**** BEGIN LOGGING AT Tue Sep 18 03:00:00 2018 Sep 18 07:19:58 alright back to fixing this pi fiasco Sep 18 07:24:48 the best way to fix a pi fiasco is by not using a pi Sep 18 07:36:08 ugh i didn't want to use the pi but i had to help this guy make a cape for it Sep 18 07:36:21 i don't know if its software or what, maybe because i'm using pb-less solder, but UGH DEBUGGING Sep 18 07:36:26 UGH PIS Sep 18 07:40:02 I avoid non-pb solder. Would only touch it for a commercial product with strict requirement for it. Sep 18 07:41:06 speaking of odd things in debugging. Sep 18 07:42:40 I used an orange-pi one last week (and promptly fried it by shitty usb-battery) and the kernel on it (4.14-somehting) would export GPIOs, but they were not active. By accident I figured out that you had to toggle back and forth the direction and then it would suddenly start working… Sep 18 07:43:22 I then used a BBB and it just worked… Sep 18 07:44:20 tbr: what's wrong with non-pb solder? Sep 18 07:44:39 the issue that _i'm having with it_ is a residue Sep 18 07:44:42 higher melting point, behaves differently for me Sep 18 07:44:47 and i dont quite know what's the right tool to clean it Sep 18 07:44:57 i have several solvents from chipquik Sep 18 07:45:13 i think i may also need a ultrasonic heated bath ? Sep 18 07:45:17 flux? I just use ethanol or isopropyl alcohol Sep 18 07:45:28 i haven't had luck with either Sep 18 07:45:33 :-/ Sep 18 07:45:36 it's not a flux residue either Sep 18 07:45:48 not a pure one anyway Sep 18 07:46:11 ayjay_t: depends on what kind of flux it uses. for non-rosin flux use alcohol (ethanol, isopropanol, etc), for rosin flux.. i don't know... probably aceton Sep 18 07:46:57 ayjay_t: what kind of residue is it anyways? Sep 18 07:47:51 it kinda looks like maybe it was flux with a small amount of white powder dissolved in it, if that makes any sense at all Sep 18 07:48:08 yeah, makes sense Sep 18 07:48:25 does the white powder appear after cleaning it, or was it there from the beginning? Sep 18 07:48:26 sn42/bi57.6/Ag.4 no-clean Sep 18 07:48:32 thats a good question Sep 18 07:48:37 i want to guess the former Sep 18 07:48:49 then it's likely you have a rosin flux Sep 18 07:49:33 okay so, i agree with that, becuase there was one batch that explicitly became difficult after cleaning with isopropanol and left a white powder Sep 18 07:49:48 but these boards have been cleaned with washable flux, water, soap, degreaser, and flux remover Sep 18 07:50:25 yeah.. but rosin fluxes are bitches to clean Sep 18 07:50:34 true Sep 18 07:50:38 flux type: synthetic no clean, RELo Sep 18 07:50:47 o_0 Sep 18 07:50:55 i mean, if its rosin, that is = tacky? Sep 18 07:51:19 not necessarily Sep 18 07:51:20 I use a no-clean flux from MG chemicals, but it's a rosin type and can be really icky to remove Sep 18 07:51:33 ayjay_t: https://www.kester.com/Portals/0/Documents/FAQs/White_Residue.pdf Sep 18 07:51:45 but since it's no-clean, I usually don't make an effort to really remove it Sep 18 07:51:56 I used some weller flux recently and it seemed to be rosin, it was in solution of some alcohol and also went away instantly with ethanol. Sep 18 07:52:42 thinkfat: i love "no clean"... you don't need to clean it until you deviate from the usual consumer grade, throw-away electronics Sep 18 07:52:56 thinkfat: for everything else, you actually need to clean it of, as it will cause problems Sep 18 07:53:26 so uh, i mean, jeez i have degreaser and flux remover Sep 18 07:53:36 i don't want to do this mechanically Sep 18 07:54:03 ayjay_t: why do you think there is problem with the residue on the board? Sep 18 07:54:38 uh i think it adds a small conductivity to the whole boards Sep 18 07:54:56 ayjay_t: degreaser does not help with flux as it is not fat Sep 18 07:54:59 at this point everything is behaving really weirdly and i'm deconstructing the clients software to do some kind of unit tests Sep 18 07:55:18 and hoping to use a multimeter to detect whats crashing the board when i isolate the board section Sep 18 07:55:20 ayjay_t: quite a few of those molecules are aromatic and polar Sep 18 07:55:41 ayjay_t: so, standard degreaser will not be able to disolve them Sep 18 07:55:42 and then i'm going to go over that area mechanically to determine the nature of the problem Sep 18 07:55:50 yeah koth if its an emulsifier... Sep 18 07:56:23 ayjay_t: that's why i recommended aceton... it's not nice to sensitive stuff, but it can disolve quite a few aromatic molecules Sep 18 07:56:50 i think i've tried it on the last batch with little success Sep 18 07:56:54 i'm using this "flux remover now" Sep 18 07:57:01 :-/ Sep 18 07:57:21 ayjay_t: what kind of board is it anyways? Sep 18 07:57:26 consumer grade Sep 18 07:57:50 actually now that i'm thinking about it, i think the residue gets underneath the IC and i need to remove the IC, clean the resude, and then solder and clean it more carefully Sep 18 07:57:51 why do you need to remove the flux then? ;-) Sep 18 07:57:58 i think that its conductive Sep 18 07:58:03 o_0 Sep 18 07:58:12 that would be weird Sep 18 07:58:21 well i mean, it disolves solder, no? Sep 18 07:58:24 rosin flux is used exactly because its non-conductive Sep 18 07:58:44 it does not disolve solder, it helps wetting surfaces by removing oxides Sep 18 07:59:01 https://en.wikipedia.org/wiki/Flux_(metallurgy) Sep 18 07:59:23 hmm i need a guide on this Sep 18 07:59:33 i wonder if that wikipedia would help Sep 18 07:59:46 it explains a few things, yes Sep 18 07:59:49 okay well the manual flux i used is a water soluable liquid Sep 18 08:00:00 but i thought we were talking about the flux the solder beads are dissolved in from the paste Sep 18 08:00:16 they are not disolved Sep 18 08:00:31 they form a paste Sep 18 08:01:17 oh okay thats a mixture Sep 18 08:02:00 the flux in the solder paste is more kind of a glue thing, that holds the solder grains together. otherwis you'd have to handle a non-sticky poweder that gets everywhere instead of sticks to the pads only Sep 18 08:02:17 powder* Sep 18 08:03:08 it is true that flux residue can become conductive over time, when it attracts dust and thus water Sep 18 08:03:28 yeah hmm Sep 18 08:03:37 but we are talking about MΩ levels here, nothing that you would see in digital circuits Sep 18 08:03:42 i really think its the residue but ill just have to deconstruct this board and figure it out Sep 18 08:04:18 temp sensors thrown off my megaohms at first, but now this board is bringing itself and the pi down when SPI gets goin don't know why need to test it Sep 18 08:04:54 i bought a whole assortment of supplies from chipquick Sep 18 08:04:55 it's more likely that you have solder bridges or unmolten solder grains froming bridges underneath ICs Sep 18 08:05:01 that might be possible Sep 18 08:05:28 i've inspected it with a microscope and dental picks and stuff tho Sep 18 08:05:29 so Sep 18 08:06:43 ¯\_(ツ)_/¯ Sep 18 08:06:46 can be anything Sep 18 08:06:56 yeah it really can Sep 18 08:07:15 the only way to do this is to see what unit tests fail, characterize that board section, and then use a process of elimination in board repairs Sep 18 08:07:24 * KotH could offer his expertise as master electronics wrangler, but ayjay_t sits on the wrong side of the big pond Sep 18 08:07:30 ayjay_t: how complex it the board? why do you think the design is susceptible to flux residue? Sep 18 08:07:31 what i'm really hoping is that if i do it for one board and then just skip to the final step on the rest i can bang these outs Sep 18 08:08:15 well thinkfat i just, i had some issues that were clearly a residue problem before, and were throwing off thermistors and stuff, and now these boards are having issues so i'm sorta raw from that Sep 18 08:08:41 if you really want to help me, you'll give me some feedback i've invested my life savings in developing this summer Sep 18 08:08:52 on a product i've invested my life savings in* Sep 18 08:09:11 i need to dispatch this contract so that i can help the two young men who have been building it for me Sep 18 08:09:16 hence why i'm awake at 4 am Sep 18 08:09:31 you're in quite a bind there are you? Sep 18 08:09:41 i want these boards out of my office Sep 18 08:10:58 i don't mean to be so negative, sorry Sep 18 08:11:13 its been an englightening conversation it's help a lot i didn't realize i still had so much to learn Sep 18 08:11:15 tends to happen after a night of fruitless tinkering Sep 18 08:11:46 i thought if i bought the right combination of "cleaners" "doers" and "washers" from chipquik i'd have the rocks-paper-scissors of soldering lol Sep 18 08:12:38 ayjay_t: but your approach is fine in principle I think, narrow down the search field by checking what actually fails Sep 18 08:12:47 turns out it's more like this: http://4.bp.blogspot.com/_wCzCY93mKq0/TFJHFjZd2jI/AAAAAAAABVI/7cAz1surslY/s1600/Advanced+Rock+Paper+Scissors.jpg Sep 18 08:13:14 ayjay_t: is it just some particular function not working or is it just generally flaky behaviour? Sep 18 08:13:48 its honestly been changing as i further treat and clean it Sep 18 08:14:02 oh... Sep 18 08:14:03 i wonder if the degreaser actually made it worse if its a rosin flux Sep 18 08:14:40 what interest do we have in creating a wikipedia page that looks at common products on the market and creates recipes for people? Sep 18 08:14:52 because i kind of want to document all this properly and make a proper guide Sep 18 08:16:31 here actually Sep 18 08:17:03 https://imgur.com/a/GUSMmgY Sep 18 08:17:42 you can also find my LCD adventures there with the beaglebone black Sep 18 08:17:44 those are the best Sep 18 08:17:57 several years old now tho Sep 18 08:18:11 i'm going to measure that white powdery bridge Sep 18 08:19:06 that looks really weird Sep 18 08:19:14 ayjay_t: well, if you want professional advice, that will cost you something... preferably high quality chocolate :) Sep 18 08:19:33 :-) Sep 18 08:19:40 hmm 10M Sep 18 08:20:54 KotH: I built an online application to automatically generate pogo-boards with power, analog in/out, DIO Sep 18 08:20:57 "I" Sep 18 08:21:19 you upload your gerbers and enumerate which points you want to test up to about 25ish, and have the pogoboard shipped to you Sep 18 08:21:35 * KotH blinks Sep 18 08:21:38 WANT! Sep 18 08:21:40 you can access it remotely and use the application to track programming/testing at a contract manufacturer, for example Sep 18 08:22:38 i'm glad you do because if no one wanted it i would be very much broke and out of luck Sep 18 08:22:38 really neat Sep 18 08:22:49 ayjay_t: https://imgur.com/a/GUSMmgY pretty much looks like what i'd expect from rosin flux being cleaned with alcohol Sep 18 08:22:58 yeah Sep 18 08:23:11 autopogo.com Sep 18 08:23:24 ayjay_t: as that is a chip from Linear, i guess it does some analog black magic? Sep 18 08:23:40 tbh i think its one of their dudes Sep 18 08:23:47 duds* Sep 18 08:23:53 and they managed to push it off onto my client somehow Sep 18 08:23:57 its an ltc2983 Sep 18 08:24:03 KotH: yup, I have a similar white residue on a board here where I tried cleaning off my flux with IPA Sep 18 08:24:24 its pretty much just a low speed fpga-come-asic that wants to measure temperature? Sep 18 08:24:35 ayjay_t: your customer has too much money! Sep 18 08:25:00 you're not wrong Sep 18 08:25:04 he's basically like a rich child Sep 18 08:25:45 analog circuitry could indeed be affected by residue-caused conductivity, depending on what impedances you're delaing with Sep 18 08:25:52 but 10M is really high... Sep 18 08:25:54 ayjay_t: i know that chip.. i saw it and thought, "dang! that's a nice chip for doing temp measurements, i'm gonna use it in my next design"... then i saw its price Sep 18 08:26:16 thinkfat: 10M is low for a 24bit ADC Sep 18 08:27:13 but who needs such insane amount of dynamic range? Sep 18 08:27:13 thinkfat: 10M means that you get 1LSB of offset, when measuring a sensor with a source resistance of >1Ω Sep 18 08:27:25 thinkfat: BRUAHAHAHAH! Sep 18 08:27:43 pardon? Sep 18 08:27:50 thinkfat: i have a DMM that has an ADC with >28bit... ENOB! Sep 18 08:28:05 Hi I need some help :) I cant find any answer on the website . Sep 18 08:28:10 yes, but we're talking consumer grade electronics, right? Sep 18 08:28:26 thinkfat: i regularly do measurements where I have ~30bits resolution (ok, it's time/frequency and not voltage/resistance, but still) Sep 18 08:28:54 Can someone tell me how long will be the beagle bone black available ? is that worth for a company to start to waork with this ? Sep 18 08:29:00 thinkfat: if you want to measure temp to 0.1°C, then you need at least 20bit of ENOB Sep 18 08:29:10 thinkfat: and a lot of consumer electronics does that Sep 18 08:29:28 bruce: it will be available at least until next week Sep 18 08:29:41 bruce: probably also the week after that Sep 18 08:30:48 KotH: it also depends on the dynamic range. I can measure .1 degree just fine with 4 bit if I don't need to cover much range Sep 18 08:31:21 mmh my question is more like : will this be available in 5 -10 years ? thanks for your answer Sep 18 08:31:38 thinkfat: true... but then you do a lot of low-noise analog circuitry that needs to be adjusted at production time.. possibly even during use Sep 18 08:31:40 bruce: the design files are all open source Sep 18 08:32:13 bruce: you can have your own boards built if you need to. Sep 18 08:32:33 bruce: so it becomes a question on how long the key components will be available Sep 18 08:33:14 bruce: "this" is a very vague descriptor Sep 18 08:33:26 bruce: the BBB has not been around for 10 years yet Sep 18 08:33:41 bruce: heck the whole BB line is barely 10y old Sep 18 08:33:53 bruce: and we went to several boards and each board through several revisions Sep 18 08:35:27 KotH: but with such requirements for residual resistance you'll have to be quite careful about the system design anyway. Like, potted sections on the pcb for the critical stuff. Sep 18 08:35:31 bruce: the processor is in the TI catalog, which means it will be available for at least 10 years _after_introduction_ Sep 18 08:35:47 bruce: for how long exactly, nobody knows Sep 18 08:35:53 KotH: I have a defective tympanic thermometer here which has that Sep 18 08:36:27 thinkfat: potting has exactly the reverse efect, it makes leakage current less predictable Sep 18 08:37:20 "flux remover" is actually doing pretty good Sep 18 08:37:27 cool Sep 18 08:37:29 KotH don't you think an uC with a nice array of ADCs would be better Sep 18 08:37:42 if you guys want me to pitch your companies into a beta program for autopogo.com i'd be happy to Sep 18 08:37:43 ayjay_t: if it works, please let me know what product you used :) Sep 18 08:37:49 ayjay_t: no, uC ADCs are shitty Sep 18 08:38:00 its still requiring a bit of mechnical action though Sep 18 08:38:17 ok thanks a lot KotH ;) will check for more information on TI website Sep 18 08:38:25 ayjay_t: the digital environment they live in makes so much noise that you lose a few bits (usually in the order of 2-3 bits) just because it's integrated with the uC Sep 18 08:38:54 yeah i guess Sep 18 08:39:01 ayjay_t: my "company" is a research institution in germany... Sep 18 08:39:20 i mean, whoever writes checks for your equipment but uh, germany might be hard for me Sep 18 08:39:27 ayjay_t: and i'm the only practicing EE on campus Sep 18 08:39:30 i really need to get my eu friend to help me Sep 18 08:39:42 HAHA KotH then you really need my help XD Sep 18 08:39:57 you must do mostly just one off designs Sep 18 08:40:04 ayjay_t: jupp Sep 18 08:40:15 ayjay_t: it's called "research" and "science" :-D Sep 18 08:43:27 a noble pursuit Sep 18 08:46:08 KotH: like, fraunhofer? Sep 18 08:47:08 thinkfat: almost Sep 18 08:47:13 thinkfat: less practice, more theory Sep 18 08:47:37 KotH: re potting, but you need to do something to prevent residue buildup, right? Sep 18 08:49:49 ayjay_t: in the german speaking world, reinhardt is pretty widely used for test adapters (because they are cheap, yet decent quality): https://www.reinhardt-testsystem.de Sep 18 08:50:18 this is my competition, eh? Sep 18 08:50:25 ayjay_t: if you add their dimensions as defaults for your app and delvier plates that fit them, that would make it great Sep 18 08:50:53 ayjay_t: nope. reinhardt does only the fixtures. they, for some unknown reason, refuse to sell drilled plates Sep 18 08:51:21 ayjay_t: so, you buy the fixture from them, a plate and then you have to find someone who can take gerber files and drill the plate and put some pogo plugs in Sep 18 08:51:57 https://www.reinhardt-testsystem.de/deutsch/adaption/pruefadapter.php Sep 18 08:52:23 ayjay_t: yes Sep 18 08:53:05 ayjay_t: if you want drilled plates, the best you can do with reinhardt is to buy their drilling system: https://www.reinhardt-testsystem.de/deutsch/adaption/reinhardt_bohrcenter.php Sep 18 08:53:31 ayjay_t: which is overkill for most SMBs as they only need just a few test fixtures per year Sep 18 08:53:59 KotH: less practive more theory explains why you're the only EE on the campus :-) Sep 18 08:54:20 oh i see, the contacts are on the bottom Sep 18 08:55:23 the idea is that we'd probably charge, say, $1500 for the first iteration, $1200 for the second, then $700, then $300 Sep 18 08:55:32 and stay at $300 Sep 18 08:55:58 thats a project price, so you can get multiple copies of the board so that you can have multiple people test them, ship them, send them to manufacturer, etc Sep 18 08:56:41 but this is the whole system i'm talking about, it includes the DIO/AIO etc Sep 18 08:58:39 for you tho, that system is pretty good? because you are the only one who needs to use the station, and you don't need to go anywhere Sep 18 09:00:44 thinkfat: campus means here an university campus where the institute is located at Sep 18 09:01:00 thinkfat: the university closed its EE program down >5y ago Sep 18 09:01:12 thinkfat: ah such kind of research institute Sep 18 09:01:28 why am I talking to myself? Sep 18 09:01:33 * thinkfat needs a coffee Sep 18 09:01:55 ayjay_t: ah.. so you are for buidling multiple test systems, not just one-offs? Sep 18 09:03:07 i guess it depends on how it will be used, but i figure more than one engineer might use it Sep 18 09:03:11 thinkfat: https://www.mpi-inf.mpg.de/ Sep 18 09:03:42 thinkfat: i'm in D1, the theoretical computer science group Sep 18 09:03:57 and i want the engineers to iterate the test process along with the product itself, so that they can use the test board immediately in the development process to unit test, impliment continuous integration, program, etc Sep 18 09:04:57 ayjay_t: hmm... most test systems I know off are one-offs. unless the company does large volume production (>1k per month) Sep 18 09:05:17 ayjay_t: multiple programming adapters and such are usually not needed, as they go onto connectors that are easily accessible Sep 18 09:06:16 ayjay_t: and be carefull with the price, 1500 is enough money to pay an engineer for at least a day, if you live in chocolate country, or a week if you live in 3rd world countries like germany Sep 18 09:07:45 what do you mean by one-offs? Sep 18 09:08:01 you build one and only one test system for production Sep 18 09:08:13 there is no need for multiple unless you are doing large volume production Sep 18 09:08:20 oh right Sep 18 09:08:32 part of that is because of the production headache of making multiple though Sep 18 09:08:33 on-desk testing is not usually done with a testadapter Sep 18 09:08:38 right Sep 18 09:08:43 this is what we're trying to enable ahha Sep 18 09:09:00 mostly because there is no need. Sep 18 09:09:13 you are building 1000 devices... total Sep 18 09:09:24 and it's usually impractical for desk testing... Sep 18 09:09:25 i'm not quite sure how you're deducing that logic Sep 18 09:09:31 you teach one guy/gal to do the testing (you don't have money to spend on teaching two) Sep 18 09:09:37 but the bottom line is that we save engineers time Sep 18 09:09:42 it then takes a week or so to east all boards Sep 18 09:09:43 and engineers time is expensive Sep 18 09:09:58 no engineers in production Sep 18 09:10:04 yup Sep 18 09:10:05 only techniciancs Sep 18 09:10:20 mostly women with very little education Sep 18 09:10:29 (at least in europe and asia) Sep 18 09:10:38 trained hands Sep 18 09:11:46 heck, if you have one guy/gal on the production floor who knows what electronics is, you are pretty lucky as a customer Sep 18 09:12:31 anyway, a test jig makes sense mostly if you have many boards to check. Sep 18 09:12:40 which usually engineers will not do Sep 18 09:12:43 thinkfat: where many can be 10 :) Sep 18 09:13:27 but yes, usually, test fixtures are not built until you have to do >100 boards Sep 18 09:13:29 KotH: you mean, at prototype phase where you need to check production spread? Sep 18 09:13:41 i mean, the test boards we make are a bit different, they have different features, they are a different form factor Sep 18 09:13:52 it takes less time to make them, you can have more of them Sep 18 09:13:58 and you can iterate on them cheaply Sep 18 09:14:06 thinkfat: i've done boards where the total production volume was 5 Sep 18 09:14:15 so i think, the idea is that there are some places where this could lead to changed behaviors Sep 18 09:14:22 thinkfat: and each of them was tested in a fixture, to have reproducible tests Sep 18 09:14:40 ayjay_t: hmm... Sep 18 09:14:48 but there are also "classic" applications where we are competitive in a traditional sense Sep 18 09:15:00 and those are easiest for us to talk to because we don't have to access the imagination Sep 18 09:15:01 ayjay_t: then you have to explain it better on your website... ok, explain it at all :) Sep 18 09:15:21 we're _mostly_ talking to people who are actively having quality control problems in contract manufacturers right now Sep 18 09:15:27 because it is VERY obvious to them that they need this Sep 18 09:16:18 ayjay_t: if you have QA problems in manufacturing, then you need proper testing. what do you offer that does not exist yet? Sep 18 09:16:50 but quite honestly, if the contractor delviers shit, change contractor Sep 18 09:17:07 besides timeframe and price, we also offer realtime monitoring of the testing process Sep 18 09:17:27 * KotH has a quite large list of contractors that are _not_ to be used, and a small list of contractors that do a teriffic job Sep 18 09:17:53 ayjay_t: ah.. you want to offer testing monitoring solutions? Sep 18 09:18:14 that is part of the package yes Sep 18 09:18:16 KotH: have you ever used one-stop-shop facilities like allpcb.com, or circuithub? Sep 18 09:18:20 ayjay_t: ie, input board data, output test system wth monitoring and reporting? Sep 18 09:18:21 you can see each board and its results as it happens Sep 18 09:18:41 we expect engineers to write some specific tests, tbh Sep 18 09:18:48 expected uart, spi results etc Sep 18 09:18:56 thinkfat: for private stuff, yes. but i wont use them anymore. they are too expensive..... thanks obama....er.. german taxes! Sep 18 09:19:19 thinkfat: for professional stuff, i recommend eurocircuits Sep 18 09:19:33 thinkfat: they are more expensive, but deliver very consistent and high quality Sep 18 09:19:34 KotH: yeah, for very small volumes they're hardly competitive against DIY Sep 18 09:20:08 thinkfat: with german taxes, they are not competitive against eurocircuit which produces in the EU and thus is "tax-free" Sep 18 09:21:28 thinkfat: e.g. I did recently 10 boards on pcbway. they were approx 2/3 of the price of eurocircuits. but i had to go to the customs office to pick it up (which takes half a day) and paid 100€ in additional taxes Sep 18 09:21:59 thinkfat: it's almost like germany does not want me to do any business with anyone outside the EU Sep 18 09:22:46 KotH: I've used jlcpcb for small boards prototypes, like < 100x100 mm and up to 4 layers with good results Sep 18 09:23:12 thinkfat: pcbway is slightly higher quality than jlcpcb Sep 18 09:23:21 KotH: very competitive price especially for 2 layer boards Sep 18 09:23:23 thinkfat: but the level is comparable Sep 18 09:24:12 thinkfat: the competitive price does not help if i still have to pay ~50% of the price in additional taxes and fees Sep 18 09:25:01 KotH: but the 2layer boards were like less than 10€ including shipping :) Sep 18 09:25:08 KotH: for 10 pieces Sep 18 09:25:30 KotH: I didn't have to pay any taxes or other duties Sep 18 09:26:27 KotH: the 4layer boards were around 30€ including shipping, I think I paid like 10€ extra in taxes and fees Sep 18 09:27:13 fits right into a hobbyist budget Sep 18 09:27:56 thinkfat: the jclpcb boards i did were 4 layers, 8.5x3cm, cost me about 70€ plus another 40€ in customs/fees (thanks DHL!) Sep 18 09:28:26 yes, I'll never use DHL again Sep 18 09:28:39 i'm in germany, there is no way around dhl Sep 18 09:28:43 pcbway + dhl = faster than most american board mfrs Sep 18 09:28:53 KotH: eh? Sep 18 09:28:56 ayjay_t: jupp, pcbway is damn fast Sep 18 09:29:08 thinkfat: the german post "outsourced" its parcel delivery to dhl Sep 18 09:29:10 KotH: depends on whether you also order a stencil I think Sep 18 09:29:31 thinkfat: no stencil, just 10 boards Sep 18 09:29:42 oh Sep 18 09:29:43 i really wish there a mfr who would let me make crazy silk screens Sep 18 09:29:44 hehe Sep 18 09:29:55 ayjay_t: crayz? Sep 18 09:31:27 KotH: interesting. The last job I did with jlcpcb was 10 pieces 4 layer 87x55mm, total cost including shipping 35€ Sep 18 09:31:39 KotH: I used the "ePacket" shipping method Sep 18 09:31:50 KotH: it basically arrived here as a "registered letter" Sep 18 09:32:01 KotH: through "Deutsche Post", not DHL Sep 18 09:32:33 KotH: no fancy options, though Sep 18 09:33:08 KotH: no ENIG coating, just HASL leaded, no other bells and whistles Sep 18 09:33:33 KotH: they really nickle-and-dime you for anything special you want to order Sep 18 09:34:30 thinkfat: i choose that as well.. but their shipper refused to handle it without a VAT number, which of course I do not have Sep 18 09:34:32 KotH: yeah multiple colors and fine resolution Sep 18 09:34:47 KotH: that is really quite interesting... Sep 18 09:35:02 thinkfat: can you spell tschörmäni? Sep 18 09:35:10 hehe Sep 18 09:35:30 thinkfat: i think .de is the most hellish country if you want to start a company Sep 18 09:35:43 oohhh really Sep 18 09:36:09 heck, if i would want to build 10 boards of something and sell it at cost to friends and family, i would need to register a small busines Sep 18 09:36:24 even for a *beeeeeep* one time job Sep 18 09:37:44 currently, my needs of selling boards are handled through switzerland, which is cheaper and less hazzle even though it's not in the EU Sep 18 09:38:02 :-O Sep 18 09:38:04 what do you sell? Sep 18 09:38:31 just small gizmos that i've designed and built which friends find usefull Sep 18 09:38:41 do you publish them? Sep 18 09:38:44 like usb temp logger and stuff like that Sep 18 09:38:56 but only a couple of them? Sep 18 09:38:57 not so far Sep 18 09:38:59 oooh Sep 18 09:39:26 the usb temp logger might end up somewhere public, as i think that everythign you can find currently is either horribly overpriced or just plain shit Sep 18 09:39:56 need to finish up the software and make it presentable, though :) Sep 18 09:41:20 ooh i see Sep 18 09:42:08 it's always the same. hardware is done quickly, software takes a long time :) Sep 18 09:43:09 oh.. and i need to replace the LDO Sep 18 09:43:58 the TLV70033 i used is a sensitive bitch... 30% of them dont survive hand soldering Sep 18 09:45:12 pollin just delivered my DIN-Rail case for the bbbw Sep 18 09:45:29 one step closer to finishing this project :) Sep 18 09:45:50 finishing projects is overrated Sep 18 09:46:06 KotH: my last project was a cape for garden irrigation Sep 18 09:46:06 just publish a paper with preliminary results and move on Sep 18 09:46:07 ;) Sep 18 09:46:18 KotH: opensprinkler compatible so I don't need to bother with software :) Sep 18 09:47:08 I mostly try to hotair solder the stuff Sep 18 09:48:08 works a treat, but a pain in the neck if you use leadfree paste Sep 18 09:48:45 maybe I'll refurbish a toaster into a reflow oven, there's plenty of kits for that Sep 18 09:51:00 hmmm, the am3358 has some temperatur sensor on the die, no? Sep 18 09:51:24 I don't find it mentioned in the device tree Sep 18 09:51:43 * KotH doesnt like IR soldering Sep 18 09:52:01 it does not lead to repeatable results unless you have time to adjust all settings Sep 18 09:54:07 roasting the board with a hotair gun even less so ;) Sep 18 09:54:34 only thing i've ever roasted was the table underneath the boards :) Sep 18 09:55:11 KotH: hehe Sep 18 09:55:12 yes Sep 18 09:55:16 me too, almost Sep 18 10:02:38 hm, will require some dremel action for additional cut outs Sep 18 10:03:33 think big, use a flex! Sep 18 10:29:00 https://www.contorion.de/elektrowerkzeug/dolmar-pc-7414-trennschleifer-10023328?gclid=CjwKCAjwxILdBRBqEiwAHL2R81BQk7sgUCPZVTgHknjVilzdN1gWGAPtWyMgh1dQ_l_uW_wmMA3gxBoCMU4QAvD_BwE&aid=260349748535&targetid=pla-474974397501&ef_id=WSNHVwAAAVD-mFQm:20180918102829:s Sep 18 10:29:03 tempting Sep 18 11:06:34 if in doubt - C4! Sep 18 12:08:44 thinkfat_: hmm... i think i still have some budget for equipment left... Sep 18 12:09:41 tbr: yes, I'm known to regularly C4 holes in my project cases Sep 18 12:15:38 the alternative are of course sharks with friggin' lasers attached to their heads Sep 18 12:39:35 tbr: i dont think i have enough budget to buy sharks with lasers Sep 18 14:32:21 m Sep 18 14:33:03 kremy-krem! Sep 18 14:33:14 Do you own the MotorCape yet? Sep 18 14:37:47 i do not Sep 18 14:37:48 was ist das Sep 18 14:37:50 Hey? Sep 18 14:37:59 What happened to my name? Sep 18 14:38:04 I used to be set_. Sep 18 14:38:07 Dang it. Sep 18 14:38:10 Please hold. Sep 18 14:38:13 you are 'milti' no Sep 18 14:38:15 now * Sep 18 14:38:20 I guess. Sep 18 14:38:38 Hey kremlin: The MotorCape was produced by GHI. Sep 18 14:38:57 beagleboard.org had them make it. Sep 18 14:39:02 It is a nice Cape. Sep 18 14:39:17 I am trying right now to make it move some motors. I am a big failure for now. Sep 18 14:41:55 oh neat Sep 18 14:41:58 ill have to check it otu Sep 18 14:46:24 It is a nice piece of machinery but I have not gotten it to work just yet. Sep 18 14:46:25 ... Sep 18 14:49:04 I can make the LED for MotorTwo turn on and off but that is it. Sep 18 15:07:02 See here:. Sep 18 15:07:03 ... Sep 18 15:07:47 I think since we are dealing w/ L298 controllers on the Cape, we need to set the pin Low, then to High, and then finally we can change direction. Sep 18 15:07:48 ... Sep 18 15:08:02 At least...that is what I did w/ an older L298 chip. Sep 18 15:08:46 But...b/c of how things are hardcoded in, there is only one GPIO channel to use instead of two (and two are needed). Sep 18 15:09:45 brb (I am going to try the other username). Sep 18 15:10:18 Yea! Sep 18 15:13:55 Can I set a pin Low somehow in userspace? Sep 18 15:23:06 Okay. Sep 18 15:24:04 How can I put a u-boot overlay onto the BBBW, i.e. this one particularly --> https://github.com/RobertCNelson/bb.org-overlays/blob/master/src/arm/BBORG_MOTOR-00A2.dts? Sep 18 15:24:37 someone in here told me to erase this Cape overlay. I did it. Now, I think I want it back. Sep 18 15:24:40 ... Sep 18 15:24:44 Send supplies! Sep 18 15:31:36 I got it. Forget about it. Sep 18 15:39:53 @jkridner: Sir! Sep 18 15:40:01 Did you get your MotorCape to work yet? Sep 18 15:40:36 hello. my motor cape has been working before they started shipping. Sep 18 15:40:45 Oh. Sep 18 15:41:09 I just used the sysfs entries for pinmux, pwm and gpio Sep 18 15:41:24 Oh. Sep 18 15:42:24 I think there needs to be two GPIO channels available per L298 chip to work w/ software. Sep 18 15:42:31 Am I close/ Sep 18 15:45:15 https://pastebin.com/rHhfiUEQ is something I typed up to test things out. I am way off so far. Sep 18 16:11:20 tbr: brag, brag, https://plus.google.com/+MatthiasWelwarsky/posts/SnChf6urYQk Sep 18 16:12:16 thinkfat_: *pat* *pat* haddugudemacht! Sep 18 16:13:44 tbr: I'm especially proud of the via placement that lets the Wifi LED shine through totally deliberately Sep 18 16:14:45 :D Sep 18 16:14:47 nice Sep 18 16:16:02 are there remote control apps for the bbb that would let me check system status and maybe power down the device, reboot it... Sep 18 16:16:05 other than ssh Sep 18 16:20:47 thinkfat_: you could probably use node-red, though it drags in nodejs... (I managed to make that stack work on a 100MB RAM device though) Sep 18 16:21:51 tbr: use as in "install and it works" or in "it's a framework you can use to build what you need" ? Sep 18 16:23:00 in the middle-ish. it has this flow plug stuff into each other thing that lets you quickly define the few things you want to do Sep 18 16:24:15 https://tech.scargill.net/wp-content/uploads/2015/09/flows.jpg Sep 18 16:24:34 https://nodered.org/images/dashboarde.png if you want some fancy dashborad too Sep 18 16:25:41 I was thinking of something like pi-control for the raspi Sep 18 16:25:48 much less mundane Sep 18 16:27:00 *nod* Sep 19 00:38:11 Still only LEDs for this fellow! Sep 19 00:38:22 Any motor enthusiasts out there w/ a MotorCape? Sep 19 00:39:55 brb Sep 19 00:54:28 set_: https://pastebin.com/ypQ3hpVb Sep 19 00:55:35 the ideal pwm frequency will probably depend on the motors used, I've picked 2 kHz fairly arbitrarily here Sep 19 00:56:16 this example assumes you've removed (or otherwise disabled) the overlay and configured the four pwm pins using config-pin Sep 19 00:56:42 Right. Sep 19 00:56:52 Do I need to configure all of them? Sep 19 00:57:13 oh and the sleep 1 should be time.sleep(1); sorry, wasn't paying attention when I typed that (I'm more used to perl than python) Sep 19 00:58:02 I would recommend configuring all four pwm pins yes Sep 19 00:58:05 I have just been using this example to run motor2: https://pastebin.com/4vpyvJg1. Sep 19 00:58:17 Okay. Sep 19 00:58:24 I was just configuring one motor at a time. Sep 19 00:58:51 The LED for MotorTwo shows running on/off, i.e. LED lights and then goes off. Sep 19 00:59:13 Okay. I am just going to copy and paste that idea. I will let you know what happens. Sep 19 00:59:53 ... Sep 19 01:00:08 I know it is arbitrary now but the else clause runs correctly but the motor does not move. Sep 19 01:00:13 On to running your software. Sep 19 01:02:05 well your code sets the duty cycle to 75% and then almost immediately sets it back to 0, the motor never has time to move. I'm assuming you're just seeing a really brief flash of the led? Sep 19 01:03:20 Yes. Sep 19 01:03:30 So, when the LED is lit/on, the Motor is supposed to move? Sep 19 01:04:02 in principle yes, but your eyes can see very very brief flashes of a light Sep 19 01:04:23 can = cannot? Sep 19 01:04:27 can Sep 19 01:04:30 Okay. Sep 19 01:04:33 ahhahaha. Sep 19 01:04:37 while a motor can't instantly begin moving Sep 19 01:04:45 Sorry. Oh. Now, I understand you. Sep 19 01:05:01 Right, it takes time and momentum, esp. for my geared motors. Sep 19 01:05:17 I could put a time.sleep(6) in there. Sep 19 01:05:33 Before making the motor turn off. Sep 19 01:06:16 Let me just try your software real quickly. I have to config-pin everything but this one motor should turn. Sep 19 01:10:41 The software ran but the motor did not move! Sep 19 01:11:06 Do you want a visual of my set up? Sep 19 01:12:09 Battery/12v to + on the Cape. GND to GND on the Cape. Motor plugged in via red on + and black on -. Sep 19 01:12:47 This is located on Motor2. Sep 19 01:12:53 On the Cape. Sep 19 01:13:59 did you change "m1" to "m2" in the tests of my example? Sep 19 01:14:23 (since your motor is connected to m2 rather than m1) Sep 19 01:14:32 Nope. Sep 19 01:14:40 Let me see. I will check. Sep 19 01:14:55 so you're controlling the motor1 output and wonder why the motor connected to the motor2 output doesn't move Sep 19 01:15:07 -.- Sep 19 01:15:15 Oh! Sep 19 01:15:54 My LED turns on and off on Motor2. Sep 19 01:16:07 when running my example? Sep 19 01:16:18 No...when running my example. Sep 19 01:16:28 I already explained why your example didn't work Sep 19 01:16:32 I know. Sep 19 01:16:55 Look. I know you are upset. Take a breather. Sep 19 01:17:11 Please hold and I will change everything to make it like you described. Sep 19 01:21:09 No-go! Sep 19 01:21:29 The LED blinks on and off, though. Sep 19 01:23:05 I changed the lower bunch of software to motor2, (m2_dir, xxxx.xxxx) respectively, but the motor2 would not go or start. Sep 19 01:24:14 Now, I know it is not the battery. I just checked it twice. I can check the motor w/ another L298 driver. Sep 19 01:24:26 Or...w/ the battery. Sep 19 01:24:43 But it was working before I hooked it up to the MotorCape. Sep 19 01:26:04 P9_19 Mode: default Direction: in Value: 0 Sep 19 01:26:33 Same for P9_20. Sep 19 01:26:37 in Value: 0 Sep 19 01:28:07 I need a smoke break. This darn issue is beating me senseless. Sep 19 01:28:11 brb Sep 19 01:33:05 believe me. I tried GPIO.LOW, GPIO.HIGH, other libraries that deal w/ the BBB, too. Sep 19 01:33:52 I dealt w/ the mfg. for a bit on their forums. They have an embedded Linux section. They told me to wait six hours before replying again. Sep 19 01:34:14 I tried all sorts of different software. Notta so far. I just wanted to let people know. Sep 19 01:34:39 It could be just the Motor2 connections. I will try others. Sep 19 01:36:37 It is not the motor or the battery. I just connected them directly. Sep 19 01:44:12 I tried motor1. No go. Sep 19 01:44:17 The LED works, though. Sep 19 01:46:02 Maybe the damn Cape is on backwards? Sep 19 01:46:24 I am totally silly. I would not doubt it. Sep 19 01:46:26 Let me check. Sep 19 01:47:04 Just for reference, I am using the BBBW. Sep 19 01:47:56 Nope. It is on correctly. Sep 19 01:48:05 P8 = P8 and P9 = P9. Sep 19 01:48:05 set_: what do you mean "w/ another L298 driver" ? Sep 19 01:48:20 I have another L298 driver here by my side. Sep 19 01:48:26 It is from some Italian Co. Sep 19 01:48:47 open-electronics.org or something like that. Sep 19 01:49:23 I don't understand what you're talking about, how are L298 drivers part of your setup? Sep 19 01:49:48 I thought this MotorCape had L298p drivers on it. Sep 19 01:50:02 I have a photo of it. Sep 19 01:50:57 I have the L298 driver on my other board also. Sep 19 01:51:06 are you suggesting you somehow managed to blow up the drivers on the motor cape (how?!) and intend to desolder and replace them? that overall sounds like a bad idea Sep 19 01:51:12 No! Sep 19 01:51:37 I am suggesting that this motorCape is faulty or both of our software examples do not work. Sep 19 01:52:55 Sir, I am not here to talk smack or get you or anyone else back for some "atrocity" caused to me. Sep 19 01:52:56 ... Sep 19 01:54:09 I just purchased this board b/c I thought it would help me, via learning and to make reality more fun, and you or whomever it would help b/c you and other people here have helped me w/ my daily ramblings and have turned those ramblings into actual learning. Sep 19 01:54:10 ... Sep 19 01:54:18 Electronics is an intersting subject. Sep 19 01:54:43 If your software does not work, what else could it be. I tested two out of the four motor connections. Sep 19 01:54:55 One and Two. Sep 19 01:55:17 if the led turns on but the motor still doesn't work, the only real explanations are: 1. you connected something wrong 2. there's something wrong with the battery or the motor 3. you somehow managed to destroy the motor driver on the cape 4. the cape has a manufacturing defect (not very likely, exclude other options first) Sep 19 01:55:29 I suggest using a multimeter to debug what's going on Sep 19 01:55:36 I did. Sep 19 01:56:05 I tested my battery. I tested the battery w/ the motor turning while plugged in to the battery. Sep 19 01:56:10 All above 12v. Sep 19 01:56:29 I tested two of the drivers already on this MotorCape. Sep 19 01:56:29 and the output voltage when the motor pwm is set to 100 ? Sep 19 01:56:51 Nope. I did not test the motor pwm when it is set to 100. Sep 19 01:57:16 So, I will do that next. Sep 19 01:57:38 zmatt: I get what you are saying. It could might as well be my fault for not knowing as much as you. Sep 19 01:58:25 But...I know just enough to not take my life in a peril state when dealing w/ electronics. Sep 19 01:58:31 ... Sep 19 01:58:57 So... Sep 19 01:59:17 Besides that garbage, I think I will test the motor like you described. Sep 19 01:59:21 I will get back to you. Sep 19 02:01:51 So, I write the PWM signal to 100 and put my leads from my Multimeter on the motor connections where the leads are connected. Sep 19 02:01:52 I got that. Sep 19 02:02:01 But...this only cancels out the connections. Sep 19 02:02:36 brb after the test. Sep 19 02:08:57 Flat = 0.0v Sep 19 02:09:08 The motor is flat at 0.0v. Sep 19 02:09:26 PWM is at 100 and running for 100 seconds. Sep 19 02:09:48 The LED is not blinking and is lit red. Sep 19 02:12:17 It is the connection or something else deeper in context. Sep 19 02:15:15 confirm you're measuring 12v across the vin/gnd terminals on the cape? Sep 19 02:15:26 Okay! Sep 19 02:15:29 Please hold. Sep 19 02:16:06 You mean where? Sep 19 02:16:10 At the leads? Sep 19 02:16:20 12.12v Sep 19 02:16:41 weird Sep 19 02:16:49 Yep! Sep 19 02:16:51 Too odd. Sep 19 02:17:06 I am not going to stay up tonight to wait my six hours. Sep 19 02:17:18 and at no point did a plume of smoke emerge from the cape? Sep 19 02:17:23 No! Sep 19 02:17:27 Sheesh, man. Sep 19 02:17:31 ;) Sep 19 02:17:34 Hhaahhaha. Sep 19 02:17:54 Plumes are for ill-fated, non-math oriented persons of the night. Sep 19 02:18:35 I am going to contact the mfg. Send me a photo of the MotorCape and how it is supposed to look, please. Sep 19 02:19:01 I will show you mine. Sep 19 02:19:23 In case you have zoom software to look around the Cape for peculiar items. Sep 19 02:27:30 https://ibb.co/n9Vkjz is a photo of my MotorCape. Sep 19 02:27:45 If you see anything out of the ordinary, let me know. Yea! Sep 19 02:44:13 Hey... Sep 19 02:44:24 I have a scope. I can test other stuff but you might want to prepare me. Sep 19 02:44:37 I do not know all. Sep 19 02:48:31 Well...if you are not going to do something or look into it, which is okay w/ me, I will just talk w/ Mouser. Sep 19 02:48:43 I am sure they can set me up w/ a new one. Sep 19 02:49:05 Or my funds back which will entail me purchasing a new one. Sep 19 02:50:06 I can send it in for testing if needed. **** ENDING LOGGING AT Wed Sep 19 03:00:00 2018