The completed board does mostly sit around but I find it useful to pull it out and see “oh how big was 0603 irl, is this going to be annoying” when choosing footprints. Hope things calm down for you!
macerated_baby_presidents [he/him]
- 0 Posts
- 19 Comments
Joined 2 years ago
Cake day: August 29th, 2023
You are not logged in. If you use a Fediverse account that is able to follow users, you can follow this user.
It’s fun! Get a practice board and once you realize SMD isn’t actually that hard you’ll be able to shrink your projects
*bonk* you already know what I’m gonna say, this board could be a quarter of the size if you packed it and put components on both sides. Even if you stay single-sided components for assembly, I totally think you could do 100x100mm if you removed the 9V holders and used pigtail-style 9V holders, or (if that causes trouble for assembly) even barrel jacks and some off-the-shelf holders for users.
I like submodules (hierarchical sheets? I’m still on kicad 6 lol) for schematic organization but I guess they make it harder to make your board dense since you end up trying to pack rectangles together.
But I looked at the JLCPCB prices and squishing it down is probably only necessary for your pride. If you start distributing these in numbers high enough for the cost to justify the extra trouble, somebody will rip off the design and start selling it on Amazon and it won’t be your problem anymore.
I will say that committing to an enclosure helps other stuff a bit. Instead of needing the big board space for the batteries, you can just have a pigtail holder and a place in the case where the batteries sit. Depending on how you did it, you’d also have more freedom in (a) less silkscreen space around the pots (b) components between the pots without users brushing them with their fingers. FreeCAD actually has a plugin to import KiCAD models from the 3d view (and then export to some other tool’s format if FreeCad isn’t your jam) so you could make a simple 3d printed case or something. Or (deja vu, sorry if I’ve already suggested this) you could put it into a guitar pedal case. Seems like a natural fit with the 1/4" jack. Once the project is solidified a bit more and I’m less busy I’m likely to order a board, try it out, and possibly design a case.
what the fuck they’re on kicad 8 already? I remember complaining about the upgrade to 6. Would you like to upload a render of the back of the board? I don’t wanna upgrade yet I’m in the middle of another project. I’m probably gonna get after you for wasted space (and those on-board 9V holders) but tbh PCBs are surprisingly cheap these days. Are you thinking about making a case for it?
at least for prototype BOM, I suggest just putting all the parts in Digikey or Mouser and using their “share cart” link. you probably have to order stuff anyway so it’s pretty painless. Good work so far!
Hell yeah, good scope, I just got a board made without test points lol
I use
make4ht(TeX). These can work but honestly you’re probably going to find more Markdown editors than LaTeX editors.Personally, I doubt this project is going to have enough documentation that you really need an automated system to do both print and HTML. It’s not like code projects where it’s automatically generated from docstrings. Stuff like how to use the device and where to order it will be manually written and not updated often. In fact I think these docs could even be a google doc. That’s how Girl I Guess has been distributed for years, it’s a big (radlib) voter guide in Chicago. Static site generator is just for fun.
yes yes get a scope! more substantial response later but I have a DSO138 and it works just fine for my purposes. you don’t have to buy a fancy one or wait for a used university scope to get on the market
oh jeez PLEASE get some flux paste and a little brush, it’s cheap and makes surface mount so much easier. If you use an obscene amount of flux surface tension will help split the solder between the legs, rosin core doesn’t have enough
schematic lgtm but analog is not my strong suit so doesn’t count for much
Sorry I didn’t do that electrical review I promised, I’ve had a busy week 😅
I assume you’re using lots of flux, maybe doing that fancy mini-wave technique? Hate to say it but the next step is probably hot air and paste.
Digikey has a cart sharing function. Ideally you’d get assembly through JLC and you’ll be the only person ever to build one by hand. But that’s tough if JLCPCB doesn’t have the right type of pot.
RC filter sounds reasonable. I’m sure there’s a chip meant specifically to trigger on falling edges (this is probably common digital logic thing) but then you’ll have to learn how it works. Other tools in your toolbox are flip flops and digital logic gates. You could do, e.g.
AND(NOT(output), NOT(NOT(output))), so that you would get a brief pulse while the output was low and had recently been high (stacking NOT delays, translate to XOR/NAND to reduce parts count). And flip flops might tolerate the output as a very irregular clock. Bunch of related Stack Exchange questions with cute Logisim pics.
lookin good! I’ll give a more substantial electrical review later. Schematic annotation for submodules is clear and readable. I am sorry to see the parts count expand but it seems like a thoughtful design and this is a medical device after all.
- If you don’t have a scope (or fast digital thing for U1 plans) you should think about getting one to check transients when pulsing and if needle loses contact. Probably fine if you’ve simulated
- test points
(though you have vias now. nature’s test points) - got an idea for what the overall device is going to look like? A chain of modules is pretty easy to pull around a desk accidentally or drop a spare needle into and bridge things. If you put it in an enclosure (guitar pedal box maybe?), you might want to either consolidate boards or put all the potentiometers and LEDs on the same board.
o7
Honestly if I was making and using it on myself I wouldn’t gaf but for distribution I guess it matters. Good luck, timing is so fucking annoying I’m learning from your work too
you could also have a failsafe of a second, fixed pulse generator that limits the max length. Either a clamp on the main pulse or an AND type thing with the trigger input. High parts cost though.
LTC6993 looks clever. I am guessing it isn’t made for manual control of RSET, because using a potentiometer as a rheostat sucks and it would be easier if you could use a fixed resistor for broad control and do fine control via a potentiometer voltage divider. With this circuit, if you have a ±20% pot (which I do think you could get down), you’re going to get a minimum time that’s set by R3 (nice and precise) and a maximum time that’s set by RV1. This is the opposite of what you would want for safety: 0.5s vs 0.6s is less important than 5s vs 6s at the top end. Worse, the failure mode - if the pot wiper comes off the track / if a spec of dust gets in there or whatever - leaves VSET floating at 1V and:
The output pulse will continue indefinitely if ISET drops below approximately 500nA
Ideally this would be backward: full disconnect yielding a 0s pulse and turning the potentiometer all the way to 0Ω (so it’s just your high-precision fixed resistor to ground) yielding the desired max time. I’m about to fall asleep but maybe there’s a way you can make this better with negative output mode and/or have the probe a current sink instead of source, or something.
I tried and failed to come up with a way to use the potentiometer as a wiper so that the total resistance of the pot doesn’t matter. Maybe you could put a regulated current sink here? They also have an example circuit for “voltage controlled pulse width” which could be helpful if you can get a voltage source to sweep across without it being too annoying. I don’t think it’ll break if you put more than 1V on it but I don’t understand what happens to the pulse with “negative” ISET. They suggest buffering the 1V for a DAC that accepts a reference voltage but IMO that’s heading into overkill territory.
sadge. Pot isn’t the end of the world (±20% strikes me as high but what do I know) since it’s probably close enough to linear but anything analog probably means users need a calibration step which sucks. Lemme go look at the post
lol i should have numbered the bits
5- i’ve only ever used those array parts once, on a board that needed a whole bunch of RC filters on input lines, and they were kind of a pain in the ass to solder without hot air. so I can’t actually recommend them but just wanted to tell you they exist
7- neat those are nice looking
11- you seem good at simulation so idk if it’ll be a problem at all, this is just what i’ve heard “through the grapevine” about non-ideal capacitors. i wonder if they make digital pulse generators that don’t require flashingsuper excited to see this project come to fruition! you have me absolutely green with envy, I want to make a widget now
OSHpark is so expensive and doesn’t do assembly but I will say that they offer a very easy “click button get PCB” distribution solution via the sharing center. I’ve used that before, on some board that the author lost the source files for.
Wanna add me to the tag list?
feedback for V2:
- round the board corners and include mounting holes. Render looks like KiCAD, from experience it’s not too annoying and JLCPCB won’t charge extra.
- good job keeping it single layer. I guess you could try to get a ground pour under the opamp but idk if it matters. Thicker traces, esp for power inputs, never hurt as long as you have room.
- This looks easy to assemble but I think that for V2 you are gonna have to choose between
- Assembled by board house, in which case keep things surface-mount and shrink board as small as you can for price
- Assembled by end user, in which case keep the board spacious and probably move to through-hole parts. You and I know that surface mount is not hard, but it intimidates a lot of noobs and many of them are working without flux or a temperature controlled iron.
- For non-power connectors, I’m a fan of putting a symbol in the schematic and just making the symbol’s footprint a connector. I had to do a bunch of head scratching to figure out that J2 is a variable voltage source. (J1 is clear, J3 I get that there’s no such thing as a probe schematic symbol.)
- It’s probably cheaper to do separate resistors like you have, but be aware that you can get resistor arrays e.g. if you want to cut down on BOM or board space.
- Do you actually need 30V for reasonable skin resistance values? The lower you can get this the better. Commodity boost converters may be cheaper than whatever parts you’re ordering in small quantities. I would again suggest a USB battery bank and not a “raw” battery. Battery management for lipos is annoying, 9Vs are not very power-dense, and standalone battery housings are surprisingly expensive (meaning a dollar or two, but every bit counts for BOM) for what they are.
- yep yep test points. they’re free, go wild. For manually-assembled stuff I expect to do a lot of debugging on, I often include two holes that I connect with a wire so I have something to clip a scope probe onto.
- you might want to include a low-value resistor in series w probe with test points on either side. Then you can measure resistance and hook up a scope to monitor the actually-applied current.
- watch out that JLC has a library of “preferred” parts, e.g. their 2904. If you haven’t already, pick parts from the library to reduce pain later
Yeah chip flashing is the bane of DIY electronics. The standard hobby solution for pulse generation is a 555, but that’s not actually very accurate over timescales of more than a few seconds. If it were me I’d want to throw in an ATTiny or something digital. It’s probably not JLCPCB price tier, but I know there are fabs that will also flash MCUs before assembly.
I built an iontophoresis machine recently, which is a fancy way of saying “15mA-ish constant current supply”. Galvanic seems like the sort of thing you could do with
constant current source (lm317 gang) -> mosfet clamp to ground (controlled by whatever pulse generator you pick) -> electrode
I think you have the right idea with prefab boost converter. I have a little USB-to-9V adapter, this type which I use with a USB battery bank to power an oscilloscope. Might want to do something like that, commodity assemblies are your friend.
I’ve used JLCPCB before but not their assembly. If you don’t get it assembled, check out https://pcbshopper.com/ in case someone else offers cheaper boards per area.
I’m not a professional EE but I’d happily review board. I also got helpful reviews from https://www.reddit.com/r/PrintedCircuitBoard/ in the past. Curious, is “multidisciplinary engineer” a real job making prototypes or are you just a specialized engineer who can also make prototypes on the side? Bc if it’s a real job I want that please let me know how you got it.
Nice work! Get your “user testing” time in and enjoy.