--- Log opened Sun Feb 06 00:00:51 2022
00:05 -!- darsie [~darsie@84-113-55-200.cable.dynamic.surfer.at] has joined #hplusroadmap
00:39 -!- Croran [~Croran@71.231.214.173] has joined #hplusroadmap
01:20 -!- webmeister [~webmeiste@192.182.148.163] has joined #hplusroadmap
01:20 -!- webmeister [~webmeiste@192.182.148.163] has quit [Changing host]
01:20 -!- webmeister [~webmeiste@user/webmeister] has joined #hplusroadmap
01:34 -!- phill [uid429774@id-429774.ilkley.irccloud.com] has quit [Quit: Connection closed for inactivity]
02:17 -!- webmeister [~webmeiste@user/webmeister] has quit [Ping timeout: 240 seconds]
03:26 -!- sivoais [~zaki@199.19.225.239] has quit [Ping timeout: 256 seconds]
03:37 -!- sivoais [~zaki@199.19.225.239] has joined #hplusroadmap
03:42 < maaku> muurkha: Perry's black ships had a maaasive impact on the day-to-day life of japanese peasantry
03:42 < maaku> not sure if that was your point
03:44 < maaku> lsneff: regarding 3d electronics, the problem is usually how do you dump heat?
03:57 -!- phill [uid429774@id-429774.ilkley.irccloud.com] has joined #hplusroadmap
04:48 -!- spaceangel [~spaceange@ip-78-102-216-202.net.upcbroadband.cz] has joined #hplusroadmap
05:24 < lsneff> In this case, you’d only use it for low-power electronics
05:25 < lsneff> But I bet it’d be possible to find a very heat-conductive substrate
05:42 -!- yashgaroth [~ffffffff@2601:5c4:c780:6aa0::f2f0] has joined #hplusroadmap
05:49 < muurkha> maaku: yes, but what *was* that massive impact?
05:51 < muurkha> nmz787_: have you been able to do basic CMOS circuits yet?  ALD and DRIE sounds like you're planning on using photolithography instead of using FIB lithography to make your transistors one at a time
05:53 < muurkha> lsneff: cordwood circuits were indeed denser than conventional PCBs, but that was before ICs
05:53 < muurkha> oh I see fenn already mentioned that, sorry
05:53 < muurkha> what kind of UWB are you doing?
05:58 < muurkha> you can make a very heat-conductive substrate by filling your resin with beryllium oxide, diamond, or other high-conductivity dielectrics, but you don't have to scale up very far before having coolant channels is a more effective strategy
05:59 < muurkha> part of the popularity of PCBs is that they make it easy to incorporate medium-sized heatsinks in the form of copper pours and (except in space) inherently incorporate "coolant channels"
06:00 < lsneff> Makes sense
06:00 < lsneff> You could also print very high density PCBs with this
06:00 < lsneff> Even without embedded components
06:08 < muurkha> ?
06:08 < muurkha> btw I would think it would be easy to get one's hands on inkjet printheads.  just wear goggles to keep the ink from staining your corneas
06:09 < muurkha> I mean at least where I live there are shops every few blocks that sell inkjet printheads
06:13 < lsneff> Oh, probably 802.15.4x
06:13 < lsneff> g/x/z
06:14 < lsneff> There aren’t shops like that here
06:14 < lsneff> And they need to good printheads: several hundred nozzles, const instant drop volume, etc
06:15 < lsneff> Not sure if off the office supply store shelf ink printheads would work well, though they might
06:18 < muurkha> oh interesting, I didn't know about 802.15.4z
06:18 < muurkha> there are piezoelectric off-the-office-supply-store-shelf ink printheads and resistive ones
06:20 < muurkha> IIRC https://depts.washington.edu/open3dp/ used to have a lot of useful information about this kind of thing
06:20 < lsneff> I was looking at something like this (https://www.xaar.com/en/products/xaar-printheads/xaar-irix/), if I can get my hands on it
06:21 < muurkha> like https://depts.washington.edu/open3dp/2011/05/diy-inkjet-printer/
06:22 < muurkha> several hundred nozzles and precise control of drop volume are pretty commonplace in office-supply printers IIRC
06:22 < lsneff> Okay, that’s good to know
06:23 -!- Llamamoe [~Llamamoe@178235178060.dynamic-4-waw-k-1-2-0.vectranet.pl] has joined #hplusroadmap
06:25 < muurkha> "know" might be an exaggeration, I've never driven one
06:25 < lsneff> Right
06:26 < lsneff> I assume I’d probably have to reverse engineer how to drive one
06:29 < lsneff> Okay, I reached out to Xaar
06:44 < muurkha> I think there are a lot of people who have already documented how to drive one or another printhead
06:45 < muurkha> but certainly, if you have the money, active vendor support would be a lot better — though I doubt the hardware quality will be competitive
06:57 < lsneff> Yeah, I just want a consistent place to buy the printheads
06:58 < lsneff> I assume I’d have to RE it anyway, to avoid breaking NDAs
07:00 < lsneff> Insane that a single printheads can cost thousands of dollars
07:06 < lsneff> This whole industry is massively overpriced
07:09 < muurkha> well, it's mostly NRE.  a mass-market printhead doesn't cost thousands of dollars, it costs US$20 if you buy it new with the rest of the printer included
07:09 < muurkha> and they may or may not be better than the thousand-dollar printhead
07:10 < muurkha> if you're building a business on powder-bed binder injection printers you're probably better off with the thousand-dollar printhead even if it's not better
07:10 < lsneff> Why would that be?
07:10 < muurkha> two reasons
07:11 < muurkha> first, your own NRE expenses are high enough per unit that the engineering time you spend on reverse engineering an HP Deskjet print head is likely to cost you thousands of dollars per unit anyway
07:12 < muurkha> second, HP can change their print head design any month, and if they find out your printers are built on their print head they might do it just because
07:13 < muurkha> at which point suddenly your supply chain collapses
07:15 < muurkha> similarly, if you're doing a 3-D printing lab on an NSF grant, it may be much more reasonable to spend a few thousand dollars to solve the problem of being able to drive an inkjet print head than to spend thousands of dollars of grad student time on the same problem with less certainty of success
07:16 < muurkha> heh, one of the specs based on 802.15.4 is called "wireles shart"
07:16 < lsneff> Ah, I see what you mean
07:18 < muurkha> spending thousands of dollars to buy from a vendor is no guarantee that the vendor won't go bankrupt or cancel the product line, but at least their incentives are somewhat aligned with yours: if they keep providing the product you're using you'll keep paying them and they'll keep making money from you
07:18 < muurkha> HP doesn't
07:18 < muurkha> they make money selling ink and probably PII
07:19 < lsneff> That’s true
07:19 < lsneff> But HP for example would have a huge problem if they changed the design of a common printheads
07:21 < muurkha> would they?  I think they do it every few months
07:23 < lsneff> You can still by hp45 printheads from hp, and those have been in use for over a decade
07:23 < lsneff> Anyhow, I don’t think the off the shelf hp printheads will work because they’re thermal inkjets
07:25 -!- phill [uid429774@id-429774.ilkley.irccloud.com] has quit [Quit: Connection closed for inactivity]
07:32 < muurkha> ISTR piezo works for a wider variety of inks; is that what you mean?
07:36 < lsneff> Yes
07:36 < lsneff> This needs a UV-curable ink and a silver nanoparticle conductive ink
07:37 < muurkha> presumably a silver-salt ink that reduces on heating would also work?
07:38 < muurkha> (not, obviously, with bubblejet type heas)
07:38 < muurkha> *heads
07:38 < lsneff> Perhaps
07:38 -!- webmeister [~webmeiste@user/webmeister] has joined #hplusroadmap
07:38 < lsneff> I haven’t seen anyone use one like that
07:39 < lsneff> It seems like the printheads in consumer brother printers would work well, but they can’t be bought independently
07:39 < muurkha> neither have I
07:40 < lsneff> Oh also, the printhead needs to be heated
07:40 < lsneff> UV curable ink is too viscous at room temperature
07:40 < muurkha> you can dilute it with solvents but maybe that introduces other problems
07:41 < muurkha> consumer brother printers are more expensive but you can still buy five of them for US$1000
07:42 < muurkha> a popular one here is apparently the DCP-T220: https://www.mercadolibre.com.ar/impresora-a-color-multifuncion-brother-inkbenefit-tank-dcp-t220-negra-220v-240v/p/MLA17354082?pdp_filters=price:*-40000|category:MLA1676#searchVariation=MLA17354082&position=2&search_layout=stack&type=product&tracking_id=fdb9f9e2-e29e-47c1-abee-8b3be23a2d19
07:42 < muurkha> oops, leaked tracking cookies!
07:42 < muurkha> now you can find the dildos I was searching for last week
07:43 < muurkha> $33597 is about US$170, probably outside Argentina they're cheaper
07:44 < muurkha> generally high-nozzle-count printheads are non-consumable, shipped with the printer
07:46 -!- webmeister [~webmeiste@user/webmeister] has quit [Ping timeout: 250 seconds]
07:48 < muurkha> https://hackaday.com/blog/?s=inkjet links to a few recent projects, maybe there's a Wiki somewhere?
08:03 < lsneff> I guess I could buy one and try to reverse engineer the communication with the printhead
08:03 < lsneff> It’d be a huge shame to have to buy a regular printer and scrap most of it just to get at the printhead
08:07 < muurkha> roughly 10% as much of a shame as spending US$1000 on a printhead
08:08 < muurkha> the other parts of the printer aren't useless though
08:09 < muurkha> typically the carriage is a light-duty high-speed linear belt drive equipped with closed-loop optical feedback with about 50 μm repeatability, and the paper advance system has similar precision
08:10 < muurkha> and the power supply is very commonly independent of the rest of the printer, usually providing 5 V plus whatever the motors run on, 36 V or something, capable of maybe 100 watts
08:13 < muurkha> sometimes you can even recover a dual H-bridge chip or something from the main circuit board.  in theory you might be able to reprogram its microcontrollers or whatever but I've never heard of anyone doing that successfully
08:14 < lsneff> Indeed
08:15 < lsneff> It seems that consumer inkjet printheads dispense much less volume than the ones designed for 3d printing
08:15 < muurkha> that's interesting!  which side of that divide does your projct fall on?
08:16 < muurkha> *e
08:16 < muurkha> also, consumer inkjet printers typically include a peristaltic pump with a motor of its own and some optointerruptors used as limit switches for homing, as well as some shafts, sometimes precision ground
08:17 < lsneff> I’ll wait for the OEMs to get back to me and perhaps they’ll sell me small quantities for cheap
08:17 < muurkha> which OEMs?
08:18 < lsneff> The 3d printing side
08:18 < muurkha> aha, cool
08:18 < muurkha> as far as I can tell the plastic parts of a consumer inkjet printer are all pretty useless, even the gears.  but that might be my lack of mechanical knowledge
08:19 < lsneff> Konica Minolta and Xaar, though it turns out the Xaar printheads can’t do conductive inks
08:19 < muurkha> huh!  that's strange!
08:19 < muurkha> might be another reason to consider silver-salt inks that reduce to metallic silver when heated
08:21 < lsneff> A salt-based ink would still be conductive, no?
08:22 < muurkha> well, water is conductive; it's all a matter of degree.  some of the silver-complex inks have pretty low conductivities
08:24 < lsneff> Yeah, they can’t print water-based inks either
08:40 < muurkha> oh! that's pretty restrictive
08:41 < muurkha> how much are consumer brother printers where you live?
08:44 < lsneff> In the range of $150 to $300 for a consumer printer
08:52 < muurkha> pretty much the same as here then
08:52 < muurkha> sounds like a pretty cheap option!
08:57 < lsneff> If I can reverse engineer the printhead communication
08:58 < muurkha> maybe pick a model someone's already done
09:01 -!- CryptoDavid [uid14990@uxbridge.irccloud.com] has joined #hplusroadmap
09:04 < lsneff> Couldn’t find one
09:05 < muurkha> yeah, I'm having trouble too
09:05 < muurkha> seems to be more work on Epson printheads
09:13 < muurkha> http://www.piclist.com/techref/pcb/etch/custom-vs.htm copied from http://techref.massmind.org/techref/pcb/etch/custom-vs.htm (Epson CX4200/4800) http://dspace.mit.edu/bitstream/handle/1721.1/85435/870681346-MIT.pdf?sequence=2_1 (Epson Workforce 30 and Epson Artisan 50) https://web.archive.org/web/20120210162510/http://jeffreygough.com/files/27c3-file-print-electronics.pdf (Epson Stylus S21) 
09:13 < muurkha> https://www.ncbi.nlm.nih.gov/pmc/articles/PMC507883/ (Epson F057020 as used in Epson Photo 700)
09:13 < muurkha> those are also piezo apparently
09:17 < muurkha> some discussion on alternative printheads at https://ytec3d.com/forums/topic/step-1-hacking-the-cn642a-inkjet-printhead/
09:22 < muurkha> https://pubs.rsc.org/en/content/articlehtml/2018/ra/c8ra00756j (Epson Stylus SX235W)
09:23 < muurkha> actually that says "Epson is the only manufacturer of piezoelectric desktop inkjet printers"
09:23 < muurkha> are you sure the Brother printers are piezo?
09:23 < muurkha> in that case they didn't reverse-engineer the printhead protocol but instead just used the entire printer
09:29 < muurkha> dunno, is that helpful?
09:30 < muurkha> they found that the Epson printhead was 0.7-3 orders of magnitude better than the expensive Microdrop one
09:34 < lsneff> Yeah, it’s too bad they just keep the whole printer in the way
09:34 < lsneff> I bet I could RE it and I think the protocol is partially in the patents
09:36 < muurkha> yes, some of the pages I linked do link to patents
09:36 < muurkha> and the Royal Society paper did call out some of the limitations of having the whole printer in the way
09:37 < TMA> I have asked a guy that was working at HP at the time; there was not a process to sell just the printhead there. buy the whole and scrap the unneeded parts was the only option; that was HP printers, though, ymmv
09:44 < muurkha> depends on which HP printers.  I think the HP C6602, for example, is widely sold as a replacement part, because that's how they sell ink
09:44 < lsneff> The drop volume is just so small for the consumer printers
09:45 < lsneff> The industrial ones I was looking at (and the printhead used by the nano-dimension printer are ~40 pL, while the consumer printers are 3 or 4 pL
09:49 < muurkha> are you looking for larger drop volume?
09:50 < lsneff> Yeah, otherwise it’ll be absurdly slow
09:50 < lsneff> I’m not using this for binderjet, the printing material will be the ink itself
09:51 < muurkha> their comparison with the Microdrop head gives the Microdrop head as having a 20 pℓ minimum drop volume, while the Epson printer was 4 pℓ
09:51 < muurkha> in the RSC paper
09:52 < TMA> that was a few years back, before the HP/HPE split, it might have changed
09:52 < muurkha> but the Microdrop head can do up to 4000 drops per second, thus 80 nℓ/s, while the Epson head can do 1.2 million drops per second
09:52 < muurkha> 4.8 μℓ/s, about 50 times faster
09:52 < muurkha> not sure if the Konica Minolta heads you're looking at have similar limitations
09:54 < muurkha> do you know what kind of ink volume do you need, to within an order of magnitude or so?
09:55 < nmz787_> muurkha: I'm not really working on fabbing transistors, outside of the work I do at my day job (Intel). Planning on using e/i beam litho for fine features, and optical litho for much more coarse stuff like fan-out to macro connections
09:56 < lsneff> Say ~20 mL for a finished product
09:57 < lsneff> I’m not sure how to determine the frequency that the consumer printheads operate at
09:58 < nmz787_> ALD for deposition of very-thin electrodes (approximating a single nucleotide of DNA in thickness) for electron tunnelling discrimination, DRIE for etching fluidic cavities and vias (which would have electrodes embedded within, like that wafer-scale paper I linked earlier shows)
09:58 < nmz787_> lsneff: some of the more-available Xaar units have all that online
09:58 < nmz787_> http://diyhpl.us › XAAR_XJ128_Guide_to_Operation.pdf
09:59 < nmz787_> errr http://diyhpl.us/~nmz787/pdf/XAAR_XJ128_Guide_to_Operation.pdf
10:00 < nmz787_> 4.25 to 5.55kHz depending on specific submodels shown there
10:02 < lsneff> Ah, that’s a great resource thanks
10:02 < lsneff> Xaar printheads won’t work for this usecase unfortunately, since they apparently cannot use conductive inks
10:02 < muurkha> so at 4.8 μℓ/s you would need a bit over an hour for a final product; that's significantly longer than printing out a page on a regular printer but well within the range of 3-D printer printing times.  on the short end really
10:03 < nmz787_> lsneff: what makes you think they can't use conductive inks?
10:03 < lsneff> Yeah, an hour would be reasonable for printing out a PCB
10:04 < lsneff> nmz787: I read it online, from some group of people reverse engineering a xaar printhead.
10:04 < nmz787_> at least, I don't see anything like that in that PDF manual, after a few keyword searches anyway
10:05 < nmz787_> they only mention "approved ink"
10:06 < lsneff> Hmm
10:07 < lsneff> If they get back to me, I’ll ask them
10:08 < nmz787_> yeah, they mention needing 5 micron filtration
10:08 < nmz787_> so possibly the folks you read about were using emulsions of microparticles that were clogging things
10:08 < nmz787_> rather than actual conduction being the problem
10:10 < muurkha> unclogging a 7-μm-wide nozzle in a US$2000 printhead sounds like a pretty bad day
10:10 < lsneff> That’s possible
10:11 < muurkha> well.  only if you afil
10:11 < muurkha> *fail
10:11 < lsneff> I don’t even know how you’d unclog this
10:11 < muurkha> ultrasonic cleaner?
10:12 < muurkha> nmz787_: so you need photolithography patterning for metal but not for doping?
10:19 < nmz787_> it's really all about feature size
10:19 < nmz787_> you're not patterning a 20nm hole with photons
10:20 < muurkha> nope
10:20 < nmz787_> well, at least not without EUV
10:20 < muurkha> right
10:20 < muurkha> #notallphotons
10:25 < nmz787_> for doping though, I probably would just choose to grow N/P type layers, rather than traditional doping
10:25 < nmz787_> it just seems like the more modern technique
10:26 < nmz787_> and if I already have ALD/CVD style equipment... it seems like it wouldn't be too much of a stretch
10:26 < nmz787_> it's already common in industry too, so I'd guess I could find a recipe in academic journals
10:34 < muurkha> interesting, I didn't realize that
10:35 < muurkha> there might be some synergy: you could use E/I beam to pattern a resist, avoiding the whole EUV problem
10:36 < muurkha> won't scale to wafer scale at reasonable speeds, of course, but still might be faster than pure FIB milling
10:37 < nmz787_> yeah, that's the idea
10:37 < nmz787_> I am not aiming for producing wafers, just proof of concept, with a plan to scale to wafers with investment
10:38 < nmz787_> so much work out there shows things that are just not clear how to scale to high volume production, which is a major risk in terms of investment (or at least a major deterrent)
10:43 < muurkha> what I'm excited about is MEMS, making me deeply unfashionable, since AlphaCode is the approved excitement object today
10:47 < nmz787_> hadn't heard of alphacode before
10:47 < nmz787_> MEMS is only unfashionable to boring people :P
10:47 < muurkha> only boring people are interested in fashion in the first place
10:49 < muurkha> (but alphacode really is pretty interesting)
11:30 -!- Malvolio [~Malvolio@user/malvolio] has quit [Quit: brb]
11:31 -!- Malvolio [~Malvolio@user/malvolio] has joined #hplusroadmap
12:31 < kanzure> f-rep geometry kernel https://libfive.com/
12:31 < kanzure> cadquery gears https://github.com/meadiode/cq_gears
12:37 -!- phill [uid429774@id-429774.ilkley.irccloud.com] has joined #hplusroadmap
13:28 < maaku> muurkha: japan went from feudal middle ages to post-industrial revolution modernity in a single generation
13:28 < maaku> so peasants went from being tennant farmers to factory line workers
13:31 < maaku> if you've eve seen The Last Samurai, it was accurate in that respect at least
13:36 < muurkha> maaku: hmm, do was that good?
13:36 < muurkha> uh, was that good?
13:37 < maaku> i mean, if you like modern health care and personal liberty
13:37 < maaku> i'm unabashedly on the side of progress always
13:38 < muurkha> did life expectancy improve?  that'd show the results of health care improvement I'd think
13:38 < muurkha> did their personal liberty improve?  factory line workers were often pretty unfree
13:39 < maaku> > Life expectancy in Japan was 36.4 in the year 1860, and over the course of the next 160 years, it is expected to have increased to 84.4, which is the second highest in the world (after Monaco)
13:39 < muurkha> 01860 is *after* the Black Ships, right?
13:39 < muurkha> not sure I want to see movies, pretty sure none of Edward Zwick, John Logan, Marshall Herskovitz, or Tom Cruise lived through the Meiji Restoration or grew up hearing their parents' stories about it.  call it a racist stereotype if you like
13:40 < muurkha> I guess only 7 years after
13:40 < maaku> meiji restoration was in 1868, so that's when the major reforms started
13:40 < muurkha> do we know what life expectancy looked like in 01888 or so?
13:41 < maaku> muurkha: there were actors and set designers who had grandparents or great grandparent who lived through the meiji restoration
13:41 < maaku> the story itself is basically dances with wolves ... IN JAPAN!
13:42 < muurkha> perhaps some of the doors and stairways would provide me valid information about Meiji-era doors and stairways then
13:42 < maaku> and they completely invented much of the characters and events as a result, but the context (the early Meiji sets and general time period) are dead on
13:42 < maaku> not a great film to be honest, but great visuals if you want to see that time period
13:43 < muurkha> I want to know what living through it was like
13:43 < maaku> you might get a reasonably accurate impression of that
13:43 < muurkha> whether most people were better off as a result, or worse off
13:43 < maaku> why would they be worse off?
13:44 < muurkha> often social instability lets people fall through the cracks, usually the poorest
13:44 < muurkha> in Britain it took generations for industrialization to improve life expectancies instead of worsening them
13:44 < maaku> peasants were disposable before the meiji period though
13:45 < maaku> i mean yeah progress and social change can leave some people destitute, but in feudal times they'd be raped and killed by the local samurai with no reprecussions
13:45 < muurkha> true
13:46 < muurkha> I guess I feel like the intense cultivation of many forms of artisanship suggests that most people in Tokugawa Japan had fairly low risk of random violence though
13:46 < muurkha> much like Switzerland at the same time
13:46 < maaku> maybe in some parts of Edo that is true
13:46 < maaku> I doubt that applied to much of the countryside
13:47 < muurkha> the advent of posthumans is likely to be a Black Ships event for Earth
13:47 < muurkha> thus my interest
13:49 < muurkha> (and hopefully not a Columbus event)
13:49 < maaku> I personally doubt the transition will be so sudden
13:50 < maaku> well actually, maybe yes. I guess a half-century or so might be right
13:50 -!- darsie [~darsie@84-113-55-200.cable.dynamic.surfer.at] has quit [Ping timeout: 256 seconds]
14:21 < muurkha> I'm thinking probably a month to a year
14:29 -!- spaceangel [~spaceange@ip-78-102-216-202.net.upcbroadband.cz] has quit [Remote host closed the connection]
14:39 < pasky> the only way that occurs to me how this would make people worse off is that perhaps it played a role in fueling japanese imperialism and made personal outcomes during second world war worse? but i actually don't know if it mattered, japan was pretty militarized to begin with
14:55 < muurkha> dunno, that's a much larger timespan; it's hard to unravel all the causes that happened in that time
16:48 < maaku> muurkha: yeah no way anything is happening that fast
16:49 < maaku> you've been drinking the singularity kool-aid
17:42 -!- yashgaroth [~ffffffff@2601:5c4:c780:6aa0::f2f0] has quit [Quit: Leaving]
18:41 -!- CryptoDavid [uid14990@uxbridge.irccloud.com] has quit [Quit: Connection closed for inactivity]
18:56 -!- phill [uid429774@id-429774.ilkley.irccloud.com] has quit [Quit: Connection closed for inactivity]
19:54 -!- phill [uid429774@id-429774.ilkley.irccloud.com] has joined #hplusroadmap
23:13 -!- darsie [~darsie@84-113-55-200.cable.dynamic.surfer.at] has joined #hplusroadmap
23:23 -!- phill [uid429774@id-429774.ilkley.irccloud.com] has quit [Quit: Connection closed for inactivity]
--- Log closed Mon Feb 07 00:00:52 2022