need is seeing if the wire deforms properly, if the anvil shape can work. 12-Dec-19 12:35 PM Better try with a cheap source. 12-Dec-19 12:37 PM Just buy a few capillaries off eBay 12-Dec-19 12:37 PM They're often less than $10 each 13-Dec-19 12:09 AM For direct-write optical lithography look at the systems from Intelligent Micro Patterning for inspiration. Mercury lamp, optics bench, optics, filter wheel, 1024x768 DMD, microscope objective, lots of ThorLabs pieces holding everything together. Plausible to reproduce them 13-Dec-19 12:20 AM btw, just broke the bank and bought a metallographic microscope for 140€ (100x to 600x ) 13-Dec-19 12:20 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/s-l400-98949.png 13-Dec-19 12:20 AM (the previous cheapest I had found was 215€: 13-Dec-19 12:20 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/ELppqegWwAAQW8K-C7C7D.png 15-Dec-19 02:54 AM I would like an inspection microscope that does like 5x up to 20x or 50x, with plenty of working distance between the stage and the objective, reflected light only, a flat stage ideally with no hole in the stage. 15-Dec-19 03:10 AM Shame mantis scopes are so damn expensive. 15-Dec-19 11:11 AM Anyone ever done RuO2 thin films? 15-Dec-19 11:11 AM Idk whether to start with Ru thermal dep and then anneal in presence of oxygen 15-Dec-19 11:11 AM Or do DC sputter of RuO2 targets off the bat 15-Dec-19 11:39 AM I started looking at Ru metal depisition a few months back 15-Dec-19 11:39 AM What do RuO2 films do? 15-Dec-19 11:40 AM From what I heard about Ru metal, it's got very good crystallinity/conformity... So you have very good conductance and low noise even with very very small/thin features 15-Dec-19 11:40 AM I feel like the oxides might be used to bond various other metal layers effectively 15-Dec-19 11:40 AM Using less thick layers, thus saving money and process steps for things like interconnects 15-Dec-19 11:40 AM But I am just going off of two days of walking around a conference poster session 15-Dec-19 12:04 PM Oh " For example, RuO2 is a metal while rutile TiO2 is an insulator, yet both have nearly identical structures. " 15-Dec-19 12:04 PM "Ruthenium has a high melting temperature of 2334 °C, is resistant to chemical attack over a wide range of temperatures, and is used industrially as a catalyst and also as a hardener in metal alloys. 2" 15-Dec-19 12:04 PM "Ruthenium has approximately five times the resistivity of silver at room temperature and is therefore a very good conductor of electric current and heat." 15-Dec-19 12:04 PM "In general, transition metal oxides have rather narrow electronic bands on the order of 1 to 2 eV" 15-Dec-19 12:04 PM "In its single crystal state, RuO2 has a higher electrical conductivity than about one-third of the pure metals in the periodic table (see Figure 1.1), although polycrystalline specimens generally show a lower conductivity" 15-Dec-19 12:04 PM "In terms of technology, ruthenium dioxide finds uses as diffusion barriers, electrodes, thick-film resistors, thin-film resistors and catalysts. An extensive summary of applications of RuO2, emphasizing catalysis but including more exotic uses such as extreme ultraviolet lithography, is given in a paper by Assmann et al" 15-Dec-19 12:04 PM From https://www.google.com/url?sa=t&source=web&rct=j&url=https://digitalcommons.library.umaine.edu/cgi/viewcontent.cgi%3Farticle%3D1258%26context%3Detd&ved=2ahUKEwjvqpznuLjmAhX0KDQIHXd2BIcQFjAAegQIAhAB&usg=AOvVaw2dq6tUj7hRGPskJ5PHQGpL 15-Dec-19 12:47 PM yeah we’re gonna be using it as an electrode for photoechem stuff 15-Dec-19 01:23 PM I was thinking to use it for nanothin electrode for electro chemical sensing 15-Dec-19 01:23 PM Liquid phase STM, basically 15-Dec-19 01:36 PM Ah ecstm is awesome lol 15-Dec-19 01:36 PM Gonna be most likely sending some samples to a group nearby to do that, test some ligand interaction stuff lol. But on topic, if anyone has experience depositing ruo2 or ru (to turn into ruo2) let me know. They sell sputter targets so I think it should be possible using just DC sputtering (especially since the resistivity is so low) 15-Dec-19 11:46 PM What toxicity does Ru have? 15-Dec-19 11:47 PM VERY carcinogenic in pretty much every form. 15-Dec-19 11:47 PM Testing is somewhat limited due to rarity, but those results sure haven't been friendly. 15-Dec-19 11:50 PM XD, not gonna try sputtering that, then, I'll keep myself into aluminium/copper/titanium 15-Dec-19 11:53 PM RuO4 is particularly nasty and it seems to be a boneseeker. Pretty common in the rare earths, but not great if it has a lingering toxicity. 15-Dec-19 11:55 PM Maybe reactive sputter with oxygen and a Ru target? Or start with a RuO2 target and sputter off that with extra oxygen. 15-Dec-19 11:55 PM Do you have access to sputter, or electron evap, or thermal evap? I think evap is a bit poor for oxide materials, I tried ITO e-beam evap once and it was poor, sputter was way better. 15-Dec-19 11:55 PM E-beam does have an advantage in that you can just get a bottle of stuff and put it in a crucible, doesn’t have to be in a special form like a sputter puck 15-Dec-19 11:55 PM I tried making up a Cr etchant with acetic acid... I’m really impressed. It’s safer, no perchloric acid, no nitric acid, works just as fast on chromium, really selective with no Ag or Cu attack which you can get with nitric acid 15-Dec-19 11:55 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/image0-311CB.jpg 15-Dec-19 11:55 PM 50 ml deionised H2O, 15 g cerium ammonium nitrate, 3.5 ml glacial acetic acid, stir to dissolve the CAN and top up to 100ml total volume with H2O. 15-Dec-19 11:55 PM Cheaper than perchloric acid too 16-Dec-19 12:13 AM from the thesis on Ru I posted earlier, it seemed like it oxidizes pretty easily, so a serially pulsed sputter chamber seems like it wouldn't be too hard to rig up 16-Dec-19 12:13 AM i.e. sputter some then stop, purge with oxygen plasma or something, repeat 16-Dec-19 12:13 AM I remember it mentioning carbon monoxide, but who wants to deal with that, lol 16-Dec-19 05:46 AM Got thermal evap and a sputter coater available 16-Dec-19 05:46 AM same facility manager in charge of both he may kill me if I go using a different target on the sputter lol 16-Dec-19 05:46 AM just because no one has ever changed it from the Au/Pd target and I’m always trying to do stuff like this lol 16-Dec-19 06:21 AM He will probably kill you, yes 16-Dec-19 10:02 AM dang, looks like all forms are pretty nasty like @funranium said. more I look into it less I feel comfortable using it in shared facilities. Was trying to avoid electrodeposition but think that’s the route I’m gonna have to take 16-Dec-19 10:37 AM [safety man salute] 16-Dec-19 11:21 AM @mike crb so electroplating? That'd be of Ru, not the oxide? 16-Dec-19 11:32 AM yeah I’m gonna electrodeposit RuO2 though 16-Dec-19 11:32 AM there’s been some work on thin film electrodep of it for super-capacitor stuff 16-Dec-19 02:08 PM @funranium You wouldn't happen to know how safe random Ruthenium plated anodes are? Because I was randomly given one years a go and it has been knocking about. 16-Dec-19 02:27 PM Should be pretty safe to handle 16-Dec-19 02:27 PM I'd use gloves 16-Dec-19 03:53 PM I was literally handed it at a hamfest years ago 16-Dec-19 03:53 PM never used gloves with it either :x 16-Dec-19 03:53 PM "Heya I recall talking with you about electrochemistry few years back and I can't even give away these ruthenium electrodes!" "You should be able to make those perchlorates with these!" 16-Dec-19 03:53 PM And thus I own one that I have never used for anything. 16-Dec-19 07:01 PM Is there any titanium thin film etchant that is free of HF ? 16-Dec-19 11:53 PM Sorry @rfs, don't have a strong feel on that. I'm a health physicist not and industrial hygienist, but I do know some stuff. I can hit up coworkers to check the NIOSH guide tomorrow. 16-Dec-19 11:56 PM Ok thanks 17-Dec-19 12:13 AM Health physicist sounds like an attempt by a gym teacher trying desperately to appeal to the nerds who fail gym class because they'd rather read than play basketball 17-Dec-19 09:30 AM The term for the profession was created specifically to obscure what we do during the Manhattan Project, which is radiation safety. The rest of the world just calls it that, but the US and Japan retain the old name. 17-Dec-19 09:30 AM @rfs I have my coworker on the case. She groaned when I asked if ruthenium has oxide shedding behavior in storage like beryllium does. My cred for bringing weird things to her is maintained. 17-Dec-19 09:40 AM XD 17-Dec-19 01:23 PM @funranium Heh. Hopefully nothing too bad. 17-Dec-19 01:41 PM It more a matter of “You, Phil. Only you come to me with this crap.” 17-Dec-19 03:42 PM ... beryllium sheds oxides in storage? 17-Dec-19 03:42 PM good to know 17-Dec-19 03:51 PM Yep 17-Dec-19 03:51 PM So one should store it next to those cadmium plated connectors and other cadmium plated military surplus 17-Dec-19 03:52 PM finely pulverized for maximally space efficient storage (/s) 17-Dec-19 06:18 PM shakes angry fist 17-Dec-19 08:58 PM Apparently, ruthenium things are a hard question. IH friend is now treating this as A Quest. I expect a whitepaper out of this. 17-Dec-19 10:49 PM yay quests :3 18-Dec-19 01:31 PM Hah 22-Dec-19 06:36 AM Btw, Just gonna start making the circuit for the electrostatic e-beam thingy 22-Dec-19 06:36 AM I'll lift off the relevant parts of this design: 22-Dec-19 06:36 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/circuit7ss-D2D89.png 22-Dec-19 06:36 AM Mainly the XY controls, and the general idea on how to drive the tube. 22-Dec-19 07:02 AM I am not sure how to do the blanking at 15Kv if I ever achieve that kind of power. ^^U 22-Dec-19 07:23 AM @Nixie I had an impression that if you do blanking by lateral shift (i.e. one more electrostatic displacement) so that e-beam miss the aperture - you can get away with very low voltages, like 200-500. 22-Dec-19 07:23 AM @niklas This is the unit that I was considering to buy for work, it does not look like it had 50kV feedthroughs) : https://deben.co.uk/wp-content/uploads/2011/08/DEB3416_DatasheetA4_PCD-Port-Beam-Blanking_v2-4.pdf 22-Dec-19 07:23 AM "400V maximum deflection voltage" 22-Dec-19 07:43 AM I...don't have an aperture to speak of...I kinda was going to use the vidicon tube as is 22-Dec-19 01:11 PM I wish it was easier to just modulate the power of say the wehnelt or the emitter itself 22-Dec-19 01:11 PM I don't like the idea that the beam is still dragging across the substrate in the process of blanking 22-Dec-19 01:11 PM My perspective is for lithography though, not just imagining 22-Dec-19 01:13 PM Use an electromagnetic beam blanker instead to defocus the beam? 22-Dec-19 01:13 PM You can have a beam blanker that sprays the beam into the walls 22-Dec-19 01:13 PM So it acts closer to fading away than dragging 22-Dec-19 01:16 PM @Nixie I found this paper a few months ago https://link.springer.com/article/10.3938/jkps.63.1287 22-Dec-19 01:16 PM Electrostatic 22-Dec-19 01:16 PM Expected resolution 10nm 22-Dec-19 01:17 PM Oh, is that the paper about the student-built Korean DIY SEM? 22-Dec-19 01:17 PM https://sci-hub.se/10.3938/jkps.63.1287 22-Dec-19 01:18 PM Says korean affiliated 22-Dec-19 01:18 PM Well, doesn't everyone know to prepend articles with sci-hub?? 22-Dec-19 01:18 PM just in case 22-Dec-19 01:18 PM It's a very short paper 22-Dec-19 01:22 PM Just found this one on electromagnetic system design 22-Dec-19 01:23 PM Pretty sure that's the same thing 22-Dec-19 01:23 PM Sorry, this one http://www.j-mst.org/On_line/admin/files/11-12483_1734-1746_.pdf 22-Dec-19 01:24 PM This one is a lot more detailed 22-Dec-19 01:24 PM I wonder what the advantages/disadvantages are for DIYing an electromagnetic system 22-Dec-19 01:24 PM I suppose you don't have to put as much hardware into vacuum? 22-Dec-19 01:24 PM Sucks that there exist no student/hobby licenses for any sort of EM/ES/FEA software 22-Dec-19 01:24 PM You can obtain slightly older versions of COMSOL through various means 22-Dec-19 01:24 PM But that's about it 22-Dec-19 01:32 PM COMSOL is pretty good for electron optics 22-Dec-19 01:32 PM It is! Just unbelievably expensive if you want to actually buy it 22-Dec-19 02:04 PM Ohhh, I should take a look at it. 22-Dec-19 02:08 PM I did use it at school to simulate some MEMS resonators, but yeah that's some weird educational license likely. 22-Dec-19 02:08 PM And COMSOL is a Finnish thing so that might have also mattered. 22-Dec-19 02:08 PM Well, was. 22-Dec-19 02:09 PM Sent them an inquiry to see if they can make a hobbyist licensing exception 22-Dec-19 02:09 PM I expect to hear back from them never. 22-Dec-19 02:09 PM lol 22-Dec-19 02:10 PM But yeah, COMSOL is the best simulation system out there 22-Dec-19 02:10 PM Oh neat, they updated the site again https://www.opensourceimaging.org/project/open-ems-a-free-and-open-electromagnetic-field-solver/ 22-Dec-19 02:10 PM the only one with such a huge multidiscipline mix as far as I'm aware 22-Dec-19 02:11 PM http://openems.de/start/index.php 22-Dec-19 02:11 PM looks like that only solves the fields 22-Dec-19 02:11 PM not particularly useful for more "empirical" DIY setups 22-Dec-19 02:11 PM you know, pew-pewing a virtual beam down a virtual SEM column design 22-Dec-19 02:14 PM I'd like to mess around with the ANSYS Workbench 22-Dec-19 02:14 PM They've got E&M as well, not suited for electron/ion optics though I think 22-Dec-19 02:16 PM I like how broken their site is 22-Dec-19 02:16 PM https://i.spirit.re/bwGAU.webm 22-Dec-19 02:16 PM Oh! They offer a full FSAE student license! 22-Dec-19 02:18 PM Yeah looks like they offer a bunch of their stuff for free 22-Dec-19 02:19 PM Neat, would be nice if COMSOL did it so we don't need to either go through the school to get a limited license or fall back to Solidworks 22-Dec-19 02:20 PM Solidworks offers no student licensing at all 22-Dec-19 02:20 PM Let alone their sim software 22-Dec-19 02:20 PM Yes they do 22-Dec-19 02:20 PM *if your school pays for it 22-Dec-19 02:20 PM *and you pay for it 22-Dec-19 02:20 PM *if they like you 22-Dec-19 02:20 PM To FSAE teams they sponsor the full simulation and CAM package 22-Dec-19 02:20 PM But yes, I will lose it after graduation 22-Dec-19 02:20 PM Autodesk's student licensing is the best in my opinion 22-Dec-19 02:20 PM Yeah for sure 22-Dec-19 02:21 PM They give away all their software under "student", but I have Inventor Pro and HSM without actually being in an institution 22-Dec-19 02:21 PM And I got into the beta program(reviewed by a human) after replying "I'm not actually a student, I use it at home" when asked something along the lines of "why are you using the student license" 22-Dec-19 02:21 PM If only they didn't switch to the subscription model :< 22-Dec-19 02:21 PM Or at least had the decency to go for the perpetual-subscription model, where when you stop paying you keep whatever you paid for 22-Dec-19 02:23 PM I've got to try out HSM/Inventor CAM 22-Dec-19 02:23 PM It's super nice 22-Dec-19 02:23 PM Like Fusion CAM but better and no training wheels 22-Dec-19 02:24 PM I use Fusion CAM now but have not yet done any serious work with it 22-Dec-19 02:25 PM I recommend switching to Inventor if you're just doing hobby work 22-Dec-19 02:25 PM Everyone I've recommended it to likes it so far 22-Dec-19 02:25 PM I've got the 5 year student license for Fusion 360 22-Dec-19 02:25 PM Sign up for a student license, no edu email is required 22-Dec-19 02:25 PM Any info entered just gets accepted 22-Dec-19 02:25 PM I have no problem with Fusion's cloud and like their integration 22-Dec-19 02:26 PM I can't accept the cloud, if I don't own my files then I don't use the tool 22-Dec-19 02:26 PM I'll check out Inventor along with HSM 22-Dec-19 02:26 PM Ah ha! I just got a license for EGN2 through my school 22-Dec-19 02:26 PM Will check it out over the break 22-Dec-19 02:26 PM Design a beam blanker for the SEM 22-Dec-19 02:29 PM EGN2? 22-Dec-19 02:29 PM Egun-igun? 22-Dec-19 02:30 PM Yeah 22-Dec-19 02:30 PM The site is straight out of 1998 22-Dec-19 02:30 PM Missing clip art 22-Dec-19 02:31 PM I know, but it could still be useful 22-Dec-19 02:31 PM I recommend obtaining comsol for home experiments 22-Dec-19 02:31 PM It's an amazingly powerful tool 22-Dec-19 02:31 PM I have the full version 22-Dec-19 02:32 PM There's no non-full version you can obtain 22-Dec-19 02:32 PM Solidworks has a non-commercial license, I have it, but it's quite restricted on what you are allowed to use it on 22-Dec-19 02:32 PM compared the fusion360 "yeah as long as you have under 250k revenue all is ok." thing 22-Dec-19 02:33 PM they will never have a similar license for inventor though :( 22-Dec-19 02:33 PM that'd make fusion obsolete instantly lol 22-Dec-19 02:34 PM Pretty sure I have everything for COMSOL Multiphysics 5.3a 22-Dec-19 02:34 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-9A452.png 22-Dec-19 02:34 PM yep 22-Dec-19 02:34 PM I may or may not have it too 22-Dec-19 02:34 PM https://i.spirit.re/xW7cm.png 22-Dec-19 02:34 PM :D 22-Dec-19 02:36 PM Yeha I've just realized how diverse it is 22-Dec-19 02:36 PM I've only used one or two modules, CFD and FEA for heat conduction 22-Dec-19 02:39 PM https://i.spirit.re/2QWlJ.png 22-Dec-19 02:39 PM Simulating electron optics is pretty interesting 22-Dec-19 02:39 PM and very complicated 22-Dec-19 02:39 PM Also very slow 22-Dec-19 02:40 PM EGN2 on modern hardware should be super fast 22-Dec-19 02:40 PM Isn't it mostly 2D? 22-Dec-19 02:40 PM Yeah 22-Dec-19 02:40 PM I like the 3D aspect of comsol 22-Dec-19 02:41 PM I don't think I'll need it? Not for a simple beam blanker 22-Dec-19 02:41 PM No, you don't 22-Dec-19 02:41 PM It's just pretty :P 22-Dec-19 02:41 PM I however dislike that it takes 5-10 minutes to compute a simple model on my toaster 22-Dec-19 02:45 PM It does molecular flow too? You can combine electron/ion simulations with vacuum in a drift tube/column 22-Dec-19 02:45 PM you can simulate everything within a single sim 22-Dec-19 02:45 PM fields, electron, thermal, stress/strain, vacuum, molecular, pumping, reflection, etc 22-Dec-19 02:45 PM it'll take you a month to compute it all at once, but you can 22-Dec-19 02:46 PM The molecular flow is not based on Monte-Carlo methods 22-Dec-19 02:46 PM Quite unique I think, but I think it won't be super useful for UHV applications 22-Dec-19 02:47 PM well, simulating vacuum isn't really necessary 22-Dec-19 02:47 PM assuming it's perfect is fine for 99% of applications 24-Dec-19 01:59 AM https://upload.wikimedia.org/wikipedia/commons/thumb/4/4d/Silicon_grown_by_Czochralski_process_1956_closeup.jpg/320px-Silicon_grown_by_Czochralski_process_1956_closeup.jpg 24-Dec-19 01:59 AM induction heater 24-Dec-19 05:02 AM Someday we'll pull our own silicon crystals 24-Dec-19 05:02 AM XD 24-Dec-19 05:55 AM I want this book but damn is shipping expensive: 24-Dec-19 05:55 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/s-l500-686E2.png 24-Dec-19 06:05 AM How's this even possible, from 3$ shippin in USA, to this: 24-Dec-19 06:05 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-EDDC6.png 24-Dec-19 06:27 AM we used a czochralski puller for a lab in undergrad, it was a fascinating machine. we grew little germanium crystals. 24-Dec-19 09:32 AM @Nixie http://www.libgen.is/search.php?req=9781574446753&open=0&res=25&view=simple&phrase=1&column=def 24-Dec-19 10:02 AM Ohh, thanks! 24-Dec-19 10:06 AM @Nixie You need a mail forwarder ) 24-Dec-19 10:10 AM XD, I got device forwarders. 24-Dec-19 10:10 AM Just happens that I just exchanged my las favor for some materials (25um Al wire and some other bits) 24-Dec-19 10:10 AM I'll ask twitter, altough the particular 1st edition is 1100 pages, so not a small book anyways. 24-Dec-19 02:06 PM Do you really want a printed book? 24-Dec-19 02:16 PM The PDF downloaded for me in about a minute, and is searchable with CTRL-F 24-Dec-19 02:28 PM libgen is great for a lot of technical texts 24-Dec-19 02:51 PM I pretty much opprefer printed copies if possible. I haven't had many technical books on my youth, so now I'm just building my library. 24-Dec-19 03:42 PM if i had a copy of all the books i wanted to reference i would be drowning in more books than i already am 24-Dec-19 03:42 PM easier to search pdf files too 24-Dec-19 05:36 PM I hoarded books as a kid, it's a lot easier to hoard PDFs 24-Dec-19 05:36 PM I even had a t-shirt in my youth that said something like "anytime i get some money, i spend it on books" 24-Dec-19 09:12 PM on the same level as libgen, https://sci-hub.tw/ for those that don't know already 24-Dec-19 10:29 PM libgen is more for legally clean stuff 24-Dec-19 10:29 PM scihub doesn't care, you input a doi and you get a doc 25-Dec-19 12:55 AM legally clean? not sure what you mean there. My interpretation would be "less illegal" - is what what you're saying? In any case, libgen is dirtier, IMO, as it's facilitating infringement of your average author whereas scihub is saying "#*@& you" to the crooked scholarly journal institution 25-Dec-19 12:55 AM libgen is blatant copyright infringement, basically.. and scihub is a protest to the journal system 30-Dec-19 03:21 AM Getting my metallographic microscope by the end of the week! 30-Dec-19 03:21 AM yay! 30-Dec-19 08:28 AM Nice. I need to pick one of those up as well 30-Dec-19 08:30 AM Also received the hot plate. It's bigger than expected, but not too much. 30-Dec-19 08:30 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/IMG_20191230_154748-ED595.jpg 30-Dec-19 02:12 PM So when are you picking up an electron microscope Nixie? 30-Dec-19 02:15 PM When he's able to make feature sizes too small for optical :P 30-Dec-19 10:42 PM When you find one nearby me. 31-Dec-19 01:53 AM You're on the top of my list for Spain 31-Dec-19 03:23 AM https://tenor.com/view/the-office-erin-hannon-ellie-kemper-text-message-blush-gif-3911645 03-Jan-20 01:27 AM @Nixie have you played with the ultrasonic scaler yet? 03-Jan-20 01:41 AM Haven't received it yet. 03-Jan-20 01:41 AM With all the christmas, it's normal shippings get slow. 03-Jan-20 01:42 AM there's kits on ebay for ~$100 so I'm tempted to see how it bonds silicon 03-Jan-20 01:42 AM got a source for wire? 03-Jan-20 01:42 AM If you have plenty of money, please, have a go, the more we try, the faster we can come up with solutions for it to work. 03-Jan-20 01:42 AM Ebay in USA 03-Jan-20 01:42 AM found a spool of 25um aluminium wire 03-Jan-20 01:43 AM heh 03-Jan-20 01:43 AM care to wind off a few meters for me? 03-Jan-20 01:43 AM shipping was killer, though 03-Jan-20 01:43 AM no luck here 03-Jan-20 01:43 AM (haven't received it either) 03-Jan-20 01:43 AM (oh) 03-Jan-20 01:44 AM Today I'm waiting for the metallographic microscope to pop up 03-Jan-20 01:44 AM DHL driver says I have to pay charges, I have proof of already paid said charges. 03-Jan-20 01:44 AM let's see how does that unwind. 03-Jan-20 04:04 AM https://twitter.com/nixie_guy/status/1213068359446728705 03-Jan-20 04:04 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/ENWvQ7iWwAE03yd-982E1.png 03-Jan-20 04:04 AM It does work, it just has the quality in par with the price. 03-Jan-20 04:04 AM But it's better than nothing, of course. 03-Jan-20 04:04 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/ENWvQWTW4AABRGu-BF851.png 03-Jan-20 06:51 AM hmm. i was able to get an old olympus school microscope to do quite well, a real metallurgy scope should be better, right? 03-Jan-20 06:51 AM this was taken using a cover slide at 45 deg as a beamsplitter for epi lighting 03-Jan-20 06:51 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/20190924_214402-DFF14.jpg 03-Jan-20 06:51 AM this as well, but at much lower magnification (the first pic is one of the small units at the lower edge of this picture, also this is at metal layers, first pic is poly) 03-Jan-20 06:51 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/EFGVUSDXsAAmHsM-98F4F.jpg 03-Jan-20 08:02 AM I can't say the one I bought doesn't work, it just could be better. 03-Jan-20 08:02 AM XDDDD 03-Jan-20 01:33 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/IMG_20200103_003421-BF7F3.jpg 03-Jan-20 01:33 PM So yeah this just fell into my lap 03-Jan-20 01:59 PM oooh 03-Jan-20 01:59 PM Yeah 03-Jan-20 01:59 PM This thing is like fancy af 03-Jan-20 01:59 PM I mean it's job is to spray hot fuming acid at things so it kinda has to be 03-Jan-20 04:46 PM neat 03-Jan-20 09:53 PM @rfs Ruthenium Oxide Update: Congratulations! You have now engaged the DOE brain trust. This continues to be an odd question that has people worried because they'd never previously considered it and the first blush is not comforting. 04-Jan-20 01:26 AM interesting... very interesting 04-Jan-20 01:26 AM maybe I ought to just look into making my low-noise electrodes with 2d materials like graphite or something 04-Jan-20 03:32 AM @mike crb got anything you can point me to re: reactivity of platinum vs tungsten when used as electrodes in electrochemical cells? I know platinum is the "standard" for non-reactivity, but idk how actually non-reactive it is... and presumably the electrode material changes things like at what voltage hydrolysis occurs? or is that independent of the electrode and totally dependent on i.e. water's properties? 04-Jan-20 05:35 AM Woah. 04-Jan-20 05:54 AM what kind potential/reaction are you looking to get? 04-Jan-20 05:54 AM Tungsten is fairly rare as far as electrochemistry goes not one I see often 04-Jan-20 05:54 AM Platinum is pretty standard especially for hydrogen evolution, but you can definitely still degrade it 04-Jan-20 05:54 AM Even tho it’s pretty stable 04-Jan-20 09:36 AM @funranium yay. 04-Jan-20 01:34 PM @rfs Just helping me maintain my cred with Industrial Hygienist community for being THAT GUY who asks the questions that generate lots of extra work. 04-Jan-20 01:53 PM Heh. Have fun. 04-Jan-20 02:35 PM @mike crb not really sure about potentials, looking to perform liquid-media electron tunnelling experiments, so the voltages I guess will be approaching 0... but I haven't been reading about the electronics/chemistry much lately, moreso on the fab side. Have tungsten and molybdenum fab capabilities for nano electrodes nearby... but not platinum at the moment. Not sure how their noise characteristics compare (i.e. dislocations in the crystal structure)... but supposedly the 2d conductors are a LOT better at quality conduction than bulk materials 04-Jan-20 05:08 PM EC STM? 04-Jan-20 05:08 PM Or just liquid STM? 04-Jan-20 05:08 PM I know some people who have worked on EC STM and we’re doing a collaboration soon if you’d like me to drop a couple questions 04-Jan-20 06:05 PM EC meaning? 04-Jan-20 06:05 PM hmm, electrochemical... I thought STM was all about electrochemistry, lol 04-Jan-20 06:05 PM https://sci-hub.se/https://avs.scitation.org/doi/10.1116/1.3081965 04-Jan-20 06:05 PM "The direct injection of liquid droplets into low pressure plasmas" 06-Jan-20 11:34 PM https://twitter.com/kenshirriff/status/1214408764217909248?s=09 07-Jan-20 02:30 AM Yay! Imaging my own etchings! 07-Jan-20 02:30 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/ENqvhn4WkAE5AHH-3A5B9.png 07-Jan-20 02:30 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/ENqvi82WkAAwv7u-EF521.png 07-Jan-20 02:30 AM That's looking pretty good! 07-Jan-20 03:00 AM https://tenor.com/view/thanks-thank-you-dwight-happy-tears-of-joy-gif-3553882 07-Jan-20 03:07 AM @AdamMcCombs did you hear anything about this? 07-Jan-20 03:07 AM https://twitter.com/rivatez/status/1145823684600320003 07-Jan-20 05:07 AM (playing with thin copper wires and hairs) 07-Jan-20 05:07 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/ENrjt2kXYAIwwhW-140FC.png 07-Jan-20 05:07 AM (someone at twitter asked for a reference photo...100um copper wire on top, yeah, not especially small features)) 07-Jan-20 07:47 AM Btw, that Silicon100 thing seems it was a closed thing for people in the fab business, so no recordings. 07-Jan-20 07:47 AM On the other hand, I got this interesting read about EUV technology. 07-Jan-20 07:47 AM https://thechipcollective.com/posts/moshedolejsi/what-is-up-with-euv/ 07-Jan-20 07:53 AM Had a chat at the bar with a fellow who worked on EUV for Intel and it crazy the dimensions they can get with it 07-Jan-20 07:53 AM Like Samsung is planning on manufacturing 3 nm node next year 07-Jan-20 07:53 AM crazy 07-Jan-20 07:54 AM Woah 07-Jan-20 09:20 AM That’s crazy. What’s theoretical limit of litho? Like 2.4 and that’s already taking ever possible lensing effect and QED considerations in. Insane. Now the game will be bonding technology to get packaging down so there can be more multi die package chips out there. With processes that small we will have integrated processors and bulk memories on the same die, via the same processes. Mind boggleing in process. Do not disturb!!!! 07-Jan-20 09:27 AM There will always be the cooling limit. You can't stack dies infinitely 07-Jan-20 09:27 AM Altough imagining a full chunk cube of silicon that is ALL processor, would be amazing. 07-Jan-20 10:21 AM I mean you could do extremely parallel but slower tasks to keep wattage down 07-Jan-20 10:21 AM just need thermal sensors in the dies to detect when things are getting warm from speed 07-Jan-20 10:21 AM you could also just cool the die stack with flourinert 07-Jan-20 10:21 AM especially if you get the solder balls big and spaced apart enough between dies to allow flow 07-Jan-20 10:38 AM Yeah, but at some thickness of stack, you can't cool efficiently the inside. Altough I remember reading about in-die cooling channels somewhere. 07-Jan-20 11:24 AM Heard from a litho chemist at Intel that EUV photons are so few, but energy so high, it's like serious monte carlo dart throwing territory... extremely challenging to ensure you get even exposure and development 07-Jan-20 11:26 AM I think they count in the 10's of photons per square nanometer or something like that. 07-Jan-20 01:34 PM From what I’ve read 1.4 nm node is the limit 07-Jan-20 01:34 PM Ridiculous how they get anything in the sub 7 nm region 07-Jan-20 01:34 PM So much energy lost in the mirrors and have to count for thermal expansions it’s crazy 07-Jan-20 01:41 PM 07-Jan-20 01:41 PM 97% energy LOST in the mirrors 07-Jan-20 01:41 PM I was like...ROFL... 07-Jan-20 01:42 PM absurd amounts 07-Jan-20 01:42 PM the ASML euv instruments are crazy too 07-Jan-20 01:42 PM Like house sized and cost like 100 million usd 07-Jan-20 01:44 PM /brain reads "ASMR EUV"/ 07-Jan-20 01:44 PM lmao 07-Jan-20 01:44 PM semiconductor processing ASMR now there’s a niche 07-Jan-20 09:49 PM Now I’m a pretty nerdy guy, but even I want to see a picture of the collection of people qualified to design/work at the level of 1.4 nm! Op please let them still have pocket protectors!!! Hahahaha 08-Jan-20 12:25 AM Give him a few years and this guy will probably be doing it: 08-Jan-20 12:25 AM https://twitter.com/lasserith 08-Jan-20 12:25 AM No pocket protector to be seen, though. 08-Jan-20 02:09 AM haha, idk, yield engineer at Intel means "excel data jockey" in my experience 08-Jan-20 02:18 AM looks like he only got hired a few months ago... so I guess he's still fresh enough that he might not settle into that job role, if it is indeed as boring as I've heard it sound like 08-Jan-20 03:22 AM Absolutely no idea. 08-Jan-20 03:22 AM XD 08-Jan-20 07:51 AM Oh yield engineer can also mean cleaning specialist. Also chemical process no-no avoidance officer as well. Hahahaha 13-Jan-20 12:24 PM https://arxiv.org/pdf/1807.07912.pdf 13-Jan-20 09:44 PM "Fabrication of low-cost, large-area prototype Si(Li) detectors for the GAPS experiment" 14-Jan-20 07:42 AM what would the photoemission wavelength be for a modern silicon CMOS process? 14-Jan-20 08:04 AM 193 nm ArF laser is usually the source 14-Jan-20 08:05 AM i mean the actual light emitted by the junctions 14-Jan-20 08:05 AM oh I’m dumb was thinking you were asking standard litho wavelength 14-Jan-20 08:06 AM nope, was thinking of doing some back of die photography 14-Jan-20 08:06 AM but not sure if i need a special camera or whether a defiltered cmos will suffice 14-Jan-20 08:07 AM what is a defiltered cmos? 14-Jan-20 08:17 AM as in CMOS cam w/o any bayer or IR-stop filters 14-Jan-20 08:47 AM @rfs I bring the wisdom from on high regarding ruthenium oxidation. RuO2 - fairly benign and the more likely room temp oxidation state. Has an LD50 comparable to TiO2, which is a lot. RuO4 - rather nasty, but also a liquid at room temp. So while it might contaminate things which you then need to throw away as hazardous waste, it isn’t going to be an inhalation hazard like BeO. 14-Jan-20 08:47 AM So, your plated probes are unlikely to be a problem just sitting in the drawer. 14-Jan-20 11:01 AM Hmm, how do you think for sputtering of less than 20nm sometimes it would be fine to use RuO2? 14-Jan-20 11:01 AM Cause seems like if you have gloves on while handling it shouldn’t be a problem 14-Jan-20 01:43 PM @funranium That is very good to hear. 14-Jan-20 01:44 PM To photograph dies I had to buy a metallographic microscope 14-Jan-20 01:44 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/IMG_jpsjrh-29D01.jpg 14-Jan-20 01:44 PM Without it it was extremely difficult. 14-Jan-20 02:09 PM nixie: i built one 14-Jan-20 02:09 PM it works, but now i want to try to capture the photoemission from a running chip 14-Jan-20 03:05 PM @pbx/peterbjornx you can see big TO-can jellybean fets emit around 800-900nm 14-Jan-20 03:25 PM Ahh, then I did not.understand.you, sorry 15-Jan-20 02:33 PM some notes on making my own photoresist from easily accessible materials, certainly not ready for prime time yet but there are encouraging results for the DIY community http://sam.zeloof.xyz/diy-pr/ 15-Jan-20 02:43 PM Sam. Is there anything I can do with a stack of boron P type bare wafers that were orphaned with me? 15-Jan-20 02:43 PM Useable for any mischief or garbage? 15-Jan-20 02:43 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/image0-1BC81.jpg 15-Jan-20 04:15 PM Well you can certainly do lithography on them, and make stampers for diffractive optics or microfluidics, holography 15-Jan-20 09:38 PM They are 1-1-0 bit difficult to use. 16-Jan-20 02:28 AM https://youtu.be/zxIiTyqEax4 16-Jan-20 04:29 AM hehe 16-Jan-20 09:25 AM I remember asking this same question loooooong ago, and the person I asked, just told me the led would not work anymore. 16-Jan-20 09:25 AM (I can't remember who I asked, sadly) 16-Jan-20 10:10 AM freeze the electrons in place and stop them from moving of-course 16-Jan-20 10:10 AM lol no it depends on the LED, some do go out I think or just start emitting something outside of human vision maybe? 16-Jan-20 10:10 AM dunno 16-Jan-20 10:19 AM but tue change is just temperature dependant, will go back with warming up 16-Jan-20 01:10 PM so whats the actual mechanism behind the colour change? this is a demo i have often performed in front of classes but no one has been able to provide me with a good explanation 16-Jan-20 01:11 PM the bandgap energy is temperature dependant 16-Jan-20 05:28 PM Yah lattice of semiconductor “shrinks” per day 16-Jan-20 05:28 PM valence electrons get closer 16-Jan-20 05:28 PM And bandgap increase due to electrostatics increasing 17-Jan-20 05:40 AM RGB projector using the same LED at different temperatures? 17-Jan-20 06:10 AM would need a really fast thermal control loop 20-Jan-20 05:44 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/P1000868-DF541.jpeg 20-Jan-20 05:44 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/P1010180-0B2CD.jpeg 20-Jan-20 05:44 PM Added a better camera to my DIY metallurgical microscope :) 20-Jan-20 06:04 PM Nice! 20-Jan-20 06:13 PM will need to make a better adapter for the camera ( this one still has the cutout for the webcam sensor, so horrible vignetting) 20-Jan-20 06:13 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/IMG_8190-C171A.JPG 20-Jan-20 06:13 PM crop from the centre of that photo, this is a 600nm IC 20-Jan-20 06:13 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-33B56.png 20-Jan-20 06:16 PM Not bad! 20-Jan-20 06:26 PM this is how I turned the bio microscope into a metallurgical one, the optics in this is no more than a cover slip, with an eyepiece used for focussing the light source 20-Jan-20 06:26 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-870D9.png 20-Jan-20 11:36 PM Yay! Well done!! 20-Jan-20 11:52 PM https://phys.org/news/2020-01-stable-semiconductor-neutron-detector.html 20-Jan-20 11:52 PM pretty neat 21-Jan-20 07:38 AM Couple this with the semiconductor based particle accelerator announced a while back and you are starting to get into some Star Trek level stuff! Getting closer to Dr McCoy’s salt shaker scanner every day! 21-Jan-20 07:54 PM I don't see any publications or info besides news 21-Jan-20 07:54 PM Ah, no, found it 21-Jan-20 07:54 PM https://i.spirit.re/5wqrA.pdf 21-Jan-20 07:54 PM Needs Li6 21-Jan-20 09:22 PM i just plugged the DOI into sci-hub heh 21-Jan-20 09:50 PM @mike crb does electrode material affect something like the (water) hydrolysis voltage? 21-Jan-20 09:50 PM What about fluid makeup? 22-Jan-20 06:22 AM Yes. Some materials are catalytic for water splitting, and as such you won’t need as intense of potential 22-Jan-20 06:22 AM what do you mean by fluid makeup 22-Jan-20 06:22 AM Like the electrolyte? 22-Jan-20 10:02 AM Yeah, electrolyte or other fluid (non conductive) 22-Jan-20 10:02 AM Being catalytic for water seems unique to a surface though, the surface topology on the angstrom or nanoscale I'm guessing, coupled with it's dipole and other electronic properties 22-Jan-20 10:02 AM But in general, are most bulk materials not affecting the hydrolysis voltage? 22-Jan-20 11:56 AM yeah usually bulk doesn’t have huge difference 22-Jan-20 11:56 AM pt group metals best for it 22-Jan-20 11:56 AM For the hydrogen evolution at least 22-Jan-20 11:56 AM and electrolyte should be either highly acidic or basic 22-Jan-20 11:56 AM most common you see h2so4 or koh 22-Jan-20 04:05 PM <__ice9#6039> For alkali hydrolysis producing mixed H2 and O2, it's best to use about 15-30% KOH with nickel electrodes. If that's too costly, 316 stainless is mostly ok, but there will be some chromium ion dissolution including Cr6+. Less so at the lower end of KOH concentrations. Electrode plates should be separated by like 1-2mm or so and operated at a cell potential around 1.7-1.8V for KOH (lower for H2SO4, about 1.3V). It's best to use DC or a square wave. AC peaks may encourage other less desirable reactions. Only the end plates are connected to power. The middle ones are at floating potentials and should be flat and evenly spaced. They can be either clamped with gaskets and holes drilled in them for fluid and gas flow ('dry cell') or just placed in a tank with some pumps to circulate the electrolyte (ABS is fine in KOH). 22-Jan-20 04:05 PM <__ice9#6039> H2 and O2 can be separated with zeolite 5A or thin walled palladium tubes 22-Jan-20 04:05 PM <__ice9#6039> Alternately, the reactor can be designed to separate the electrodes for each cell to collect the gases separately, but practical production rates will be much lower. In industry, usually Nafion or similar membranes are used, but they are pretty costly. 22-Jan-20 04:05 PM <__ice9#6039> Obviously mixed oxyhydrogen is explosive. It's important to build the reactor in a way that overpressure can be safely dissipated and use flashback arrestors at the start and end of the tube/hose conveying the gas mix (usually to some kind of torch or similar) 22-Jan-20 04:05 PM <__ice9#6039> Current density for flat metal plates should not exceed about 0.5A/in2 roughly 22-Jan-20 04:05 PM <__ice9#6039> Using nickel foam electrodes can improve current density substantially 23-Jan-20 12:46 AM so i'm working on reverse engineering a CMOS IC, what tools would you guys recommend for storing the geometry? 23-Jan-20 12:46 AM atm i'm just tracing micrographs on inkscape but that's obviously not ideal 23-Jan-20 05:55 AM I can't find what Ken Shirriff uses, but Robert Baruch also uses Inkscape 23-Jan-20 08:56 AM @__ice9 actually I'm not interested in producing H2 or O2... they'd be undesirable reactions .. I was just wondering about the phenomenon in general and how/if it depended on the bulk electrodes and fluids 23-Jan-20 08:56 AM In my case I'm interested in micro and nanofluidics, where electrophoresis may be used to move ions around, and tunneling current (or some form of voltammetry/impedance spectroscopy) may be used for analysis. 23-Jan-20 08:56 AM So a bubble would basically break conductivity and present major problems in terms of clogging and overall device functionality 23-Jan-20 09:06 AM https://pubs.acs.org/doi/10.1021/ac0007940 23-Jan-20 09:07 AM Yeah pedot:pss is one 23-Jan-20 09:08 AM I’m not sure how it’d be able to work with redox buffers though if you’re trying to do CV or EIS 23-Jan-20 09:08 AM hm 23-Jan-20 09:09 AM But they don't last forever 23-Jan-20 09:09 AM Yeah I guess just need lower voltages 23-Jan-20 09:09 AM Below reactivity levels 23-Jan-20 09:09 AM I've been thinking EIS, theres a nice integrated one-cip EIS solution from Analog Devices I believe 23-Jan-20 09:10 AM that’s nice 23-Jan-20 09:10 AM you can build cheap potentiostat for cv too 23-Jan-20 09:11 AM ad5933 23-Jan-20 09:11 AM Dev board https://www.digikey.com/product-detail/en/digilent-inc/410-246/1286-1093-ND/4090068 24-Jan-20 05:55 AM so i've been having some trouble understanding this mask rom cell 24-Jan-20 05:55 AM all metal layers removed 24-Jan-20 05:55 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-56271.png 24-Jan-20 05:55 AM metal 1 24-Jan-20 05:55 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-BC5A0.png 24-Jan-20 07:50 AM Hmm, I've got no experience with reading semiconductors micrographs, but do you think you have enough resolution? 24-Jan-20 07:56 AM Only slightly 24-Jan-20 07:56 AM The problem is that the cells seem to have two independent gates 24-Jan-20 07:56 AM The structure (horizontally) in this picture is P G ? G P where P is a power or ground bus, G is poly ? is only present on some places and has a via to the readout lines 24-Jan-20 08:44 AM Two independent gates? isn't that a logic AND gate? 26-Jan-20 04:21 AM @pbx/peterbjornx What microscope is this and what camera adapter? It looks like it will benefit from lateral chromatic aberration correction using tca_correct. 27-Jan-20 03:00 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/IMG-20200127-WA0024-DADF7.jpg 27-Jan-20 03:00 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/IMG-20200127-WA0023-C387B.jpg 27-Jan-20 03:00 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/IMG-20200127-WA0022-5993E.jpg 27-Jan-20 03:00 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/IMG-20200127-WA0021-E685E.jpg 27-Jan-20 03:00 PM I made a.phone holder for my microscope. 27-Jan-20 03:00 PM I have to do a revision, but seems to be on the right track. 27-Jan-20 03:00 PM (yes, I'll add some foam between the eyepiece and the phone.) 27-Jan-20 03:00 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/IMG_-cory2g-4D1C8.jpg 27-Jan-20 03:00 PM Quick photo taken before running to bed. (Not the best focus, 150x mqgnification) 27-Jan-20 03:45 PM looks pretty decent 27-Jan-20 03:47 PM A bit blurry, can do better, but the phone was hanging by a thread, so to speak. 27-Jan-20 03:47 PM I have to make adjustments to the model. 27-Jan-20 03:48 PM i usually just hold my phone because ive been too lazy to make a holder for my scope 27-Jan-20 03:48 PM thankfully what im usually looking at doesnt really need terribly sharp photos 27-Jan-20 03:49 PM I definitely need the holder. And it's chaper than a c-mount camera, in any caae. 27-Jan-20 03:49 PM That reminds me I have to do a wheel adapter for fine focusing. The thumbwheels are sorta small. 28-Jan-20 08:40 AM V1.5 is a go. 28-Jan-20 08:40 AM https://twitter.com/nixie_guy/status/1222196727236775938 28-Jan-20 08:40 AM https://twitter.com/nixie_guy/status/1222196784073793536 28-Jan-20 05:05 PM beware that phone lenses have horrible geom. distortion 28-Jan-20 05:08 PM Yeah but a good OEM will calibrate that out with a simple checkerboard pattern 28-Jan-20 05:08 PM I.e. https://docs.opencv.org/2.4/doc/tutorials/calib3d/camera_calibration/camera_calibration.html 28-Jan-20 05:08 PM https://youtu.be/ViPN810E0SU 28-Jan-20 05:22 PM my samsung S10 was not calibrated so i dont know which phones are 28-Jan-20 10:39 PM My budget says that it's either the phone camera, or nothing for the moment. 29-Jan-20 12:53 AM On a different note: 29-Jan-20 12:53 AM How bad will this be: 29-Jan-20 12:53 AM https://twitter.com/nixie_guy/status/1222440673481314304 29-Jan-20 12:53 AM I'm hopping for 0,1mm steps 29-Jan-20 12:53 AM Not trying to take nice die photos, but to be able to scroll whatever NOT by hand. 29-Jan-20 05:52 AM https://www.aliexpress.com/item/32790147861.html 29-Jan-20 07:06 AM Yeah, someone also posted that on twitter too. 29-Jan-20 07:06 AM Bought one for the lulz. 29-Jan-20 07:06 AM But the travel is very small at 2.5mm 29-Jan-20 09:00 AM That was me 29-Jan-20 10:53 AM Ohh, XDDD 29-Jan-20 10:53 AM Sorry 31-Jan-20 06:42 AM https://www.aliexpress.com/item/33016168515.html 31-Jan-20 06:42 AM https://www.aliexpress.com/item/4000081241884.html 31-Jan-20 06:42 AM these are nice though 31-Jan-20 07:15 AM psht those are precision, if you can see the thread pitch its not fine enough yet! 31-Jan-20 07:15 AM :P 31-Jan-20 07:15 AM I have a little (maybe motor controlled?) linear x-y stage that came in my latest vacuum chamber acquisition, so even good for vacuum! 31-Jan-20 07:15 AM and man the positioning thread have a polished round on the end and it rests against a glass? optical flat for moving the stage <3 31-Jan-20 07:24 AM well, vacuum compatible is sort of a low bar for motors 31-Jan-20 07:24 AM my buddy has a lathe in his FIB and just uses cheap-o hobby motors that he soaks in alcohol to remove any oils before using in his chamber 31-Jan-20 07:24 AM the rest of his apparatus takes care of mechanical confinement though, so he doesn't need to worry much about the motor runout or stiction, etc 31-Jan-20 07:38 AM for an x-y linear stage moreso than motors.. 31-Jan-20 02:21 PM @nmz787 Major consideration is thermal. making sure you can pull heat away from those windings I'd say! D: 31-Jan-20 02:22 PM Guy's been using them for years and years... Only thing to die is the brushes, which is normal for a brushed motor 31-Jan-20 02:25 PM huh 31-Jan-20 02:25 PM brushed motors? o.o 31-Jan-20 02:25 PM aeeek, really wonder how hot the rotor gets, given no way to really get rid of the heat 31-Jan-20 03:09 PM well it's mounted to a piece of metal, which is mounted on the stage... so conduction 31-Jan-20 03:09 PM doubt it gets very hot, it doesn't see much use... usage is like, pulse the motor to spin the workpiece, then work on the piece for 5 to 10 minutes I think 31-Jan-20 11:21 PM are there decent DIY steppers out there? like a reverse optical comparator is what I am thinking (for PR exposure) 31-Jan-20 11:21 PM oi, actually, wasn't a common computer projector a decent solution? What if you want to target UV light (say, for SU-8).. you would then maybe need to switch to a DLP with a UV light source? 01-Feb-20 07:08 AM Sam Zoloff ( think his on this server too) IIRC uses a DLP projecter modified with UV led as the light source to do photolitography 01-Feb-20 08:32 AM @Noxz https://pdfs.semanticscholar.org/b888/ab11bb242607fd66f4ff4d1d51c4e5bf549d.pdf 01-Feb-20 08:32 AM http://diyhpl.us/~nmz787/pdf/Mask-less_Photolithography_System.pdf 01-Feb-20 08:32 AM That one is much much longer 01-Feb-20 08:32 AM In either case you progressively navigate and expose... The first paper calls this out at least from my memory by referring to using multiple colors of light... One for non-exposures so you can get accurate alignment between fields of view, and the other to actually expose with once aligned 01-Feb-20 08:32 AM re: vignetting, for lithography setups... where your light source brightness might not be totally even... you determine the spread of unevenness with a flat-field image (average of multiple images, so you cancel out shot noise), then brighten the areas that were dimmer (or visa versa) 01-Feb-20 08:32 AM https://en.wikipedia.org/wiki/Flat-field_correction 01-Feb-20 08:43 AM also, I specifically stated SU-8 because I want to do tall features, a la UV-LIGA... think tiny gears and such, MEMS sorta things 01-Feb-20 09:12 AM The aspect ratio of su8 isn't as good as liga 01-Feb-20 09:12 AM You might also check out chemically amplified resists, like ZEP 01-Feb-20 09:33 AM Oh, i see they're using su8 in this uv-liga, at least in one paper. And zep aspect ratio is worse than su8 normally... I guess just easier to deal with due to lower internal stresses 01-Feb-20 09:33 AM Hmm, yeah seems like a ton of optimization vs DRIE 01-Feb-20 09:33 AM I've basically been doing a less-deep version of liga, e-beam photoresist exposure, electroplating, then molding/imprint 01-Feb-20 09:58 AM neat, that's sort of what I am after 01-Feb-20 11:08 AM also, checking the dates on those articles, 2005 and 2009.. I could search for newer.. but I am wondering if anyone has any bookmarked (already filtered through all the repeats, etc) 01-Feb-20 11:08 AM I understand how DLP and so forth would create 'pixels' which is not adventageous to round objects like gears, without an aditional movement + new image (like rasterizing or anti-aliasing?) .. the idea of having a very large version of the produc,t say, 100x, that gets beamed through reducing lenses was one thought train... but this does not align with the articles much as it is far from "rapid prototyping" but more towards exact replica production 01-Feb-20 11:32 AM glanced through both articles.. mostly what I was expecting from students and such... not that I dont have a limitted budget, but the results also seemed a bit less on the precision engineering side which I have been reading heavily on 01-Feb-20 11:51 AM Precision engineering just means standardize the crap out of your reagents and processes,mostly 01-Feb-20 11:51 AM To hide the pixelization, just over-represent your features... I.e. use more pixels per feature size 01-Feb-20 11:51 AM This is just another aspect of engineering for your needs 01-Feb-20 11:51 AM The flat fielding is another optimization, just as stitching fields of view with a non-exposing wavelength is.. using camera feedback 01-Feb-20 11:51 AM Precision engineering on that front means using machine vision, but for a student or one-off project you could likely just do it manually 01-Feb-20 11:51 AM Using a better illumination source would be another "precision engineering" upgrade 01-Feb-20 11:51 AM I had those bookmarked myself, not sure if there's anything newer... Not sure much has really changed to be honest 01-Feb-20 01:23 PM thanks for the input 01-Feb-20 01:33 PM Other enhancements are with the stage, i.e. using an interferometer feedback, but that's mostly when optical feedback can't work due to target feature sizes being less than optical resolution 01-Feb-20 01:33 PM If you're aiming for mems stuff, then you're probably above the optical limit for your smallest features, i.e. corners and curves 01-Feb-20 01:35 PM by above, you mean within the operating limits or beyond 01-Feb-20 01:35 PM sub micron is the goal, but certainly no 15nm needed 01-Feb-20 01:45 PM Yeah like IC steppers doing anything less than like 400nm features 01-Feb-20 01:45 PM If you can't accurately index fiducials to within your feature size, you get stitching errors that are intolerable 01-Feb-20 01:45 PM I've been planning to use an algorithm that panorama photo generators use... I.e. using the scene as the fiducial, rather than engineering a fiducial outside of your features 01-Feb-20 01:45 PM Like OpenCV's SIFT or SURF algorithms 01-Feb-20 01:56 PM you don't need interferometric feedback 01-Feb-20 01:56 PM all you need is a very high pitch, preloaded screw, a small DC motor, and a really high resolution magnetic strip encoder 01-Feb-20 01:56 PM the better ones can get down to nanometer per count resolution 01-Feb-20 02:46 PM You don't need it, but the fabs use them 01-Feb-20 02:46 PM There's no doubt with them 01-Feb-20 03:17 PM What travel range are you needing for a stage? There are cheap piezo flexures bundled in Z axis objective positioners on eBay these days if your range is very small but needs to be very accurate. 01-Feb-20 03:17 PM https://www.ebay.com/itm/283391364525 and a few other styles similarly priced. I have several of various styles if anyone needs me to measure travels of the lone flexure elements (should be the same as the overall assembly/product specs according to their geometry I see) 01-Feb-20 03:17 PM They have inbuilt strain gauges to compensate for the piezo hysteresis for closed loop positioning 01-Feb-20 03:17 PM https://youtu.be/JE8yJfXbmrQ gives a feel for them. That's at 150vpp, -20 through 130 sweep 01-Feb-20 03:26 PM a mask stepper implies large precise motion 01-Feb-20 03:26 PM so way beyond what a piezo actuator can do, unless it's a walker 01-Feb-20 03:27 PM Fair enough 02-Feb-20 12:23 AM Sounds like @Noxz can just use normal steppers and lead screws, and use optical feedback to align fields-of-view 02-Feb-20 09:13 AM I'm more concerned about getting the image/pattern smaller 02-Feb-20 09:13 AM and vertical light rays 02-Feb-20 09:13 AM stepping seems like a non-issue, especially since t his will be aone layer exposure so no need to really allign any fudicials 02-Feb-20 10:03 AM Alignment I'm talking about is in XY, not Z 02-Feb-20 10:03 AM If you have a 1080p DLP... that's 1920 pixels in the X, if you want a 25 micron wide gear, the teeth would be what, 3 microns tall and wide?? Then you need to decide how much definition you want on the edges and corner... Because each tooth is not going to be in line with a row of DLP pixels, so you're going to get stair-stepping and blockiness if you just said i.e. 1 pixel==1 micron 02-Feb-20 10:03 AM So is 10 pixels per micron enough? Or will it be 100?? 02-Feb-20 10:03 AM If it's 10 pixels per micron, you're already over-representing your features by around 4x the wavelength of light you're exposing with (assuming 405nm) 02-Feb-20 10:03 AM Which is fine, because the airy-disc blur can still be moved with less than wavelength precision 02-Feb-20 10:03 AM So 1920/10 = 192 microns in the field of view at a time, in the X... In the Y it's 108 microns 02-Feb-20 10:11 AM this is why you want a DC motor on a super fine pitch, preloaded screw, but the servoing should be done using a magnetic linear scale 02-Feb-20 10:11 AM they're accurate down to <1um and cheap(<$100) 02-Feb-20 10:12 AM So if you're only making a few gears, and your whole device is within those bounds, you don't need any XY alignment 02-Feb-20 10:12 AM @Spirit seems better to me to use optical feedback, but both in combination would be better 02-Feb-20 10:12 AM optical feedback for that accuracy will cost you thousands 02-Feb-20 10:12 AM No way 02-Feb-20 10:12 AM It's just a trinocular microscope 02-Feb-20 10:13 AM wat 02-Feb-20 10:13 AM With a decent camera and microscope objective 02-Feb-20 10:13 AM no 02-Feb-20 10:13 AM ? 02-Feb-20 10:13 AM that would have all kinds of distortion and weirdness 02-Feb-20 10:13 AM Umm, yes 02-Feb-20 10:13 AM absolutely not suitable for repeatable stepping 02-Feb-20 10:13 AM a set of high precision digital glass scales would cost you a serious amount of money 02-Feb-20 10:13 AM I'm not suggesting any scales 02-Feb-20 10:13 AM Use SIFT/SURF algorithm 02-Feb-20 10:13 AM How your digital smartphone camera makes panoramas 02-Feb-20 10:13 AM Use the scene as the fiducial 02-Feb-20 10:14 AM that is ridiculously crutchy 02-Feb-20 10:14 AM Millions or billions of people use it everyday 02-Feb-20 10:14 AM plus you'd need a separate microscope for feedback 02-Feb-20 10:14 AM No 02-Feb-20 10:14 AM You use a splitter 02-Feb-20 10:15 AM you're not going to have micron scales all over your work 02-Feb-20 10:15 AM The articles i posted above do basically that, minus the pano algo 02-Feb-20 10:15 AM very inconvenient and requires full manual operation 02-Feb-20 10:15 AM Not sure what micron scales you're talking about feature sizes? 02-Feb-20 10:15 AM Not if you're a programmer 02-Feb-20 10:15 AM Opencv is free and easy from Python 02-Feb-20 10:16 AM it even sounds ridiculous 02-Feb-20 10:16 AM all you need is a pair of magnetic scales which are cheap 02-Feb-20 10:16 AM Ok, hope your backlash is insanely low 02-Feb-20 10:16 AM ... back at you? 02-Feb-20 10:16 AM you're still going to have backlash regardless of what turns the screw 02-Feb-20 10:17 AM Sounds more expensive than a decent smartphone camera and microscope that will already be present to beam the DLP down 02-Feb-20 10:17 AM $50 for a meter of high grade magnetic tape, plus $20 per axis 02-Feb-20 10:17 AM Yes backlash will exist but you compensate with the optical feedback 02-Feb-20 10:17 AM and you get sub-micron 02-Feb-20 10:17 AM which is automatically compensated by the zero backlash magnetic encoder 02-Feb-20 10:17 AM the actuators could be pneumatic for all you care, you're servoing based on the position of the axis 02-Feb-20 10:17 AM not the screw 02-Feb-20 10:19 AM Interesting, that technique never jumped out at me 02-Feb-20 10:19 AM it's how all high end CNC gear works 02-Feb-20 10:19 AM except when you get high end enough, it goes to glass scales 02-Feb-20 10:19 AM https://www.rls.si/en/products/linear-magnetic-encoders/rlb-linear-component-magnetic-encoder-system 02-Feb-20 10:19 AM Select 8192 interpolation, 15mhz, connector only, periodic, none, ms05 scale, class b, 98mm each, with VHB 02-Feb-20 10:19 AM you get 83.22 eur for 2 axes 02-Feb-20 10:19 AM 244nm resolution 02-Feb-20 10:19 AM @Noxz this is what I was talking about a while ago 02-Feb-20 10:19 AM I believe this encoder will also work with their ring? 02-Feb-20 10:32 AM "contact for quote" 02-Feb-20 10:32 AM Never is a good sign 02-Feb-20 10:33 AM I can order right away 02-Feb-20 10:33 AM did you select the right components? 02-Feb-20 10:35 AM Oh now I see that 02-Feb-20 10:35 AM What does the "select actuator" mean? 02-Feb-20 10:35 AM It shows +/- 40 microns 02-Feb-20 10:36 AM per meter 02-Feb-20 10:36 AM it's the magnetic scale strip you can glue on 02-Feb-20 10:43 AM Optical can still be better than 244nm res, but whatever is easier and more appropriate for the application 02-Feb-20 10:43 AM I guess I've been in the mindset of pulling as much performance out of: A) as cheap equipment as possible, and B) retrofitting tools that you can't physically modify (i.e. like a shared SEM, where you could easily plug in a video splitter and USB device to act as a keyboard and mouse) 02-Feb-20 10:45 AM you can't index properly using optical 02-Feb-20 10:45 AM because your work won't have optical scales etched precisely all over it 02-Feb-20 10:45 AM your X/Y travel would be fiducial+- some few microns 02-Feb-20 10:45 AM the further you get, the worse your error gets 02-Feb-20 10:45 AM and you lose tracking as soon as you go out of the field of view 02-Feb-20 10:45 AM which is useless for stepping 02-Feb-20 10:50 AM Ah, yeah I've only really been concerned about stitching fields of view 02-Feb-20 10:50 AM was walking the dog, on that site for mag scales now 02-Feb-20 10:50 AM and yeah, one "concern" is the +/- 40um, that I also saw.. but there is a way to calibrate it, basically (interferrometry would be the expensive solution, but basically just mapping the distance across known standards.. like calibrated gauge blocks) 02-Feb-20 10:51 AM it's +-40 microns PER METER 02-Feb-20 10:52 AM But is that +- in the backlash sense? 02-Feb-20 10:52 AM static error 02-Feb-20 10:52 AM 4 nanometers per micron error, essentially 02-Feb-20 10:52 AM negligible 02-Feb-20 10:53 AM Ben Krasnow did a nice video on using Blu-ray lasers for self mixing interferometery 02-Feb-20 10:54 AM not bluray, old red 02-Feb-20 10:54 AM The Z height is also a concern for upper levels of performance 02-Feb-20 10:54 AM @Deleted User i could swear it was a combo laser package 02-Feb-20 10:54 AM for Z you can use piezo as @Metanoic suggested 02-Feb-20 10:54 AM @nmz787 it is, but red, not blue 02-Feb-20 10:54 AM Ah 02-Feb-20 10:54 AM photodiodes were in the red laser era 02-Feb-20 10:54 AM so CD/DVD 02-Feb-20 10:55 AM I have piezo stages for some of that fine tuning 02-Feb-20 10:55 AM How would pieze work to sense the height of the tool from the photoresist film face? 02-Feb-20 10:55 AM Presumably the lead screw will have some wobble over the travel distance 02-Feb-20 10:55 AM Which would affect focus 02-Feb-20 10:55 AM you don't care about that 02-Feb-20 10:55 AM dont measure with leadscrew 02-Feb-20 10:55 AM yeah 02-Feb-20 10:55 AM that is 1920's thought train 02-Feb-20 10:55 AM Hmm? 02-Feb-20 10:55 AM you measure on the position of the actual axis 02-Feb-20 10:55 AM servoing is done based on the linear scale feedback 02-Feb-20 10:56 AM you use a mag/glass scale... or displacement sensor, etc 02-Feb-20 10:56 AM I mean you step in XY, and your Z changes 02-Feb-20 10:56 AM sometimes both of them together 02-Feb-20 10:56 AM no, it won't 02-Feb-20 10:56 AM Why not? 02-Feb-20 10:56 AM if your Z changes with XY, you're doing something very wrong 02-Feb-20 10:56 AM remember that the XY axes ride on linear rails 02-Feb-20 10:56 AM yeah, soemthign with your structural loop 02-Feb-20 10:56 AM Your linear rail isn't perfect, right? 02-Feb-20 10:56 AM for this application, it is 02-Feb-20 10:56 AM you can map it, with a flat standard 02-Feb-20 10:56 AM and a displacement sensor 02-Feb-20 10:57 AM not to mention the rail conforms to the surface 02-Feb-20 10:57 AM you bolt it to granite 02-Feb-20 10:57 AM Well the bolted granite != Your substrate 02-Feb-20 10:57 AM You're not gonna be touching the photoresist 02-Feb-20 10:57 AM At anything but the edges 02-Feb-20 10:58 AM your substrate is to be assumed perfectly flat 02-Feb-20 10:58 AM Or with a vacuum Chuck 02-Feb-20 10:58 AM Assumptions make ass out of u and me 02-Feb-20 10:58 AM it should be well within focus 02-Feb-20 10:58 AM not really an issue 02-Feb-20 10:59 AM (interesting, is the adult filter disabled these days??? I'm surprised that comment got through) 02-Feb-20 10:59 AM I guess it's an animal??? 02-Feb-20 10:59 AM I think it is hardly up there.. not an a-hole 02-Feb-20 11:00 AM Mapping the displacement change makes sense 02-Feb-20 11:00 AM I was planning to do that with a crappy xy stage 02-Feb-20 11:00 AM Before i put that on hold 02-Feb-20 11:00 AM my main question was more along the lines of image generation and shrinkage.. I am pretty sure I know how to do some fanxy XY travel (and mapping for Z/titl if need be) 02-Feb-20 11:01 AM putting a DLP image backwards through a microscope 02-Feb-20 11:01 AM Shrinkage will just be a matter of experimentation 02-Feb-20 11:01 AM especially since I want very vertical features in SU-8 02-Feb-20 11:01 AM @spirit yeah see the last pdfs I posted in here 02-Feb-20 11:01 AM trinocular would be the way to go for that? 02-Feb-20 11:01 AM you won't get them, your surface will be vertical but your etch will not be 02-Feb-20 11:01 AM not etching 02-Feb-20 11:01 AM not for large heights 02-Feb-20 11:01 AM not etching? 02-Feb-20 11:02 AM electro(less?) forming 02-Feb-20 11:02 AM I want metal gears, not silicon 02-Feb-20 11:02 AM you want to form gears? 02-Feb-20 11:02 AM yeah buddy 02-Feb-20 11:02 AM This is just called aspect ratio 02-Feb-20 11:02 AM just go for ion beam milling :P 02-Feb-20 11:02 AM well, aspect ratio also has a width to it too? 02-Feb-20 11:02 AM Su-8 is like 20 on a good day, 40 max with that uv-liga 02-Feb-20 11:02 AM that's how Grand Seiko do it 02-Feb-20 11:03 AM @Spirit ion beam is not appropriate 02-Feb-20 11:03 AM Rolex does UV-LIGA 02-Feb-20 11:03 AM Scale much much too large 02-Feb-20 11:03 AM galvanic 02-Feb-20 11:03 AM @nmz787 tell that to Seiko 02-Feb-20 11:03 AM And aspect ratio is insane 02-Feb-20 11:03 AM @Spirit i have a seiko FIB 02-Feb-20 11:03 AM Seiko watches 02-Feb-20 11:03 AM nothing to do with Seiko FIB 02-Feb-20 11:03 AM They if anything use fib/sem then DRIE 02-Feb-20 11:03 AM I just worked with a brand new Rolex 3250, and the escape wheel and pallet fork was UV-LIGA formed 02-Feb-20 11:03 AM Or just liga 02-Feb-20 11:03 AM nickel-phosphor based 02-Feb-20 11:04 AM Nickel sulphamate is least internal stress electroform based on nickel 02-Feb-20 11:04 AM Seiko use MEMS techniques to make their parts 02-Feb-20 11:04 AM for the ridiculously overkill watches 02-Feb-20 11:05 AM That's what this whole discussion is about 02-Feb-20 11:05 AM they don't grow them afaik 02-Feb-20 11:05 AM Not sure what you mean 02-Feb-20 11:05 AM I dont know enough about seiko to comment 02-Feb-20 11:05 AM one dude in class lieks them, but I dont think he has anythign too new 02-Feb-20 11:05 AM oh, one instructor has a few grand seikos.. 02-Feb-20 11:05 AM he'd be the one to ask 02-Feb-20 11:06 AM We're talking about etching plastic, then sputtering metal, then electroforming and pulling the layers apart 02-Feb-20 11:06 AM grow = electroform 02-Feb-20 11:06 AM etching plastic? 02-Feb-20 11:06 AM Photoresist is basically plastic 02-Feb-20 11:06 AM looks like Seiko use UV-LIGA 02-Feb-20 11:07 AM everyone uses it 02-Feb-20 11:07 AM many places used DRIE for some parts.. I think for silicon parts 02-Feb-20 11:07 AM ouch, SU-8 would be illegal here 02-Feb-20 11:07 AM The only growth in this discussion is the eldctroforming, which is just using the topography in the exposed/developed photoresist as a mold 02-Feb-20 11:07 AM but the metal ones are grown, mostly 02-Feb-20 11:08 AM due to GBU solvent 02-Feb-20 11:08 AM illegal? really? 02-Feb-20 11:08 AM hah 02-Feb-20 11:08 AM Su-8 not needed for DRIE 02-Feb-20 11:08 AM I know 02-Feb-20 11:09 AM If I can find an alternative to SF6, i think r134a air conditioner gas will suffice for replacement of C4F8 02-Feb-20 11:10 AM really? neat.. btu can you get the plain gas without added lubricants? 02-Feb-20 11:10 AM But c4f8 probably isn't too hard to get realistically 02-Feb-20 11:10 AM why look for SF6 alternatives? 02-Feb-20 11:10 AM you can buy it pretty cheap 02-Feb-20 11:10 AM Yeah maybe not needed to find alternates 02-Feb-20 11:10 AM I haven't asked praxair or whoever 02-Feb-20 11:10 AM Still need to get my rf sputtering system up and running first 02-Feb-20 11:10 AM I guess wiki says SF6 is non-toxic so that's good 02-Feb-20 11:12 AM suffocating 02-Feb-20 11:15 AM No worse than my last girlfriend 02-Feb-20 11:18 AM you'll probably want a higher purity version of it 02-Feb-20 11:18 AM since they sell bulk sf6 for transformers/insulators/etc 02-Feb-20 11:29 AM so have we all settled on DLP through a trinocular microscope as the best way to expose a fast prototyping pattern? .. what if I had a large (100x?) physical copy of the item (thus removing the idea of pixels, but aberation and whatnot of the lenses would still exist) 02-Feb-20 11:29 AM then you become a fab 02-Feb-20 11:29 AM har har 02-Feb-20 11:29 AM if you have an accurate stepper, you can just keep increasing the microscope magnification to get rid of the pixels 02-Feb-20 11:29 AM but ALSO 02-Feb-20 11:30 AM just thinking for cases like gears, you certainly dont want to come into staircases of tooth profiles 02-Feb-20 11:30 AM remember that the pixels are not binary(well, electrically they are but otherwise not) 02-Feb-20 11:30 AM so you can do antialiasing, as you would in games, for masks 02-Feb-20 11:30 AM right, I brought that up as an option a day or two ago 02-Feb-20 11:30 AM a few SLA 3D printers do it and it works really well 02-Feb-20 11:31 AM if not prototyping (ie, if in production), then the physical copy sort of makes sense 02-Feb-20 11:31 AM and thus why I was sort of thinking of an optical comparator in reverse 02-Feb-20 11:32 AM well, you could always replace the projector with a mask through some optical voodoo 02-Feb-20 11:32 AM sure.. I would likely prototype with projector first.. but a physical thing makes a bit of sense as well 02-Feb-20 11:33 AM if you can get good enough resolution with a projector, you can probably just stay with that 02-Feb-20 11:33 AM with a good enough optical stepper you can just keep increasing the resolution 02-Feb-20 11:34 AM yeah, that also makes sense 02-Feb-20 11:34 AM do it in quadrants, etc 02-Feb-20 11:34 AM yup 02-Feb-20 11:35 AM unfourtenately I always have this... drive? feeling? of trying to stay "traditional" towards watchmaking and away from computers.. but it really doesnt matter in the end, I know that.. 02-Feb-20 11:35 AM you're trying to do "traditional" watchmaking with #semiconductor methods 02-Feb-20 11:35 AM you've crossed the line long long ago 02-Feb-20 11:36 AM yeah yeah... 02-Feb-20 11:36 AM so, I love chemistry and all.. so PR arent exactly not non-approved in my book 02-Feb-20 11:36 AM and they've been doing plating for a long long time, so electroforming shouldnt be... 02-Feb-20 11:36 AM obviously the answer here is to make semiconductor scale watches 02-Feb-20 11:36 AM what are those chips going to wear to tell the time? 02-Feb-20 11:37 AM right.. so I want to beat "the smallest mechanical watch" but at what oint will "they" cross the line 02-Feb-20 11:37 AM and be like: that isnt made in the traditional sense 02-Feb-20 11:37 AM like, mems will overrule it, sure, but how to stay within the bounds 02-Feb-20 11:37 AM who makes the bounds 02-Feb-20 11:37 AM but the "drive"? this time is because I worked on a brand new rolex with UV-LIGA escape wheel + pallet fork.. and remembered that I looked into doing that sort of thing a while ago and somehow got pulled away due to tradition 02-Feb-20 11:37 AM but... my end goal is basically "proof of precision" 02-Feb-20 11:37 AM a tag line I want to stay with 02-Feb-20 11:37 AM instead of "yeah this works and tells time good enough" 02-Feb-20 11:40 AM patek straight up do silicon wafer mechanicals 02-Feb-20 11:40 AM pure semicon litho 02-Feb-20 11:40 AM yeah dude 02-Feb-20 11:40 AM and they were pretty darn traditional 02-Feb-20 11:54 AM @Noxz might check into how good quality shrinky dinks shrink into 02-Feb-20 11:54 AM If you have a 1x scale copy you could just try copying by molding 02-Feb-20 11:54 AM Coat your part with anti stick (say FDTS) then cast silicone, separate, sputter and then electroform 02-Feb-20 12:07 PM https://scholar.google.com/scholar?hl=en&as_sdt=0%2C38&q=shrinky+dinks+uc+khine&btnG= 02-Feb-20 12:33 PM hah.. I have a vintage box of shrinky dinks.. 02-Feb-20 12:33 PM Nosy Bears 02-Feb-20 12:33 PM I am missing one out of the 24 nosy bears.... argh! 02-Feb-20 12:33 PM havent seen it posted on ebay for 2+ yr 02-Feb-20 12:36 PM Lol 03-Feb-20 01:40 AM @Noxz if you sputter and electroform a given topography (that say, looks like a gear) the electroform will be a plane with a gear raising out of it... Any papers on how to detach this gear from that plane/sheet? Or any papers on making a gear on a spindle, such that it can rotate? 03-Feb-20 01:40 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/Screenshot_20200203-014248-9E3F2.png 03-Feb-20 01:40 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/Screenshot_20200203-014131-7F06B.png 03-Feb-20 01:40 AM These are two different attempts at nickel electroform of ZEP photoresist exposed with e-beam, the black is remaining photoresist that I need to clean up 03-Feb-20 01:40 AM I'm not sure why the second pic attempt has such poor aspect ratio and "chunkiness" 03-Feb-20 01:40 AM I guess maybe you could have two separate electroforms, one with the gears, and another with pins/spindles/axles... Mate them together, then etch off the backside of the gear sheet? 04-Feb-20 02:23 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/IMG_20200203_180152-26C76.jpg 04-Feb-20 02:23 PM Photogrphing a CD. 04-Feb-20 02:23 PM And also 04-Feb-20 02:23 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/IMG_20200204_201426-9601F.jpg 04-Feb-20 02:23 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/IMG_20200204_201437-F23BF.jpg 04-Feb-20 02:23 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/IMG_20200204_202312-C4438.jpg 04-Feb-20 02:27 PM that's getting to zeptobars levels of quality 04-Feb-20 02:27 PM :P 04-Feb-20 02:30 PM Just run it through tca_correct and it will be much better ) tca_correct saved me until I got Olympus... 04-Feb-20 02:30 PM I am just having fun, there is no need for the irony @Spirit 04-Feb-20 02:30 PM What is tha? some software? 04-Feb-20 02:31 PM hey I'm not being ironic 04-Feb-20 02:31 PM it's a compliment 04-Feb-20 02:32 PM @Nixie Yes, part of Hugin package, corrects chromatic aberration. You calibrate it once per lens, and it significantly reduces lateral chromatic aberrations at the corners of the frame: https://wiki.panotools.org/Tca_correct 04-Feb-20 02:34 PM Ohh, noted! 04-Feb-20 02:34 PM Thanks. 04-Feb-20 02:34 PM @Spirit if you say something and then add " :P " that makes it sound like you are making fun of the person. 04-Feb-20 02:38 PM aw, sorry 04-Feb-20 02:38 PM it genuinely does look good 04-Feb-20 02:38 PM next step is a stepping stage and some automation with hugin! 04-Feb-20 02:43 PM @ZeptoBars does fulla correct for distortion by using a photo of a checkerboard pattern? 04-Feb-20 02:43 PM At first glance of the page on the panotools site, it doesn't immediately appear to 04-Feb-20 03:09 PM I will be making a crude stepping stage with a 5mm and 7mm linear rail. 04-Feb-20 03:32 PM @nmz787 I am getting correction coefficient automatically when stitching panoramas. Hugin can derive distortion coefficients from points optimization. 04-Feb-20 03:32 PM @nmz787 I.e panoramas of any subject with large number of trackable features (like random logic microchip) 04-Feb-20 04:08 PM Like with sufficient overlap? 04-Feb-20 04:08 PM That seems less obviously correct than just calibrating a given lens system once with a checkerboard 04-Feb-20 04:08 PM Or at least it feels like it's less intuitively correct or obvious, not to mention I bet it steals some performance of the stitching 04-Feb-20 04:44 PM @nmz787 Yes, 30-50% overlap is good enough for such correction. It does not visibly reduce performance. During panorama stitching there are N*2 variables where N is number of photos. So, for 50 photos you get 100 variables. If you add lens model - it adds just 3 more variables (a,b,c), so around 3% of speed. I typically start optimization without abc, and add it at the very end, when panorama is already good enough. It allows to slightly reduce overall average error. Checkerboard targets is tricky and expensive with microscopes... 04-Feb-20 04:46 PM But even using a complex scene with a lot of overlap, seems like you'd only need to do this modelling once, not every new panorama stitching run 04-Feb-20 04:46 PM @nmz787 Yes, you can do it once. But as it is so easy I don't even bother to save profile. 04-Feb-20 04:47 PM And 3% more variables doesn't necessarily mean only 3% worse performance 04-Feb-20 04:47 PM Hmm 04-Feb-20 04:47 PM Maybe it is so easy because my lenses have quite low distortions 04-Feb-20 06:01 PM while it's on my mind (I'm heading out for a few hours, but just to remember) LIGA originally used xray for the very small wavelength and thus more vertical sidewall features, yeah? and UV-LIGA + SU-8 was developed as a way to do it more easily with readily available UV sources.. but I was thinking.. does SU-8 expose with microwaves? single mode microwaves could be a way to obtain more vertical walls/aspect ratio than UV.. but first page of google results didnt point me in the direction I wanted to go, anyways, just on my mind and needed to note it before I forgot 04-Feb-20 07:16 PM Mm I'd assume SU-8 relies on the photoelectric effect so microwave of say 2.4ghz is orders of magnitude too weak. There's lots of the microwaves but none of the photons carry enough energy to excite whatever atom in the resist needs to be excited to kick off the chemistry. 04-Feb-20 08:14 PM Also the wavelength of 2.4GHz is 12.5cm, gonna be some big diffraction limited features 04-Feb-20 09:17 PM oh wow, I completley mis-stepped and went to a bigger wavlength, doh! 04-Feb-20 09:17 PM for some reason I was thinking microwave was faster than visible light, guess not, just forgot 04-Feb-20 09:17 PM so yeah, I guess stick with UV since I am not making xrays anytime soon 04-Feb-20 09:17 PM the thought is what counts, even if it was wrong 05-Feb-20 05:58 AM E-beam seems easier than x-ray 05-Feb-20 10:19 AM <__ice9#6039> Has pretty similar characteristic limitations too 05-Feb-20 10:19 AM <__ice9#6039> But yes definitely 11-Feb-20 07:51 AM So... I have a theory on some cheap low resolution photoresist... 11-Feb-20 07:51 AM I ordered some, easily obtainable and cheap. The real test will be seeing how small of a feature size you can push it to... 11-Feb-20 07:51 AM But the thickness when coated on something apparently goes down to like 10-15um which is pretty thin considering its intended application. I am going to keep quiet about it until I do some tests to make sure I don't drive people to buy useless stuff :) 11-Feb-20 07:51 AM If it can do 100um even though... it'll be great for home hacking hobby stuff. 11-Feb-20 07:51 AM I started hacking up a DLP projector last night and ordered a UV LED and such, so I am working on getting towards a DLP litho setup. Just seems like the easiest in terms of obtainability for people. Also I can't even try e-beam litho unless I can buy e-beam photoresist which lol yea... 11-Feb-20 07:51 AM Probably going to just use a trinocular microscope with the DLP for now to get started and use a camera in one eye-piece to do focusing 11-Feb-20 07:51 AM I have some x-y tables I need to rig up with some motors to do some stepping. I have one stage that is marked to do 1/1000th of an inch increments, each rotation is 1/40th of an inch with 25 divisions for 1/1000th With microstepping I should be-able to push that pretty far for precise alignment which I will use for the alignment of multiple exposures for getting more than 1024x768 for a single die... 11-Feb-20 07:51 AM but that x-y stage can only move 1 inch in either direction total... So I will be-able to do up to a 1 inch die size max... Then I want to place that stage on a more course less precise/tight stage. For stepping to the next "die" on the wafer. 11-Feb-20 07:51 AM I feel this avoids the cost of a super expensive massive precise stage, with the trade-off of more complexity and a cheaper secondary x-y. 11-Feb-20 07:51 AM I think it will be a good budget way of doing precise entire wafer stepping, if each die isn't precise with one another so long as your dicing saw can fit between them it shouldn't be a big deal if they are shifted a little bit amungst one another. 11-Feb-20 08:01 AM @Conmega Isn't ebeam photoresist is simple PMMA? Which is everywhere? 11-Feb-20 08:01 AM Yea... but I have seen people try to make it at a level that it works well for consistency and well yea... 11-Feb-20 08:01 AM its never seems to pan out very well 11-Feb-20 08:01 AM its also that simple to do the physics but not that simple to make a thin even coating that sticks well to a wafer... 11-Feb-20 08:01 AM As far as I know commercial stuff has a lot of fancyness to it that makes it good and they also generally recommend other chemicals they sell to make the wafer more conducive to taking that resist layer... 11-Feb-20 08:01 AM But either way, if what I found works out decently, its like... really cheap and available and doesn't require super nasty things to develop and remove. And it seems pretty good for standing up to other stuff too. Like HF acid. 11-Feb-20 08:37 AM Pretty sure commercial pmma resist is just solvent and pmma, of a certain chain length range 11-Feb-20 08:37 AM And 10-15 microns is quite thick for a photoresist layer! 11-Feb-20 08:37 AM For some recent e-beam/FIB litho I was recommended to use a 50nm thick resist! 11-Feb-20 08:37 AM Get yourself a spincoater too, they come up on eBay now and then for cheap... SCS (specialty coating systems) 11-Feb-20 08:37 AM Or just use a hobby motor and a cdrom spindle case 11-Feb-20 08:37 AM Cheap spincoater build here https://www.joshuataillon.com/pdfs/2017_jtaillon_MandM.pdf 11-Feb-20 08:40 AM yea I'll probably be making one... Most spin coaters seem to go for hundreds even used... :/ 11-Feb-20 08:40 AM Using a speed controlled fan hub seems like a great choice that many have done. 11-Feb-20 08:41 AM A convenient, and very cost-effective solution, built from simple off-the-shelf components, is presented in this work. This “rotary coater,” as shown in Figure 2, is quickly and easily assembled from a benchtop vise, a rotary tool, and a plastic shield. With just these three pieces, samples can be coated in a matter of seconds, just before FIB/SEM observation. Assembly of the rotary coater is simple. On a stable lab bench, a rotary tool (preferably with variable speed control) is clamped into a fixed vise such that its axis of rotation is normal to the bench. A plastic shield can be easily fashioned from a recycled CD or DVD container with a removable cover, that has had a hole drilled in its base to allow attach- ment to the rotary tool. With the collar nut of the rotary tool removed, the base of the shield is slipped over the shaft of the rotary tool, and the collar nut reattached to affix the shield base to the tool. Use of a level to verify the angle of the assembly is helpful, but not strictly necessary. For further convenience, a foot pedal switch may be added to apply power to the rotary tool. 11-Feb-20 08:41 AM Yeah I paid $272 shipped for my SCS P6708D 11-Feb-20 08:43 AM yesh yea I can make a nice speed controlled thing with a readout for less than that hehe 11-Feb-20 08:43 AM The seller said "doesn't power up" but that's because there's a low pressure purge cutout... Just have to provide air, or wire around the pressure switch 11-Feb-20 08:43 AM I just didn't want to deal with making a vacuum chuck 11-Feb-20 08:44 AM heh 11-Feb-20 08:44 AM Mostly 15-Feb-20 11:19 AM any known sources for SU-8 for smaller bottles than the 500ml from Fisher? 15-Feb-20 11:19 AM also.. has anyone here used it before? 15-Feb-20 02:22 PM I’ve used SU-8 before 15-Feb-20 02:22 PM It was a lot thicker than positive resists, and a little more of a hassle to get going 15-Feb-20 02:22 PM Had mixed results using it to make 5um features, but I know people who use it all the time and have no problems (with good aspect ratio and crit dim size) 15-Feb-20 02:22 PM As far as suppliers no clue for it besides fisher at least for sizes. I got my positive resist as a liter and that was smallest I could get from microposit. 15-Feb-20 02:28 PM @Noxz how much do you want? I might have some really old stuff might be able to send... I can check when I head to the lab I part-time at 15-Feb-20 03:42 PM actual quantity unknown, enough to do some preliminary tests? no more than 100ml? 15-Feb-20 04:30 PM is this all "expired" stuff that is not used because you have fresher? 15-Feb-20 04:44 PM I think it's expired, but not bectwe have fresher, just because the employee doing projects with photoresist left the company. I was able to successfully use some 10 year old PMMA, but that is basically just plastic solution, no active chemistry going to happen as far as I know during storage 15-Feb-20 05:06 PM I was thinking PMMA would be a relatively simple route.. but I am unaware of how well it works with easily obtainable UV sources, as well as aspect ratios.. I understood PMMA was better with xray 15-Feb-20 05:29 PM pmma is good for ebeam 15-Feb-20 05:29 PM but it will work with short uv 15-Feb-20 05:56 PM Yeah it works via chain scission, so you need a good bit of focused energy to break the bonds, since there's no catalyst... If a laser cutter can cut it, then that wavelength should work assuming you can focus small enough. Maybe a bluray drive optical sled would do it 15-Feb-20 05:56 PM I know my 10 micron co2 laser will cut acrylic 15-Feb-20 05:56 PM Have a bluray sled driver board (diyoupcb) but never hooked it up to try 16-Feb-20 10:16 AM also note, I still need to setup t he UV exposure rig.. will presumably be a DLP + Microscope 16-Feb-20 10:16 AM initial tests will be tiny roman numerals (via electroless nickel plating of the developed PR) 16-Feb-20 10:18 AM so as I found out the microscope isn't necessary for very basic like... 20um per pixel feature size... 16-Feb-20 10:18 AM you can just remove the outer most optic, generally the focus optic, by removing a screw and unthreading the optic ussually 16-Feb-20 10:18 AM then it focuses to about a half inch square about 2 inches from the opening 16-Feb-20 10:18 AM pretty decent for getting started and testing 16-Feb-20 10:18 AM then you don't have to worry about aligning a bunch of optics... 16-Feb-20 10:19 AM I want the end result (say a small square with a roman numeral "cut out" to be on the order of, oh, 2.00mm square? 20um per pixel would equate the res of that to 100pixels? which is likely plenty 16-Feb-20 10:19 AM yup so just do what I did, look in #lasers_and_optics 16-Feb-20 10:20 AM oh, nice 16-Feb-20 10:20 AM Just removed the bulb and ballast, had to remove the controller board from the ballast to fake out the mainboard to think the ballast is still there... moved the color wheel but left it connected 16-Feb-20 10:20 AM d'you replace the lightsource already? 16-Feb-20 10:20 AM and remove the outer optic 16-Feb-20 10:20 AM I don't have the UV led but I was using a bench top lightsource for testing yes 16-Feb-20 10:21 AM ah, nice 16-Feb-20 10:21 AM so "replaced" yes... but no lol 16-Feb-20 10:21 AM still waiting on my UV LED and driver from the slow boat :/ 16-Feb-20 10:21 AM would you reccomend that brand/model of projector? 16-Feb-20 10:21 AM but in the mean time I can build up a frame for everything 16-Feb-20 10:21 AM I mean I just happened to have a friend selling a bunch for like 20 bucks but I mean, I guess? Its more complex than needed though like it has ethernet and ton extra 16-Feb-20 10:21 AM I'd personally look for something simpler and smaller if I could 16-Feb-20 10:21 AM its a bit unwieldy so its going to be a pain to mount it up 16-Feb-20 10:21 AM I'd just try to keep above 1024x768 which is the common like... cheap size below 1080p at this point 16-Feb-20 10:23 AM sub $100 for a 1080p projector 16-Feb-20 10:23 AM this isn't 1080 this is 1024x768 16-Feb-20 10:23 AM which is 4:3 16-Feb-20 10:23 AM which honestly I think is going to be easier to deal with 16-Feb-20 10:23 AM I don't think 16:9 will be fun to deal with since basically all of your dies will end up rectangular to a degree unless you do some weird patterning 16-Feb-20 10:23 AM and 1024x768 projectors are cheap and available 16-Feb-20 10:24 AM what about micro projectors designed for phones and such? smaller optics? 16-Feb-20 10:24 AM so better hacking 16-Feb-20 10:24 AM soooo Sam Zeloof experimented with those but moved away from them 16-Feb-20 10:24 AM I dunno why but maybe he has some insight. 16-Feb-20 10:24 AM nods 16-Feb-20 10:25 AM so maybe shoot him an email 16-Feb-20 10:25 AM but yea honestly I'd expect you to pay 40 bucks shipped for a 1024x768 projector with a working but maybe not long lasting bulb off ePay 16-Feb-20 10:25 AM just make sure it doesn't have dead pixels and such which will be the worst thing to deal with 16-Feb-20 10:25 AM but also hardest to check for honestly 16-Feb-20 10:25 AM but yea, again, not super expensive, I'll probably end up with more projectors to hack on at some point as I see them for cheap at hamfests or swap meets and such 16-Feb-20 10:25 AM but ultimately for home litho I think the smaller projectors and micro DMDs will be king for just the compactness and yea lack of need to mount this big bulky stupid thing ontop of some other big bulky stupid thing 16-Feb-20 10:25 AM I am just curious as to why Sam moved away from it... maybe resolution, maybe just did on a whim 16-Feb-20 11:14 AM went afk.. someone was at the door I had to chat with 16-Feb-20 11:41 AM np :P 16-Feb-20 11:41 AM can continue conversations hours, days, weeks later, not like the internet is going anywhere, hopefully :) 16-Feb-20 11:46 AM yeah, I get you.. good to know that a microscope isn't really needed, for larger features 16-Feb-20 11:46 AM I've been on the fence for a while w/re to what microscope I want (for general use) so removing that from the equation is nice (even if it were to be dedicated) 16-Feb-20 11:46 AM we got a pretty nice new Zeiss at school, a bit pricey though.. I love the brand for various reasons, so sort of looking at them 16-Feb-20 12:01 PM A personal microscope is a different thing but for optics for litho I think you don't need a WHOLE microscope just some of the optics 16-Feb-20 12:01 PM so I think you can find bits and pieces 16-Feb-20 12:02 PM I was thinking that as well 16-Feb-20 12:02 PM but honestly a cheap chinese microscope is great for home stuff 16-Feb-20 12:02 PM if you want anything better than that you want a nice metallurgy microscope 16-Feb-20 12:02 PM like an old Leica Ergolux or something (what I have but I got for free so...) 16-Feb-20 12:03 PM we have a few microscopes at shcool that I have become accustomed to 16-Feb-20 12:03 PM always keep an eye out for used stuff but yea just for like 16-Feb-20 12:03 PM a general bench scope for basic magnification and like PCB work and vacuum part handling and such 16-Feb-20 12:03 PM getting a binocular cheap from china for like 200-300 bucks is great 16-Feb-20 12:03 PM I have one and it works well 16-Feb-20 12:04 PM yeah, for me it would inspection of watch parts 16-Feb-20 12:04 PM oh yea I thought you meant for silicon fab stuff 16-Feb-20 12:04 PM likely for measuring production 16-Feb-20 12:04 PM well, I want to do that as well 16-Feb-20 12:04 PM get a chinese one 16-Feb-20 12:04 PM but how many rigs do I need, or rather, can I get away with one microscope (no) 16-Feb-20 12:04 PM they are decent and cheap but if you can find used stuff cheaper go for it 16-Feb-20 12:04 PM any of the omax or similar ones then? 16-Feb-20 12:05 PM I am not sure what brands, I just followed the link to the company that StrangeParts on youtube uses 16-Feb-20 12:05 PM I used one in person at HOPE and it was nice 16-Feb-20 12:05 PM I figure anything new is going to be decent 16-Feb-20 12:06 PM yea 16-Feb-20 12:07 PM oh wow.. I think a linear scale I have been hunting for may have become available for cheap-ish 16-Feb-20 12:07 PM goes back to ebaying 16-Feb-20 12:08 PM hehe 16-Feb-20 01:49 PM @AdamMcCombs had a bunch of good inspection microscopes for sale 3-4 months ago... Like a whole pickup truck full 16-Feb-20 01:49 PM Olympus i think, i got one 17-Feb-20 12:38 AM Is Adam the microscope guy at the Ham Swap? 17-Feb-20 11:03 AM No, he's the creator of this chat group, a Seattle area resident, which is where noxz is too 17-Feb-20 03:13 PM Okay, cool. 17-Feb-20 03:21 PM hah, countdown clocks says I have ~6mo left here 17-Feb-20 03:21 PM or rather, until school ends and I can finish up fixing the house and sell 18-Feb-20 08:43 AM So wondering if anyone digging into silicon fab have found a good way to get an n-dopant? I think p-dopants are fairly simple. But I am having a hard time finding a good source for phosphoro-silica for n-doping. Any ideas? I would like to try some nmos first since I already have some p-doped wafers but if need be I will find some n-doped wafers and go for pmos since its supposedly simpler? 18-Feb-20 08:43 AM I suppose one way would be to make it using some off the shelf source of phosphorous. I am not sure how crucial it is to have a perfect source or what not. I would assume impurities would cause issues. 18-Feb-20 08:43 AM er maybe I am mixing things up here... I need to go back and read more stuff heh 18-Feb-20 08:53 AM @Conmega Back in the days I planned to use POCl3 - but it is nasty, even got some. Now I would really prefer to use simple phosphoric acid if possible. Do you think it is possible? 18-Feb-20 08:54 AM I am actually looking into that now... It looks like using Phosphoric Acid is possible? It seems to have had some academic use for solar panel stuff? 18-Feb-20 08:54 AM Suppose its at-least worth a try 18-Feb-20 08:54 AM https://www.tandfonline.com/doi/abs/10.1179/026708408X370203?journalCode=ysue20 18-Feb-20 08:54 AM The thing is finding research into using it as standard silicon chip fabrication will be nil due to that stuff being so high in the sky with yaknow like 7nm and stuff... 18-Feb-20 08:54 AM but it seems they have promise in the solar industry so I will probably grab some and try to use it 18-Feb-20 08:54 AM only problem is it seems you need to cut it with something since water cut stuff will be hydrophobic to bare silicon 18-Feb-20 09:04 AM @Conmega I've seen references that you add acetic acid to improve vetting. I am using acetic acid to HF to improve vetting to etch metalization on microchips. Anyway, doping will likely go in gaseous phase, as temperature will be much higher than boiling point of phosphoric acid. 18-Feb-20 09:05 AM right, I am not sure how that breaks down, if it just all evaporates off or if it leaves behind a layer that can do the doping... or if you just have to keep the entire enviroment around the wafer with that gas so that it can dope the wafer... 18-Feb-20 09:06 AM @Conmega The answer will depend on target doping profile. limited vs unlimited source... 18-Feb-20 09:08 AM hrm, yea I wonder what works out best for this, probably a limited source since its probably easier to control 18-Feb-20 09:08 AM and repeat 18-Feb-20 09:08 AM hrm I wonder if you can use a similar method to POCl3 18-Feb-20 09:08 AM https://books.google.com/books?id=REQkwBF4cVoC&pg=PT314&lpg=PT314&dq=doping+profile+limited+vs+unlimited&source=bl&ots=JO3YIooVOV&sig=ACfU3U20HXkTgZxmMg_A2fQRT1BAKIv2fQ&hl=en&sa=X&ved=2ahUKEwjfuvGSy9vnAhVEba0KHaz9AE0Q6AEwAHoECAoQAQ#v=onepage&q=doping%20profile%20limited%20vs%20unlimited&f=false 18-Feb-20 09:08 AM Ah I suppose that nitrogen bubbling gets the gas there but then the reaction from there is dependant on the fact that its POCl3 18-Feb-20 09:08 AM I wish I didn't flunk chemistry in highschool ‍♂️ I really need to brush up on chemistry 21-Feb-20 02:02 AM I did buy phosphoric acid, but my wafers where N-doped so I had to experiment with boric acid instead. 21-Feb-20 08:45 AM @Nixie And, did it worked? 21-Feb-20 09:49 AM Yes, I mean, to an extent because I just put it directly on top of the silicon, but the diode-ish behaviour you can see on my profile pic is from that process 21-Feb-20 09:50 AM hrmmm cool ok 21-Feb-20 09:50 AM Afterwards I tried on passive diffusion (putting the SiO2 window near the boric acid source) and it also worked, but I was not controlling time, si it was not as good. 21-Feb-20 09:50 AM I'll have to get both types so I can do nmos at first since I have p type silicon atm then I can try cmos stuff later 21-Feb-20 09:50 AM Thanks for the info Nixie :) 21-Feb-20 09:51 AM For my next experiment, I have boron dust and I'll sit the silicon upside down on top of it, a few millimeters apart, that should work wonderfully) 21-Feb-20 09:51 AM I didn't say much @Conmega 21-Feb-20 09:55 AM heh well you confirmed it works at-least partially 21-Feb-20 09:55 AM I mean personally I was wondering if baking with a phosphoric acid will just make it evaporate away and just not really do anything else or if it would actually drive phosphor into the wafer 21-Feb-20 09:57 AM It does work, it's a proper technique, to have the source some way apart from the wafers and allow diffusion through recirculating gas (nitrogen with a bit of oxigen, I'll try argon) 21-Feb-20 09:57 AM There's the expensive way, wich is glass seal the source and wafers and bake that (then you have to break the glass seal, XD) or the one with recirculating gas 21-Feb-20 11:33 AM I mean you can seal the glass tube furnace too right? 21-Feb-20 11:53 AM Yes, but the diagram I remember fully sealed the glass with itself, instead of capping the ends to have the whole thing at temp. 21-Feb-20 11:53 AM (too many books now to easily find the schematic, sorry) 25-Mar-20 09:07 AM https://www.youtube.com/watch?v=YAPt_DcWAvw 13-Apr-20 08:29 AM slightly off topic, but wrote a tool for decoding/parsing mask roms/PLAs with non-independent cells: https://home.strw.leidenuniv.nl/~pbosch/pladec_x4l_and.png 13-Apr-20 08:29 AM haven't quite finished it but it's on github: https://github.com/peterbjornx/pladecode 13-Apr-20 09:19 AM That's pretty cool. 13-Apr-20 09:19 AM I wish I had a trinocular head for my leitz ergolux metallurgical scope so I could do die imaging. I got the thing for free and buying a 500 dollar trinocular head off ebay just feels... wrong... 13-Apr-20 09:19 AM oh actually there may be one around the 250~ mark... with a make offer... that may be more worth it hrm... 13-Apr-20 05:04 PM @Conmega You don't absolutely need trinocular head. You can insert camera in one of eyepieces. I don't use eyepieces at all. 13-Apr-20 05:08 PM to be honest, the imagery i am working off at the moment was made for me by someone else. i'd love to get my own (especially now i need to look at some of the logic), but daren't work with HF yet 13-Apr-20 05:09 PM Yea but its a bit more cludgy, especially for using a big camera like a mirrorless. 13-Apr-20 05:09 PM was thinking of trying hot concentrated sulphuric or phosphoric to get rid of the passivation layer and then tackling it mechanically 13-Apr-20 05:09 PM I mean I didn't ask for the ergolux... a friend was like, so yea this fab is throwing stuff out, want to dig through the pile with me? sure! and yea it was sitting there... 13-Apr-20 05:09 PM lol 13-Apr-20 05:10 PM on an older Intel chip I managed to get down to polysilicon using qtips and brass polish, but didn't seem to work on this one 13-Apr-20 05:10 PM don't know about the older chip but this one has .7 microns of nitride as passivation 13-Apr-20 05:10 PM my aluminum etch did seem to work quite nicely though: 13-Apr-20 05:10 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/IMG_8853-328CA.JPG 13-Apr-20 05:10 PM used a nitric+phosphoric+acetic acid mixture for that 13-Apr-20 05:18 PM was thinking of trying hot concentrated sulphuric or phosphoric to get rid of the passivation layer and then tacking it mechanically @pbx/peterbjornx Never seen boiling sulfuric getting through passivation, until it's very late and causes lots of damage. HF etching is not very hard. That's the only chemical processing step that I have automated. Epoxy etching is harder to automate, and less predictable. Surely HF is scary, you can get full body suite, mask, gloves and work with small volumes. I use HF solution with HCl and Acetic acid - it make smell much more significant, and harder to miss if something goes wrong. 13-Apr-20 05:19 PM @ZeptoBars i only (when not in lockdown) have access to the DIY fume hood at the hackerspace, which I wouldn't trust with HF. 13-Apr-20 05:19 PM but, if the passivation is nitride, it should be etched by hot phosphoric or sulphuric, right? 13-Apr-20 05:21 PM I was almost sure passivation is typically SiO2. Inter-layer dielectric is definitely SiO2, so without HF there won't be much luck to get to the bottom. 13-Apr-20 05:22 PM the chip i'm working on has a report on it from smithsonian and it says the passivation should be nitride under polyimide 13-Apr-20 05:25 PM Surely HF is not for public work. Small hermetic glovebox with baking-soda vapor neutralization will make it somewhat safe for home use. I might get something like this soon: https://www.alibaba.com/product-detail/Acrylic-Vacuum-Glove-box-for-hard_60678100724.html?spm=a2700.galleryofferlist.0.0.7fba636fcx5fot Or one could DIY 13-Apr-20 09:26 PM tried a sort of brute force decap one of the fake max3232 chips i got on some eBay boards and (aside from the chipped die) things actually turned out pretty well for a first attempt 13-Apr-20 09:26 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/IMG_20200413_204942-3DF2D.jpg 14-Apr-20 01:48 AM @ZeptoBars my fume hood is open, but I got a 300W fan that absolutely sucks the air from all the chamber/whole area inlet. 14-Apr-20 01:48 AM https://twitter.com/nixie_guy/status/1026528285382066176 15-Apr-20 09:32 AM Soo, hot sulphuric did get rid of the polyimide layer and brittled the nitride passivation to the point where polishing could remove it. I polished down to metal 1 and then applied a metal etch to remove metal 1, which seemed to work nicely. 16-Apr-20 04:49 AM photos, photos! 17-Apr-20 12:15 PM Does anyone here have any familiarity with perovskite crystal semiconductors? I've been reading about them and they simultaneously seem very interesting for radiation detection and readily accessible on a small scale. 17-Apr-20 12:24 PM did a synthesis/characterization of CsPbBr3 nanocrystals for PVs once but thats it 17-Apr-20 12:40 PM I’d been reading about CsPbBr3 and MAPbI3 but as larger crystals 17-Apr-20 12:40 PM It’s seeming like a really amazing balance of capability and low cost 18-Apr-20 12:05 PM trying to use KOH to etch sio2 now 18-Apr-20 12:05 PM just put a drop on the die, so it should not touch the exposed si 24-Apr-20 05:22 AM How did it end? 24-Apr-20 12:42 PM Cool, yea supposedly it etches fine, just slow and in certain crystal directions... 24-Apr-20 12:42 PM so its a bit weird to work with and supposedly not great for standard chip fab stuff with the 45 degree angles it makes.... but if you design your process around it I am sure you could get it to WORK 24-Apr-20 12:42 PM just might not be too fun 24-Apr-20 12:42 PM but for simpler stuff with the right crystal orientation should be pretty easy to work with. 24-Apr-20 10:33 PM https://onlinelibrary.wiley.com/doi/abs/10.1002/anie.202002861 24-Apr-20 10:33 PM Selective Enzymatic Oxidation of Silanes to Silanols 25-Apr-20 04:08 AM That...sounds like some posh swearing. 25-Apr-20 04:08 AM XDDD 29-Apr-20 04:38 PM hey all, any suggestions on documents on how to properly and safely keep chemicals at home? specifically what chemicals can be kept together vs not together, types of containers they should be in, etc? 29-Apr-20 04:53 PM Just based on a quick skim but these seem like pretty decent guides. Page 2 has a nice chart of incompatibilities: https://ehs.oregonstate.edu/sites/ehs.oregonstate.edu/files/pdf/si/chemical_storage_guidelines_si.pdf This is more in depth: https://www.ehs.berkeley.edu/sites/default/files/lines-of-services/hazardous-materials/chemicalstoragebooklet.pdf 29-Apr-20 07:15 PM Those are good choices. @samy, happy to help with advice if you tell me what you've got. 29-Apr-20 08:27 PM thanks! 29-Apr-20 08:27 PM Hydrochloric Acid 32%, Nitric Acid 65%, Methyl Ethyl Ketone (MEK) (Butanone), Sulfuric Acid < 51% (UN2796), Sodium Hydroxide (Lye), Tetrachloroethylene, Ferric Chloride, Ethyl Acetate, Potassium Permanganate, Vinegar, Acetone 99%, Isopropyl Alcohol 99%, Isopropyl Alcohol 70%, Ethyl Alcohol (Ethanol), Ammonia, Borax, Baking Soda, Washing Soda 29-Apr-20 10:14 PM Looking at that list, you would like an acid cabinet and a flam cabinet, both with a vent line to the outside world, preferably not right next to each other. The flam cabinet you want to ground. The HCl, HNO3, and H2SO4 all go in the acid cabinet. MEK, TCE, ethyl acetate, acetone. IPA, and ethanol go in the flam cabinet. Lye, KMnO4, Borax, Baking Soda and Washing Soda are solids that should go on a shelf on the other side of the room, just in case of earthquake whoops. Vinegar and Ammonia can stay in your kitchen. All the liquids go in buckets that can contain 110% of the liquid in their original containers in the event of losing the bottle. 29-Apr-20 10:15 PM Is the HCL lab grade/is it in a good bottle? 29-Apr-20 10:15 PM Poor quality bottle and you'll get off gassing and rust everything nearby. 29-Apr-20 10:18 PM On a positive note, acid and flam cabinets can be come by pretty cheaply. Especially university surplus sales. 29-Apr-20 10:18 PM small cabinets are the hard part 29-Apr-20 11:11 PM https://twitter.com/nixie_guy/status/1255741130223882241 29-Apr-20 11:22 PM thank you @funranium! 30-Apr-20 12:00 AM Slightly getting out of hand: 30-Apr-20 12:00 AM https://twitter.com/nixie_guy/status/1255752503158812673 30-Apr-20 01:30 PM I've never seen any academic or professional labs using buckets in their cabinets 30-Apr-20 01:30 PM I can't remember if I ever saw any insides of cabinets when I interned at the national lab 30-Apr-20 01:33 PM I had our cabinet moved far away from my stuff 30-Apr-20 01:33 PM because, for example, no buckets. 30-Apr-20 04:51 PM https://ehs.msu.edu/news/2015-07-10-fridge-explosion.html 30-Apr-20 04:51 PM https://www.ebay.com/sch/i.html?_from=R40&_trksid=m570.l1313&_nkw=explosion+proof+refrigerator&_sacat=0 pretty affordable actually 30-Apr-20 06:43 PM Wow, ethanol is on the list of no nos 30-Apr-20 06:57 PM Well, yeah...? 30-Apr-20 06:57 PM Funranium would not enjoy a safety tour of my university but even we store ethanol in self-sealing metal containers in a flam cabinet 30-Apr-20 06:57 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/stainless-steel-lab_can-7D744.png 30-Apr-20 09:55 PM [shrug] I have no off-switch but I also do my best to keep in mind who I am paid to be responsible for. At other people's institutions, I tend to only speak up at Immediately Dangerous to Life & Health situations unless I have been specifically invited to tear someone a new one. 30-Apr-20 10:04 PM As far as never seeing secondary containment @nmz787, I'm gonna guess you aren't in earthquake country. Regulators get rather excited about spills and get quite insistent about the 110% containment. They often crank it higher than that for petrochemical industry. They do have a bit of history and all. 30-Apr-20 10:22 PM @funranium went to school in upstate NY, internned at Berkeley Natl Lab , living in Oregon for almost 8 years and seen a few labs around here. One I part time at... Has lots of chems in (I think) non vented (to outside) cabinets... glass gallon of DCM jug just sits in fume hood 30-Apr-20 10:22 PM Def earthquake country up here 30-Apr-20 10:22 PM I haven't been in any of the Intel labs long enough to look at chem cabinets 30-Apr-20 10:50 PM [sigh] Fume hood storage. We try to train people, we really do. But this is why I do radiation and laser safety, so I can have a totally different set of "Really? SERIOUSLY?!" basic safety problems to deal with. 01-May-20 07:10 AM @funranium would defiantly stand at a good distance from my garage whilst having a conversation! Hahahaha. Actually it’s not that bad. Just too much in it. I am still trying to build my new place but SARS CoV 2 had other plans! 01-May-20 03:14 PM Likely. 01-May-20 04:51 PM Defiantly or definitely? 01-May-20 04:51 PM The implications are much different 01-May-20 06:06 PM Hahahaha... I stand by my statement. 01-May-20 06:06 PM I think that was an autocorrect choice. 01-May-20 06:06 PM Actually the most dangerous thing in my garage is the Datsun roadster. And maybe the barium carbonate. 28-May-20 01:25 PM https://old.reddit.com/r/ElectronicsList/comments/grngvm/must_go_warehouse_of_abandoned_semiconductor/ 28-May-20 01:25 PM Not sure if any of that is of interest to anyone but I just saw that being talked about... 31-May-20 01:19 AM I saw that pop up on Twitter. I think that has been swarmed over already. 01-Jun-20 05:26 AM Yeah, makes me terribly sad and envious. XD 17-Jun-20 08:07 AM Finally working on the XY stage, the motors took more than two months to get here: 17-Jun-20 08:07 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/Eat60v1WsAM-AHC-23669.png 17-Jun-20 08:36 AM @Nixie is this for a pick and place machine? :o 17-Jun-20 11:11 AM Noo 17-Jun-20 11:11 AM it's an XY stage to move a chip under the microscope and see what is there. 17-Jun-20 11:15 AM Ah yes, it's a chip 21-Jun-20 01:50 PM @Nixie If you want to go even smaller, there are these things: 21-Jun-20 01:50 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-2B801.png 21-Jun-20 01:50 PM Two phase stepper motor with planetary gearbox 21-Jun-20 01:50 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-BB8E0.png 21-Jun-20 01:50 PM They're only $2 + shipping from China, pretty incredible that these things can be made this cheap 21-Jun-20 01:50 PM 8.5 mm travel 21-Jun-20 01:50 PM They are for moving a zoom lens in compact cameras 21-Jun-20 03:46 PM Wow. My eyes hurt just looking at those! 21-Jun-20 04:36 PM WOW! those look gorgeous O_O 21-Jun-20 04:36 PM Definitely will buy some later. 21-Jun-20 04:36 PM I will wait for the M2 fine pitch tap to arrive first 21-Jun-20 04:36 PM With the motors I showed, I have a step of 0.0125mm (full step), I am not sure it will work well at half step (theoretical 0.00625mm) 21-Jun-20 04:36 PM with no backlash in the motor (that small unit looks like it has a reduction) 21-Jun-20 06:42 PM There are surprisingly a lot of similar units on Aliexpress: https://www.aliexpress.com/item/4000884241895.html 21-Jun-20 06:42 PM Some have extremely fine threads 21-Jun-20 06:42 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-D56DF.png 21-Jun-20 08:23 PM I think I finally got some of something like those after waiting 3 or 4 months 21-Jun-20 08:23 PM Oh, no I had some of those since like 7 years ago, recently I got some of the XY gantry gizmos that seemed like for imaging chip stabilization or something 24-Jun-20 12:32 AM building uv litho steppers out of old blue ray drives? 24-Jun-20 12:32 AM jk 26-Jun-20 04:22 PM I've started making some thin films and I'm looking for some relatively simple DIY methods for measuring film thickness. Any ideas? I'm not opposed to buying some old eBay equipment, but I was wondering if anyone has had luck with homebrew methods 27-Jun-20 11:33 AM @ericdalgetty i haven't tried, have you checked Sam Zeloof's vid? https://www.youtube.com/watch?v=vOijNeDcX-0 27-Jun-20 12:21 PM Thanks, I've been trying out that method, but with a home-made webcam spectrometer. I haven't had any luck actually seeing the interference fringes 27-Jun-20 01:32 PM ahh 27-Jun-20 02:10 PM @ericdalgetty try a dSLR 27-Jun-20 02:10 PM Or if you want to DIY a bit more, pick up a CCD array and a grating 27-Jun-20 02:10 PM TCD1304AP is one I've used before 27-Jun-20 02:10 PM I have code online (for parallax propeller) to talk to it 27-Jun-20 02:10 PM Would probably use timer peripherals today if was doing it again, or fpga 30-Jun-20 11:27 PM I know it's not much, as she follows 2.6k people, but this made my day: 30-Jun-20 11:27 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/f7bcea85-f48d-4cac-bf53-5cb13ce3df87-ED45A.png 02-Jul-20 04:12 AM Nice 02-Jul-20 04:12 AM So I was looking around for encoders to add to my XY stage but couldn't really find any, is there a search term I should be using? 02-Jul-20 04:12 AM I need someting lightweight/small with about 1 or 2cm travel and 1-5um resolution 03-Jul-20 04:51 AM Someone might be interested in this: 03-Jul-20 04:51 AM https://fossi-foundation.org/2020/06/30/skywater-pdk 03-Jul-20 07:48 AM That is awesome, I have been working with the open-source EDA tools as of late 03-Jul-20 07:48 AM so its great to see the PDK side being built up too 03-Jul-20 07:48 AM though man that video is tough to watch... all the UMMMs 03-Jul-20 07:48 AM 4 times speed through this video 03-Jul-20 07:48 AM The shuttle program is pretty cool 03-Jul-20 07:48 AM I may need to see if I can squeak into that.... I doubt I have the current ability or skill to. And I am sure a ton more people are setup and ready. But it would be awesome to have my own silicon that isn't like 2 transistors. 10-Jul-20 11:19 AM Anyone have any experience with "Intellemetrics" QCM thickness monitors? I just bought one off eBay but can't find a manual for it. 14-Jul-20 10:22 AM Nope ^^U 28-Jul-20 09:39 PM This is pretty amusing https://www.youtube.com/watch?v=ihkRwArnc1k&feature=youtu.be&t=1156 28-Jul-20 09:39 PM Proper old school way 29-Jul-20 01:02 AM Man, I can't say how often I watched this documentary 29-Jul-20 01:02 AM so well done 01-Aug-20 12:58 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-1DA70.png 01-Aug-20 12:58 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-C3FA7.png 01-Aug-20 12:58 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-27791.png 01-Aug-20 12:58 AM on chip bodgewires! 01-Aug-20 12:58 AM from: http://www.kip.uni-heidelberg.de/lhcb/Publications/ASIC_Mar2004.pdf 01-Aug-20 07:14 AM Amazing wizardry! But how did they allign the ASIC inside the FIB? That surgery must require really accurate positioning of the specimen, right? 01-Aug-20 07:43 AM Well I mean if it’s a good FIB then it’s probably a cross beam, so they position it with the SEM and then mill or deposit 01-Aug-20 07:46 AM I wonder how the hole down to the die was made, is that what coarse fib machining looks like? 01-Aug-20 07:50 AM Looks like they made a ton of crude overlapping passes 01-Aug-20 07:50 AM Coarse Fib machining will leave you with square shaped hole because your fib is scanning over the whole scan window 01-Aug-20 07:50 AM Fib milling, even at high beam currents is fairly accurate though, you just get some chamfer on the edges 01-Aug-20 07:50 AM This is my favorite thing I ever milled, a sub micron version of my uni logo 01-Aug-20 07:50 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/image0-7634A.jpg 01-Aug-20 07:54 AM neat! 01-Aug-20 09:14 AM Well in the last picture it seems that the semiconductor structures were buried prior to milling. the area around the milled spot looks very homogennous to me and the milled area is hardly larger than the modified area. I think there must have been some kind of fiducial or something that was read to get accurate positioning, but if that fiducial isn't close to the etched area, accurate mechanical movement must have been necessary afterwards. Now I wonder, would it also be possible to change the masks which were used in the production of that ASIC with a FIB to get new corrected chips made in a fab? 01-Aug-20 12:40 PM maybe, depending on what size features are 01-Aug-20 12:40 PM Cause some masks are phaseshifted to get smaller features 01-Aug-20 12:40 PM But I don’t think ASICS would be produced with phase shifted masks so idk 01-Aug-20 12:54 PM I've interviewed for a FIB circuit edit (rework) position before, basically you have the CAD for the silicon to know how to approach the edit 01-Aug-20 12:54 PM And if you don't then you'd better have some sacrificial chips on hand to get a feel, first 01-Aug-20 12:54 PM But you can definitely do that with a single-beam device 01-Aug-20 12:54 PM I've got a platinum GIS now that I want to get bolted onto my FIB and try out some trace creation 01-Aug-20 12:54 PM As-is I've only got carbon deposition, which is useful to protect areas from being milled as you scan around imaging 01-Aug-20 12:54 PM @analogmultiplizierer my stage is very very accurate and precise 01-Aug-20 05:45 PM GIS is so sweet lol, I’ve only used carbon too. Our Pt source was out. Crazy seeing the FIB working in an additive way though 06-Aug-20 10:17 PM did you guys try ebid? 07-Aug-20 03:06 AM Not I 07-Aug-20 09:40 PM @nmz787 for the additive FIB, is there an actual CAD / transistor layout format that's used? 07-Aug-20 09:41 PM Unfortunately for my FIB it just takes BMP files and is buggy :( 07-Aug-20 09:41 PM It's got analog inputs for the beam scan, which I hope to some day interface a DAC to 07-Aug-20 09:46 PM ahh 11-Aug-20 12:41 AM ok want to try some EBID some day (if I get all the precs) 12-Aug-20 06:41 PM has anyone attempted a build similar to Sam Zeloof's photolithography build? 13-Aug-20 01:10 PM I started to but abandoned it because attention deficit and not enough resolution for my needs at the end of the day 13-Aug-20 01:10 PM Also, those sort of systems are more and more common in industry contractors/services 13-Aug-20 01:10 PM Got some stuff like that done in the past year from 3 diff companies for free 13-Aug-20 01:13 PM I want to try some ufluidics DEP stuff where I don't need ridiculously low feature size 13-Aug-20 01:13 PM Hehe, yeah all my goals are moonshot sort of stuff, systems integration of macro/micro/nano all at once 13-Aug-20 01:14 PM just need maybe some 10 um electrodes and some 100s of um wide well 13-Aug-20 01:14 PM I guess I've been focusing on the nano and figuring the larger stuff will be easier 13-Aug-20 01:14 PM I'd agree with that lol 13-Aug-20 01:14 PM If you want a few services, I can recommend 13-Aug-20 01:14 PM New Light Industries in Spokane WA 13-Aug-20 01:14 PM excelsioroptronics.com 13-Aug-20 01:14 PM Heidelberg Instruments 13-Aug-20 01:14 PM All 3 gave freebies 13-Aug-20 01:15 PM oh nice 13-Aug-20 01:16 PM The first is probably the most flexible to talk with 13-Aug-20 01:18 PM yeah I'll look into that 15-Aug-20 10:10 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/image0-DEF36.png 31-Aug-20 10:42 PM https://hackaday.com/2020/08/31/an-analog-ic-design-book-draft/ 31-Aug-20 10:42 PM I hope it gets finished some day 31-Aug-20 10:42 PM Then it might be worthier of #resources 31-Aug-20 11:04 PM Huh - cool! 31-Aug-20 11:04 PM that being said 31-Aug-20 11:04 PM I wonder when we'll ever get to a point where say 31-Aug-20 11:04 PM you can get chips made at I dunno 31-Aug-20 11:04 PM a hundred or two for x number of chips? 31-Aug-20 11:04 PM something that a hobbyist could save up a bit for and afford 31-Aug-20 11:09 PM sorry if that's a dumb question, but what advantage would a chip have over a "chip like pcb"? 31-Aug-20 11:09 PM I'm talking about stuff like those GPS modules 31-Aug-20 11:09 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/Img-gps-module-mc-1010-1-6213D.png 31-Aug-20 11:10 PM custom high speed ASICs? custom analog ASICs for various experiments, like Josephson Junctions? 31-Aug-20 11:10 PM you can design those PCBs yourself and get them for way cheaper than a chip 31-Aug-20 11:10 PM custom high speed ASICs? custom analog ASICs for various experiments, like Josephson Junctions? @SleepyOwl Joyce ok, that level of hobbyists 31-Aug-20 11:14 PM there's just so much you can do even with a high speed FPGA like a Kintex-7 or Virtex-7 31-Aug-20 11:14 PM at the end though - you're not beating an ASIC especially in speed for computational applications 31-Aug-20 11:14 PM FPGAs are good if maybe you want to try something like RISC-V or other open source processor design and verify it 31-Aug-20 11:14 PM but to actually get something practical and performant - I am not sure about that 31-Aug-20 11:14 PM unfortunately - semiconductor fabrication costs are still out of the reach of most hobbyists 31-Aug-20 11:14 PM let me see what's the costs again 31-Aug-20 11:14 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-99514.png 31-Aug-20 11:14 PM older process tech is cheaper for sure 31-Aug-20 11:14 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-03DBC.png 31-Aug-20 11:14 PM still probably out of most of our budgets at that cost per mm^2 31-Aug-20 11:57 PM the Skywater 130nm process will hopefully be more accessible to hobbyists in the future 31-Aug-20 11:57 PM There is/was the MakeLSI thing from Japan, but not much details about it thesedays. 01-Sep-20 12:01 AM I see 01-Sep-20 12:01 AM MakeLSI was like 200-500usd and you'd get some packaged chips 01-Sep-20 12:01 AM In a Hitachi or Mitsubishi fab in Japan. There ware some silicon art with local memes made at least. 01-Sep-20 12:01 AM And at least one led flasher. 01-Sep-20 12:02 AM oh 01-Sep-20 07:36 AM @GigaSquirrel I had the same thought. A guy on The Amp Hour a few weeks back had something like that. It was just a built up mini circuit with flip pack IC’s on it. I can see the moduleized boards for things like radios and such where there is a very fiddly or pre approved layout or going vertical or mezzanine for parts density but just for general layout I have my issues with it. 01-Sep-20 10:16 AM FWIW the company that mithro/google is working with had prices of closer to $70k for an IC 01-Sep-20 10:16 AM some advantages of ASIC vs chip on board: lowest possible power, lower size 01-Sep-20 10:16 AM There are projects (indoor temp/humidity/etc monitoring) where saving a few tens of uA can add weeks to a several month lifespan. Approaching the point of small solar panels being feasible to use as well. 01-Sep-20 10:16 AM (Jeelabs was playing around with this maybe 5 years ago) 01-Sep-20 10:16 AM They couldn't get small solar panels to provide nearly enough power indoors, so continued swapping batteries out every 6-9 months or something like that 15-Sep-20 02:11 PM https://twitter.com/Chips4Makers/status/1305831021594763267 15-Sep-20 02:19 PM those have to be the longest bonding wires I have ever seen 16-Sep-20 08:01 AM something I've been noodling over lately: it feels like there's a market for "really old" process chips targeted at hobbyists. I'm thinking like 500nm or maybe even larger. Would be pretty forgiving (relatively speaking), could target smaller wafer sizes due to high mix / low volume production, equipment is cheap (again, relatively) because it's old and liquidated/retired. And you can make a Pentium with 500nm so still halfway decent in the scheme of things. Would really need an ecosystem of software to support it (route/place, synthesis, simulation, etc) 16-Sep-20 08:01 AM honestly even super old stuff like 1um would be fun to play with as a hobbyist 16-Sep-20 08:02 AM yeah, but it's pretty time consuming to get a reproducable process up & running imho 16-Sep-20 08:02 AM even if it's 'only' 1um-ish 16-Sep-20 08:02 AM with such old machines you would need a team to yield quantities 16-Sep-20 08:02 AM you said "market" 16-Sep-20 08:02 AM implies selling chips 16-Sep-20 08:02 AM my guess is that a capable team would cost alot of money - all the time, especially the first year of figuring out 16-Sep-20 08:02 AM even if you've access to cheap students like an university, they have always some regulary paid technicians and engineers in those facilities to keep those old machines up & running 16-Sep-20 08:02 AM but if you can akquire the funding, y not 16-Sep-20 08:02 AM I would totally use it for some analog stuff ️ 16-Sep-20 08:02 AM such a (cheap??) service 16-Sep-20 08:59 AM yeah those are fair points i guess the real sweet spot would be old process node with new equipment, but that gets you back to being rather expensive 16-Sep-20 08:59 AM and then you're battling low volumes overall, because even if hobbyists love it there is still limited volume (and realistically, no one knows how to design chips unlike electronic design) so the uptake would be slow and probably never very large 16-Sep-20 08:59 AM well a guy can dream 16-Sep-20 09:07 AM not just the process itself 16-Sep-20 09:07 AM but the air filtering, etc 16-Sep-20 09:07 AM the chemical processes 16-Sep-20 09:07 AM costs a fair bit 16-Sep-20 09:07 AM "Cheap" chip fab will come 16-Sep-20 09:07 AM but I don't know how soon and what feature sizes 16-Sep-20 09:07 AM there are software tools that can take FPGA designs and do the floorplanning and and layout for you 16-Sep-20 09:07 AM I also am not sure how well and how efficiently those work 16-Sep-20 09:33 AM I'm not going to deny that people would buy those chips if they were to be produced, but I have a lot of trouble imagining how you could ever make them at a reasonable price while not losing money. Everything about the process is still expensive and requires pretty significant expertise. 16-Sep-20 09:34 AM reminder that there's a company making reproduction motherboards for the Apple Lisa 16-Sep-20 09:39 AM I forsee this as something you'd do as part of an already existing fab system 16-Sep-20 09:39 AM like MOSIS or other companies that serve scientific or academic customers 16-Sep-20 09:41 AM The 130nm openpdk stuff and all the risc-v work seems to be helping on the software tooling front 16-Sep-20 09:41 AM yup 16-Sep-20 09:41 AM RISC-V imo, will drive a lot of open, low cost ASIC work leading up into the future 16-Sep-20 09:42 AM And possibly later other commercial fabs will offer openpdk/mosis style access to really old process nodes 16-Sep-20 09:42 AM especially since RISC-V is an open source processor design 16-Sep-20 09:42 AM Hopefully so 16-Sep-20 09:42 AM there's also work using the POWER isa 16-Sep-20 09:42 AM oh? Cool! 16-Sep-20 09:43 AM one thing about older process is that they should be more robust to things like particle contamination, etc. there's probably a sweet spot where (more) modern equipment and knowledge could be applied to older process for a net win. And prices in general have come down.... i was reading an article from the 90s saying that prime 200mm wafer were $3k. You can get prime 200mm from university-oriented sources for $30 now 16-Sep-20 09:43 AM As apparently from their point of view it's more open than risc-v, or just some opensource purity thing 16-Sep-20 09:43 AM But they are working on one 16-Sep-20 09:43 AM (the wafer is probably the cheapest thing, compared to all the process chemicals, but the point probably holds across the board ) 16-Sep-20 09:43 AM Yep 16-Sep-20 09:43 AM And for hobbyist grade stuff (thinking about hackerspace grade semiconductor process) it would need to be relatively green 16-Sep-20 09:43 AM Less arsine, silanes and phospine 16-Sep-20 09:44 AM yeah that's true 16-Sep-20 09:45 AM I mean the works of folk like Jeri Ellsworth and Sam Zeloof have shown that you could in theory DIY a lot of these processes 16-Sep-20 09:45 AM Yes 16-Sep-20 09:45 AM i wonder how far you could go with metal gates using modern litho tools 16-Sep-20 09:45 AM even up to pretty advanced lithographic techniques 16-Sep-20 09:45 AM like maskless projection based litho 16-Sep-20 09:45 AM there is also the open process from that university in Hong Kong 16-Sep-20 09:45 AM But yeah maskless litho coould be the thing that would allow a jlcpcb of chips to exist 16-Sep-20 09:46 AM yup 16-Sep-20 09:46 AM I just wonder about some of the other processes 16-Sep-20 09:46 AM like ion implantation/doping 16-Sep-20 09:47 AM yeah feels like you'd have to use thermal diffusion doping to keep it cheap, and that has a whole host of problems 16-Sep-20 09:47 AM I don't remember what Sam Zeloof did 16-Sep-20 09:47 AM posphine \o/ 16-Sep-20 09:47 AM he did thermal 16-Sep-20 09:47 AM but Jerri used some sort of film based diffusion doping 16-Sep-20 09:47 AM using tantalumfilm and stuff like that 16-Sep-20 09:47 AM for implanting I mean 16-Sep-20 09:48 AM I should pull out my semiconductor fab textbook and look at some of the doping processes 16-Sep-20 09:48 AM ion implanting is crucial 16-Sep-20 09:48 AM easy mode: spin-on boric or phosphoric acid, bake to drive in. hard-mode: ion implanting 16-Sep-20 09:48 AM I am ngl 16-Sep-20 09:48 AM this Discord will likely have someone with the resources and crazy enough to actually do it 16-Sep-20 09:49 AM we can hope! 16-Sep-20 09:50 AM https://libresilicon.com/index_en.html Scroll down to "The process" 16-Sep-20 09:50 AM well you talked abozt processes of 199x processors 16-Sep-20 09:50 AM right. 16-Sep-20 09:50 AM The last time I skimmed it it looked pretty clean and usually had alternatives to nasty chemicals 16-Sep-20 09:50 AM Most of the time it was plasma ashing 16-Sep-20 09:50 AM IIRC 16-Sep-20 09:51 AM I wanted to do this too, but I have little space and Singapore's a pain of a place to get chemicals of sorts in 16-Sep-20 09:51 AM it's impossible without an ion implanter 16-Sep-20 09:51 AM even the libresilicon process needs one 16-Sep-20 09:51 AM yeah the industry switched from diffusion doping really quickly iirc? 16-Sep-20 09:51 AM yes 16-Sep-20 09:52 AM for good reasons 16-Sep-20 09:52 AM DIY ion implantation device, anyone? 16-Sep-20 09:52 AM heh 16-Sep-20 09:52 AM brb building a particle accelerator 16-Sep-20 09:52 AM but it's definitely not 'only' the implanter 16-Sep-20 09:52 AM all those steps in the libresilicon are not trivial 16-Sep-20 09:53 AM yup 16-Sep-20 09:53 AM and the "high tech" process calls for ion implantation anyhow 16-Sep-20 09:53 AM I just did a really really quick skim, so feel free to correct me if I said something inaccurate 16-Sep-20 09:55 AM I would totally participate in trying 16-Sep-20 09:56 AM needs to plan to move somewhere where she can get a sizable garage 16-Sep-20 09:56 AM bootstrapping such an oldschool foundry 16-Sep-20 09:58 AM Deep reactive-ion etching 16-Sep-20 09:58 AM wew. 16-Sep-20 09:58 AM that's another fun part 16-Sep-20 09:58 AM (from the high tech process) 16-Sep-20 09:59 AM i wonder if some alternate techniques could be stolen from related fields. e.g. modulation doping with aluminum instead, or laser-assisted doping, that sort of thing 16-Sep-20 09:59 AM the low tech process calls for KOH and TMAH 16-Sep-20 09:59 AM I was thinking about the aluminum based doping a number of years back when I was reading up into this whole thing 16-Sep-20 09:59 AM but I don't know how well it would work and what exactly is needed for the process to be done 16-Sep-20 10:01 AM i just stumbled on it the other day and lodged it under the "nifty" category to learn more about soon 16-Sep-20 10:01 AM ah, cool 16-Sep-20 10:01 AM is now really tempted to pull out the wafers she bought back in 2013, but realizes she can't do much anyway 16-Sep-20 10:01 AM bummer. 16-Sep-20 10:01 AM my process would have been very crude, similar to that of Jerri Ellsworth 16-Sep-20 10:01 AM she used Armor Etch for the etching process 16-Sep-20 10:01 AM and I am sure that might introduce all sorts of other things into the mix 16-Sep-20 10:01 AM iirc, the transistors she made wouldn't turn off even when the gate is biased below cutoff 16-Sep-20 10:04 AM yeah getting good, pure process chemicals is probably pretty important to avoid contamination and defects 16-Sep-20 10:06 AM yup 16-Sep-20 11:43 AM okay, 1 prob is physical distance 16-Sep-20 11:43 AM or is it feasible to split the process in its steps and ship the wafers around the globe? 16-Sep-20 11:43 AM large semicon vendors do that for sub-stacks (multiple consecutive steps per site) 16-Sep-20 11:47 AM I mean why not - but you'd have to ensure that every step of the way is clean 16-Sep-20 11:48 AM yeah - good secure containment 16-Sep-20 11:50 AM I would be excited to see an open source fab though 16-Sep-20 11:50 AM run by a hackerspace or something 16-Sep-20 11:50 AM and I would totally help contribute even in terms of monthly funds, albeit in small amounts 16-Sep-20 06:47 PM My FIB will implant... Gallium 16-Sep-20 06:47 PM And apparently people have used FIB to inplant for making some devices 16-Sep-20 10:09 PM yeah 18-Sep-20 04:30 AM @N00N let's talk about the topic here, since this would be a good place for this haha 18-Sep-20 04:30 AM but yeah 18-Sep-20 04:30 AM RE: lithographic processes 18-Sep-20 04:30 AM https://www.youtube.com/watch?v=Nxz_ENnmgtI <-- projection litho 18-Sep-20 04:30 AM he's also done E-beam litho 18-Sep-20 04:30 AM https://www.youtube.com/watch?v=SB94rQtKlKI 18-Sep-20 04:32 AM yeah that's the way to go for small quantities 18-Sep-20 04:33 AM DLP based litho is interesting 18-Sep-20 04:33 AM yep should be faster 18-Sep-20 04:33 AM I just wonder what the smallest feature sizes you can achieve are 18-Sep-20 04:33 AM compared to EBL 18-Sep-20 04:33 AM and besides, it elminates the need for masks 18-Sep-20 04:33 AM yes 18-Sep-20 04:33 AM but for larger quantities you can also do some masks this way 18-Sep-20 04:34 AM yup 18-Sep-20 04:34 AM I just wonder what it would take to build a small feature size process 18-Sep-20 04:34 AM so for reference 18-Sep-20 04:36 AM hm EBL could be pretty small, right? 18-Sep-20 04:36 AM but slow 18-Sep-20 04:37 AM the Pentium used a 0.8 um process 18-Sep-20 04:37 AM EBL can 18-Sep-20 04:37 AM tbh 18-Sep-20 04:37 AM a good goal to reach for anyone wanting to attempt DIYing a fab process 18-Sep-20 04:37 AM would be to even reach the same feature sizes as the Pentium and so on 18-Sep-20 04:37 AM I mean, you can do QUITE a lot of stuff with such a process 18-Sep-20 04:41 AM so it seems as a stepper you could use 405nm with a mid-NA optics to reach 800nm 18-Sep-20 04:41 AM 405nm would be h-line of a Hg-lamp 18-Sep-20 04:41 AM but an excimer laser would be faster 18-Sep-20 04:41 AM as a light source 18-Sep-20 04:42 AM yeah 18-Sep-20 04:42 AM I am thinking about handling and moving also 18-Sep-20 04:42 AM process steps in cabinets with airlocks between them? 18-Sep-20 04:42 AM that way 18-Sep-20 04:42 AM you don't have to air filter the whole room 18-Sep-20 04:42 AM I wonder how feasible such a setup is 18-Sep-20 04:42 AM given what we're talking about 18-Sep-20 04:42 AM is a Hackerspace run initiative 18-Sep-20 04:45 AM yep, guess glove boxes are a good idea 18-Sep-20 04:46 AM still, doping and stuff is quite a challenge 18-Sep-20 04:46 AM yep 18-Sep-20 04:46 AM even the litho 18-Sep-20 04:46 AM reproducable results, usable resists etc 18-Sep-20 04:46 AM which is why I would rather discuss this here - since I am pretty sure that the folks here might be able to think of a physical realization of a implanter 18-Sep-20 04:47 AM yeah, would be nice 18-Sep-20 04:47 AM I mean it IS after all, vacuum hackers 18-Sep-20 04:48 AM that's what they're living for 18-Sep-20 04:48 AM etching is going to be interesting too 18-Sep-20 04:48 AM especially since you want a controlled etch 18-Sep-20 04:48 AM and you want to avoid over or under etching 18-Sep-20 04:49 AM yep, KOH or plasma stuff, but I guess plasma is the way to go 18-Sep-20 04:49 AM dammit brb I am going to pull out my semiconductor fab textbook 18-Sep-20 04:49 AM for most cases 18-Sep-20 04:49 AM I should be working on an open source server design - but as you can see, this topic is too interesting to pass on 18-Sep-20 04:50 AM with low energies and more time you can also manage isotropic results with plasma, right? 18-Sep-20 04:50 AM for wanted underetching 18-Sep-20 04:50 AM for MEMS or similar things 18-Sep-20 04:52 AM I would think so 18-Sep-20 04:52 AM I also wanted to build a small ion implanter, but time & moneyz 18-Sep-20 04:52 AM just give me a moment 18-Sep-20 04:52 AM digging out books 18-Sep-20 04:55 AM oh man those feelz oscillating between 'could be possible' and 'never happen' 18-Sep-20 05:01 AM fudge 18-Sep-20 05:01 AM I hate living in one room apartments for a reason 18-Sep-20 05:01 AM book is buried under my table so deep in that I just said fudge it 18-Sep-20 05:01 AM it'll take an hour of searching and putting everything back in 18-Sep-20 05:01 AM let's do this just by using the internet as a reference then 18-Sep-20 05:01 AM anyway plasma etching let see what gases are typically used 18-Sep-20 05:01 AM https://www.linde-gas.com/en/images/Silicon%20Semiconductor%20China%20-%20Enabling%20Electronics%20Manufacturing%20Etching%20Relies%20on%20Electronic%20Special%20Gases%20-%20English_tcm17-482187.pdf 18-Sep-20 05:01 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-7A98F.png 18-Sep-20 05:09 AM yeah - depening on what kind of etching they're reactive or simply ions like Ar+ 18-Sep-20 05:09 AM yes those are for RIE 18-Sep-20 05:09 AM not thaaat nasty stuff 18-Sep-20 05:11 AM I wonder if this can be adapted 18-Sep-20 05:11 AM https://www.youtube.com/watch?v=vYnl-wNq0JE 18-Sep-20 05:12 AM the nastiest garden-variety stuff would be the PH3 for doping, right? 18-Sep-20 05:12 AM yeah :/ 18-Sep-20 05:14 AM yes that could be used 18-Sep-20 05:14 AM but it has the coil inside the chamber 18-Sep-20 05:15 AM yeah - you'd need to change that 18-Sep-20 05:15 AM what could be okay, but in general is not perfect because of contamination - the coil material sputters all over 18-Sep-20 05:16 AM something like this conceptually 18-Sep-20 05:16 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/3-s2-6AA0F.png 18-Sep-20 05:17 AM mostly you've the coils outside behind a viewport or ceramics 18-Sep-20 05:17 AM yeah, that's capacitive 18-Sep-20 05:17 AM the thing in the video is inductive coupled 18-Sep-20 05:17 AM inductive coupled (ICP) is alot faster (more power) afaik 18-Sep-20 05:19 AM it's also actually pretty small 18-Sep-20 05:19 AM anyway gases in ion implantation: P, B, BF3, PH3, and AsH3 18-Sep-20 05:19 AM checks out - but phosphine is nasty lol 18-Sep-20 05:19 AM so is boron trifluoride 18-Sep-20 05:21 AM ok AsH3 and PH3 are instand killers, right? 18-Sep-20 05:21 AM ah 18-Sep-20 05:21 AM yeah maybe it's somehow possible to use differend harmless precursors to "feed" the ion gun? 18-Sep-20 05:21 AM but I guess if that would be easy 18-Sep-20 05:21 AM it would be the standard 18-Sep-20 05:22 AM perhaps 18-Sep-20 05:23 AM naively I would try elementary B and P 18-Sep-20 05:23 AM and sputter it 18-Sep-20 05:24 AM again - describe a process of getting elementary P and B into such a system 18-Sep-20 05:24 AM boron, like elementary boron is rare on Earth 18-Sep-20 05:26 AM yeah but metalloid boron is available 18-Sep-20 05:26 AM like white phosphorus 18-Sep-20 05:27 AM yeah, so how would one get it into a DIY ion implanter setup for example 18-Sep-20 05:27 AM I'm just curious 18-Sep-20 05:27 AM and via an arc discharge or a sputter as a feeder for the ion gun it could be possible? 18-Sep-20 05:27 AM you've only to vapourise/sputter and ionize the elementary stuff 18-Sep-20 05:27 AM and then you can suck it into the gun 18-Sep-20 05:30 AM touching back on the litho resolution discussion from earlier, that's probably the easiest part really. A cheap 20x plan apo microscope object will let you hit ~300nm @405nm but obviously has a small field of view (see: https://youtu.be/_w0Z2Y5vaAQ?t=310). 60x air or 100x immersion objectives would let you go even smaller. The key with that route is really nailing the stepper mechanics since you're using such a small FOV: need excellent encoders, excellent servos, no vibrations, temperature compensation, etc etc 18-Sep-20 05:30 AM but as I said - naively thinking 18-Sep-20 05:30 AM @polyfractal yeah 18-Sep-20 05:30 AM the DLP (or LCoS) pixel size becomes a limit as well 18-Sep-20 05:31 AM I think you should also try to build a large stepper 18-Sep-20 05:31 AM with normal masks 18-Sep-20 05:31 AM yup - that's what I was thinking about DLP 18-Sep-20 05:32 AM and use the dlp or ebl setup only for feature/cell testing 18-Sep-20 05:32 AM because it's slow 18-Sep-20 05:32 AM and for making masks 18-Sep-20 05:32 AM can find some good reduction lenses on ebay for cheap too, if you want to go with masks, or even maskless (although pixel size is a killer there). i picked up a 10:1 405nm zeiss litho lens a while back for a great deal 18-Sep-20 05:32 AM how great? 18-Sep-20 05:33 AM $200 USD iirc? something in that ballpark 18-Sep-20 05:33 AM ah cool 18-Sep-20 05:33 AM cool! 18-Sep-20 05:33 AM for a large stepper we should use mirrors 18-Sep-20 05:33 AM esp because it has lesser wavelength limitations 18-Sep-20 05:35 AM i bet slow dlp/lcos/laser/ebl are probably fine for "hackerspace" style setups though, since the rest of the steps are gonna be a lot slower anyway (furnace, etches, etc) and it's low volume 18-Sep-20 05:35 AM yeah but I mean for the story of scaling up 18-Sep-20 05:35 AM yup 18-Sep-20 05:36 AM to akquire invest 18-Sep-20 05:36 AM using EBL is pretty convenient of course 18-Sep-20 05:36 AM it's simple to re-allign to fiducials 18-Sep-20 05:36 AM simply using the SEM capability of the rig 18-Sep-20 05:36 AM an optical stepper is a lot of PITA to realize in a good way 18-Sep-20 05:39 AM ah so i was noodling on that the other day... i wonder if a "just in time" model would work for small scale. picture one complete fab assembly line: has a maskless litho, small RTP furnace, resist/etch/wash stations, automation to move things between the stations, etc. Can handle one wafer from start to finish. Then clone that a dozen times. It's obviously a lot lower volume than optimized batch processes, but also allows fast turnover (like waiting for a PCB order to fill up), optimized for yes sounds reasonable for me 18-Sep-20 05:39 AM yeah seems workable 18-Sep-20 05:39 AM I just wonder about another question 18-Sep-20 05:39 AM size of wafers one would deal with 18-Sep-20 05:40 AM I would start with 4" 18-Sep-20 05:40 AM (because it fits in my stuff) 18-Sep-20 05:40 AM yeah, 100mm is probably the sweet spot. small and easy to handle, cheap, can still get a decent number of dies per wafer (https://caly-technologies.com/die-yield-calculator/). Although the 50mm wafers are really cheap so I could see an argument there, if you had a "just in time" proccess really dialed in 18-Sep-20 05:42 AM ok smaller is better for me 18-Sep-20 05:42 AM I have 50.8 mm wafers on hand haha 18-Sep-20 05:42 AM (more homogenity with the processes) 18-Sep-20 05:42 AM but I was thinking of yield optimization 18-Sep-20 05:42 AM of course 18-Sep-20 05:42 AM larger wafers and smaller sizes per die means higher yield 18-Sep-20 05:42 AM yes 18-Sep-20 05:42 AM but larger machines 18-Sep-20 05:43 AM since there are defects in the material that cannot be avoided 18-Sep-20 05:43 AM so I guess 2" would be better 18-Sep-20 05:43 AM a smaller die and larger wafer means lesser defective chips 18-Sep-20 05:43 AM for the first steps 18-Sep-20 05:43 AM yup 18-Sep-20 05:43 AM I found 50.8 mm wafers to be a size that was comfortable to work with for a DIY/hackerspace process 18-Sep-20 05:44 AM 4" is the max of what I think I could handle 18-Sep-20 05:44 AM yeah, and i think the industry trend towards large wafers doesn't apply at small scale, since you're not optimizing for batches of millions of chips (with months of lead time) 18-Sep-20 05:45 AM yep 18-Sep-20 05:45 AM yup 18-Sep-20 05:45 AM are there 1" wafers? 18-Sep-20 05:45 AM yes it seems 18-Sep-20 05:45 AM yep... although they start getting expensive the smaller you go amusingly 18-Sep-20 05:45 AM ah ok 18-Sep-20 05:45 AM https://order.universitywafer.com/default.aspx?cat=Silicon 18-Sep-20 05:46 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-E23A1.png 18-Sep-20 05:46 AM i think lots of universities use the 50mm size, so that price point has stabilized due to volume. smaller is probably less commonly used 18-Sep-20 05:46 AM yeah 18-Sep-20 05:47 AM heh 1969 18-Sep-20 05:47 AM 26" sweet lord 18-Sep-20 05:47 AM that's enormous 18-Sep-20 05:47 AM drop one of those and you're having a bad day 18-Sep-20 05:48 AM the 1960s 18-Sep-20 05:48 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/image0-4637E.jpg 18-Sep-20 05:48 AM drop? 18-Sep-20 05:48 AM what do you mean? 18-Sep-20 05:48 AM we automate humans out of the process 18-Sep-20 05:48 AM machines build themselves 18-Sep-20 05:48 AM haha fair 18-Sep-20 05:48 AM with scara bots? 18-Sep-20 05:49 AM I imagine a fab with little to no people 18-Sep-20 05:49 AM except just to resolve machine issues 18-Sep-20 05:49 AM until machines get smart enough that human operators themselves are removed 18-Sep-20 05:49 AM would be okay 18-Sep-20 05:50 AM and robots fix themselves 18-Sep-20 05:50 AM when that happens 18-Sep-20 05:50 AM yah, i think most of the personel in modern fabs are Q&A towards the end of the process, and fixing broken stuff 18-Sep-20 05:50 AM https://giphy.com/gifs/bbccapital-3o7abFpd91G18NYtpe 18-Sep-20 05:50 AM you don't want to have employees in germany 18-Sep-20 05:50 AM in any case, 2 inch or even 4 inch wafers might be something you want to set as standard 18-Sep-20 05:53 AM our gov has plans to abolish the possibility of ending employment 18-Sep-20 05:53 AM since all your processes are going to be based on that 18-Sep-20 05:53 AM I would think that some level of automation and movement between airlocks, etc are going to be needed 18-Sep-20 05:54 AM yeah small belt conveyors 18-Sep-20 05:54 AM or at least placement into airtight sealed wafer containers 18-Sep-20 05:54 AM then moved to another process with air cleaning and airlocks 18-Sep-20 05:54 AM with viton belts 18-Sep-20 05:54 AM the ordinary chip fab stuff 18-Sep-20 05:54 AM scara bots and belt conveyors 18-Sep-20 05:54 AM you can get them cheap if some photovoltaic cell production facility is modernizing or dead 18-Sep-20 05:54 AM those smallish dual belt stuff 18-Sep-20 05:54 AM pretty similar to the pcb handling systems 18-Sep-20 05:54 AM but smaller of course 18-Sep-20 05:59 AM that would work, just in some sort of air filtered or airlocked system 18-Sep-20 05:59 AM yeah 18-Sep-20 05:59 AM with a laminar flow 18-Sep-20 05:59 AM like a modular system of closed laminarflow boxes with locks 18-Sep-20 06:00 AM yup 18-Sep-20 06:00 AM air locks 18-Sep-20 06:00 AM you could repurpose say 18-Sep-20 06:00 AM and some gloves for manual intervention 18-Sep-20 06:00 AM hooded work tables for that 18-Sep-20 06:00 AM yep 18-Sep-20 06:00 AM maybe those cheap chinese hepa filtered laminar flow box kits for smartphone repair shops are a good base 18-Sep-20 06:03 AM I am thinking like 18-Sep-20 06:03 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/830_ABC-7ECF7.png 18-Sep-20 06:03 AM maybe those cheap chinese hepa filtered laminar flow box kits for smartphone repair shops are a good base @N00N pics? 18-Sep-20 06:06 AM wouldn't be too hard to knock together a basic "clean box" cleanroom thing. You can buy large volume HEPA filters for moderately cheap (they are used in medical facilities so pretty easily obtained), then just build a frame and use plastic sheeting for most of the walls (ideally anti-static). That's roughly equivalent to "soft-wall clean rooms" that you can purchase. i believe those can get down to like class 1000 or so, and if it's only robots in there and not humans, even better 18-Sep-20 06:06 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/Softwall-cleanroom-85727.png 18-Sep-20 06:06 AM https://www.aliexpress.com/item/4000592369110.html 18-Sep-20 06:06 AM yeah 18-Sep-20 06:07 AM and if it's only robots in there and not humans, even better you want to increase the probability of death by AI 18-Sep-20 06:07 AM building a larger room around that would be smart 18-Sep-20 06:08 AM yeah 18-Sep-20 06:08 AM that is another way about it 18-Sep-20 06:09 AM although, heh, i see HEPA prices are way up due to ye olde pandemic 18-Sep-20 06:09 AM they were cheap a year or three ago when i last looked 18-Sep-20 06:10 AM in all cases - you could have people in the "fab" 18-Sep-20 06:10 AM so a larger HEPA filtered soft cleanroom as you said 18-Sep-20 06:10 AM but also with laminar flow boxes for extra assurance of cleanliness 18-Sep-20 06:10 AM but a soft cleanroom would make a lot of things easier 18-Sep-20 06:10 AM like processes which can't be fed into an airlock so easily 18-Sep-20 06:10 AM like DRIE 18-Sep-20 06:10 AM or implantation 18-Sep-20 06:13 AM yeah 18-Sep-20 06:13 AM heh speaking of cleanrooms, hyugens optics just posted a video on making a laminar flow hood for litho 18-Sep-20 06:13 AM https://www.youtube.com/watch?v=bkmVmsMUsJU 18-Sep-20 06:14 AM yeah I briefly looked 18-Sep-20 06:19 AM other than ion implant, the other major hurdle is the polysilicon layer Might be able to push metal-gates further than they did in the 60s with modern tools (better alignment, linear motors, etc)... but not having self-aligned gates is probably going to hurt. and i dont think any hackerspace wants to deal with silane gas 18-Sep-20 06:19 AM hmm 18-Sep-20 06:20 AM spin-on-glass can probably be used for all the dielectric layers "on top" (around all the routing wires basically), but that one polysilicon layer on the gate is tough 18-Sep-20 06:20 AM unless you totally change chemistry and go with copper/copper oxide FETs or something 18-Sep-20 06:20 AM (or high-k insulator/metal gates but i dont know much about that area tbh) 18-Sep-20 06:28 AM could you make Poly-Si via e-beam evap? 18-Sep-20 06:29 AM hmm, i bet you could 18-Sep-20 06:30 AM but no idea if those layers are good enough 18-Sep-20 06:30 AM but maybe better 18-Sep-20 06:30 AM because no H 18-Sep-20 06:30 AM involved 18-Sep-20 06:30 AM H impurities are not good for the poly-Si s conductivity right? 18-Sep-20 06:30 AM so maybe PVD like E-Beam evap can make the better layers 18-Sep-20 06:30 AM not that it makes things easier 18-Sep-20 06:33 AM hehe 18-Sep-20 06:33 AM definitely in the realm of doable though, so that's a plus 18-Sep-20 06:51 AM yeah 18-Sep-20 06:51 AM I think everything has to be in the realm of something someone here could do 18-Sep-20 06:51 AM with as little hazardous processes as possible 18-Sep-20 06:51 AM probably look at the LibreSilicon project 18-Sep-20 06:51 AM and let's take a look at their processes and stuff like that 18-Sep-20 06:51 AM identify what would be doable with little risk? 18-Sep-20 07:07 AM https://www.helmholtz-berlin.de/media/media/angebote/bibliothek/Examensarbeiten-Master/dissertation-sontheimer-hzb-b-23.pdf <-- page 33 talks about a E-beam process 18-Sep-20 07:07 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-B3A0F.png 18-Sep-20 07:23 AM yep 18-Sep-20 07:23 AM that would be cool 18-Sep-20 07:14 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/20200918_184514-637AF.jpg 18-Sep-20 07:14 PM This is cute 18-Sep-20 07:14 PM Apparently they make flat gate valves for passing wafers into machines 18-Sep-20 07:14 PM ah I love it 18-Sep-20 07:14 PM I've seen them before 18-Sep-20 07:14 PM I wonder why they're scrapped 18-Sep-20 07:15 PM I wonder why they're scrapped @idmb A company went out of business and I've been going through all their waste 18-Sep-20 07:15 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/20200918_113018-43A34.jpg 18-Sep-20 09:55 PM ooooh! 18-Sep-20 09:59 PM nasty lol 18-Sep-20 10:02 PM I wish there were dumps I could go to and look at stuff lol 18-Sep-20 10:02 PM uni waste 18-Sep-20 10:03 PM here, they are sold off to asset recovery companies 18-Sep-20 10:03 PM there's no waste or scrap dumps of this sort in Singapore 18-Sep-20 10:06 PM that sounds like something management higher up decided 18-Sep-20 10:06 PM without really considering the value of the time used to do so 18-Sep-20 10:06 PM I wonder if it has a net positive or negative effect on research output 18-Sep-20 10:07 PM I mean hey, generally there are no dumps of these sort here likely because of regulations and stuff like that 18-Sep-20 10:07 PM I am looking through the listings of one such company 18-Sep-20 10:07 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-5C99C.png 18-Sep-20 10:07 PM "Please self inspect. Can be powered on but no further testing has been done." 18-Sep-20 10:07 PM As the case with most of the listings 18-Sep-20 10:11 PM lol cinema display logic board 18-Sep-20 10:14 PM a lot of it is uninteresting 18-Sep-20 10:14 PM or stuff I can't store lol 18-Sep-20 10:14 PM but check this out 18-Sep-20 10:14 PM one doesn't look like the other 18-Sep-20 10:14 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-A6B86.png 18-Sep-20 10:14 PM wew 18-Sep-20 10:14 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-426B2.png 18-Sep-20 10:14 PM I wonder what an excimer laser would cost brand new 18-Sep-20 10:21 PM from what I've seen the new costs of lasers go from ~2x to ~20x the used prices 18-Sep-20 10:21 PM the 2x stuff just doesn't really come up that often 18-Sep-20 10:21 PM stuff like this maybe: https://www.ebay.ca/itm/Spectra-physics-Vanguard-355-350-UV-Laser-DPSSL/171935551636?hash=item2808285094:g:27AAAOSwHjNV-P9w 18-Sep-20 10:24 PM I see 18-Sep-20 10:24 PM I am looking at stuff that might be useful for some of the processes talked about much earlier in this channel haha 18-Sep-20 10:28 PM For what it's worth, entire systems get decomissioned from one faulty part 18-Sep-20 10:28 PM A ~300k-new system went for <75k a few months ago, because the power supply of one part was dead. We've replaced that same power supply ourselves for $300. 18-Sep-20 10:29 PM I see 18-Sep-20 10:38 PM right now I am thinking about lithography and photoresist application 18-Sep-20 10:38 PM and how one might do that on a wafer 18-Sep-20 10:38 PM iirc, fab processes do spin coating 18-Sep-20 10:38 PM for photomask writing, there is DLP it seems 18-Sep-20 10:38 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-AA515.png 18-Sep-20 10:38 PM of course this is expensive stuff 18-Sep-20 10:38 PM but I am wondering if one could build such a thing on their own as part of a DIY process 18-Sep-20 10:38 PM I am looking at this list of various photoresists too 18-Sep-20 10:38 PM https://www.cleanroom.byu.edu/resistmanufacturers 18-Sep-20 11:04 PM so in Sam Zeloof's process 18-Sep-20 11:04 PM Positive photoresist (AZ MiR 701 for SiO2 patterning and AZ 4210 for Al layer) is spun on at around 3000rpm yielding a film of about 1.5μm for the AZ MiR 701 or 3.5μm for the AZ 4210 which is soft baked at 90C on a hotplate. 18-Sep-20 11:04 PM I am not sure that's easy to get though 18-Sep-20 11:04 PM Sam was also working out a solution on DIY photoresist 18-Sep-20 11:04 PM http://sam.zeloof.xyz/diy-pr/ 18-Sep-20 11:24 PM http://sam.zeloof.xyz/developing-patterning-process-for-homemade-microelectronics/ <-- his process is also quite well documented 19-Sep-20 03:13 AM I found this: https://www.nanowerk.com/spotlight/spotid=23714.php 19-Sep-20 03:39 AM that being said - in terms of projection based litho 19-Sep-20 03:39 AM I wonder if LCD projectors can be repurposed 19-Sep-20 03:39 AM since they're supposedly sharper than DLP 19-Sep-20 06:50 AM yep, they can this was a test I did a while ago with my resin printer as a contact-litho process. Divisions are 0.001", so the lines are ~200um and the gaps are ~500um (contrast is low since the resist was on a transparent glass slide so hard to image). can also use LCoS projectors, I played around with them a few years ago with moderate success. In both cases, since they need polarizers you have to use non-coherent light sources (LEDs) but otherwise it's fine. Contrast is a little poorer than DLP since the darkened cells still let some light through. And they get grumpy with UV light over time... liquid crystals and polarizers begin to degrade 19-Sep-20 06:50 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/resolution-A0768.png 19-Sep-20 06:50 AM this is from an electroplating project I was fiddling with. Different resist, but still using my printer for contact litho. the skinny "legs" on the right should be ~200um although I didn't measure those 19-Sep-20 06:50 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/vlcsnap-2020-09-15-19h40m08s433_2-828B0.png 19-Sep-20 07:04 AM Ah I see 19-Sep-20 07:04 AM I wonder how far we can push this 19-Sep-20 07:04 AM If we can get like <1.0 um sort of features 19-Sep-20 07:16 AM one of my concerns again is the resist also 19-Sep-20 07:16 AM where to get resist that has a very good resolution 19-Sep-20 07:20 AM theoretically doable if you have enough reduction power. Main issue is that LCDs designed for displays have relatively large pixel size. E.g. my resin printer has 47um pixels, so with a 10:1 reduction lens you're still at ~5um pixels and features would need several pixels to resolve well without aliasing. that's the main perk of tech designed for projection as opposed to displays, they tend to have much smaller pixels (DLP are usually only a few microns, I believe LCoS/LCD projectors are similar) 19-Sep-20 07:20 AM the first photo was with Datak ER-71 negative resist. Sam used it initially and had good results, although i haven't found a good solvent to fully dissolve it yet. The second resit was Injectorall PC 197 positive resist, novolac based and similar to "real" positive resists. Develops and fully dissolves easily in alkaline solutions... but the company appears to be gone now 19-Sep-20 07:20 AM I contacted http://micromaterialstech.com/ a few years back and they will sell to individuals, but the smallest quantity was a quart of resist and cost $600 USD iirc 19-Sep-20 07:24 AM I'm not surprised 19-Sep-20 07:24 AM that's still a fair amount of resist for smallish wafers 19-Sep-20 07:24 AM like 2" or 4" 19-Sep-20 07:24 AM yeah might work for a group buy or something too 19-Sep-20 07:25 AM tbh - if someone was serious in making this work - I wouldn't mind haha 19-Sep-20 07:25 AM I would even ship over my own wafers 19-Sep-20 07:25 AM fwiw i'm working on a laser direct writing setup right now. I played around with projection a few years ago and ran into enough issues to just quit and go with a laser in some ways simpler, in other ways more complicated. we'll see how it goes 19-Sep-20 07:25 AM else they'd be doing nothing 19-Sep-20 07:25 AM I see 19-Sep-20 07:26 AM LDW makes the optics super easy (focus laser, done), and puts all the burden on the xy stage and alignment process 19-Sep-20 07:27 AM I'd assume so 19-Sep-20 07:27 AM for a different idea I had (well PCB material ablation) 19-Sep-20 07:27 AM I was thinking of like using a set of mirrors and moving those to "sweep" the laser beam across a PCB 19-Sep-20 07:27 AM like how a stage laser would do things 19-Sep-20 07:27 AM I just wonder how well it would work in this application 19-Sep-20 07:29 AM should work well in theory. Need a decent f-theta lens to make sure it stays in focus across the whole "image", and then it boils down to how well you can control the mirror galvos since that will determine the resolution 19-Sep-20 07:30 AM yeah 19-Sep-20 07:30 AM of course as we talked about yesterday 19-Sep-20 07:30 AM vibrations start to be a problem at the feature sizes we talk about 19-Sep-20 07:31 AM hehe yep although mounting everything on a big block of granite will get you pretty far, all things considered 19-Sep-20 07:31 AM i wonder if you could DIY an optical air table... hmm 19-Sep-20 07:34 AM or can you find one cheap? 19-Sep-20 07:42 AM so photoresist application 19-Sep-20 07:42 AM I assume its just the usual spin coating 19-Sep-20 07:45 AM yep. i made mine from an old hard drive + 3d printed enclosure + RC PWM controller. crude but works fine (more details: https://hackaday.io/project/25260-makerfoundry/log/60808-spincoater-from-old-hdd) 19-Sep-20 07:45 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/1164981496621773342-86192.png 19-Sep-20 07:45 AM The Datak stuff is very thin, I think you could get pretty good resolution with it 19-Sep-20 07:48 AM cool! 19-Sep-20 07:48 AM I remember when Jerri was using a 12V fan as a spin coater 19-Sep-20 07:48 AM it certainly works - although this looks a lot more professional 19-Sep-20 07:48 AM oh cool - I saw on the comments a link to this 19-Sep-20 07:48 AM https://hackaday.io/project/9205/logs 19-Sep-20 07:48 AM this would be very useful for checking patterns out of photolithographic processes 19-Sep-20 07:54 AM ++ 20-Sep-20 03:19 AM nice 20-Sep-20 05:04 AM in the first iteration I'll focus on EBL 20-Sep-20 05:14 AM and then tryin to organize a stepper for 2" - 4" wafers 20-Sep-20 06:17 AM a...and then we build a eco friendly chip foundry 20-Sep-20 06:17 AM (don't want to use the terminus 'fab' because) 20-Sep-20 06:38 AM @N00N I am trying to figure out what I can do with my own wafers and what I can do to help haha 20-Sep-20 06:38 AM here's a question - if I shipped 6 pieces of N-type, <110> (iirc - it's been way long since I touched them) wafers, would it help, even as a test media? 20-Sep-20 06:39 AM for me? everything helps 20-Sep-20 06:39 AM I'll have to dig it out, since its deeply buried in a stash of stuff under my desk 20-Sep-20 06:39 AM but when I get it out, let's see if I can reseal it properly for shipping 20-Sep-20 06:39 AM you might need to also figure a means of cleaning them, etc 20-Sep-20 06:39 AM cause it was opened before and there is likely contaminants and particles 20-Sep-20 06:41 AM would try some masking and etching with KOH and TMAH 20-Sep-20 06:41 AM I see 20-Sep-20 06:41 AM @polyfractal and I were chatting earlier, we were talking about documenting our processes, etc 20-Sep-20 06:41 AM also @polyfractal woops I got the cost of the HKUST gofundme wrong 20-Sep-20 06:41 AM I mixed it up - it's $50k USD 20-Sep-20 06:48 AM ok that's not extreme 20-Sep-20 06:48 AM https://www.gofundme.com/f/libresilicon-cleanroom-rent <-- said project 20-Sep-20 06:48 AM Currently we rented a Clean Room at the Hong Kong Univeristy of Science and Technology. While working there, the contract finishes soon in June. Well, for the next rent period of one year, we need $ 50.000. This will pay a years worth of cleanroom access at the HKUST. 20-Sep-20 06:48 AM so that's not too bad 20-Sep-20 06:49 AM ah yeah, still a lot of money but a bit more reasonable 20-Sep-20 06:49 AM however, having only raised ~600euro doesn't look so good for them... 20-Sep-20 06:50 AM yeah 20-Sep-20 06:50 AM what surprises me though 20-Sep-20 06:50 AM is that they can rent an academic institution's lab 20-Sep-20 06:50 AM at least where I am, that's unheard of 20-Sep-20 06:54 AM it is a little bit of an odd arrangement (being a private organization), but at least here in the US most universities are setup so the professors "rent" lab space and equipment from the university despite being part of the same organization. e.g. the uni charges professors and they draw money from their grant to pay the uni. so not too unusual in that light, i could see a deal being cut somehow 20-Sep-20 06:55 AM I see 20-Sep-20 06:55 AM if you knew the right people i guess 20-Sep-20 06:56 AM also - you want to start a github/wiki or something for documentation purposes? 20-Sep-20 07:03 AM yeah it probably makes sense to have something like that. we could either start using the semi@home github, the repo is a bit inactive but some of the folks seem to be active on their discord. or we could setup a wiki. i'm about to go on a camping trip, but i could spin one up when i get back 20-Sep-20 07:03 AM Cool! 20-Sep-20 07:03 AM Have fun btw! 20-Sep-20 07:03 AM @N00N how might you be doing EBL? 20-Sep-20 07:03 AM I'd like to know the setup you're intending 20-Sep-20 07:07 AM using an old SEM 20-Sep-20 07:07 AM with DAC attached for the deflection and some ADCs for alignment via SE and BSE detection I guess 20-Sep-20 07:07 AM and a little bit of opencv scripting for the fiducial detection purpose 20-Sep-20 07:10 AM I see 20-Sep-20 07:10 AM fiducial as in a mark on the wafer? 20-Sep-20 07:10 AM yes 20-Sep-20 07:10 AM or are you intending some sort of holder/template~ oh 20-Sep-20 07:10 AM speacial features outside the region with function 20-Sep-20 07:10 AM to align the different process steps/layers 20-Sep-20 07:11 AM I was thinking that a fiducial would eat into the wafer space itself 20-Sep-20 07:11 AM though then again we're not batch processing chips 20-Sep-20 07:11 AM without scanning (illuminating) the regions of interest 20-Sep-20 07:12 AM can't we use the flat and notches of the wafer as a fiducial? 20-Sep-20 07:12 AM yeah I guess so 20-Sep-20 07:12 AM but you want features nearby to have a bilinear reagion (for the DAC) 20-Sep-20 07:12 AM maybe it's somehow possible to make proper calibratin for wider field illumination 20-Sep-20 07:12 AM but I think it could make sense to keep it simple as possible 20-Sep-20 07:12 AM for the first experiment iterations 20-Sep-20 07:14 AM yeah 20-Sep-20 07:18 AM and I would use a PMMA based resist - the standard EBL stuff, I think 20-Sep-20 07:18 AM and maybe it's possible to figure out a reproducable recipe for that - as an alternative to the good stuff from microchem 20-Sep-20 07:21 AM I see 20-Sep-20 07:21 AM generally my worry is the photoresist 20-Sep-20 07:21 AM cause it's usually expensive and hard to procure 20-Sep-20 07:26 AM yeah I want to experiment more with PMMA based stuff and NC based stuff which is kinda self developing, what helps for learning/training of compensation models for charging (electron optical distortion) and proximity effects 20-Sep-20 07:27 AM ah 20-Sep-20 07:40 AM I've 'discovered' the NC self developing by accident a few months ago - but it was of course an old idea from the 1980s 20-Sep-20 07:40 AM I mean it's part of experimentation right? 20-Sep-20 07:40 AM you sometimes "discover" things that people already were aware about 20-Sep-20 07:42 AM yes, it's always the case 20-Sep-20 09:33 AM http://www.emresist.com/products/pmma-resist/ <-- hmmm 20-Sep-20 09:33 AM I wonder how deep a hole this will burn in one's pocket 20-Sep-20 09:41 AM not sure, but I'll ask the this week 20-Sep-20 09:41 AM the good stuff (non DIY) I have is microchem labled from our uni 20-Sep-20 09:42 AM I see 20-Sep-20 09:47 AM my gueds is that's stupidly expensive in small amounts 20-Sep-20 09:49 AM Perhaps 20-Sep-20 09:49 AM as polyfractal said 20-Sep-20 09:49 AM a quart of whatever resist he was working with cost $600? 20-Sep-20 09:49 AM so I won't be surprised if that is the ballpark for resist prices 20-Sep-20 09:49 AM I should probably also look at tools like this 20-Sep-20 09:49 AM http://opencircuitdesign.com/magic/index.html 20-Sep-20 09:49 AM and of course, translating the output of Magic 20-Sep-20 09:49 AM to something that you can do lithography with 20-Sep-20 10:00 AM ah 20-Sep-20 10:00 AM yes 20-Sep-20 10:00 AM totally makes sense 20-Sep-20 10:00 AM but we need also a set of standard cells 20-Sep-20 10:00 AM and the question if polysilicon is feasible or we've to stick to metal etc 20-Sep-20 10:00 AM that also has influence on the standard cells geometries, right? 20-Sep-20 10:04 AM I would think, yeah 20-Sep-20 10:04 AM thing about polysi is that the gases needed for CVD of polysi are quite... noxious 20-Sep-20 10:04 AM and dangerous to handle 20-Sep-20 10:04 AM unless there are other processes of depositing polysi which I don't know about? 20-Sep-20 10:08 AM yeah I thought about the EBE process 20-Sep-20 10:08 AM electron beam evaporation 20-Sep-20 10:08 AM of the silicon 20-Sep-20 10:09 AM I'll need to take a read at the process 20-Sep-20 10:09 AM admittedly some of the processes are new to me 20-Sep-20 10:09 AM and also evaporate the dopant 20-Sep-20 10:10 AM also dopant sources 20-Sep-20 10:11 AM But I dunno if that's practical for circuits 20-Sep-20 10:11 AM just read briefly about the process 20-Sep-20 10:11 AM maybe it's usable with more research 20-Sep-20 10:11 AM the same with ion implantation sources, using sputtering of solid state stuff comes from the heavy ion research groups afaik 20-Sep-20 10:14 AM oh 20-Sep-20 10:14 AM But if it works for them it also should be okay for doping of semiconductors 20-Sep-20 10:14 AM the only difference is that doping needs less energy 20-Sep-20 10:14 AM 1MeV 20-Sep-20 10:15 AM I am wondering how well would a low tech process work, like diffusion for example 20-Sep-20 10:15 AM instead of > GeV 20-Sep-20 10:15 AM >> 1MeV 20-Sep-20 10:15 AM ah 20-Sep-20 10:16 AM http://www.emulsitone.com/alphacatalog.htm <-- using the same film... 20-Sep-20 10:16 AM ah scratch that 20-Sep-20 10:16 AM < 1 MeV instead of >> 1GeV 20-Sep-20 10:16 AM I don't think they are selling any of it anymore 20-Sep-20 10:16 AM I am not even sure the company is still operational 20-Sep-20 10:16 AM humm 20-Sep-20 10:16 AM read about application of such doping films via a sliding blade 20-Sep-20 10:19 AM I just looked up Zeloof's process 20-Sep-20 10:19 AM Doping is then carried out by either solid or liquid source. The solid source is a Boron Nitride disk that is placed in proximity (<2mm) from the wafer in the tube furnace. Alternatively, spin-on liquid sources can be prepared from Phosphoric or Boric acid in water or solvents and doping is carried out in a standard pre-deposition/HF dip/drive-in/deglaze process. I obtained Phosphoric acid in pure form on Amazon and Boric acid from Roach & Ant killer. Since the starting wafer for PMOS here is N-type, I am doing P diffusions of Boron for the source/drain regions and am targeting a sheet resistance in diffused regions of 100 to 250 Ω/sq. 20-Sep-20 10:20 AM yeah but it's not deep enough, right? 20-Sep-20 10:20 AM let's see if I can find data 20-Sep-20 10:21 AM too shallow 20-Sep-20 10:24 AM I am reading this at the moment 20-Sep-20 10:24 AM https://nptel.ac.in/content/storage2/courses/113106062/Lec24.pdf 20-Sep-20 10:29 AM ah 20-Sep-20 10:32 AM I would think the depth is dependant on temperature, concentration, time in process, etc 20-Sep-20 11:46 AM @SleepyOwl Joyce I think making your own pmma solution seems feasible for larger feature sizes... Just grab the smallest pmma available on ebay or whatever (i.e. chips or pellets or powder) and the get some of the solvent (anisole) 20-Sep-20 11:54 AM hmmm... I wonder 21-Sep-20 01:58 AM yeah 21-Sep-20 02:06 AM for high molecular weight PMMA, right? 21-Sep-20 04:18 AM I wonder what's the yield or purity though 21-Sep-20 04:18 AM and how well it would work especially if you want to be working at smaller feature sizes 21-Sep-20 06:07 AM @Foxx UV for writing to a resist is something that is typically done 21-Sep-20 06:07 AM @polyfractal is working on a direct laser writing system for this 21-Sep-20 06:07 AM yes, this is why on I was talking about vibrations and stuff on our hangout 21-Sep-20 06:07 AM because I brought up the idea of moving mirrors to sweep the beam 21-Sep-20 06:07 AM like how show lasers do 21-Sep-20 06:09 AM Well keep in mind I havent studied this subject in 20 years... 21-Sep-20 06:09 AM Plus I was kinda high just pondering fun with uv lasers 21-Sep-20 06:09 AM Cuz I'm thinking an SLA 3d printer 21-Sep-20 06:09 AM But if it could do moar! 21-Sep-20 06:09 AM but the vibrations of the machine as always is a concern 21-Sep-20 06:10 AM Then you mentioned the semiconductor fabrication 21-Sep-20 06:10 AM it might make the image blurry especially at the edges due to vibrations of the beam 21-Sep-20 06:10 AM That's actually my 1st concern since I'm not familiar with laser physics as much... 21-Sep-20 06:10 AM Keeping the beam uniform as it travels 21-Sep-20 06:11 AM hey you have #lasers_and_optics for laser discussion and questions 21-Sep-20 06:11 AM And what sort of tolerances in error are acceptable 21-Sep-20 06:11 AM and I am not too knowledgeable on lasers myself 21-Sep-20 06:11 AM I know like just some basics 21-Sep-20 06:12 AM Same. Just the core concepts 21-Sep-20 06:12 AM @N00N is possibly looking at electron beam lithography 21-Sep-20 06:12 AM Noice 21-Sep-20 06:13 AM there are many challenges 21-Sep-20 06:13 AM Stupid question.... 21-Sep-20 06:13 AM ask away 21-Sep-20 06:13 AM Is electrolysis possible via an electron gun?? 21-Sep-20 06:14 AM electrolysis as in? 21-Sep-20 06:14 AM Electroplating metals 21-Sep-20 06:14 AM I am not sure if electroplating is the idea here 21-Sep-20 06:14 AM but we have an issue that we're discussing 21-Sep-20 06:15 AM Na, just random thought popped in. 21-Sep-20 06:15 AM see, for the gate layers of transistors, we use polysilicon layers 21-Sep-20 06:15 AM the problem here is that the materials used for deposition of polysilicon 21-Sep-20 06:15 AM uses a gas called silane 21-Sep-20 06:16 AM Just thinking if there could be a way to electro plate but via beam to draw stuff. But you could just use a layer mask which is easier 21-Sep-20 06:16 AM which is noxious and toxic 21-Sep-20 06:16 AM so N00N brought up something called electron beam deposition 21-Sep-20 06:17 AM How is the layering done? 21-Sep-20 06:17 AM I would think by the usual processes of growing or depositing your layers, lithography, doping, etc 21-Sep-20 06:17 AM rinse, repeat till done 21-Sep-20 06:17 AM but see 21-Sep-20 06:17 AM We were talking about a process that a group of folks from a hackerspace could do and maintain 21-Sep-20 06:17 AM and noxious gases are something we want to avoid as much as possible 21-Sep-20 06:17 AM so there's going to be some interesting methods discussed here 21-Sep-20 06:17 AM tbh 21-Sep-20 06:17 AM I didn't know there was something called electron beam deposition 21-Sep-20 06:17 AM it wasn't something I come across in my materials or study 21-Sep-20 06:17 AM but basically the setup looks like this 21-Sep-20 06:17 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/1280px-Electron_Beam_Deposition_001-276BE.png 21-Sep-20 06:17 AM when electrons strike the ingot/material 21-Sep-20 06:17 AM the kinetic energy of the electrons are converted to thermal energy which sublimates the material 21-Sep-20 06:17 AM at least that's what I think I understand 21-Sep-20 06:22 AM What sort of scale is possible? 100nM? 60nM? 21-Sep-20 06:22 AM I am not sure that's easy to say 21-Sep-20 06:22 AM wait 21-Sep-20 06:22 AM are you talking about layer depth or feature size? 21-Sep-20 06:23 AM I'm really not familiar with the bare metal itself. I need to do more homework. But not a lot of places that teach you semiconductor research and design 21-Sep-20 06:23 AM Layer depth 21-Sep-20 06:24 AM if you're talking about layer deposition thickness, I am not sure about EBL 21-Sep-20 06:24 AM but there should be some data available about the process in literature 21-Sep-20 06:24 AM I don't have the numbers off my head 21-Sep-20 06:24 AM so I wiill need to comb through the literature for process data 21-Sep-20 06:25 AM That's where I need some hand holding. My eyes glaze over after a while 21-Sep-20 06:25 AM I daresay that papers get dry after a while 21-Sep-20 06:25 AM Seems like a very interesting project..keep me posted 21-Sep-20 06:26 AM in terms of lithography 21-Sep-20 06:26 AM Polyfractal thinks that with a proper reduction lens, we can hit 300 nm feature sizes or even smaller 21-Sep-20 06:26 AM but as always 21-Sep-20 06:26 AM dust is a concern 21-Sep-20 06:26 AM so a soft cleanroom like this was brought up 21-Sep-20 06:26 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/32ca2026-83e1-4c1f-8e23-d07f9eead28b-BE018.png 21-Sep-20 06:26 AM I was also suggesting laminar flow boxes with airlocks 21-Sep-20 06:27 AM It's not hard to make an epic filter 21-Sep-20 06:27 AM Got one right next to me 21-Sep-20 06:28 AM for me, the way I see this 21-Sep-20 06:28 AM is that this is probably hard to do and run as an individual, this whole thing, because of the processes, chemicals, etc 21-Sep-20 06:29 AM Iirc this does 360cfm with activated charcoal and HEPA filter 21-Sep-20 06:29 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/20200921_092820-34A69.jpg 21-Sep-20 06:29 AM but it might be doable by a group of people like a Hackerspace 21-Sep-20 06:29 AM Has a 6" port 21-Sep-20 06:29 AM I see 21-Sep-20 06:29 AM gimme a sec, someone did build their own cleanbox 21-Sep-20 06:30 AM You can pick them up at indoor garden supply shops. And this is a small one 21-Sep-20 06:30 AM https://www.youtube.com/watch?v=bkmVmsMUsJU 21-Sep-20 06:30 AM he explains the challenges of handling wet etchants and stuff in such laminar flow boxes 21-Sep-20 06:30 AM You can get some very sizable fans and filters designed to keep a whole greenhouse of plants safe with filtered air 21-Sep-20 06:31 AM I wonder how clean is clean though 21-Sep-20 06:31 AM Not cheap though. That setup cost me about $220 21-Sep-20 06:31 AM like a soft cleanroom for example has a certain class rating for example, class 1000 21-Sep-20 06:32 AM They have various rated filters for them. 21-Sep-20 06:32 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-F8EC5.png 21-Sep-20 06:32 AM which basically set a limit of the number of given size particles in a given volume of air 21-Sep-20 06:32 AM I just happened to find what I needed in Amazon's indoor greenhouse section when I needed this thing 21-Sep-20 06:32 AM I see 21-Sep-20 06:33 AM It does a decent job when set up properly. It was designed to filter the air in my room 21-Sep-20 06:33 AM Damn chinchilla poof bath, yo! Gets EVERYWHERE 21-Sep-20 06:34 AM I wonder how much might a air particulate sensor cost 21-Sep-20 06:34 AM one that can measure the number of particles in air 21-Sep-20 06:34 AM but of course we're talking really fine particles here 21-Sep-20 06:35 AM This was the smallest one I could buy, too. Its sitting on a small bucket right now 21-Sep-20 06:35 AM It doesnt do much of a dent in my basement, as it's too large of an area 21-Sep-20 06:35 AM But it does help with a fume hood 21-Sep-20 06:36 AM I wonder how well this would work 21-Sep-20 06:36 AM https://www.ebay.com/itm/PMS5003-High-Precision-Laser-Dust-Sensor-Module-PM1-0-PM2-5-PM10-With-Fan-Top/153999063044 21-Sep-20 06:36 AM as a particulate matter sensor 21-Sep-20 06:36 AM for the purposes we want to use it for 21-Sep-20 06:37 AM This uses all typical HVAC parts sooooo.... 21-Sep-20 06:37 AM Probably not that hard depending how it works 21-Sep-20 06:38 AM I mean I do kinda want to quantify things like how well the filter works, the amount of suspended particulate matter in air, etc 21-Sep-20 06:38 AM cause even at 1 um feature sizes 21-Sep-20 06:38 AM dust can ruin your day 21-Sep-20 06:38 AM also, the quantification of things like this does matter 21-Sep-20 06:39 AM Wish they had the size listed 21-Sep-20 06:39 AM especially since you want to be coming up with like a set of specs, how to build instructions, etc 21-Sep-20 06:39 AM so that others can replicate the setup 21-Sep-20 06:39 AM and with as little guesswork and ambiguity as possible 21-Sep-20 06:40 AM Cuz I'd suppose you could just mount it in a duct, or even coupler. Punch a small hole for the wires 21-Sep-20 06:40 AM That's a 6 inch duct on my fan, and that's the smallest you can find 21-Sep-20 06:40 AM 8" is more available 21-Sep-20 06:41 AM perhaps, although it would be dependent on if you're looking at a cleanbox with laminar flow, or a larger soft cleanroom like in the photo 21-Sep-20 06:41 AM so those are yet to be worked out 21-Sep-20 06:41 AM I'm worried the chinchilla will fall in at times. It's quite a wide area. 21-Sep-20 06:42 AM I can easily reach in there to detach the blade and clean it. 21-Sep-20 06:42 AM grill of sorts then? 21-Sep-20 06:42 AM How big is that sensor? 21-Sep-20 06:42 AM let me search 21-Sep-20 06:42 AM 50×38×21 mm 21-Sep-20 06:44 AM Oh, it's set up so the top is an exhaust. So the fan stays clean, its pulling in from the filter. Ninja does like to put his face over the blast of air. 21-Sep-20 06:44 AM datasheet: https://www.aqmd.gov/docs/default-source/aq-spec/resources-page/plantower-pms5003-manual_v2-3.pdf 21-Sep-20 06:44 AM Oh, that's easy peasy to mount in a duct coupler. 21-Sep-20 06:44 AM I would mount it somewhere inside the work area 21-Sep-20 06:44 AM where I actually want to measure the particulate matter 21-Sep-20 06:45 AM Right off the fan output 21-Sep-20 06:45 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/20200921_094501-84FE6.jpg 21-Sep-20 06:45 AM You need one to couple any ducting, regardless. 21-Sep-20 06:45 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/20200921_094558-7E1E1.jpg 21-Sep-20 06:46 AM yup 21-Sep-20 06:47 AM Could make a few with that particulate sensor built into the couplers 21-Sep-20 06:47 AM yeah that might be one means for a laminar flow box of sorts 21-Sep-20 06:47 AM when you were in NYC Resistor, what sort of space did you folks have for a hackerspace? 21-Sep-20 06:48 AM You would be surprised how much of what you need for a clean room can be found in garden supply section 21-Sep-20 06:48 AM It was old loading wearhouses 21-Sep-20 06:48 AM Large, old, brick flats 21-Sep-20 06:48 AM Or old industrial manufacturing 21-Sep-20 06:48 AM Like old sewing farms and such 21-Sep-20 06:49 AM call me ambitious 21-Sep-20 06:49 AM but see, I am wondering if a sort of Elecrow or JLCPCB for chips could work 21-Sep-20 06:49 AM where a run of say, 20-50 chips might cost a designer a few hundred dollars 21-Sep-20 06:49 AM even in the upper hundreds would be considered cheap 21-Sep-20 06:49 AM where the process can be run and maintained by a hackerspace or volunteer group 21-Sep-20 06:52 AM Maybe. Then later on do a thingaverse for bare metal designs 21-Sep-20 06:52 AM Perhaps run a quarterly contest in forms of some weird design challenge 21-Sep-20 06:53 AM that brings me to the software tooling - one of the most popular VLSI design tools that's FOSS is Magic VLSI 21-Sep-20 06:53 AM You know what I regret??? 21-Sep-20 06:53 AM what? 21-Sep-20 06:53 AM Tossing all my old caller ID boxes 21-Sep-20 06:53 AM Didnt realize they had rare bell 103 spec afsk chips which can be used for ham radio 21-Sep-20 06:53 AM I'd looooove to see some 1200 baud afsk again!! 21-Sep-20 06:55 AM software methods for that? 21-Sep-20 06:55 AM like DAC/ADC and signal processing? 21-Sep-20 06:55 AM Or take all the fun out of it and just run fldigi on my netbook. 21-Sep-20 06:55 AM The fun is running on real hardware. I have some radio bbs stuff with dead afsk chip 21-Sep-20 06:56 AM fair 21-Sep-20 06:56 AM No way to fix it. Obsolete part for 20+ years 21-Sep-20 06:57 AM I mean you could probably recreate it if you could fab your own chip - I see a possible appeal there 21-Sep-20 06:57 AM designing and fixing stuff that parts no longer exist for 21-Sep-20 06:57 AM And there is also power drain competition 21-Sep-20 06:57 AM Like how little current drain can you design a 7-seg driver? 21-Sep-20 06:58 AM that brings me to the software tooling - one of the most popular VLSI design tools that's FOSS is Magic VLSI regardless, such an open sourced fab project would need good software tools 21-Sep-20 06:58 AM we should look at standardizing cell sizes, etc 21-Sep-20 06:58 AM like design rules that can be plugged into something like Magic VLSI 21-Sep-20 06:58 AM there are a few projects similar to this btw 21-Sep-20 06:58 AM https://libresilicon.com/index_en.html <-- this for example 21-Sep-20 06:58 AM they document their intended process steps, etc 21-Sep-20 06:58 AM but their project stopped short last year 21-Sep-20 06:58 AM https://www.gofundme.com/f/libresilicon-cleanroom-rent <-- they failed to raise cash for the rent of the HKUST cleanroom 21-Sep-20 06:58 AM which is why I am thinking that actually working with the process and tooling at a more DIY level might be the best approach 21-Sep-20 06:58 AM since fab facilities like this, even academic fabs are going to be expensive 21-Sep-20 07:03 AM You gave me an insane idea 21-Sep-20 07:03 AM I have an old Tandy TRS Model M100 that needs repair 21-Sep-20 07:03 AM Bad gate on one of the LCD driver logic chips leaving a solid line in one quadrant 21-Sep-20 07:03 AM Now I'm thinking.... 21-Sep-20 07:03 AM Modern chips take up so much less current.... 21-Sep-20 07:03 AM Why not replace all of them!? 21-Sep-20 07:05 AM effort, cost. etc? 21-Sep-20 07:05 AM also you should check this out 21-Sep-20 07:05 AM The thing already peaks 12 hours on AA cells 21-Sep-20 07:05 AM Sam Zeloof not only described a chip making process 21-Sep-20 07:05 AM he bought a wire bonding device 21-Sep-20 07:05 AM and 21-Sep-20 07:05 AM https://www.youtube.com/watch?v=mvZ1dJuvenw 21-Sep-20 07:06 AM Then again I could probably find someone that's already cloned the trs m100 mainboard.... 21-Sep-20 07:06 AM I just have no means of finding a new display 21-Sep-20 07:06 AM They used a custom fabbed 4 panel in parallel 21-Sep-20 07:07 AM spare parts online? 21-Sep-20 07:07 AM or is that an avenue that has been exhausted? 21-Sep-20 07:07 AM Oh it's all core logic. 7400 21-Sep-20 07:07 AM Oh the LCD has been exhausted for years 21-Sep-20 07:07 AM Turns out it's just a blown logic gate on the LCD controller. 21-Sep-20 07:07 AM 4 displays In parallel, and one output is always left high. Leaving a line of pixels across one quadrant 21-Sep-20 07:07 AM Iirc just a quad op amp dip 21-Sep-20 07:09 AM any equivalent parts? 21-Sep-20 07:10 AM Yeah. Easy. Still in use today 21-Sep-20 07:10 AM if it is a quad op-amp, you probably can find a suitable replacement 21-Sep-20 07:11 AM Yeah, a replacement isnt a problem. But I was gonna do a diagnostics tutorial & repair. 21-Sep-20 07:11 AM So it's on hold 21-Sep-20 07:11 AM ah 21-Sep-20 07:11 AM But with how little current modern chips use... 21-Sep-20 07:12 AM I mean it has something to do with the size of transistors, process technology and fabrication technique 21-Sep-20 07:12 AM Wonder if it's possible to recreate the device with modern parts and see how how long it lasts 21-Sep-20 07:12 AM which affects the power draw of these things 21-Sep-20 07:12 AM well you probably could 21-Sep-20 07:12 AM although I think for a hackerspace fab sort of process 21-Sep-20 07:12 AM Now I'm pondering the possibility of a clone 21-Sep-20 07:12 AM you're looking at like 21-Sep-20 07:12 AM 1990s process tech 21-Sep-20 07:12 AM But use an ips display 21-Sep-20 07:12 AM The whole schematic came with the trs m100, its online 21-Sep-20 07:13 AM ah 21-Sep-20 07:14 AM Back when your computer came with schematics and source code 21-Sep-20 07:14 AM You know more about it than I do... 21-Sep-20 07:14 AM Take a look at the trs m100 schematic and tell me... whatd that cost to do in FPGA? 21-Sep-20 07:14 AM https://archive.org/details/m100service 21-Sep-20 07:17 AM I just did a quick look 21-Sep-20 07:17 AM Not including display 21-Sep-20 07:17 AM I am pretty sure many of the chips could be reimplemented in a FPGA 21-Sep-20 07:17 AM Uses like 10 freaking encoders 21-Sep-20 07:17 AM It's mostly 7400 logic 21-Sep-20 07:18 AM even the 80c85 could be redone in a FPGA 21-Sep-20 07:18 AM oh you reminded me 21-Sep-20 07:18 AM Magic VLSI is made by open circuit design 21-Sep-20 07:18 AM and they have something called Qflow 21-Sep-20 07:18 AM that can take your RTL files from a FPGA design 21-Sep-20 07:18 AM and do routing and floorplanning for you 21-Sep-20 07:19 AM Nice 21-Sep-20 07:19 AM http://opencircuitdesign.com/qflow/ 21-Sep-20 07:19 AM DM.that to me plz 21-Sep-20 07:20 AM done 21-Sep-20 07:20 AM again all of this cool stuff depends on the implementation of a DIY process and tooling 21-Sep-20 07:20 AM which is really what you'll see come up a lot 21-Sep-20 07:20 AM tackling the actual fab steps 21-Sep-20 07:20 AM and how to translate CAD files to an actual thing that we can do litho with 21-Sep-20 07:20 AM interesting 21-Sep-20 07:20 AM so on Qflow's page 21-Sep-20 07:20 AM there's a custom 32 bit RISC-V microprocessor 21-Sep-20 07:20 AM and they use this process 21-Sep-20 07:20 AM https://www.xfab.com/technology/cmos/018-um-xh018/ 21-Sep-20 07:20 AM I kinda wonder how small of a feature size we could reliably achieve 21-Sep-20 10:54 AM hmm 21-Sep-20 10:54 AM I guess pretty smol 21-Sep-20 10:54 AM but dust kills it 21-Sep-20 10:58 AM hence the disucssion on particulate matter detection and soft cleanrooms and laminar flow cleanboxes? 21-Sep-20 10:58 AM I wonder if we could work on something that's like 21-Sep-20 10:58 AM Class 100 or Class 1000 at least? 21-Sep-20 10:58 AM something quantifiable 21-Sep-20 10:58 AM https://www.ebay.com/itm/PMS5003-High-Precision-Laser-Dust-Sensor-Module-PM1-0-PM2-5-PM10-With-Fan-Top/153999063044 21-Sep-20 10:58 AM I wonder if this is useful for particulate detection for semiconductor stuff 21-Sep-20 10:58 AM datasheet: https://www.aqmd.gov/docs/default-source/aq-spec/resources-page/plantower-pms5003-manual_v2-3.pdf 21-Sep-20 11:30 AM hmmmm I didn't even know 21-Sep-20 11:30 AM There are materials that cannot be wet etched, for example, SiC, GaN, TiC and diamond. These materials, can, however, be plasma etched. Some materials cannot be etched even by plasmas because no suitable source gas/volatile product combination exists. In that case, purely physical etching, known as ion milling or ion beam etching (IBE), can be used: argon ion bombardment will erode any material. Many solid-state laser garnets and magnetic materials (of the type Gd3Ga5O12 , gadolinium gallium garnet) are etched by ion milling. It is, however, difficult to find suitable non-eroding masking materials: if anything can be etched by argon bombardment, this applies to masking materials as well. Typical ion milling rates are 10–100 nm/min, an order of magnitude less than in plasma etching. 21-Sep-20 11:30 AM I wonder if that might be worth a shot 21-Sep-20 11:30 AM also it seems like with this process 21-Sep-20 11:30 AM you can guarantee that many materials Argon after that 21-Sep-20 11:30 AM runs away 21-Sep-20 02:05 PM https://www.youtube.com/watch?v=1kQUXpZpLXI 21-Sep-20 02:05 PM Some really interesting footage 21-Sep-20 04:22 PM cool! 21-Sep-20 04:50 PM @SleepyOwl Joyce I've got a pms5003 and recently took it into a clean-enough-room with overhead HEPAs and plastic skirts all around the room 21-Sep-20 04:50 PM oh? What did you measure? 21-Sep-20 04:50 PM or rather what did you manage to characterize? 21-Sep-20 04:51 PM Also azonenberg said he converted/calculated the output to equivalent ISO terms and said it was ISO 8 when no smokepocalypse going on, ISO 7 with 21-Sep-20 04:51 PM I measured all 0s 21-Sep-20 04:51 PM For pm1 , 2.5, and 10 21-Sep-20 04:51 PM But I was only in there a minute or so 21-Sep-20 04:52 PM I see 21-Sep-20 04:52 PM Not sure what azonenberg's method was 21-Sep-20 04:52 PM Presumably some long term sampling 21-Sep-20 04:52 PM hmmmm 21-Sep-20 04:52 PM I wonder what's the smallest particles these things can measure 21-Sep-20 04:52 PM especially for lower ISO rating cleanrooms 21-Sep-20 04:52 PM I mean they are measuring PM1 , 2.5, and 10 21-Sep-20 04:52 PM I would kinda like to get an idea of the smallest particles it can measure 21-Sep-20 04:52 PM in the datasheet 21-Sep-20 04:52 PM it does say that 21-Sep-20 04:52 PM Range of measurement 0.3~1.0; 1.0~2.5; 2.5~10 um 21-Sep-20 04:52 PM if mean if it does what it says on the tin 21-Sep-20 04:52 PM you could probably really use it as a measurement reference for a class 100 cleanroom for example 21-Sep-20 11:21 PM hmmm 22-Sep-20 12:43 AM hmmm? 22-Sep-20 12:46 AM mmhm. 22-Sep-20 12:47 AM https://giphy.com/gifs/owl-head-hcRjyAXqihoqY 22-Sep-20 01:24 AM anyway 100nm features are probably achieveable 22-Sep-20 01:24 AM quantifying the dust is important though 22-Sep-20 01:47 AM I found this graph 22-Sep-20 01:47 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-CEF3B.png 22-Sep-20 01:47 AM definitely, if you have more dust, the more dense the chip is, the worse the yield is expected to be 22-Sep-20 01:47 AM and its interesting what they do in the cleanrooms in terms of particulate contamination 22-Sep-20 01:47 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-7B9CF.png 22-Sep-20 02:48 AM yeah 22-Sep-20 05:07 AM this might be a better sensor: https://cdn.sparkfun.com/assets/2/d/2/a/6/Sensirion_SPS30_Particulate_Matter_Sensor_v0.9_D1__1_.pdf 22-Sep-20 05:07 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-E8718.png 22-Sep-20 05:07 AM it has its own PM0.5 measurement 22-Sep-20 05:07 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-B875F.png 22-Sep-20 06:18 AM right now I am speaking to someone who's actually worked with cleanroom design and stuff like that 22-Sep-20 06:18 AM to try and get an idea for reference measurements, etc 22-Sep-20 06:55 AM so here's what I got from a brief conversation 22-Sep-20 06:55 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-4E8F3.png 22-Sep-20 07:40 AM so gloveboxes might have to be a thing 22-Sep-20 07:40 AM with airlocking? 22-Sep-20 07:40 AM so for context I was trying to push the envelope and see if we could go down to Class 10 22-Sep-20 07:40 AM but Class 100 seems like what might be achievable 22-Sep-20 07:40 AM or somewhere between ISO class 4 and class 5 22-Sep-20 07:44 AM the part about operators is very important 22-Sep-20 07:44 AM It is extremely easy for one person to ruin a clean room. Ones students have access ato at universities, for example... 22-Sep-20 07:45 AM yeah - that's why gloveboxes with airlocks might be a thing to consider here 22-Sep-20 07:45 AM got to separate the grubby hoomans from the processes 22-Sep-20 07:45 AM inb4 we get ideas for 22-Sep-20 07:45 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/zEMuVUliEIn8d3-KlV-vYR5J_k4Wlg3_M-rNEKVrmt-B5AA6.png 22-Sep-20 07:45 AM AMHS 22-Sep-20 07:45 AM although I wonder if a sealed transport box between processes might work 22-Sep-20 07:45 AM where wafers can be put in and moved from one isolation box to another 22-Sep-20 07:56 AM that is a thing 22-Sep-20 07:56 AM they are called FOUPs 22-Sep-20 07:56 AM https://en.wikipedia.org/wiki/FOUP 22-Sep-20 07:59 AM yeah, I remember seeing those, though my concern would be like, how a DIY process can ensure cleanliness between processes, especially when loading into certain processes that might be unwieldy to enclose, like plasma etching, sputtering/deposition, etc 22-Sep-20 07:59 AM the whole process from head to tail has to be taken into consideration to minimize contamination 22-Sep-20 07:59 AM https://en.wikipedia.org/wiki/SMIF_(interface) <-- might be a thing for the smaller wafers? 22-Sep-20 08:15 AM one more thing I didn't think about 22-Sep-20 08:15 AM wafer dicing 22-Sep-20 08:15 AM once you're done making chips, how do your dice them up and bond to your package? 22-Sep-20 08:23 AM dicers seem to show up surplus time to time for not a ton of money... just kinda rarely 22-Sep-20 08:23 AM Another way is if your wafer has the correct alignment you can sheer it and get pretty straight lines. Just pattern the wafer with your dies far enough apart to have a good chance at getting a clean line between them. 22-Sep-20 08:25 AM yeah, though this would be one of the processes with a lot of dust generation too 22-Sep-20 08:27 AM Well generally by the point your dicing you have placed a protective insulator layer across the top of the die and have only the metal of your bonding pads visible. Meaning once they are diced you can safely and carefully wash them. At-least this is how I assume they do it in industry? Correct me if I'm wrong. 22-Sep-20 08:28 AM I mean yeah, but in the context of a enclosed DIY fab space 22-Sep-20 08:28 AM I mean particulate is really an issue in-between steps as far as I know. 22-Sep-20 08:28 AM well sure don't keep your dicing machine in the cleanroom space. 22-Sep-20 08:28 AM But its also generally a wet cut 22-Sep-20 08:28 AM so not much gets airborn 22-Sep-20 08:28 AM the process will likely release a lot of dust if you're trying to keep your real estate small 22-Sep-20 08:28 AM well true 22-Sep-20 08:29 AM I have seen dicing machines fully enclosed to small desktop manual tools 22-Sep-20 08:29 AM I see 22-Sep-20 08:29 AM obviously the small desktop machines are most accessible for the home gamer 22-Sep-20 08:29 AM some are pretty manual, some programmable for distances and servo controlled, etc 22-Sep-20 08:29 AM building one probably also wouldn't be out of the question, 22-Sep-20 08:29 AM its objectively the least precise process lol 22-Sep-20 08:30 AM yeah 22-Sep-20 08:30 AM just have to have really good high speed bearings and a nice brushless high speed motor for the dicing blade 22-Sep-20 08:30 AM but dicing blades are cheap and available surplus... 22-Sep-20 08:30 AM surplus store near me probably has/had a few hundred of them 22-Sep-20 08:32 AM hmmm 22-Sep-20 08:32 AM I wonder how many processes and machines are going to be DIYed at this point heheheh 22-Sep-20 08:32 AM and how much real estate is expected 22-Sep-20 08:35 AM Its honestly time vs money, it always has been :) 22-Sep-20 08:35 AM I suppose your personal skills and knowledge play in there at some point too 22-Sep-20 08:35 AM yeah 22-Sep-20 08:35 AM although at this point I can do none of the processes on my own where I am, so the best I can do is to discuss the processes, source materials, etc for others who are doing these sort of thing 22-Sep-20 08:36 AM I have been trying to flush out the cruff of my vintage computing hobby to try and get organized here to try and continue on semiconductor stuff... I have a lot of the bits and pieces. Just need to get the rest of the bits and pieces and throw it all together. 22-Sep-20 08:36 AM ah cool 22-Sep-20 08:37 AM Well, depending on where your located you can also colaberate and share workspace 22-Sep-20 08:37 AM Singapore is NOT a good place for this xD 22-Sep-20 08:37 AM Real estate is expensive, even common chemicals like 99% ethanol are restricted and may need licenses to keep 22-Sep-20 08:37 AM that sort of thing 22-Sep-20 08:37 AM Ah jeez, well if your ever in the US near NY area, your always welcome by my mess lab 22-Sep-20 08:37 AM But doing reading/learning of processes is definitely a large part of the battle 22-Sep-20 08:38 AM I recall there being a hackaday article about some dude building a saw for cutting alumina substrate. 22-Sep-20 08:39 AM this is why I am thinking of sending N00N my wafers, I realize I can do nothing with them, and they're sitting since 2013 22-Sep-20 08:39 AM From an old HDD motor and ebay'd semiconductor dicing blade. 22-Sep-20 08:39 AM There is a fair bit of wafer tinkering you can do with very little. I believe it was Jeri Ellsworth? who did some transistors at home with nothing more than some wafers, some glassware for chemicals, and like tape... 22-Sep-20 08:40 AM yeah I remember 22-Sep-20 08:40 AM Yep 22-Sep-20 08:40 AM @rfs lol actually a hard drive motor is a pretty good idea 22-Sep-20 08:40 AM I have so many dead hard drives... hrmmm 22-Sep-20 08:40 AM I had a bigger lab space back in 2013 22-Sep-20 08:40 AM but now I am lucky to keep my electronics workbench and have servers 22-Sep-20 08:40 AM that being said, I think the Emulsitone films Jeri uses aren't produced anymore 22-Sep-20 08:41 AM Fair heh. I just got a new (free) rack that I can now rack up a bunch of the random servers I have kicking around 22-Sep-20 08:41 AM ah 22-Sep-20 08:41 AM well where one product dies generally some form of replacement exists 22-Sep-20 08:41 AM Or there is a way to make it 22-Sep-20 08:42 AM I haven't been able to find any replacement for those to the best of my knowledge 22-Sep-20 08:42 AM the emulsitone company index page itself 404s now 22-Sep-20 08:42 AM http://www.emulsitone.com/alphacatalog.htm <-- this is one of the online pages 22-Sep-20 08:42 AM MSDS: for one of the films http://www.emulsitone.com/tafMSDS.htm 22-Sep-20 08:45 AM I'm just happy I was able to get rid of my old hobby rack. 22-Sep-20 08:45 AM It was taking up space in my parents garage 22-Sep-20 08:45 AM http://www.emulsitone.com/bsifii.html <-- borosilicafilm 22-Sep-20 08:45 AM its MSDS: http://www.emulsitone.com/bsifiimsds.html 22-Sep-20 08:45 AM Ah jeez, well if your ever in the US near NY area, your always welcome by my mess lab @Conmega I'm ngl, I am thinking of moving out to the Americas lol 22-Sep-20 08:45 AM so that I can get involved in things like this and beyond 22-Sep-20 08:45 AM on the real estate side 22-Sep-20 08:45 AM @Foxx said that for NYC Resistor, they used old warehouses for their hackerspace 22-Sep-20 08:45 AM I wonder if finding some lowish cost space for people interested to do this might be possible 22-Sep-20 08:45 AM especially if you're looking at a sort of hackerspace fabrication for small batches of chips 22-Sep-20 08:45 AM I admit - I am kinda ambitious and I want to see how far the envelope can be pushed on DIY processes 22-Sep-20 08:52 AM I think the US is great for people looking for making up big workshops and doing fun stuff with tons of surplus equipment... We have other problems... of-course... but yea 22-Sep-20 08:52 AM The US is massive though, there are places you can get a 3,000 sqft commercial building for what I paid for my house up here in NY state 22-Sep-20 08:52 AM but that's also generally more rural, meaning less high paying professional jobs 22-Sep-20 08:52 AM its a lot of decision making and such 22-Sep-20 08:53 AM though I do think that even hitting something like a 0.18 um or 0.13 um process would be a big step for makers and hackers wanting to get involved with chips, etc 22-Sep-20 08:53 AM its a lot of decision making and such oh yeah of course 22-Sep-20 08:53 AM I am probably, if I stay here at my house, going to get a shop built on the back of my property 22-Sep-20 08:53 AM but hey, if you ever want to visit, see what its like, when things are back to more normalish your always welcome to visit! its nice to have like minded people around :) 22-Sep-20 08:55 AM Sure haha 22-Sep-20 08:56 AM But yea, my dream has always been to get a bunch of people together and go in on an abandoned mall, basically each store is like a lab for something, biology, chemistry, cleanroom, whatever 22-Sep-20 08:56 AM would be really cool 22-Sep-20 08:56 AM Always heard storys about what places like bell labs was like back in the day and its amazing 22-Sep-20 08:57 AM I just believe in a lot of the hacker ideals, especially with the sort of tech we deal with and use 22-Sep-20 08:57 AM I have an on hold project atm 22-Sep-20 08:57 AM called verifiable computing project 22-Sep-20 08:57 AM goal was to use FPGAs; synthesize and run RISC-V processors and things like that 22-Sep-20 08:57 AM capable of running Linux 22-Sep-20 08:57 AM this is in response to things like Intel ME and other co-processors in chipsets 22-Sep-20 08:57 AM so I mean chip fab is still not exactly within most of our reaches 22-Sep-20 08:57 AM so FPGAs are just next best 22-Sep-20 08:57 AM but I now I realize there's so many folks interested in this 22-Sep-20 08:57 AM that maybe a more formalized process, etc might lead to something useful 22-Sep-20 08:57 AM and microfabrication would open up other things for citizen science too 22-Sep-20 08:57 AM like for instance, microfabricated single photon sources 22-Sep-20 09:02 AM Heh your talking to one of those people btw :P 22-Sep-20 09:02 AM https://spoolqueue.com/new-design/fpga/migen/litex/2020/08/11/acorn-cle-215.html 22-Sep-20 09:04 AM Cool! 22-Sep-20 09:04 AM I am a big fan of RiscV and hope it goes far. I am waiting to receive my Microchip Icicle to play around with that too <3 22-Sep-20 09:04 AM also can I make a suggestion? 22-Sep-20 09:04 AM would it do good to have a dedicated #electronics channel? 22-Sep-20 09:07 AM Hrm. I believe something may have been discussed in the past. I think we generally try to stick to physics/sciency stuff in the server. But I don't think there is harm discussing stuff like this here or in general or anything. I just think we are the minority in this server in wanting to discuss FPGA for CPU design and not like data acquisition systems for sensor arrays ;) 22-Sep-20 09:07 AM Definitely something that we can discuss though. I'll mention it in the admin channel lol 22-Sep-20 09:08 AM but in any case 22-Sep-20 09:08 AM how's the performance of your RISC-V processor on FPGA? 22-Sep-20 09:08 AM I was thinking of getting some Kintex dev board for this 22-Sep-20 09:08 AM and you can probably guess why it's on hold 22-Sep-20 09:10 AM I am using the litex libraries which I think uses the Vexriscv core? I haven't pushed anything but I believe people can get 100+Mhz out of the cores on this Artix-7 chip 22-Sep-20 09:10 AM These little SQRL Acorns were a thing for crypto currency mining that kinda flopped.. so they are like 200 dollar chips for 60 bucks... 22-Sep-20 09:10 AM really good bang for buck and basically the largest FPGA fabric chip usable with the free Vivado tools 22-Sep-20 09:11 AM right! 22-Sep-20 09:11 AM Oh and its speed grade 3, the top one 22-Sep-20 09:11 AM I forgot about that 22-Sep-20 09:12 AM So its a great entry 22-Sep-20 09:12 AM But... not much I/O 22-Sep-20 09:12 AM I forgot that Vivado is license locked 22-Sep-20 09:12 AM basically only a handful, 12, besides the PCIe x4 Gen2 22-Sep-20 09:12 AM So its not a greatttt board for I/O but for RiscV processor playing with its great 22-Sep-20 09:13 AM I see 22-Sep-20 09:13 AM I have made that board I show in my blog in hopes to get PCIe root complex designed and going. 22-Sep-20 09:13 AM So I can attach devices to it and use them under linux 22-Sep-20 09:13 AM and hopefully give a cheaper linux development enviroment than a 500 dollar dev board 22-Sep-20 09:13 AM for PCIe that is 22-Sep-20 09:14 AM what if, hear me out, riscv is not such a nice architecture? 22-Sep-20 09:14 AM but I am either dumb or dumb because I can't figure out why it refuses to link train 22-Sep-20 09:14 AM Please explain? its better in pretty much every metric compared to most other archs... Code density, sanity of instructions, not having a 5 mile long errata... 22-Sep-20 09:14 AM Single cycle execution with things architected correctly like barrel shifter... 22-Sep-20 09:15 AM So you have consistently known execution times which matters a LOT in realtime applications 22-Sep-20 09:15 AM oops 22-Sep-20 09:15 AM but I looked through your site 22-Sep-20 09:16 AM Its one thing Microchip/Microsemi touts with their RiscV FPGA SoCs 22-Sep-20 09:16 AM lol yea you can't escape the bot on this server sometimes 22-Sep-20 09:16 AM I didn't know you actually ARE the mainframe "kid" 22-Sep-20 09:16 AM lol... yeaaa that be me 22-Sep-20 09:16 AM he is 22-Sep-20 09:16 AM but now he's a grown man 22-Sep-20 09:16 AM with beard and stuff 22-Sep-20 09:17 AM though unofficially I am now the mainframe engineer. finally got a promotion from technician 22-Sep-20 09:17 AM and I thought a half filled 42U rack of servers behind me was power hungry enough and all 22-Sep-20 09:17 AM Congrats? 22-Sep-20 09:17 AM its not announced to my team yet... but its done and done heh 22-Sep-20 09:17 AM I see 22-Sep-20 09:17 AM you work for IBM now - Mainframe Kiddo? 22-Sep-20 09:18 AM Its really great to be an engineer without having to go and spend time at college like everyone told me I absolutely had to 22-Sep-20 09:18 AM I have for nearly 5 years yes 22-Sep-20 09:18 AM ok kiddle, y not on IBMs transmon QC stuff 22-Sep-20 09:18 AM I don't make enough money to afford the mortgage I've had for 3-4 years with just doing talks :P 22-Sep-20 09:19 AM the moneyz? 22-Sep-20 09:19 AM ok 22-Sep-20 09:19 AM I actually applied for Quantum, got offered a position but Z wanted to keep me more 22-Sep-20 09:19 AM hm hm 22-Sep-20 09:19 AM I may still consider Q in the future 22-Sep-20 09:19 AM but for now Z is still my home 22-Sep-20 09:19 AM where the future is 22-Sep-20 09:20 AM Well... I am sure it will take off, its still in its growing pain stages 22-Sep-20 09:20 AM its kinda crazy over there from the sounds of it 22-Sep-20 09:20 AM yep 22-Sep-20 09:20 AM I like crazy in my house, not at work which I bring home 22-Sep-20 09:20 AM lot's of cocain I hearf 22-Sep-20 09:20 AM heard 22-Sep-20 09:20 AM cocain and witches 22-Sep-20 09:20 AM lol I don't think so but maybe, I dunno how the physicists do it 22-Sep-20 09:21 AM Oh boy this whole Q thing is a deep rabbit hole lol 22-Sep-20 09:21 AM and coders pair coding in knee socks 22-Sep-20 09:21 AM Remember, math, not even once right @GigaSquirrel 22-Sep-20 09:21 AM knee socks are good for programming 22-Sep-20 09:21 AM you can't tell anyone otherwise 22-Sep-20 09:21 AM you're right, they also do math, dang, those bstrds 22-Sep-20 09:21 AM god 22-Sep-20 09:21 AM I remember back in highschool on pi day, stupid word puzzle games... 22-Sep-20 09:21 AM one of the... lets say dimmer bulbs in the class... got an equation wrong and got the word answer for one of them which he thought was right and said the sentence out loud in excitement "COOL KIDS LIKE METH" then immediately realized his mistake... 22-Sep-20 09:21 AM lol 22-Sep-20 09:23 AM lol 22-Sep-20 09:23 AM in any case, what I was thinking for verifiable computing project 22-Sep-20 09:23 AM was to take a Pinebook 22-Sep-20 09:23 AM remove the motherboard 22-Sep-20 09:23 AM and replace it with a FPGA RISC-V based system 22-Sep-20 09:23 AM the reason I chose the Kintex-7 is really for the higher speed fabric 22-Sep-20 09:23 AM but now you reminded me about the license thing for Vivado 22-Sep-20 09:27 AM p h y s i c s 22-Sep-20 09:27 AM Gotta brag again I’ve been in a d wave computer 22-Sep-20 09:28 AM brb let me invite my physicist friend 22-Sep-20 09:28 AM yes, I would love to build a RiscV laptop myself to just build one 22-Sep-20 09:28 AM I was frankly thinking of using the Microchip Polarfire SOC 22-Sep-20 09:28 AM I seen those Microsemi parts, but using the SoC kinda defeats the point 22-Sep-20 09:28 AM since it has a lot of FPGA fabric and the 4 600Mhz RiscV cores already 22-Sep-20 09:29 AM in 22-Sep-20 09:29 AM @idmb was it cold inside? 22-Sep-20 09:29 AM Hrm, I suppose maybe, but depending on the way you archetect it and the capabilities of microsemi's tools you can spin up a core, load it into your fabric and play with it 22-Sep-20 09:29 AM I also just want a RiscV laptop because I think it would be cool ^^ 22-Sep-20 09:29 AM Cooler than an Arm laptop anyway heh 22-Sep-20 09:30 AM It would, but again, VCP would mean that I don't have to trust that the core is doing only what it says it does 22-Sep-20 09:30 AM and nothing more 22-Sep-20 09:30 AM Another chassis I was considering using for a platform like that was the MNT Reform, if he will sell the chassis only to me 22-Sep-20 09:31 AM It was off 22-Sep-20 09:31 AM Sure, I definitely think it would be nice to say make it somewhat modular 22-Sep-20 09:31 AM that MNT reform looks clunky 22-Sep-20 09:31 AM I mean if you make a processor board that fits on a dimm and have an open design standard for that interface you can drop whatever in 22-Sep-20 09:31 AM they did that with the MNT Reform but the I/O is pretty limited in terms of PCIe lanes 22-Sep-20 09:31 AM I don't like it 22-Sep-20 09:31 AM But the MNT Reform itself is extremely well designed I think 22-Sep-20 09:31 AM is it larger and chunkier sure, but I would like to see a similar sized team of people make something sleeker in the time period 22-Sep-20 09:31 AM I respect the blood sweat and tears he put into that machine 22-Sep-20 09:32 AM @idmb they're pretty expensive annealers, right 22-Sep-20 09:32 AM And I think its really well built quality wise 22-Sep-20 09:32 AM is it a macbook air, no lol 22-Sep-20 09:32 AM but the price point and build materials and design I think are extremely impressive 22-Sep-20 09:33 AM I see, I mean it probably isn't that more clunky than Bunnie Huang's Novena Laptop 22-Sep-20 09:33 AM which was based on NXP i.MX6 (iirc) 22-Sep-20 09:33 AM I think we all have our own tastes too 22-Sep-20 09:33 AM Yes annealers 22-Sep-20 09:34 AM I personally like old chunky laptops like the old stinkpads thinkpads 22-Sep-20 09:34 AM They do do stuff 22-Sep-20 09:34 AM I have near like 50 old thinkpads lol 22-Sep-20 09:34 AM They breed I warn you all 22-Sep-20 09:34 AM Most physics quantum computations are generally doing problems that can be better done classically with a couple valid assumptions, according to the theorists I’ve heard 22-Sep-20 09:34 AM (The ones currently proposed) 22-Sep-20 09:35 AM its never been about why, its been about if we can ;) at-least I like it that way hehe 22-Sep-20 09:35 AM I meant the ones people talk about 22-Sep-20 09:35 AM there are plenty of whys 22-Sep-20 09:36 AM I remember the older Thinkpads, I generally prefer the designs of the older ones than my current ThinkPad L460 22-Sep-20 09:36 AM in terms of keyboard layout and all 22-Sep-20 09:36 AM idk there was some talk about some US gov project funding a bunch of quantum computer research because they had something they claimed would take an eternity classically and be quick quantum, then this person showed you could reduce the problem to like a week on a regular classical supercomputer 22-Sep-20 09:36 AM and did it 22-Sep-20 09:37 AM yeah I mean an annealer would pretty decent fidelity to be universal 22-Sep-20 09:38 AM Oh sure, there are lots of things we don't need quantum computers for that people think we can use it for... Javascript seems to be very similar in that sense shivers 22-Sep-20 09:38 AM since no error correction is known for them, right? 22-Sep-20 09:38 AM @SleepyOwl Joyce yes, many people like the older keyboards and design. The current T series isn't horrible IMO but yea. My ideal cute portable formfactor is like the old X40... wonderful little thing 22-Sep-20 09:38 AM gate based is easier "easier" theoretical 22-Sep-20 09:38 AM I remember being some error correction algorithms proposed in literature 22-Sep-20 09:39 AM I wish 4:3 screens were still available... :( best we have right now is 3:2 22-Sep-20 09:39 AM I remember being some error correction algorithms proposed in literature @SleepyOwl Joyce I mean for quantum annealers 22-Sep-20 09:39 AM Using an old T60 with a nice IPS 4:3 screen then back to 16:9 it feels like I have so little room :( 22-Sep-20 09:39 AM not for gate baed QCs like the IBM, google etc ones 22-Sep-20 09:40 AM right, oops 22-Sep-20 09:40 AM my uncle did his masters on quantum error correction in 1999 22-Sep-20 09:40 AM oh! 22-Sep-20 09:42 AM I'm pretty sure quantum computing has been in the works since quantum mechanics has existed 22-Sep-20 09:43 AM I beg to differ 22-Sep-20 09:43 AM I would say probably only in the 60s and 70s after the Feynman lectures on Computation and stuff like that 22-Sep-20 09:43 AM did the concept of quantum computation really start to become a thing 22-Sep-20 09:43 AM if my understanding is right 22-Sep-20 09:45 AM yeah the first explicite QC paper from feynman was 1982 22-Sep-20 09:46 AM that's not a great metric 22-Sep-20 09:46 AM cause that excludes anything that didn't get published 22-Sep-20 09:46 AM yes of course 22-Sep-20 09:46 AM he had such ideas earlier and of course he was not alone 22-Sep-20 09:47 AM hence I put it to around the 60s and 70s 22-Sep-20 09:47 AM where quantum mechanics was a thing wayyy before that 22-Sep-20 09:48 AM yuri manin also published sbout such stuff 22-Sep-20 09:48 AM with the whole discussion on the nature of light and stuff like that back in the 1800s 22-Sep-20 09:48 AM I mean explecitely using quantum systems to simulate other stuff 22-Sep-20 09:49 AM I am saying that that's when quantum mechanics started to come in the picture 22-Sep-20 09:49 AM in the whole story of humanity's exploration into physics 22-Sep-20 09:49 AM but I think what you're talking about in the 60s/70s/80s 22-Sep-20 09:49 AM is when people presented fleshed-out ideas about it 22-Sep-20 09:49 AM well the explicite QC computing stuff is later 22-Sep-20 09:50 AM I am still amazed people can work on stuff like this when I am still working on being an adult 22-Sep-20 09:50 AM more like, when questions such as: "can classical computation simulate the quantum nature of things" started to be thought about 22-Sep-20 09:50 AM but it went hand in hand with reversible computing (fredkin, toffoli et al) 22-Sep-20 09:50 AM like if you look at the '79 paper, it's a fully mathematical "here is a quantum turing machine" 22-Sep-20 09:50 AM not "oh huh, I wonder what a quantum turing machine would be like" which must have happened wayyyyyyyyyyyy before that 22-Sep-20 09:51 AM which the idea of even a classical turing machine only existed in the 1930s 22-Sep-20 09:52 AM formulation, not idea 22-Sep-20 09:52 AM you mean those benioff 22-Sep-20 09:52 AM papsers? 22-Sep-20 09:52 AM for qtm 22-Sep-20 09:53 AM Yeah, 1979: "The Computer as a Physical System: A Microscopic Quantum Mechanical Hamiltonian Model of Computers as Represented by Turing Machines " 22-Sep-20 09:53 AM I am just saying that computation (as in turing machines) as a concept probably didn't really take place until the 1900s at least 22-Sep-20 09:53 AM that's not "here's an idea I'm working on" that's "here it is" 22-Sep-20 09:53 AM about a century after we started exploring the quantum nature of the universe 22-Sep-20 09:54 AM oh 22-Sep-20 09:54 AM I consider "quantum mechanics" to be like the early 1900s formulation 22-Sep-20 09:54 AM copenhagen and all 22-Sep-20 09:54 AM I consider that period to be where physicists were arguing about the interpretation of the findings in experiments, etc 22-Sep-20 09:55 AM not the "oh hey here's some discreteness" 22-Sep-20 09:55 AM bbl (fooood) 22-Sep-20 09:55 AM I would think that the experiments done to test the nature of things were much earlier than 1900s 22-Sep-20 09:55 AM the interpretation of findings is essentially what QM is? 22-Sep-20 09:56 AM I mean at this point we're just arguing semantics and language 22-Sep-20 09:56 AM Because there were already experiments in the field before the whole idea of the Copenhagen Interpretation of QM 22-Sep-20 09:56 AM and even till today, no one is really settled on the interpretation of what we observe in experiments 22-Sep-20 09:58 AM Depends how you mean that? 22-Sep-20 09:58 AM in the sense that everything is quantum, even the particles that make up you and I 22-Sep-20 09:58 AM so one of the things like what my friend is working on and trying to understand 22-Sep-20 09:58 AM is how QM and the macroscopic world intertwine 22-Sep-20 09:58 AM and we're not exactly sure what interpretation makes the best sense 22-Sep-20 09:58 AM Like if you go with Copenhagen, you're going to deal with the measurement problem and all - "What does it mean to have measured?" You go with Many Worlds interpretation, and you get another bunch of questions 22-Sep-20 10:03 AM My experience is most physicists aren't too fussed until someone proposes a method of testing stuff like that 22-Sep-20 10:07 AM At which point they get physical 22-Sep-20 10:07 AM But I digress. I define the time where experiments exploring the quantum nature of things were conducted as the time quantum mechanics started to take shape 22-Sep-20 10:07 AM Interpretation of the results we get in experiments and how they all tie together are just a part of the process 22-Sep-20 10:15 AM I think more of the "it's all quantum, always has been" findings 22-Sep-20 10:15 AM rather than the "oh look, this one specific thing is quantized" 22-Sep-20 10:15 AM but sure 22-Sep-20 11:27 AM https://twitter.com/FossiFoundation/status/1308420839092768769 22-Sep-20 10:59 PM walks like a quantum mechanician 24-Sep-20 04:55 PM you know i was thinking last night, an open source power mosfet project could have more utility than an open source 1980's vlsi semiconductor process 24-Sep-20 04:55 PM power mosfets can get away with large feature sizes 24-Sep-20 04:55 PM and still be commercially viable 24-Sep-20 05:23 PM a lot less fun though (and a lot less room to innovate too. like a hacker vlsi could let people play with clockless processors, or analog neuromorphic chips, etc) 24-Sep-20 06:29 PM Oh I disagree, I think the field of germanium and carbide power mosfets is under explored 24-Sep-20 06:29 PM They've been silicon for so long because everybody was optimizing for price but nowadays there are other optimizations to consider like switching speed, and minimizing the amount of bulk capacitance necessary in an automobile 24-Sep-20 06:31 PM that's fair! 24-Sep-20 06:31 PM feature size on those is pretty enormous, isn't it? relatively speaking anyhow 24-Sep-20 06:34 PM though I would say that SiC power FETs have been out for a fair while now 24-Sep-20 06:34 PM and there's already GaN based power FETs out there 24-Sep-20 06:34 PM of course, it would be much cool to explore transistor switching speeds and things like that 24-Sep-20 06:36 PM What kind of dielectrics are used with GaN? 24-Sep-20 06:36 PM I assume SiC uses SiO2 24-Sep-20 06:38 PM I am not sure 24-Sep-20 06:38 PM seems like 24-Sep-20 06:38 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-D6CB8.png 24-Sep-20 06:38 PM https://www.researchgate.net/publication/263056786_Recent_advances_on_dielectrics_technology_for_SiC_and_GaN_power_devices <-- paper 24-Sep-20 06:38 PM there's a whole table of insulators and stuff 24-Sep-20 06:38 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-9D326.png 24-Sep-20 06:45 PM on the topic of alternative stuff, I've been binging on literature about metal–semiconductor field-effect transistors (MESFETs) lately. simpler overall process (no need for insulator under gate, or polysilicon gate) and faster switching speed, but increased gate leakage and difficult to make complementary. but still, might be interesting from a DIY point of view especially if you embrace the gate leakage aspect and use it for analog chips 24-Sep-20 06:47 PM Cool! 24-Sep-20 07:24 PM One thing that would help tremendously in the power electronics industry is building some kind of charge shuttling gate driver right into the transistor 24-Sep-20 07:24 PM Which would eliminate the need for special high side gate driving circuitry 24-Sep-20 07:24 PM and also allow for much faster switching speeds since you don't have to worry about inductance between the driver and the transistor 24-Sep-20 07:24 PM I think you could actually do this with a photogate 24-Sep-20 07:24 PM You can use an LED to illuminate the photogate, 24-Sep-20 07:24 PM And that would provide very good voltage isolation 24-Sep-20 07:24 PM I know this is used on some power thyristors but I've never seen it on a mosfet before 24-Sep-20 07:39 PM Aren't those made? 24-Sep-20 07:39 PM GPU's and computer mobo's seem to use something similar in the DC-DC's feeding Vcore on those. 24-Sep-20 07:39 PM mosfet+gatedriver and sometimes some additional logic in a single package. 24-Sep-20 07:44 PM For low voltage yes 24-Sep-20 07:44 PM But not 1200V power transistors 24-Sep-20 07:44 PM It would be very well received to have such a thing in the solar and EV world 26-Sep-20 01:15 AM henlo 26-Sep-20 01:15 AM @SleepyOwl Joyce I think we should try a FPGA instead 26-Sep-20 01:15 AM with a resilient bus hierarchy insPirated by radiation hard architectures 26-Sep-20 01:15 AM to handle the poor yield - if a block is broken by dust sht simply don't use it 26-Sep-20 01:15 AM and it's much much easier to design 26-Sep-20 02:53 AM Sure, the thing though is that no matter what 26-Sep-20 02:53 AM FPGA designs aren't going to be as fast as any ASIC 26-Sep-20 02:53 AM Although open source FPGA sounds cool 26-Sep-20 03:24 AM yes of course, but I think it's easier to get one logic block working compared with a complete microprocessors core 26-Sep-20 03:24 AM I mean it also would make sense to start with simple logic gates, I guess 26-Sep-20 03:24 AM but a provable working CLB would get some attention 26-Sep-20 03:26 AM Well yeah - although if you read certain materials I shared 26-Sep-20 03:26 AM RAM might be a good test of a process 26-Sep-20 03:26 AM provable in the meaning of sending some of them to community members willing to test them 26-Sep-20 03:26 AM it's easier to identify faults and failures with RAM, etc 26-Sep-20 03:27 AM yeah you're right 26-Sep-20 03:27 AM provable in the meaning of sending some of them to community members willing to test them oh I certainly don't mind testing stuff out 26-Sep-20 03:27 AM I mean for the PR attention 26-Sep-20 03:27 AM even if the stuff is pretty crappy compared with commercial it would gain attention to the whole project 26-Sep-20 03:31 AM oh yeah - although I would say getting the process done right, and releasing the tools, libraries, etc so that people can make stuff with whatever processes we're developing is also an important thing 26-Sep-20 03:31 AM so one working - even if slow - CLB would somehow indicate the scalability of the project 26-Sep-20 03:31 AM imho 26-Sep-20 03:32 AM oh yeah - although the really important and interesting parts are in the feature size, yield, etc data 26-Sep-20 03:33 AM yes, more important actually 26-Sep-20 03:33 AM but for PR it's different imho 26-Sep-20 03:34 AM ¯\_(ツ)_/¯ 26-Sep-20 03:34 AM As advertised on TV! 26-Sep-20 03:34 AM yes 26-Sep-20 03:34 AM "they build their own FPGA in the shed!!!!" 26-Sep-20 03:34 AM even of only 1 of 100 CLBs works for a couple of minutes 26-Sep-20 03:37 AM looks good on the outside... 26-Sep-20 03:37 AM on the inside... ahem 26-Sep-20 03:37 AM check this out though 26-Sep-20 03:37 AM DIY FPGA design from 7400 series logic 26-Sep-20 03:37 AM http://blog.notdot.net/2012/10/Build-your-own-FPGA 26-Sep-20 03:37 AM hmmmm 26-Sep-20 03:37 AM what if you synthesize this design on a FPGA 26-Sep-20 03:37 AM "Yo! I heard you like FPGAs so I synthesized a FPGA in your FPGA" 26-Sep-20 03:41 AM :)) 26-Sep-20 04:21 AM again though - processes have to be figured out and nailed 26-Sep-20 04:21 AM and of course, documented 26-Sep-20 04:21 AM with the tools, cell libraries for VLSI design, etc published 26-Sep-20 04:21 AM so that people can make use of that process 26-Sep-20 07:24 AM yep 28-Sep-20 05:54 AM pmma is also UV sensitive, right? 28-Sep-20 05:54 AM and we can build cool structures out of it 28-Sep-20 08:21 AM yeah but only deep uv 28-Sep-20 08:21 AM need like sub 250 nm, so must general purpose exposure systems won't work 28-Sep-20 08:35 AM will try 28-Sep-20 08:35 AM (because I read it could be usable for 365nm (i-line)) 28-Sep-20 08:35 AM but primarily I want to use it for EBL stuff 28-Sep-20 08:45 AM oh EBL it works no problem 28-Sep-20 08:45 AM https://ieeexplore.ieee.org/document/4233076 I have doubts about anything above 250 though.. other papers might have been using some sort of chemically amplified PMMA, or perhaps optical manipulation (optical nanojet stuff maybe) to get the most out of i-line 28-Sep-20 08:45 AM but I've seen quite a few papers that put 250 as the upper limit 28-Sep-20 08:45 AM okay yeah you can do it but need to add a sensitizer 28-Sep-20 08:45 AM https://iopscience.iop.org/article/10.1088/0268-1242/31/2/025010/meta 28-Sep-20 09:30 AM yeah a crosslinker or how they're called 28-Sep-20 09:30 AM would be awesome because it's cheap and pretty resilient 28-Sep-20 09:30 AM but you're right, pure it needs 200nmish rad 28-Sep-20 10:52 AM someone with more knowledge about lasers would know for sure, but you might be able to hit deep UV with nonlinear crystals. E.g. third harmonic of ND:YAG lasers is 355, maybe a similar arrangement that can get you down to deeper UV (using a less-far source laser). I'm sure it's crazy inefficient, but so are excimer lasers so 28-Sep-20 10:52 AM mercury vapor has a emission line around 250, which is probably the easier route 28-Sep-20 10:54 AM there are lasers like these: https://photonsystems.com/products/deep-uv-sources/lasers/ 28-Sep-20 10:54 AM seems to be these sort of laser 28-Sep-20 10:54 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-D6742.png 28-Sep-20 10:54 AM E.g. third harmonic of ND:YAG lasers is 355, maybe a similar arrangement that can get you down to deeper UV (using a less-far source laser). Nd:YAG lasers also seem to have a fourth and fifth harmonic at 266 and 213 nm respectively 28-Sep-20 11:00 AM thinking on it... just blasting with a big ol mercury vapor lamp (with or without a lowpass filter, probably doesn't matter) would almost certainly be cheapest and easiest. Not overly efficient but meh 28-Sep-20 11:01 AM lasers though 28-Sep-20 11:01 AM check this paper out 28-Sep-20 11:01 AM https://laser-crylink.com/wp-content/uploads/2018/12/1.A-LD-side-pumped-deep-ultraviolet-laser-at-266nm-by-usin_2014_Optics-Laser.pdf 28-Sep-20 11:01 AM Under the pump power of 115.1 W, a laserwith an average output power of 1.27 W, a pulse width of 60 ns and the repetition rate of 12.5 kHz is obtained. 28-Sep-20 11:01 AM really "efficent" though 28-Sep-20 11:04 AM we're getting 226nm by using the third harmonic of a YAG to pump a dye and then doubling that 28-Sep-20 11:04 AM buckets of 226nm 28-Sep-20 11:05 AM oooh 28-Sep-20 11:05 AM $$$ though. 28-Sep-20 11:05 AM Big yag, two dye cells, a bunch of time spent aligning, and the dye basically needs to be changed every week 28-Sep-20 11:09 AM I see 28-Sep-20 11:09 AM https://giphy.com/gifs/swag-money-make-it-rain-2u11zpzwyMTy8 28-Sep-20 11:24 AM 200ish nm is about the limit of crystals, but you really want to use an excimer if you need laser powers in that range and don't need tunability. In general you want to use a lamp unless you specifically need something only a laser can provide. And if the efficiency of that deep UV YAG scares you, don't look into the wall to output efficiency of a typical unstable resonator q-switched YAG. Something like 3%. 28-Sep-20 11:24 AM Probably more like 210 actually 28-Sep-20 11:25 AM watt-to-watt our 10Hz 50mJ centrifuge wins tbh 28-Sep-20 11:30 AM Looks like the wall to output efficiency for my VUV design is like 0.048% lol 28-Sep-20 11:30 AM Not including the heating tape 28-Sep-20 11:30 AM The "power" of multiple conversions 28-Sep-20 12:05 PM If hes doing litho he probably doesn't want a laser. I'd assume it's contact litho and not using a stepper? Probably easier to just adjust the resist formula instead of trying to put a different light source in the aligner 28-Sep-20 12:05 PM @N00N https://www.sigmaaldrich.com/catalog/product/aldrich/196118?lang=en®ion=US this is the initiator in the second paper I linked.. not terrible pricing 28-Sep-20 12:07 PM question 28-Sep-20 12:07 PM will they sell to you though? 28-Sep-20 12:07 PM like outside of academia or research orgs 28-Sep-20 12:08 PM I'm assuming if he's using it for ebeam stuff then it is through one of those... though knowing this server maybe not 28-Sep-20 12:08 PM Theres chinese suppliers but I've never used a website like this and don't know if it's trust worthy lol 28-Sep-20 12:08 PM https://www.made-in-china.com/productdirectory.do?word=Irgacure+651&file=&subaction=hunt&style=b&mode=and&code=0&comProvince=nolimit&order=0&isOpenCorrection=1 28-Sep-20 12:08 PM also similar photoinitiators on ebay but idk how well they'd work with pmma, would have to mess with it a bit 28-Sep-20 12:08 PM https://www.ebay.com/i/202792375041?chn=ps&mkevt=1&mkcid=28 28-Sep-20 12:12 PM Sigma definitely won't sell to you 28-Sep-20 12:12 PM None of the big boys will 28-Sep-20 12:13 PM slight clarification: sigma will sell to anyone... but you need a commercial address. ultimately the same end-effect though 28-Sep-20 12:14 PM From what I hear they're also picky there. If you buy certain things while not being in the right business for that chemical, you won't get the thing. Nor do you want to pay Sigma prices unless you're a big buyer with a good account with them. 28-Sep-20 12:15 PM ah interesting, didn't know that aspect 28-Sep-20 12:15 PM and definitely on price, sigma is expensive 28-Sep-20 12:32 PM @N00N https://www.ebay.com/i/202792394153 try 5-10% pmma, 15% initator, dissolved in ethyl lactate (w/w)... it's the same type of initiator (norrish 1 reaction) which will work with i-line 28-Sep-20 12:32 PM everything you need should be available on ebay, I'd suggest the 907 initiator. If you don't need to get stuff off of ebay and can use aldrich then just use the 651 because it looks like that paper got decent resolution. Then I guess just use whatever solvent as a developer since itll be a negative resist 28-Sep-20 12:46 PM @N00N https://www.sigmaaldrich.com/catalog/product/aldrich/196118?lang=en®ion=US this is the initiator in the second paper I linked.. not terrible pricing @mike crb ah nice 28-Sep-20 12:46 PM @N00N https://www.ebay.com/i/202792394153 try 5-10% pmma, 15% initator, dissolved in ethyl lactate (w/w)... it's the same type of initiator (norrish 1 reaction) which will work with i-line @mike crb will buy that 28-Sep-20 12:46 PM sry for bad reply - my cat is attacking me 28-Sep-20 12:46 PM friendly cuddle attack with needles 28-Sep-20 12:48 PM I would try the 907 first, if it isn't working and you can get the aldrich one, then use that. I don't know if the different radicals will make too much of a difference, but they could 29-Sep-20 01:30 AM is p+ doping feasible with diffusion? 29-Sep-20 04:12 AM why not? 29-Sep-20 04:12 AM it wouldn't be P+ though 29-Sep-20 04:12 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-DCD42.png 29-Sep-20 04:17 AM I don't think he was talking about positively charged phosporus doping, but strong p doping 29-Sep-20 04:17 AM Small numbers of dopant atoms can change the ability of a semiconductor to conduct electricity. When on the order of one dopant atom is added per 100 million atoms, the doping is said to be low or light. When many more dopant atoms are added, on the order of one per ten thousand atoms, the doping is referred to as high or heavy. This is often shown as n+ for n-type doping or p+ for p-type doping. 29-Sep-20 04:18 AM fair 29-Sep-20 04:18 AM I just use the term p-type doping than p+ more often 29-Sep-20 04:19 AM p type is the general name, and p+ is stron p type 29-Sep-20 04:20 AM but yeah, Boron for example can be diffusively doped into Silicon 29-Sep-20 04:20 AM this is what Sam Zeloof's process was, isn't it? 29-Sep-20 04:20 AM I requote 29-Sep-20 04:20 AM Doping is then carried out by either solid or liquid source. The solid source is a Boron Nitride disk that is placed in proximity (<2mm) from the wafer in the tube furnace. Alternatively, spin-on liquid sources can be prepared from Phosphoric or Boric acid in water or solvents and doping is carried out in a standard pre-deposition/HF dip/drive-in/deglaze process. I obtained Phosphoric acid in pure form on Amazon and Boric acid from Roach & Ant killer. Since the starting wafer for PMOS here is N-type, I am doing P diffusions of Boron for the source/drain regions and am targeting a sheet resistance in diffused regions of 100 to 250 Ω/sq. 29-Sep-20 04:20 AM and the solubility of dopants is a function of temperature 29-Sep-20 04:20 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-9C7CD.png 29-Sep-20 04:20 AM of course, how deep the atoms diffuse would be dependent on time also 29-Sep-20 04:26 AM p+ is heavy doping, yes 29-Sep-20 04:31 AM me is just groggy and sleepy 29-Sep-20 04:31 AM @GigaSquirrel ikr 29-Sep-20 04:31 AM coffee is ❤️ 29-Sep-20 04:34 AM coffe is life, coffee is love 29-Sep-20 04:34 AM the fun part of it all was that I just awakened not too long ago to a customer support incident 29-Sep-20 04:34 AM and spending time trying to figure out why even example code doesn't even run 29-Sep-20 04:35 AM and I've to check how deep the p+ must go 29-Sep-20 04:35 AM yeah coffee is a good idea 29-Sep-20 04:36 AM reminds me of those 5 hour energy stuff 29-Sep-20 04:39 AM and I've to check how deep the p+ must go snrk 29-Sep-20 04:39 AM oh god customer support is my personal nightmare 29-Sep-20 04:42 AM the best part 29-Sep-20 04:42 AM my boss originally set up WhatsApp chats for support 29-Sep-20 04:42 AM which means that my clients have my personal number 29-Sep-20 04:42 AM (since we don't have work phones - too cheap for that) 29-Sep-20 04:42 AM so they'll call me, maybe 29-Sep-20 04:42 AM (more like surely - not the first time I got direct calls from clients) 29-Sep-20 04:44 AM oh no 29-Sep-20 04:44 AM time to get a new number 29-Sep-20 04:49 AM hmm there is a newer red pitaya with 16bit dac 29-Sep-20 04:49 AM but now I started my own 16 bit dac board 29-Sep-20 04:49 AM for EBL 29-Sep-20 04:49 AM so I guess it makes sense to support both 29-Sep-20 04:51 AM cool! 29-Sep-20 04:51 AM tempting 29-Sep-20 04:51 AM since red pitayas are pretty common 29-Sep-20 04:51 AM and it of course scales down to work with lower resolutions 29-Sep-20 04:56 AM I really enjoy my pitaya 29-Sep-20 04:56 AM all the talk about it makes it tempting to get as a DAQ instrument and all 29-Sep-20 04:56 AM but isn't the 16 bit one just rf? 29-Sep-20 04:56 AM but I already wanna get a FPGA dev board 29-Sep-20 04:57 AM the red pitaya also has an FPGA on board 29-Sep-20 04:57 AM I need a FPGA for other purposes the Red Pitaya can't do 29-Sep-20 04:57 AM but isn't the 16 bit one just rf? - Tailored for SDR & other RF applications, HF and 50MHz bands > - No DC coupling ! minimum RF Input frequency (-3dB): 300kHz > - Comes with two 16 bit ACDs 50 ohm inputs; 14 bit DACs 50 ohm outputs 29-Sep-20 05:00 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-E4238.png 29-Sep-20 05:02 AM isit? 29-Sep-20 05:02 AM phew 29-Sep-20 05:02 AM so it makes sense again 29-Sep-20 05:04 AM was kinda paralysed already 29-Sep-20 05:04 AM you're welcome 29-Sep-20 05:04 AM yeah tnx! 29-Sep-20 05:04 AM soooo I think then it makes sense to have a "highres" cheapo dac and also support for the red pitaya with lower resolution 29-Sep-20 05:06 AM yeah it doesn't need high speed, right? 29-Sep-20 05:07 AM yeah 125MSps should be plenty 29-Sep-20 05:08 AM I am not familiar with data acquisition for EBL 29-Sep-20 05:08 AM what are the smallest signals you expect to deal with? 29-Sep-20 05:08 AM you mean for the DAC? 29-Sep-20 05:08 AM like the deflection of the beams? 29-Sep-20 05:09 AM facepalms 29-Sep-20 05:09 AM god I need coffee 29-Sep-20 05:09 AM see 29-Sep-20 05:09 AM I was looking at the ADC instead of the DAC 29-Sep-20 05:09 AM if you're monitoring the beam current though it's usually low nanoamp- picoamp beam currents 29-Sep-20 05:10 AM yeah but ADC is also handy to locate fiducials etc 29-Sep-20 05:10 AM argl 29-Sep-20 05:10 AM the ADc 29-Sep-20 05:10 AM so for the deflection of the beam 29-Sep-20 05:10 AM how would the signal look like? 29-Sep-20 05:10 AM in terms of frequency spectra 29-Sep-20 05:11 AM ah well like MHz-ish 29-Sep-20 05:11 AM because EBL is slow 29-Sep-20 05:11 AM and < 10 MHz? 29-Sep-20 05:11 AM or higher? 29-Sep-20 05:11 AM < 10MHz 29-Sep-20 05:11 AM would be enough 29-Sep-20 05:11 AM but maybe there are some exposure strategies benefiting from higher signal bandwidths 29-Sep-20 05:13 AM so let's look at the DAC then 29-Sep-20 05:13 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-8221E.png 29-Sep-20 05:13 AM which is only 14 bit at most for the DAC 29-Sep-20 05:14 AM yeah 'my board' would use a 200MSPS DAC 29-Sep-20 05:15 AM question would also be how much resolution do you really need? 29-Sep-20 05:15 AM depends by the SEM 29-Sep-20 05:15 AM the noise of its amps and so on 29-Sep-20 05:15 AM I don't want to build the complete control electronics 29-Sep-20 05:15 AM yet 29-Sep-20 05:16 AM ah 29-Sep-20 05:17 AM more like hooking the DAC and ADC behind the normal electronics 29-Sep-20 05:18 AM Huh would you look at that apparently today is national coffee day 29-Sep-20 05:18 AM must be a sign 29-Sep-20 05:18 AM fine! 29-Sep-20 05:18 AM makes coffee 29-Sep-20 05:18 AM replacing their built-in signal generation (mostly ramps) and signal processing units 29-Sep-20 05:56 AM I'd like to see what the circuitry used in SEMs are like 29-Sep-20 06:43 AM hum I have schematics of old SEMs 29-Sep-20 06:47 AM oooh gib! 29-Sep-20 07:40 AM it's like 100s of pages of dina2 paper stuff ️ 29-Sep-20 07:41 AM oh haha 29-Sep-20 10:40 AM I've been working on a transimpedance amp with a 16bit ADC and +-250nA range 29-Sep-20 10:40 AM I guess the general architecture/learnings could be applied to measuring beam current, hmm 29-Sep-20 10:40 AM I'd been thinking of using an LPC4370 29-Sep-20 10:40 AM Has an 80MSPS ADC on board 29-Sep-20 10:40 AM No community behind it though outside of the NXP forums 29-Sep-20 10:40 AM I got some stuff compiling and running on it years ago tho 29-Sep-20 11:23 PM yeah, I guess NXP is kinda exotic for hobbyists 29-Sep-20 11:23 PM for whatever reason 29-Sep-20 11:23 PM maybe because of those cheap disco boards from STM? 30-Sep-20 09:49 AM I'd like to see what the circuitry used in SEMs are like @SleepyOwl Joyce I've been working on a Hitachi S-2300 and later also an S-2400 SEM. Their Schematics are very hard to work with. The S-2300 Schematic can be found here: https://manualsdump.com/en/manuals/hitachi-s-2300/253351/1 Those SEMs generate analog raster signals which were optionally processed by some Raster Rotation circuitry or similar stuff. The Deflection Signals were produced with an +/-10V Signal amplitude and were fed into a DAC0832 8-bit DAC reference input. The output is fed into a DAC1006 10-bit DAC reference input to scale the raster signals down to magnify the image. I have built a data aqcuisition device that simply applies X and Y +/-10V raster signals to the SEM and converts the SE-Detector Signal with an ADC. I also plan to provide my design as open Hard/Software 30-Sep-20 10:00 AM I simply use an stm32 30-Sep-20 10:00 AM but it sucks so I switched back to a red pitaya 30-Sep-20 10:00 AM but I hope that I'm able to complete a cheaper zynq based board 30-Sep-20 12:01 PM @analogmultiplizierer I think those are called MDACs? 30-Sep-20 02:14 PM I simply use an stm32 @N00N STM32s are pretty decent microcontrollers 30-Sep-20 02:14 PM also @analogmultiplizierer that's cool! I've never really looked into SEMs and their acquisition electronics 30-Sep-20 02:14 PM wew 30-Sep-20 02:14 PM that schematic is dense 30-Sep-20 04:11 PM that schematic is dense @SleepyOwl Joyce It's actually all handwritten and -drawn! 30-Sep-20 05:37 PM Yeah 30-Sep-20 11:00 PM yeah they are, but the integrated adc and dac are not bit-wide enough - but it was the cheapest option to implement an USB to DAC/ADC thingy with a stm32 disco board 30-Sep-20 11:01 PM oh yeah 30-Sep-20 11:01 PM but you can say that with any MCU 30-Sep-20 11:01 PM and the ADC/DAC are not fast enough in sample rate 30-Sep-20 11:02 PM the fastest I've seen on a MCU is the NXP LPC4370 part 30-Sep-20 11:02 PM with 80 MSPS sample rate 30-Sep-20 11:03 PM what bit depth? 30-Sep-20 11:04 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/unknown-38F68.png 30-Sep-20 11:04 PM so yeah, 12 bit 30-Sep-20 11:04 PM that's the fastest I have ever encountered 30-Sep-20 11:16 PM hm cool 30-Sep-20 11:35 PM Heh, I think the airspy series of SDR receivers uses NXP LPC series MCU's 01-Oct-20 12:07 AM it does 01-Oct-20 12:47 AM Ahhahha 01-Oct-20 12:47 AM 12 bits in an MCU 01-Oct-20 12:47 AM Try 8 if you're lucky 01-Oct-20 12:47 AM The on die flash memory controller usually makes the LSBs just noise. 01-Oct-20 12:47 AM Charge pumps aren't friendly to your nosie floor 01-Oct-20 12:50 AM Yeah 01-Oct-20 12:50 AM Hence you do have something called ENOB 01-Oct-20 12:52 AM Yup. To get better than 8 bits you really need an external adc 01-Oct-20 12:52 AM In my experience most MCU have ENOB of 9 no matter what they say on the PDF 01-Oct-20 01:14 AM Airspy's do get great performance out from it. 01-Oct-20 01:14 AM But that also requires good frontend design and some good DSP 01-Oct-20 01:27 AM 9 was the ENOB I got from the stm32 01-Oct-20 01:27 AM that's y I'm now building a board with 16 bit adc/dac ^^ 01-Oct-20 01:29 AM :) 01-Oct-20 01:29 AM as an intermediate solution I use a red pitaya 01-Oct-20 01:29 AM currently the one with 10 bits but I'll upgrade to 14 01-Oct-20 01:29 AM my 14 bit pitaya is integrated in an osa currently - don't want to rip that apart 01-Oct-20 01:29 AM I hate that to build other experiments - killing working rigs 01-Oct-20 01:29 AM but on the other side the pitaya is kind of a waste for that purpose 01-Oct-20 01:29 AM (not using the dac) 01-Oct-20 01:29 AM but I could - for stabilization of the reference 01-Oct-20 01:29 AM but then I would need another ADC 01-Oct-20 01:46 AM but for EBL stuff I want more than two input channels for different sensors of course 01-Oct-20 08:14 AM just get Red Pitayas for everything 01-Oct-20 08:14 AM use them like Arduinos 01-Oct-20 09:50 AM yeah 01-Oct-20 12:49 PM Well you can always use oversampling to increase ENOB 01-Oct-20 12:49 PM 80MSPS affords quite a bit over oversampling 01-Oct-20 12:49 PM Also, you can use things like the timers and DMA in conjunction with deep sleep mode to sample when most of the digital side is off and quiet 01-Oct-20 01:27 PM http://ww1.microchip.com/downloads/en/appnotes/doc8003.pdf 01-Oct-20 10:44 PM "China currently produces only 5 per cent of the photoresists it uses in KrF ultraviolet lasers for 8-inch silicon wafers, and relies entirely on imports for more advanced ArF photoresists for 12-inch wafer production, according to the company" 01-Oct-20 10:44 PM https://amp.scmp.com/tech/innovation/article/3103674/smic-supplier-moves-china-one-small-step-closer-chip-self 01-Oct-20 11:41 PM @nmz787 yeah - that's what I did for slower scan rates of course 05-Oct-20 09:27 PM so after 7 years I end up closing an experiment as unsuccessful 05-Oct-20 09:27 PM and this is probably the last time I will see these things 05-Oct-20 09:27 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/20201006_121914-015BC.png 05-Oct-20 09:27 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/20201006_122005-427C0.png 05-Oct-20 09:27 PM So these are N-type Silicon Wafers, <100> orientation 05-Oct-20 09:27 PM Diameter: 50.8 mm Thickness:0.300 mm 05-Oct-20 09:27 PM Single side surface polished 05-Oct-20 09:27 PM unfortunately didn't get to do anything of my own with them 05-Oct-20 09:27 PM now I ship it to someone else 05-Oct-20 09:27 PM @N00N I should be shipping them over within this week 05-Oct-20 09:27 PM will DM the tracking number 05-Oct-20 10:18 PM Wafers that small sell for a premium since they're so "old" 05-Oct-20 10:43 PM shrug 05-Oct-20 11:34 PM I mean more like, they're a nice find/trade/donation to keep with hackers 05-Oct-20 11:34 PM In this community 05-Oct-20 11:34 PM They're a lot easier to work with than the 6 or 8 inch wafers I got a few years ago 05-Oct-20 11:34 PM /me still hasn't opened those 05-Oct-20 11:34 PM I think mine are doped tho 05-Oct-20 11:34 PM Mostly just bought them for general lithography and etching practice 05-Oct-20 11:56 PM hii 05-Oct-20 11:56 PM @SleepyOwl Joyce thank you 07-Oct-20 01:09 AM Wafers that small sell for a premium since they're so "old" @nmz787 aren't 4 or 6 in the standard sizes for PV wafers nowadays 07-Oct-20 05:18 AM yeah they’re M4 or even M6 now but if feels weird calling them wafers since they’re squares lol 07-Oct-20 05:18 AM wafers like the ones Joyce posted are cool, I never see anyone use less than 100 mm for most research work 07-Oct-20 06:49 AM hi 07-Oct-20 06:49 AM is there a standard test pattern esp for EBL? 07-Oct-20 06:49 AM ok I think I have an idea 07-Oct-20 12:06 PM Most papers working on development of methods or resist just use parallel lines, like a grating 08-Oct-20 08:07 AM @mike crb Wait, square wafers? Where are those a thing outside of PV? 08-Oct-20 08:28 AM nowhere lol 08-Oct-20 08:28 AM that’s why I think PV “wafers” are weird lol 08-Oct-20 09:41 AM hmm 08-Oct-20 09:41 AM I thought OASIS is a modern format 08-Oct-20 09:50 AM It is, isn’t it? More modern that GDSII at least 08-Oct-20 09:58 AM yeah but I mean ... started reading the format description 08-Oct-20 09:58 AM they have different types of polygons etc 08-Oct-20 09:58 AM y? 08-Oct-20 09:58 AM those point lists 08-Oct-20 09:58 AM it's like with the CAD file formats IGED and STEP 08-Oct-20 09:58 AM but worse 08-Oct-20 09:58 AM IGES 08-Oct-20 09:58 AM I hope I find a parsers lib for GDSII and OASIS 08-Oct-20 09:58 AM like opencascade for STEP 08-Oct-20 09:58 AM ok cool? 08-Oct-20 09:58 AM python-gdsII and fatamorgana for oasis? 08-Oct-20 09:58 AM or any better recommendations? 08-Oct-20 09:58 AM (only want the polygons per layer) 08-Oct-20 10:08 AM Beyond anything I know lol, I just knew oasis was kinda new standard so confused why not modern 08-Oct-20 10:08 AM without parsing those braindead formats by myself 08-Oct-20 10:08 AM yeah have a look 08-Oct-20 10:08 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/image0-B78B5.jpg 08-Oct-20 10:08 AM w t f 08-Oct-20 10:08 AM every mediocre game engine has smarter formats 08-Oct-20 10:08 AM or do I miss something? 08-Oct-20 10:08 AM ok - need to buy some beer to cope that 08-Oct-20 11:24 AM I just used pyGDS 08-Oct-20 09:05 PM ah ok, will try 08-Oct-20 09:09 PM There's also a BMP to GDS script out there 08-Oct-20 09:09 PM I did some work in GIMP, used the conversion script, then used pyGDS to copy/paste (duplicate) the data to fill a field of view 08-Oct-20 09:10 PM another question: is there a free alternative to magic? 08-Oct-20 09:10 PM I mean for having a GUI 08-Oct-20 09:10 PM Sam Zeloof has a writeup 08-Oct-20 09:10 PM That's for semiconductor drawing? 08-Oct-20 09:11 PM yes 08-Oct-20 09:11 PM Or design simulator? 08-Oct-20 09:11 PM for drawing 08-Oct-20 09:11 PM Not sure 08-Oct-20 09:12 PM and 'simulation' and keeping track of net lists 08-Oct-20 09:12 PM like an EDA for VLSI 08-Oct-20 09:12 PM I wrote a design synthesis tool once using a SAT solver 08-Oct-20 09:12 PM But that was more for netlist generation when reusing standard parts 08-Oct-20 09:12 PM http://www.opencircuitdesign.com/magic/ 08-Oct-20 09:13 PM Like, kind of what Cadence FPGA System Planner does 08-Oct-20 09:14 PM yeah, I'm looking for a magic style eda for analog circuits 08-Oct-20 09:14 PM or mixed 08-Oct-20 09:15 PM All the real analog design stuff I've heard of uses ANSYS 08-Oct-20 09:15 PM I haven't looked for open stuff tho 08-Oct-20 09:15 PM so with standard stuff like caps and coils and dividers as parametrical building blocks 08-Oct-20 09:15 PM Oh you don't mean analog transistor stuff? 08-Oct-20 09:16 PM also analog transistor stuff 08-Oct-20 09:16 PM cadence and mentor are providing suites for doing that 08-Oct-20 09:16 PM but $$$ 08-Oct-20 09:16 PM https://www.awr.com/awr-software/products/awr-design-environment 08-Oct-20 09:16 PM I mean it's quite a different topic 08-Oct-20 09:16 PM compared with digital vlsi eda stuff like magic 08-Oct-20 09:16 PM but it's also interesting 08-Oct-20 09:16 PM but first I've to expose simple polygons reproducable of course 08-Oct-20 09:47 PM do you also made photonics/optics stuff? 08-Oct-20 09:47 PM talking about test patterns led me to the "idea" of putting gratings on oddly shaped surfaces 08-Oct-20 10:54 PM what do you mean by oddly shaped 08-Oct-20 10:54 PM like curved surface? 09-Oct-20 12:02 AM yes 09-Oct-20 05:40 AM You can do detachment lithography for curved surface depending on feature size 09-Oct-20 05:48 AM https://onlinelibrary.wiley.com/doi/abs/10.1002/adfm.201090000 09-Oct-20 05:48 AM They got down to 2 um 09-Oct-20 05:50 AM ah that looks cool 09-Oct-20 05:50 AM so I mean ... what's the best developer for PMMA? 09-Oct-20 05:50 AM MIBK + IPA? 09-Oct-20 05:55 AM looks like it 09-Oct-20 05:55 AM But if you have sonication you can just do ipa/water 09-Oct-20 05:55 AM ok, simply acetone would not work? 09-Oct-20 05:55 AM because it simply strips off everything? 09-Oct-20 05:55 AM tried acetone very short - cleaned the surface completely :)) 09-Oct-20 05:55 AM but maybe the dosage is to low 09-Oct-20 05:58 AM Yeah do ipa/water 09-Oct-20 05:58 AM Looks like it’ll work without sonication actually 09-Oct-20 05:58 AM Is this with ebl? 09-Oct-20 06:00 AM yeah 09-Oct-20 06:01 AM Yeah try maybe upping dosage a little and using the 3:1 ipa to water 09-Oct-20 06:03 AM ok 09-Oct-20 06:03 AM is there any way to see something pre development? 09-Oct-20 06:03 AM and water + ipa should be slower, right? 09-Oct-20 06:03 AM but I also have mibk 09-Oct-20 06:03 AM so what's more convenient? 09-Oct-20 06:03 AM I mean maybe my pmma/anisol mix is a failure 09-Oct-20 06:10 AM idk with ebl of pmma.. with normal UV litho you can see the pattern slightly 09-Oct-20 06:10 AM (Not UV of pmma, gen positive resist) 09-Oct-20 06:11 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/image0-CF398.jpg 09-Oct-20 06:11 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/video0-9BE4B.mov 09-Oct-20 06:11 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/image0-85547.jpg 09-Oct-20 06:11 AM need to buy some pipette flasks 09-Oct-20 06:11 AM what would be a more convinient alternative to PMMA? thing is, I want reproducable cheap and easy recipes 09-Oct-20 06:11 AM but maybe that's a silly idea 09-Oct-20 06:15 AM For EBL? 09-Oct-20 06:16 AM yes 09-Oct-20 06:17 AM Nothing that’s going to be easily available 09-Oct-20 06:17 AM hmm ok 09-Oct-20 06:18 AM What type of pmma do you have 09-Oct-20 06:18 AM is it 450kDa or 900kDa 09-Oct-20 06:18 AM Not that it matters too much just need to adjust recipe for it 09-Oct-20 06:18 AM https://static1.squarespace.com/static/57b26cc76b8f5b7524bf9ed2/t/59272752d2b857857ae994c6/1495738198009/PMMA_Process_SOP_May2017.pdf 09-Oct-20 06:45 AM 900-1500kDa 09-Oct-20 06:45 AM some cheap stuff ^^ 09-Oct-20 06:54 AM @N00N 4:1 ethanol, and yes you can see it after development with ebl or fib litho https://iopscience.iop.org/article/10.1088/0957-4484/27/3/035302/pdf 09-Oct-20 06:57 AM yeah after 09-Oct-20 06:57 AM now I used MIBK:IPA 1:3 09-Oct-20 06:57 AM stopped the process with pure IPA 09-Oct-20 06:58 AM 4;1 ethanol is basically liquor store grade 09-Oct-20 06:58 AM and rinsed with DI water 09-Oct-20 06:58 AM Do your litho then take shots 09-Oct-20 06:59 AM well now it kinda worked but I've to figure out (used a piece of copper as substrate) cause of charging etc 09-Oct-20 06:59 AM you mean vodka would work? 09-Oct-20 06:59 AM how many seconds in vodka? 09-Oct-20 07:00 AM That paper has it all 09-Oct-20 07:01 AM okok 09-Oct-20 07:01 AM "We have found that ethanol/water mixtures at a 4:1 volume ratio are an excellent, high resolution, non-toxic developer for exposed PMMA. " 09-Oct-20 07:25 AM will try 09-Oct-20 07:40 AM I guess my dosage was to high since no line resolution 12-Oct-20 05:15 AM here an image from the resist I made friday, but now it's strangely grainy compared to friday - but maybe simply too much anisol vaporized? 12-Oct-20 05:15 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/image0-753A5.jpg 12-Oct-20 01:44 PM That looks very odd 12-Oct-20 01:44 PM What was your procedure? 13-Oct-20 04:14 AM pmma 13-Oct-20 04:14 AM the resist was kinda shtty stored 13-Oct-20 04:14 AM so it lost much of the solvent 13-Oct-20 04:14 AM currently stirring new one 13-Oct-20 04:14 AM that's also why the thickness was app 10 micometers or so (high viscosity) 13-Oct-20 04:14 AM that's y two of the three test patterns liftet off unintended 13-Oct-20 04:14 AM so it was more like writing in a thick pmma film with an e-beam 13-Oct-20 04:14 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/video0-04886.mov 13-Oct-20 04:14 AM stirr 13-Oct-20 04:55 AM Welcome to @N00N lab ASMR 13-Oct-20 05:23 AM ^^ 13-Oct-20 05:31 AM do you have a sonicator? Could you sonicate the resist before spinning? 13-Oct-20 05:31 AM Oh it was the thickness issue that makes sense. 13-Oct-20 07:25 AM more detail 13-Oct-20 07:25 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/image0-060A2.jpg 13-Oct-20 07:25 AM but maybe I'll stick to expensive pmma 'from' merck 13-Oct-20 07:29 AM cop out 13-Oct-20 07:29 AM but yeah, I would think making your own PMMA resist is going to be quite a process on its own 13-Oct-20 07:31 AM yeah ️ 13-Oct-20 07:32 AM yeah I guess it would be a better endeavor to focus your time on getting the litho right 13-Oct-20 07:32 AM than messing around with PMMA that may or may not work well and may or may not be reproducible 13-Oct-20 07:38 AM yep you're right 13-Oct-20 07:38 AM I mean 13-Oct-20 07:38 AM synthesizing PMMA or a resist of sorts is probably also worthwhile if photoresist is hard to get 13-Oct-20 07:38 AM but probably later stage thing? 13-Oct-20 07:38 AM then recording the steps to synthesize somewhere would be helpful for others too 13-Oct-20 07:57 AM You could try buying lower MW pmma powder, would give you less viscous solution 13-Oct-20 07:57 AM If you can find 495k da pmma I would use that 13-Oct-20 07:57 AM What does the film look like when you spincoat 13-Oct-20 07:57 AM Does it look fine? What speed are you coating at? What is your percent PMMA in anisole? 13-Oct-20 08:28 AM i've to filter it 13-Oct-20 08:28 AM and weight % wise it's app 5% 13-Oct-20 08:28 AM I made another batch - stirred longer at 80 deg C or so 13-Oct-20 08:28 AM 60 mins 13-Oct-20 08:28 AM will test tomorrow 13-Oct-20 08:28 AM and I ordered some 350 k pmma 13-Oct-20 08:35 AM Okay, you’ll get lower viscosity as you lower MW so just something to be aware of 13-Oct-20 09:35 AM yes 13-Oct-20 09:35 AM I simply bought what they have 13-Oct-20 09:51 AM my goal is something like 10€ per l of process chems 13-Oct-20 09:51 AM the anisol I'm using is cheap fpod grade 13-Oct-20 09:51 AM food 13-Oct-20 09:51 AM mikb and ipa are dirt cheap 13-Oct-20 09:51 AM etc 13-Oct-20 09:51 AM unknown pmma also 13-Oct-20 09:51 AM but then such things are happening 13-Oct-20 02:36 PM @N00N your process is "pmma"??? That isn't a process, it's an ingredient 13-Oct-20 02:36 PM Was this just old commercially sourced, or you took ebay/hardware store plastic and dissolved it? If the latter, did you centrifuge it after dissolving and maybe filter with a submicron filter? Looks like your resist solution is very dirty. 13-Oct-20 08:51 PM ? 13-Oct-20 08:51 PM that's what I want to use a vacuum filtration for 13-Oct-20 08:51 PM to get those shtty grains out of the resist 13-Oct-20 08:51 PM @nmz787 the intention is: making cheap process chemistry 13-Oct-20 08:51 PM for that I'm experimenting with multiple sources for the ingredients 13-Oct-20 08:51 PM and yes - the first try from monday has it's issues 13-Oct-20 08:51 PM I ordered a vacuum filtration kit and a set of different < 1 micrometer filter packs yesterday to see if I can get rid of those grains from the cheap pmma powders 13-Oct-20 08:51 PM yesterday 13-Oct-20 08:51 PM but I also ordered a couple of - hopefully cleaner - pmma powder samples from different suppliers 13-Oct-20 08:51 PM @nmz787 what you see on the image above is only the result of the first try mixing my own pmma resist and applying it on a piece of Si wafer @SleepyOwl Joyce sent me on monday 13-Oct-20 09:10 PM I got ya now 13-Oct-20 09:10 PM How are you coating? 13-Oct-20 09:11 PM and after spinning the resist on it I couldn't resist to EBL it 13-Oct-20 09:11 PM wow 13-Oct-20 09:11 PM Real spin coater, hobby airplane motor? 13-Oct-20 09:11 PM Dremel? 13-Oct-20 09:11 PM even if the coating was very unpretty 13-Oct-20 09:11 PM what pun. 13-Oct-20 09:11 PM hobby king bldc 13-Oct-20 09:11 PM "real spin coater" pffff 13-Oct-20 09:12 PM In my experience to prevent streaking, I had to sort of buff the wafer with a nylon cleanroom towel, while squirting acetone as it spun 13-Oct-20 09:12 PM Not applying pressure to the wipe, just draping it on the spinning wafer 13-Oct-20 09:12 PM I think I may have alternated acetone and isopropanol 13-Oct-20 09:12 PM @N00N well if you have an EBL then you may also have a real spinner 13-Oct-20 09:14 PM ok - I have old SEMs for EBL but no real spinner 13-Oct-20 09:14 PM didn't found a cheap one yet 13-Oct-20 09:14 PM ok acetone towel first? 13-Oct-20 09:15 PM Ideally it's a clean room wipe, something totally dust and lint free 13-Oct-20 09:15 PM Idk what the diy substitute would be 13-Oct-20 09:16 PM ok, I guess streaking could be an issue 13-Oct-20 09:16 PM I paid $272 for my "real" spin coater 13-Oct-20 09:16 PM Yeah streaking is a pita 13-Oct-20 09:16 PM "Comet tails" 13-Oct-20 09:16 PM 272$ new? 13-Oct-20 09:17 PM SCS P6708D 13-Oct-20 09:17 PM Nah, ebay, used 13-Oct-20 09:17 PM Luckily I'd already repaired one at my part time job so had a little confidence 13-Oct-20 09:17 PM I think it only had a gummed up bearing... Spinning it by hand and squirting solvent seemed to take care of that 13-Oct-20 09:20 PM ok - let's see when I get one on ebay 13-Oct-20 09:20 PM but til then I've to live with my crppy setup 13-Oct-20 09:20 PM and now a little bit more sleep 13-Oct-20 09:28 PM From what I've seen and read in papers your spinner should be fine 13-Oct-20 09:28 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/Screenshot_20201013-213018-A892D.png 13-Oct-20 09:28 PM From https://www.joshuataillon.com/pdfs/2017_jtaillon_MandM.pdf 13-Oct-20 10:16 PM yes 13-Oct-20 10:16 PM no need for sleep 13-Oct-20 10:16 PM that's how my first spin coater set up looked like 13-Oct-20 10:16 PM the dvd spindle I'm still using 13-Oct-20 10:16 PM also the sem stubs 13-Oct-20 10:16 PM the main difference is my bench top vice is red 13-Oct-20 10:16 PM ah and the dremel is a proxxon 13-Oct-20 10:16 PM in my rig it was a proxxon I wanted to say 17-Oct-20 11:41 PM Clean the sample and dry it, make sure it’s clean, mount on spin coater chuck, then give it a good nitrogen blow off just before the photoresist goes on. 17-Oct-20 11:41 PM Especially if not in a cleanroom 18-Oct-20 09:49 AM Anyone have any good way to remove organic surfactants from nanoparticles? 18-Oct-20 09:49 AM I have been washing them with ethanol ipa methanol, etc 18-Oct-20 09:49 AM But it tends to remove them from my aubstrate 18-Oct-20 09:49 AM Can they be baked to remove the organics without changing the nanoparticle distribution 18-Oct-20 09:54 AM lol do not bake 18-Oct-20 09:54 AM will most likely ruin your distribution for sure 18-Oct-20 10:02 AM what type of particles are they 18-Oct-20 10:15 AM Metal oxides 18-Oct-20 10:20 AM if it's metal oxides then may be a bit more resistant to ripening 18-Oct-20 10:20 AM what type of surface? 18-Oct-20 01:22 PM on one hand for SEM i was planning on doing a Si Surface 18-Oct-20 01:22 PM for TEM i was planning on doing a Copper/Carbon-coated grid mesh 18-Oct-20 04:12 PM okay. I'd try ozone cleaning 18-Oct-20 04:12 PM Should get rid of the surfactant without changing the distribution 18-Oct-20 07:40 PM Is that a thing?? 18-Oct-20 07:40 PM That may chance my stoichiometry 18-Oct-20 08:40 PM Yes. And maybe, but if you want to have your particles stay relatively monodispersed you can’t use heat. Especially if they’re sub 30 nm particles 21-Oct-20 12:10 PM hmm ok ill loon into 21-Oct-20 12:10 PM thanks mik @mike crb 22-Oct-20 10:22 AM Anyone have some suggestion for alignment mark designs for photomasks? 22-Oct-20 10:22 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/examp-82BAE.png 22-Oct-20 10:22 AM Have to align those little black squares (5 um by 5 um) in the correct position on those wires 22-Oct-20 10:22 AM This is on a glass slide, other wise I'd use the template 4" wafer alignment marks for the aligner im using (either karl suss MA-6 or EV620, have access to both) 22-Oct-20 10:22 AM Black will be a brightfield mask so should make things easy 27-Oct-20 11:16 AM nice 27-Oct-20 11:16 AM but no idea 27-Oct-20 11:30 AM I figured it out, just will need to put glass slide long ways so I get both alignment marks on fov of the aligner 27-Oct-20 11:45 AM https://leadingedgetech.io/floating-silicon-method/ 27-Oct-20 11:45 AM This is HUGE 27-Oct-20 04:41 PM Oh wow, it is 27-Oct-20 04:41 PM I wonder how that will scale for stuff like solar panel manufacturing 27-Oct-20 06:07 PM I think it can slash costs by 50% or more 27-Oct-20 06:31 PM NICE 27-Oct-20 07:32 PM Found the appropriate grant technical report for anyone interested 27-Oct-20 07:32 PM https://www.osti.gov/servlets/purl/1126212 27-Oct-20 08:09 PM wait 27-Oct-20 08:09 PM this is pog 27-Oct-20 08:09 PM Cz is Czochralski right 27-Oct-20 08:15 PM I assume so 28-Oct-20 11:24 AM do you recall thee eevblog/blab about the laser slicing silicon into wafers, and how much bull it was with the savings it may save in comparison to the normal diamond saw or whatever 28-Oct-20 11:24 AM and now I cant find it, gave it a few mins of searching, no luck 28-Oct-20 01:51 PM Yup, I can recall that one and cannot find it either. I think you have a good chance of getting a quick response from Dave if you ask him about it/ why it can't be found (he might have not noticed that) 28-Oct-20 06:35 PM I tried searching for it again and failed, heh 28-Oct-20 06:35 PM checked eevblog, eevblog2, eevblab 01-Nov-20 02:40 PM @rfs I'm pretty sure they don't grow boules of silicon crystals for solar, so I don't know that this method would help cut costs as much as you think. 01-Nov-20 02:40 PM I think a lot of solar panels are cast 01-Nov-20 02:40 PM I should say solar cell* 01-Nov-20 02:40 PM nvm, it sounds like some are cast, some are still grown into boules 01-Nov-20 02:46 PM There's also the polycrystaline vs. monocrystaline cells. 01-Nov-20 02:52 PM There are cast monocrystalline cells, but my quick googling showed that tech isn't in mass production. 01-Nov-20 03:18 PM @rfs I'm pretty sure they don't grow boules of silicon crystals for solar, so I don't know that this method would help cut costs as much as you think. @transorbs I think the method's big benefit is you get monocrystal performance from something super cheap 01-Nov-20 03:25 PM maybe, there's a lot of this sort of stuff out there that never goes anywhere though. 01-Nov-20 03:32 PM I'm not exactly convinced that it'll be cheap enough to warrant the change, but this one is very real. Like, it actually is an order of magnitude faster way to make monocrystalline silicon 01-Nov-20 03:33 PM But for solar it /has/ to be cheap enough or it is worthless 01-Nov-20 03:33 PM actually, that's true of the whole semiconductor industry. Cost is king. 01-Nov-20 04:43 PM Sure, but I also doubt they would have bothered scaling up the method if they didn't think they'd have buyers. It's not like this is just a grant at this point. 02-Nov-20 12:46 PM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/PXL_20201102_204544710-DFB63.jpg 02-Nov-20 12:46 PM New book! 02-Nov-20 12:49 PM @transorbs I think it depends on what your market you're in. If you're trying to do utility scale solar of course, cost is king and you have no pricing power. 02-Nov-20 12:49 PM But if you're trying to do something like ultra high reliability power, for a hospital or data center, cost is not the biggest factor. 02-Nov-20 07:05 PM Ooooooooooo.... that’s a cool sounding book 04-Nov-20 07:03 AM boooooooooooooks 08-Nov-20 12:26 AM Is it possible to epitaxially grow crystalline silicon on metal, say, aluminum, instead of, on a wafer with a pre-existing lattice? 08-Nov-20 12:26 AM or is it doomed to always be amorphous 08-Nov-20 07:36 AM Doubt it the lattice mismatch will be way to high 08-Nov-20 07:36 AM GaAs, Si, Ge, InP, pretty much all semiconductor a=5.6-6 08-Nov-20 07:36 AM Metals much lower like 3.5-4 08-Nov-20 06:54 PM what about, vapor deposition of a-si, followed by remelting and slow cooling to get polycrystal? 08-Nov-20 06:58 PM Can vapor deposit poly-si 08-Nov-20 06:58 PM So no need to do in two steps 08-Nov-20 06:58 PM thought you were initially asking about like epitaxially growing 100 single crystal layer of Si on 100 Au or something 08-Nov-20 07:13 PM Nah just looking to make thin film solar cells with reasonable efficiency 08-Nov-20 07:13 PM the a-si ones seem to max out at 7% 08-Nov-20 07:13 PM but poly-si seems to do better, but, I have not seen poly-si yet that didn't start out with a diamond sawed wafer 08-Nov-20 07:29 PM cause poly-si cvd pretty slow and pretty expensive 08-Nov-20 07:52 PM do you know of any good materials I can read about the process? 08-Nov-20 07:54 PM think these might be something your looking for 08-Nov-20 07:54 PM https://matmatch.com/blog/making-cheaper-silicon-for-photovoltaics/ 08-Nov-20 07:54 PM https://www.tf.uni-kiel.de/matwis/amat/elmat_en/kap_6/backbone/r6_1_1.html 08-Nov-20 07:54 PM https://www.tf.uni-kiel.de/matwis/amat/elmat_en/kap_6/backbone/r6_3_3.html 08-Nov-20 07:54 PM https://ieeexplore.ieee.org/document/7503950 08-Nov-20 07:54 PM https://www.sciencedirect.com/science/article/abs/pii/S0927024897001104 08-Nov-20 07:54 PM siemen process is main process used in industry but would not be as easy at home 08-Nov-20 07:54 PM in general will need silanes for most Si cvd so that's something to consider as well 08-Nov-20 08:17 PM was looking for this one: https://aip.scitation.org/doi/abs/10.1063/1.332717?journalCode=jap you'd get amorphous but could process to poly si? idk it produces silanes in situ which would be a huge plus as you don't have to have a tank of silane 08-Nov-20 08:21 PM yeah that's what I am not sure of, can you remelt a-si in an infrared zone heater, then cool it slowly to get poly-si. it seems like in principle, you probably could as long as the silicon wetted the substrate and didn't bead up 10-Nov-20 08:37 PM @Haru What about O2 plasma? Should remove Organics selectively. What are the NPs made from? 10-Nov-20 08:37 PM Or UV/Ozone or piranha? 10-Nov-20 10:15 PM Metal oxides nps 10-Nov-20 10:15 PM Don't want to change surface stoich 10-Nov-20 10:15 PM Is a headache 12-Nov-20 02:40 AM oh boy i found the right keywords 12-Nov-20 02:40 AM https://en.wikipedia.org/wiki/Metal-induced_crystallization 12-Nov-20 04:11 AM https://1366tech.com/about-1366/ 18-Nov-20 11:45 AM What are people using to etch wafers at home? I'm more intereesteed in MEMS stuff, so polycrystaline - which may mean I need to result to DRIE? if single crystal I have a ref or two on vertical sidewall etching using a surfactant (iPrOH) with NaOH/KOH.. My use case would be for Si, not oxides/whatever 18-Nov-20 11:45 AM the ref I was ref'ing: https://people.eecs.berkeley.edu/~pister/147fa15/Resources/Brockmeier12.pdf 18-Nov-20 04:05 PM anisotropic etching of 110 wafers? 18-Nov-20 04:05 PM 111 plane should give vertical walls yeh? 19-Nov-20 12:01 AM Good old oxide etch 19-Nov-20 12:01 AM Vacuum%20Hackers%20-%20Text%20Channels%20-%20semiconductor%20%5B597249671353597953%5D.html_Files/image0-99FFF.jpg