this post was submitted on 18 Mar 2024
39 points (100.0% liked)

diy

21641 readers
1 users here now

Finally, a comm for that one user who hand-makes longbows. This ones for you, comrade.

founded 4 years ago
MODERATORS
[email protected]
[email protected]
[email protected]
[email protected]
39
Sphynx electrolysis development journal entry #3: More boards, more safety, and we're on sourcehut ๐ŸŽ‰ (hexbear.net)
submitted 7 months ago* (last edited 7 months ago) by [email protected] to c/[email protected]
13 comments fedilink hide all child comments
 

Hi friends - still sick, feel like ass, but I did a lot of work recently, and if everything goes according to plan, within 2 more journal entries, I'll be able to report on my very first hair removal test!

What I've been up to

This week was mostly board and circuit design. I just ordered two more circuit boards I designed, here they are:

Current pump, mk2

Schematic:

PCB:

3D PCB:

I talked about this one in Journal Entry #1 - since then, I received the board and the parts for mk1, and it works duck-dance

Technically, this means I can test hair removal on myself, but I'm going to try and wait until I have the new board, for a lot of reasons. Here's what I changed:

  • Before, the current went out a current pump and went in to ground. If the current pump were to malfunction, it would cause more current than expected to leave the probe, which is a safety issue (not electrocution level, but possibly a scarring issue - very bad). I added a current sink hard-wired to 2 mA on the return of the current probe so that even if the current pump fails, the current sink will still cap the max current at 2 mA.
  • I added feedback so that I can measure the output current from a microcontroller. This is pre-work for the Sphynx Uno, and it can also be used as a safety measure - if the current goes too high, I can cut board power and flash an angry red LED as an error light.
  • I added a current knob! This is one more thing to test for the final version.
  • I broke out and very neatly labeled parts of the schematic so that this is an easier resource to learn from as an outsider.
  • I changed up some resistor values to make the current pump and current sink more stable.

My highly scientific outlook on this one is that there's a 75% chance it works as intended first shot.

This is the board I'm going to use to test hair removal on myself (if it works)! If it works, there aren't really any other current pump related changes to make and this one can get incorporated into the Sphynx Lite!

Digital timing, mk1

Schematic:

PCB:

3D PCB:

@[email protected] game me some really good feedback on this one in Journal Entry #2. I since figured out a better way to make the timer work by actually reading the manual for the part I'm using. This means the knob I'm using will actually be accurate and not have a dangerous failure mode! yay!

This one is a little less likely to be perfect on the first try - it is all new, unlike the current source, it's also a little more complicated, and because of that, I'm giving it 40% odds to work on the first try.

When I get these two boards, if they both work as designed, I will be able to plug in a benchtop power supply, rig up some kind of weird holder for an electrolysis needle, and go for a spot of hair on my thigh I've been growing out for this moment! Journal entry 4 will probably be me designing and ordering the battery submodule, so journal entry 5 will be my first report of actual hair removal!

New Developments

We're on git! Specifically sourcehut! https://git.sr.ht/~_410bdf/sphynx

It's somewhat empty right now, it's also especially clunky because I just set it up and all of my commits are just dumping in all of my files, but now people can actually look at my files and mess with them for themselves, or even contribute if there's anything that anyone feels comfy adding!

Next up

As mentioned above, steady progess, waiting for boards to come in, making the battery board, and probably before the end of March I'll be starting to get rid of some body hair. End of April is a very reasonable timeline for a beta version of the Sphynx Lite to be available for enthusiastic individuals to order, build, try, and review! We're getting there!

Any ways to help?

Calls are out for a Sphynx logo for sure! Design review is always appreciated as well! Besides that, I'm just working through the early stuff, things are a little too turbulent at the moment for me to ask for much because things are moving too fast. Once I'm working on the Lite, it'd be SO sick if anyone artistically inclined would want to design some cool silkscreen for the Lite, maybe with a Blahaj, some trans/commie logos, possibly some original artwork - the possibilities are endless party-blob

This one's a little terse because I'm sick and tired, but as usual, from last time - If you have any questions, please ask below! It doesn't matter if they're technical or non-technical, it doesn't matter if you think they're basic, dumb, not worth my time, or anything else - I want people to get excited about this and I would love to take the time to communicate the inner workings of this to y'all so that everyone can be included! I can't guarantee I'll reply to everyone but I'll do my best to reply to comments that are asking something directly or I have something to add to!

top 13 comments
sorted by: hot top controversial new old
[โ€“] [email protected] 14 points 7 months ago

aaaaAAAAA I noticed a design error almost immediately after posting, the knob in the current source board is supposed to go from 0.1 mA to 2 mA, but because I didn't think very hard and connected VR1 straight from ground to 5V, the power now ranges from 0 mA to 2 mA. Pretty minor, but this means if you turn the knob all the way to the left, the board will do nothing and act like it's working. It's too late to cancel the order. Making a comment here so I remember and also to facilitate discussion/give y'all an opportunity to point-and-laugh-1point-and-laugh-2

[โ€“] [email protected] 11 points 7 months ago* (last edited 7 months ago) (1 children)

Tags:

@[email protected] @[email protected] @[email protected] @[email protected] @[email protected] @[email protected] @[email protected] @[email protected] @[email protected] @[email protected] @[email protected] @[email protected] @[email protected] @[email protected] @[email protected] @[email protected] @[email protected] @[email protected] @[email protected] @[email protected] @[email protected]

If you'd like to be on the tag list, or if you're on the tag list and would like to be removed, let me know here!

[โ€“] [email protected] 4 points 7 months ago (1 children)

I'd love to be on the tag list, if it's not too much trouble!

[โ€“] [email protected] 2 points 7 months ago (1 children)

added! cat-trans

[โ€“] [email protected] 1 points 7 months ago

thanks much!

[โ€“] [email protected] 7 points 7 months ago* (last edited 7 months ago) (1 children)

rat-salute-2 rat-salute-2 rat-salute-2

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 bonk (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.
[โ€“] [email protected] 7 points 7 months ago* (last edited 7 months ago) (1 children)

the parts count growing is a little bit of a bummer, but you'd be surprised at how cheaply I've been finding things, my 5V LDO is like $3, the LTC6993 is like $4.50, and then everything else combined is less than $10, so we're not breaking the bank yet. Total BOM cost can probably stay under $40 ๐Ÿคž

in order:

  • really good point on the transients, I think they should be pretty tame given that every device on the way up the chain has bypass caps, the voltage is all regulated by LDOs and there's no inductors around, not to mention that a sufficiently brief transient doesn't really pose a health risk (too low voltage to cause electrocution, too brief to cause over-electrolysis), but that all said there's definitely no harm in adding some output capacitance with a bleed resistor!
  • test points kitty-cri-screm i'm used to poking legs of ICs with needle probes but I thought I was gonna be better but I forgot boohoo this is what I get for working while I'm sick kitty-cri
  • the end product will be one single PCB! I'm just debugging with separate ones so I can iterate on different parts of the topology without having to rely on an omni-board for testing. Once the three boards (battery/power, timing, current pump) are done, I'm just gonna merge all the designs and route it as one very dense PCB, so in total I'm thinking smaller than a 3x5 card!
[โ€“] [email protected] 2 points 7 months ago* (last edited 7 months ago) (1 children)

one very dense PCB

Is there a reason to make it small and dense vs larger and more spread out? I would guess that the tighter the board is, the more difficult it will be to solder on components. And will heat management be a problem? How much heat do you think the components will generate? If there is a buildup of heat, it could warp or melt a 3d printed enclosure. PLA softens at 60C. So if it does generate heat, perhaps a header for a small fan would be a good addition.

The small size does make it easier to find a suitable off the shelf enclosure, or print one. I think most printers have a 8"x8" build area at a minimum. So that gives you some room to make a larger PCB so long as cost isn't a concern.

[โ€“] [email protected] 1 points 7 months ago

Hi thanks for engaging! trans-heart

You honestly have a pretty good point about the board size, my original motivation was that ordering PCBs goes into price tiers - if I can keep it in a 100mm by 100mm square, it'll be cheaper, but I think going up to 100mm by 200mm doesn't add much more than a few bucks, so if it gets too cramped you're right and I should definitely go up. Solderability is a real thing and you'll hear some more about it when I make another post in a couple hours

As for thermals, I think we'll probably be okay in that regard, the whole board - LEDs, beeper, electrolysis front end, everything - should probably come in in the neighborhood of ~ 100 mW peak and ~20 mW average, so thermally it's pretty chill and heat buildup won't be too much of an issue - good looking out though!

[โ€“] [email protected] 5 points 7 months ago* (last edited 7 months ago)

This is awesome work! I can't wait to see the final form. fidel-salute-big

[โ€“] [email protected] 4 points 7 months ago

So cool to see this!

[โ€“] [email protected] 4 points 7 months ago

Im so glad this is coming along so well! And im so happy you liked the sphynx name! swole-chonk

Ive been checking out each of these posts and love seeing them, i just havent been commenting cause lifes been busy (week from homelessness and kinda sorta seeing someone new who ive fallen head over heels for), but ill just say it again: this is seriously game changing!

Ill ask some of my design friends if theyre interested in making a logo, and maybe someone will make something for the project!

And as always, you fucking rock! I dont have much to contribute right now, so im just cheering you on from the sidelines! Youre amazing and wonderful and making my dreams of being hair-free come within reach! cat-trans bridget-vibe

[โ€“] [email protected] 2 points 7 months ago* (last edited 7 months ago)

Hello! The needles and pencils I ordered have all come in now!

from preliminary testing it seems like the idea of using the pencil lead collet as a holder will work! out of 6-7 pencil types tested I've come up with 2 that seem to work well for the 2 needle sizes (F and K shank)

For the larger F shank needles, I'm using this 1.1mm "Yasutomo Artist Pencil"

And for the smaller K shank needles, I'm using this 0.7mm Pentel Quicker Clicker

Both came apart reasonably easily and accommodated a small piece of wire running down the length of the pen and being soldered into the base of their brass collets, allowing for great electrical connection to the needles, and a reasonably strong physical grip. The pentel has an extremely grippy textured rubber grip, while the yasutomo is just plain plastic, but the pentel also required me to strip like 2.5 inches of insulation off the wire, since the insulation wouldn't fit down the lead tube otherwise. Both are probably usable without any further modification, through trimming the end off of the yasutomo would probably be a good idea so that the needle can engage further into the collet without disappearing down in the pen tip.

Quick notes on other pencils I tried/what else could work:

The factors I was looking at were:

  • collet material. You want the part gripping the lead to be metal, ideally brass (kind of a dull gold color). Often there is a little brass sleeve, but the actual collet is just 2 pieces of plastic, which won't work. Unfortunately hard to tell this from a product listing.
  • size:
    • the K shank is 0.8mm so a 0.8 or 0.7mm pencil is probably what you want (0.8mm doesn't seem to even exist really)
    • the F shank is 1.25mm, so a 1.1mm (used in some antique and specialty pencils) or 1.3mm (I found one for construction use,
  • ease of disassembly/reassembly/ general dimensional compatibility. If its too fiddly/fragile, or won't fit a wire, has a long tip that's integral to the mechanism somehow, etc. that's ofc bad

I also looked at:

  • staedtler 1.3mm jumbo construction pencil (crummy plastic collet)
  • bic velocity "Side Clic" (crummy plastic collet)
  • pentel Side FX 0.7 and twist-erase CLiCK 0.7 (plastic collet. hard to reassemble
  • Zebra M-301 (kinda crummy overall, removing the brass collet was tough and seemed to damage the plastic it clips into, plus the integral screw-on tip had to be cut down)

I will post pictures/instructions on how to modify the 2 I liked soon! I haven't gotten to the point of attaching any plugs or anything yet, just a wire and a continuity test. But there may be enough space for a 2.5mm aux jack at least, if that's the route we want to go.