>>2989599
>suspiciously placed huge walnut with an eye in the middle

I saw a guy 3D print normal plastic with the traces of a PCB layout embossed into it. He then stuck down copper foil/tape and sanded off the high-spots, leaving just the embossed traces. Soldering it would definitely soften the plastic, but if you do it quickly it shouldn’t be an issue even with PLA.

But the best way is definitely a fibre laser. They can punch holes through the FR-4 that are charred, allowing you to electroplate them with copper. I think dual-colour lasers can also make un-charred holes and slots and edge cuts using an ultraviolet or blue laser, but I’m not sure. Blue is definitely better for cutting through copper than IR.

>>2989609
>huge walnut
looks regular-sized to me.
wtf are pogo pins for?
also, anyone noticed a fuckup?

>>2989623
Pogo pins connect to the antenna, the thin metallic thing stuck to the inside of the shell. Obviously not a big deal considering it worked without the lid on.

Maybe some smoke started quickly offgassing when the last of the batts were soldered on? Hard to say.

>>2989624
Shorted the batteries with his tweezers turned the tip red hot.

>>2989610
>even with PLA
Doubt, but maybe with a lot of flux and high temperature one can do it.

>>2989674
It takes like 5 seconds max to make a joint to a non-power component. Even if the plastic under the component melts, if you don't deform it it won't be an issue, and it won't heat up enough to impact the structure of the whole board. Unless you're making 10x10mm mini boards. Making your board like 3mm+ thick probably helps.

You can always anneal your PLA.

File: BabbysFirstCircuit.jpg (1.7 MB)

>>2989599
>Where to get started?
Babby's first circuit.

File: images.jpg (5.8 KB)

>>2989690
>Babby's first circuit.

>>2989676
Pla is half-step up from hot melt glue… it melts in your car.
The whole premise is ridiculous.
Real PCB material is probably cheaper than the pla.

File: 1749581539951388.jpg (62.7 KB)

>>2989624
>antenna
ahhhh, right! ty.
>>2989628
>Shorted the batteries with his tweezers turned the tip red hot.
yep. space is so cramped that when he released the pressure on the tweezers, he ended up shorting positive to the ground.

wtf do you even find lipo cells this small? in vapes or some shti like that? I think I've seen a vape battery in a video someone did and they looked bigger tho...

>>2989706
If you anneal it, the heat deflection temperature goes up past ABS or even polycarbonate, to 150C or so. To anneal something flat, just turn your bed up to 100C after it’s printed and chuck a towel over it for an hour. If you’ve ever tried to plastically weld with a soldering iron, you’ll have a good idea of how poor a thermal conductor plastics are, only a few millimetres around your iron tip gets molten. If thats beneath a layer of copper, then it will hold its shape.

Fibreglass FR4 can’t be 3D printed, so its cost is moot when discussing this method of prototyping. You could print a negative then apply a thermoset resin like epoxy, maybe with fibreglass, and do the same technique on that instead. But epoxy is definitely more expensive per kilogram than PLA.

>>2989720
>watching some man from africa read a chatgpt script
ishygddt

>>2989728
sounds like you watched it, senpai

>>2989720
nigger

>>2989725
> PLA
Is literal waste from making greek yogurt, they didn’t know what to do with all the lactic acid from the processes until 3D printing came along.

>>2989745
Pretty sure it’s corn/maize derived. Biodegradable if you own a 50C+ bioreactor!

>>2989745
>Is literal waste from making greek yogurt
Hey! They could mix it with some fluoride (waste from steel making) and put it in our toothpaste.

>>2989747
> Biodegradable if you own a 50C+ bioreactor!
No, 50 deg C won’t do it any longer… I’ve been tempering it my PLA. Over and over. So now it’s harder than a diamond, denser than osmium, and melts 5 times higher than tungsten.

Actually, I’m thinking I can make PCBs from Turds if I temper them….

I can’t afford a $2 phenolic FR2 PCB after I bought a $3000 printer and a $1500 PLA drier, and a PLA subscription service. I’m glad I can eat the PLA though… thanks for letting me know.

File: Screenshot 2026-05-07 230647.png (52.2 KB)

Do I recall correctly that surface mount electrolytic capacitors are difficult to hand-solder? I made my layout with thru-hole ones instead. I need a 1 uF bipolar capacitor to remove the DC offset from an audio signal and I've heard that ceramics have "microphonics".
Also does anyone know a good SMT equivalent of the 1N4148 diode? I found one with a SOD-523 package but I think that might be a bit on the small side, all my other components are 0805.

>>2989730
10 seconds of it, Big Chaks.

>>2989757
>capacitors
What's the cap doing? Doesn't look like it's a cap to ground, for any series caps in an audio circuit I'd want to use film caps. Though if it's under constant DC bias, an electrolytic isn't the end of the world, they can actually be pretty low leakage if you overspec their voltage rating. Polymer electrolyte caps should be better than conventional liquid electrolyte caps by any relevant metric.

>a good SMT equivalent of the 1N4148 diode?
1N4148W is what I'd go with. The WS is a small and somewhat delicate SOD-323 package (still larger than a SOD-523), but the W is in a larger SOD-123. There's also always MELFs, which are bigger still. Pic related shows MELF, 123, 323, 523.

File: retard moment.png (143.3 KB)

>>2989780

File: 1747473417981695.webm (3.9 MB)

Question for /ohm/: is there any money in electronics repair? Do people even look to repair stuff in 2026? Can you even make a living wage with it?

>>2989790
Yes, but it seems tough. I know a guy that specialises in television and home audio, he does both home-call installation and repair. So I guess he's already effectively filtering for whate repairs are and aren't worth attempting. Because of the labour cost, you can only really effectively repair things that are sufficiently valuable. If that's new electronics, then it's often difficult to repair thanks to tiny SMDs, custom programmed MCUs and ASICs, and proprietary software tools. You might have a chance with those for items popular enough to heve been reverse-engineered, like smartphones, but microsoldering doesn't seem very appealing.

The market I'd aim for is analogue audio, especially vintage stuff but not only. Recapping amps and crossovers is easy on the eyes and the hands, while you're at it you can sell them bluetooth receivers and pre-terminated speaker cables. The harder fixes are still relatively easy to troubleshoot, by measuring Vbe and bias currents and such, and with repeating audio signals it's pretty easy to use even the shittiest scope. Even newer stuff can be pretty easy to work on, unless it's class-D or some sort of fancy multimedia HDMI amp.

I volunteer locally with a "repair cafe" once a month, there's definitely a lot of people out there looking to get broken things fixed. But not many people willing to pay a lot for that. If you can replace a micro-USB socket in a bluetooth speaker in ten minutes, maybe it's worth it. But more often than not it takes half an hour just to take it apart.

Is easyeda stealing all my data?

File: circuit.png (158.4 KB)

>>2989780
>>2989781
Thanks, the MELF package looks like what I want, in fact I had some in a smt "practice kit" but didn't know what they were called.
The circuit has 3 caps in series at various places to remove DC bias that is introduced from OTAs, or to centre my waveform around 0 V after applying a DC offset. I think it's important to remove the DC around the filter feedback loop.
To be honest, this is not high-fidelity audio, just square and sawtooth waves, and I wonder if film caps would be overkill. I'm just using basic bipolar aluminium electrolytics on my breadboard which are about 1/8th the cost of the equivalent film cap. Easiest of all would be a smt ceramic cap but I've heard that it's "conventional wisdom" not to use them to decouple audio.

>>2989797
>The market I'd aim for is analogue audio, especially vintage stuff but not only
Interesting. Thanks! I guess that makes sense.
I was hoping for some computer stuff since I know a lot more about it.
>But not many people willing to pay a lot for that. If you can replace a micro-USB socket in a bluetooth speaker in ten minutes, maybe it's worth it. But more often than not it takes half an hour just to take it apart.
Yeah, 90% of things are unprofitable since the replacement is cheaper than the time it takes to open and diagnose something.
Items that have nostalgia and sentimental value are the only things people are willing to pay to have them repaired.

>>2989812
Yeah I’d just use electrolytic caps, probably 50V rated or more. Though in your case it looks like biases will be going positive and negative, so you’ll want to consider bipolar electrolytic caps.

File: Screenshot 2026-05-08 174705.png (458.9 KB)

>>2989856
Out of interest I tried swapping my 1 uF bipolar aluminium electrolytics with 220 nF polypropylene film capacitors. Now my square wave slopes downwards, but it still sounds the same. I'll probably stick with the electrolytics.

File: Screenshot 2026-05-08 175207.png (267.1 KB)

>>2989867
this is with the 1 uF
resonance is looking nice :^)

>>2989720
Youtube is so cancerous right now, how they manage to get people to listen is beyond me.

File: Screenshot 2026-05-08 212937.png (274.3 KB)

I aligned everything to the grid this time (sort of) and used a consistent trace width. I feel like the layout could be more compact, but it still comfortably fits within the 10x10cm square that seems to be some price break point at JLCPCB

>>2989890
Gonna cram in some pin headers for input/output connectivity? Both for power and signals.

Also looks like a few of those blue traces could be a bit shorter, and some of the pads and vias could be spaced a bit more from nearby traces for clearance reasons. I also like to ensure vias of different nets (and any traces or footprints on the reverse layer really) are spaced apart from one another, such that the ground fill can get between them. You could also thicken up the power traces if you want, but nots not a big deal for this kind of circuit.

You can also use Mitxela’s PCB trace melting KiCAD plug-in if you want to make your boards look cooler.

File: Screenshot 2026-05-08 221422.png (72.2 KB)

>>2989894
Yes this is about 95% done, I'll connect the pin header tomorrow. I know I should try and fit as much as possible on the top plane, so is it fair to say that the *number* of vias isn't so important, but I should optimize to minimize the length of traces on the lower layer?
Once I am satisfied with the connectivity, I'll turn grid snapping off to space out the traces a bit more nicely. Good pointer about the ground fill.

I'm a bit annoyed I have to waste an entire half of a dual flip-flop IC to make my frequency divider, but the single flip flop I ordered turned out to have much faster edge transition requirements than the rest of my circuit can provide. Maybe I could wire it up as a second frequency divider, to provide a sub-sub-octave....

>>2989900
>Maybe I could wire it up as a second frequency divider, to provide a sub-sub-octave....
I'd just wire it up with pads to solder wires onto, and a pull-down resistor or jumper on each input if needed. That way you can wire it up to whatever you want if you ever need to.

File: schematic.png (16.4 KB)

>>2989599
i want to build a dtmf tone generator to revive a rotary dial phone without janky ass microcontrollers doing shenanigans and use a tp5088n tone generator chip, i got everything figured out except interfacing the chip with the actual phone line, the datasheet gives this very simplistic schematic, would this actually work? wouldnt the dc current of the phone line fry the transformer? for the ringing ac im not concerned since i want to place this thing after the on/off hook switch

>>2989936
Maybe the transformer has a high enough ESR to handle continuous DC?

>>2989599
why haven't IC manufacturers capitalized on photovoltaic inverter ICs the way they have with SMPS ICs? you can find hundreds of different ICs for buck converters, boost converters, flyback converters, full bridge converters, etc. but it seems solar is always done with an MCU.

File: secretary phone.jpg (87.7 KB)

>>2989936
>wouldnt the dc current of the phone line fry the transformer?

of course not
you'd use a 600:600 ohm transformer which is the load that a typical phone puts on the line
however, you def need a hook switch or else the phone is forever busy
like every asian tranny hooker

but i'd suggest stuffing in a portable miniature secretary phone instead, as it has all the circuitry and buttons already

>>2989941
ill order the thing from aliexpress and hook it up, at worst i'll fry my modem

>>2989938
Because a proper MPPT algorithm is a looping process. Measure voltage and current, multiply them together, and vary the duty-cycle until the product is maximised. Even without the seeking logic, multiplication like that is a pain to do analog.

I’ve seen some cheap mock-MPPT ICs, where you program in a maximum power voltage via resistive divider, and its feedback loop attempts to keep that constant. Might be 5% less efficient than a proper MPPT algorithm, but it’s much simpler and cheaper. It was a Chinese brand chip like this that I used.

File: Screenshot 2026-05-09 184852.png (296.9 KB)

finished routing and attached the edge connector. I think this is looking better than my previous attempts. Tried to keep the jumps on the bottom layer as short as possible, it will otherwise be ground, +/- 12 V and +5 V

File: Screenshot 2026-05-09 204159.png (455.8 KB)

>>2989999
I more or less didn't bother with ground fill on the top layer because there are so many traces that I end up with a lot of isolated areas. The back ground plane is fairly unbroken, though
thanks for reading my blog

>>2990009
Stitching vias. Needs more stitching vias to cover the long traces going to the pin header.

File: 5v perimeter.png (454.7 KB)

>>2990009
>>2990017
Yeah I think either a bunch of stitching vias, or at least go out of your way to push the power traces to the very perimeter, like I've done here with the 5V rail. Such that no ground-referenced pads are outside them, ideally no signal traces either. R12, R67, etc. have a pretty roundabout return path.

>>2990009
Also C25 is backwards, if it's polarised.

File: Illustration2.png (1.4 MB)

>>2990017
>>2990041
>>2990042
thanks anons, it's useful to be able to run these by people who know what they are doing. I tried harder to route the power traces and I think they break up the ground plane a lot less now?

File: IMG_5944.jpg (385.8 KB)

What is this mystery pin?
I spotted a shiny thing on the carpet. It's not from any project of mine. I don't recognize this type of pin at all. It must have fell out of a device or appliance. I have no clue what though. Nothing around the house appears to be malfunctioning.

File: A_plug[1].jpg (484.1 KB)

>>2990128
NEMA connector?

>>2990131
Far too small.
0.4mm thick
13.88mm length
2.54mm at the wider part
1.41mm at the thinner part
Not magnetic at all.

>>2990128
>>2990136
USB A contact?

>>2990138
Looking closer, 2.54mm seems way to wide for that. It still looks like a connector from some sort of plug. Especially given the plating, wear, being non-ferrous, and the indent on the smaller side.

>>2990062
Yeah that looks good. C29 looks a bit isolated, but it doesn't matter since the top-side ground-plane connects it to C24. It's not necessary anymore, but if you wanted to, you could run the traces from the pin-header to the THT caps on the top-side of the board, then run bottom-side traces from the THT caps to the rest of the board.

>>2990142
I hope it won't matter too much because I just ordered the board, lel. At that point in the board the top layer ground plane leads right back to a ground, pin so I'm hoping it should be fine.
The boards were about 50 cents each for 10. I don't know how JLCPCB do it. I suppose they're subsidized by their government as part of a plan to achieve global dominance or something.

Got hired by my friend to do a smart meter lora project for him for several buildings. This is great. I love doing shit like that, plus i can do it literally 10x cheaper than the cheapest commercial solution.
All those chinkshit parts i impulsepurchased from ali over the years and never used will finally come in handy. I have such a stock of them i don't even need to buy anything else kek

>>2990179
Check the current prices for the things you need to replenish and include/mark up that cost in your invoice.

File: led1.png (21.0 KB)

>>2989599
Which button do you need to press to turn on the LED?

>>2990264
2 or 5

File: img.png (538.7 KB)

>>2990128
I figured it out. It's a placeholder for the analog ground on an HDMI to DVI-D cable.

File: Jlead.jpg (67.3 KB)

>>2989757
For SMT caps, the lead is on the bottom J-lead. You don't want to tack the edges with an iron. You want the whole lead to be soldered to the board. Best for a hot air station work. See image.

>>2990425
Flux and capillarity will cover the whole lead just fine as long as you have enough protruding pad to soak the heat
Don't want to heat caps too much

>>2990425
>>2990427
interesting anon. I'm glad I decided to go with through-hole electrolytics though, they will probably be less painful to work with.
Just waiting for all the parts to arrive now......

File: Make_It_BOOTYFULL.jpg (601.1 KB)

>>2990062
Looks like its coming together nice.
Some points I have found for your consideration.
1 and 2: No right angles, the board manufacture may or may not have improved etching processes controls in place and its best to design your board to be manufactured from anyone. Etching chemical may pool in the corners and the trace may fail over time. Its just good manufacturing process to eliminate fail points for a process.
3. Now that you have used the grid to line up all your components, you can now break-it to line shit up.
4. You got the space here. Why are the traces bunched up? Move the bank of resistors north and force yourself here to have the bottom traces for the resistor start going south and fan it out nice.
Make it beautiful.

>>2990427
One end for SMT caps is typically tied to a ground plane. ISO-fine for one time fix, ISO-not-so-fine for manufacturing.

>>2990432
Thanks for the pointers, although I'll have to use them to inform future designs, as this board is already ordered.
I was dimly aware of the "90 degree corner etchant trap" concept, but is it really still a consideration in current year? My traces are 0.3 mm.

I see that one of those traces in area 4 goes uncomfortably close to a pad, but hopefully I can get away with it.

>>2990435
From experience: First prototypes never make the cut for manufacturing. You have more knowledge now then when you started and best practices are best. Hunt down the IPC standard pdf files that have been posted on the net and learn from the process and manufacturing engineers experience from the past to make you a better designer.
Are you having a fun time doing this?

>>2990432
>Etching chemical may pool in the corners
Never heard of that issue with any fab house, and even my own etched boards don't suffer from it.

>>2990434
Is roasting an electrolytic cap with hot air ISO-fine?

>>2990435
In future you can change your design rules to force larger clearances, might make avoiding situations like R34's pad 2 easier. In principle these should be changed to be at least as restrictive as your chosen fab-house, but I've never had any issues relating to fab house clearances.

File: Force_Convection_Nitrogen_Reflow_Oven_20190201100148292-1708209528.jpg (71.3 KB)

>>2990437
>Is roasting an electrolytic cap with hot air ISO-fine?
They are designed for hot air [re]flow ovens.

File: Screenshot 2026-05-12 195319.png (21.3 KB)

>>2990436
This is only ever going to be a hand-soldered hobby project thing, but it will still make sense to look at the standards for future projects. I guess it is fun - although for one-off circuits, I'm still imagining that strip board and through hole components are my best option. But maybe I am living in the last century.

>>2990437
0.2 mm spacing between that pad and trace, but at least the trace will have the solder mask covering it...............

>>2990439
>but at least the trace will have the solder mask covering it...............
This is where you need to know your board shops tolerances.
This passed my light-cones (eyes) today. Maybe something you would like.
https://hackaday.com/2026/05/12/another-gift-to-the-world-from-cern-their-entire-set-of-kicad-libraries/
https://gitlab.com/ohwr/cern-kicad-libs
Looks like 18k footprints library.

File: Screenshot 2026-05-12 201746.png (30.9 KB)

>>2990441
Looks like I am in the clear, but I'll tweak my design rules settings anyway for next time.
Thanks for the link, 18k footprints sure is a lot.

>>2990441
Cmon man, my package manager already takes like 5 minutes to confirm that there are no updates to my footprint, symbol, and 3d libs each.

>>2990442
Now what is the tolerances for solder mask registration? Is there a chance the mask maybe offset from the pad that the trace is exposed? I'm not trying to be a dick. These are things for consideration when you manufacture boards. Good manufacturing practices.

>>2990443
>5 minutes to confirm
Sounds like you got five minutes for games now. Your welcome.

>>2990444
this is all useful stuff that I wasn't specifically aware of, so thanks. I can't actually find figures on the tolerance for the alignment, so I have decided to believe that it will be fine. Guess I'll find out when they arrive.
If it all turns out to be a disaster and that trace and pad somehow get connected, the only consequence is that I lose the use of 1 out of 8 board addresses, no big deal.

>>2990447
No thank you. I had to leave a career in contract electronic manufacturing for personal moral reasons. I loved the work. Keep posting I would really like to see the progress.

File: 1774391282752703.jpg (21.2 KB)

>>2990425
>J-lead.
>>2990431
If you go down the SMT path, as you should, then definitely go for this type of a cap and not the through hole. If you're doing SMT, a soldering iron should not be your primary soldering tool.
When I do SMT, I always use a stencil. I can't even explain how much faster spreading solder with a stencil is vs. a soldering iron. Heck, sometimes it's OK to just squirt some paste out of syringe and spread it that way if you don't have a stencil.
Also, oftentimes you don't even need electrolytics and can get away with MLCC or tantalums.
tl;dr: if you're doing SMT, everything should be SMT including the electrolytics. That simplifies the BOM, cost, and speeds up assembly.

>>2990520
Here is an AI blurb on esr.
The ESR values can vary significantly based on the type of capacitor:
Capacitor Type Typical ESR Range
Ceramic Capacitors 0.01 - 0.1 Ω
Aluminum Electrolytic Up to several Ω
Tantalum Capacitors Higher than ceramics

Understanding and managing ESR is essential for optimizing the performance and longevity of capacitors in electronic circuits.

>>2990522
ESR is not that big of an issue in modern circuits if you know how to avoid oscillation and instability. Modern caps will outlive device's useful lifespan if you know how to manage the temp. Modern caps will easily do 10k hours at max temp. Orders of magnitude at lower operating ranges.
MLCCs almost always fail because they were soldered with a soldering iron or the board was flexing too much due to temp changes.
The reasons why you'd choose electrolytics over MLCCs or tantalums is the need for ca cap to operate at higher voltages, or you need higher capacitance or want to reduce cost.
PS: and don't use AI for this shit. Use your fucking head., Don't outsource thinking to a machine.

File: Screenshot_20260513_1.jpg (218.2 KB)

So how much should I trust these counterfeit Scyllac capacitors (normally $10k+ each) being sold for $1.50?

File: hmm.png (1.0 MB)

>>2990537
Even less than the ebay sellers quoting $120 for an m4 mac mini.

>>2990524
>device's useful lifespan
We come from a different school of thought. Your being a dick. I don't outsource thinking to AI. I'm just not investing time-energy into replying to you for your shitty attitude.

>>2990576
>shitty attitude
Says a guy who using an AI to craft responses and when he's proven wrong, he claims the other guy is being a dick and has a shitty attitude. Your attitude sucks and that's why you know so little.

>>2990520
If you’ve got stencils and hot air, you’re probably right. But I think this anon is stuck with a soldering iron, also stencils are kinda expensive. To the point where it’s cheaper to get them to solder the components on for you.

>>2990524
MLCCs also have a distinct voltage dependence to their capacitance to watch out for. And microphonic behaviour. Electrolytics don’t have that. But you wouldn’t catch me dead using a tant.

>>2990576
You posted shit about ESR unprompted, nobody asked about ESR in the first place. Why would you go out of your way to do that?

File: 1764878681515805.jpg (406.7 KB)

>>2990613
>If you’ve got stencils and hot air, you’re probably right. But I think this anon is stuck with a soldering iron, also stencils are kinda expensive. To the point where it’s cheaper to get them to solder the components on for you.

NTA but you don't even need hot air. You can do it with a cheap hotplate. I've used a syringe to dispense solder paste onto pads and have placed components with tweezers. Surface tension just pulls all of the components in place and I rarely had bridges or soldering failures.

Hot plate element is like $15 (or less) and you need a temp PID controller or even better, an Arduino, so you can program the temperature ramps.

Soldering SMD with an iron is not only slow but also damages components.

>>2990615
isn't it difficult to get the entire surface to heat up uniformly without hot spots?

>>2990615
>cheap hotplate
Eh, if you're doing SOICs and 0805s and don't have stencils, I think I'd rather use an iron. Hand-dispensed solder paste would be worth considering if it were a bunch of 0402s, or DFNs or other leadless chips. Number of parts is another factor of course.
>Hot plate element is like $15 (or less) and you need a temp PID controller
Looked into this myself, those hotplates have a PTC heating element, so an off-the-shelf PID controller won't work very well. PID algorithms assume the input duty-cycle will produce a proportional dT/dt regardless of state temperature, but here that proportionality constant changes significantly as a function of state temperature. I expect you'd end up with a very sluggish system, but it might still be fast enough for normal reflow curves. With an arduino you'd be able to use a PID algorithm followed by a correction equation/lookup table.

Even cheaper and easier would be to manually control the temperature with a TRIAC dimmer while watching a thermocouple. Going blindly open-loop probably wouldn't be the end of the world either. Not like manual hot-air reflowing follows a proper cooling curve.

>Soldering SMD with an iron is not only slow but also damages components
Source? Seems like using an iron can be significantly quicker, putting less heat into the components. I can't imagine the thermal expansion/contraction of components is a sufficient factor unless you're somehow blasting heat into the component but not the board. MLCCs seem the most likely to break from thermal stress, but even then they're designed to handle some amount of it. Never had a hand-soldered MLCC break on me in my years of hand-soldering 0805s and 0603s.

File: 1767368845924969.jpg (146.7 KB)

>>2990624
>isn't it difficult to get the entire surface to heat up uniformly without hot spots?

No, they're made to be uniform. You can find pictures of people testing them with a thermal camera.

If you want to go third-world, a clothes iron with an arduino can also work. There's so many people showing this method online (YT, Instructables etc).

File: Screenshot 2026-05-13 220629.png (1.0 MB)

endless options on AliExpress to burn your house down, it seems.
Why on earth do they make USB-powered hot plates, that seems like a lot of juice to pull through a USB connection

>>2990439
>at least the trace will have the solder mask covering it.
Just FYI, solder mask is incidentally an insulator (bonus!) but is not intended as one and should not be used deliberately as one. Safety standards are explicit about this, but it's best to consider it a general principle even if you're not going to be going through that kind of evaluation.

>>2990643
The USB ones I've seen before are very small, 5x5cm sort of thing. On a 65W PD supply that's slow, but not completely unreasonable. Those Alientek ones are bigger (104x64mm), but they can do 140W PD and 200W from DC, and claims 120s heat up time to 260C.

I was going to make my own reflow plate, since there is a big price difference between the little PTC hotplates you wire up yourself and those. But taking a closer look, these things solve a lot of the problems that are difficult to solve with a soldering iron and a 3D printer, namely the heat resistant housing. House fires aren't really a risk considering you turn these things on, use them for 5-20 minutes, then turn them off again. Not like a 3D printer that you leave for hours unsupervised. My only problems are as follows:
- Lots of them say they're for preheating, and alientek claims many other hotplates only go up to 260C, so may struggle reflowing a board on their own.
- Most of the reflow plates look to use simple temperature setting interfaces, which are likely constant temperature regulators, instead of temperature-time curve controllers.
- Why do I have to buy two different hotplates for soldering and chemistry, can I not get a 350C hotplate with magnetic stirring?
And now I'm back to thinking I should make one or modify one. I guess adding magnetic stirring to a soldering hotplate is easier than the alternative. Assuming the hotplate itself isn't magnetic.

>>2990643
>endless options on AliExpress to burn your house down, it seems.
They ship tens of thousands of those and how many houses burn down from them?
>Why on earth do they make USB-powered hot plates, that seems like a lot of juice to pull through a USB connection
because USB-C PD exists and you can do a fair bit with 240W? But yeah, I'd prefer something around 400W.

File: temperature curve pid.png (208.5 KB)

>>2990712
hmmmm

>>2990714
Even then I've never seen any 56V 5A soldering hardware, only ever 20V/28V 5A or lower.

File: sddefault.jpg (30.7 KB)

>>2990643
>endless options on AliExpress to burn your house down
you preheating these things then going for a nuggie run?

>>2990633
Haha… using an iron for everything was common practice in the 70s and 80s. If we had to put ends on a bunch of phono plugs we’d put them in the steam holes on the bottom of the iron and start tinning.

>>2990585
Ya i don't keep the scale for esr in memory by capacitor type. I worked in medicine and engineering without a degree for I am competent.
EAT DICK FAGGOT.

You lack comprehension for the post thread.

>>2990712
>only go up to 260C
isn't there something to be said for low temperature solder for hobby projects though?

Sn42/Bi57.6/Ag0.4 allegedly melts at about 140 C

https://cpc.farnell.com/chip-quik/ts391lt/solder-paste-low-temp-15g/dp/SD02706

File: Screenshot 2026-05-14 224712.png (301.6 KB)

>>2990643
I am tempted. Even if it's not very good it will let me recover $20 worth of ICs from other boards that I messed up in the past

>>2990842
I’ve used it before, it solders nice. But it’s also somewhat weaker, much less common, and forms very brittle joints when mixed with leaded solder. Also not very useful for existing boards without special flavour solder.

>>2990850
I'll have to look up review videos of these. Because they just regulate for a constant temperature, unlike the smarter anemic USB models.

File: file.png (822.8 KB)

Use case?

>>2990906
they can be used to replace Compton valves in older designs

File: Vid20260516150459.webm (3.6 MB)

Why is my solder being outright repelled by my iron?
>clean ur tip
I've used tip cleaner and flux and it hasn't made any difference.

>>2991156
>used tip cleaner and flux

completely inadequate for the situation
to get rid of the black crud, you scrape it with the back side of an x-acto knife, or box cutter
as soon as you see a hint of shininess, apply solder to tin it
keep going around until the tip front is all shiny and chrome

>>2991156
If a good hard scrub (with something no harder than copper) doesn't bring back a tinnable surface, your plating is probably fucked. But I'd guess it's just too much crud. I find a scrub with a facial tissue or paper-towel (while hot) often works where copper wool doesn't.
Citric acid powder might also work, idk.

>>2991156
When you get a new tip, tin it and clean it a few times before you use it. Keep it clean while you work, and load the tip with a fresh blob of solder before you turn it off. Only use RMA flux and leaded 60/40 or 63/37 solder with your iron, and keep the iron at a moderate temperature.

>>2991156
If the plating is really gone, tip cleaner will not help, and you will need a new tip. Tip cleaner is very aggressive and you should only use it in last-ditch maybe-I-can-rescue-this situations like, well, that black tip that won't take tinning for love or money. If it doesn't work you need a new tip.

If you have used plumbing flux on the tip, it can completely destroy the plating. That's just one of many reasons soldering with flux is at least intermediate-level. Because you can choose the wrong flux.

>>2991156
>>2991178
>keep the iron at a moderate temperature
I would be more explicit and say to keep it at the lowest temperature practicable. Hot tips get eaten by oxidation and burn off flux way too quickly.

>>2991156
Scrub off the junk and hope the plating isn't fucked. If the plating is damaged, the tip is effectively dead and it will erode very fast.
Always clean the tip properly (use brass wool, NOT steel wool - check with a magnet to test for counterfeits) and keep it tinned to prolong the life. Properly cared for, a tip can last years with daily use.
May be worth looking into tip care, they generally don't go bad like that unless you're doing something to damage it. Using overly aggressive fluxes, leaving it untinned (especially at high temperatures), or scraping with something abrasive can all very quickly degrade it.

File: Screenshot From 2026-05-16 23-59-03.png (103.7 KB)

>>2989599
Why is a diode needed on the relay here?

>>2991279
To prevent the inductive effect caused by the relay, so the reverse flowing current wont damage the transistor

Does anyone have some suggestions for some projects I could do to improve my my ability?

I started an avionics apprenticeship a few months ago and I'm learning but slowly and I need my hand held most of the time. Would like to try and get better on my own but I'm not much of a self starter and don't know where to start.

>>2991290
What exactly is it you want to train by making a project?
>your ability to design circuit schematics
>your ability to lay out circuits on a PCB or protoboard
>your ability to solder and assemble circuits
>your ability to test and troubleshoot circuits
Or something else?
And what fields are you specifically trying to learn, I don't know shit about avionics.

Here's some lists of projects:
https://github.com/Rocheez/4chan-electronics-challenges/blob/master/list-of-challenges.png
https://github.com/74HC14/ohmOP/blob/main/ProjectCategories.md
But some are "follow the instructions" type challenges, others are "design X from scratch" type challenges, with various stages in between. To some extent, the projects you build can be useful, especially instrumentation projects like a milliohm meter or a capacitor/inductor ring-tester.

>>2991295
Well in my job I've gotta make/repair a lot of wiring looms and troubleshoot components/circuits, interpret wiring diagrams.
I've got a bit of practice putting together looms now though I'm hardly a pro. Probably learning to understand how they work would be the most useful thing for me.

For instance, the other day I was given this device, not sure what you call it but you plugged your headset in and 3 position switch to transmit on intercom and a radio PTT button, and I had to work out what all the wires were doing so I could alter these other ones that had a volume knob instead of the button.

Anyway I'll take a look at the projects, thanks. Building useful stuff always appeals to me.

File: waveform.png (357.6 KB)

>>2989599
how would i do pic related? it doesn't matter if it's edge or level triggered.

>CLK is 200 khz, variable duty cycle
>OUT gets triggered "on" with high CLK
>OUT gets triggered "off" with low CLK (OR) high RST
>if OUT is triggered "off" it remains "off" until the next high CLK
>RST can be ignored with low CLK

the basic component would be a reset dominant SR latch, where CLK=S and RST=R. but i'm not understanding how to get the RST event to latch OUT "off" until the next positive CLK cycle.

File: Untitled.png (1.9 MB)

hey guys, I need some help with USB stuff
I have a charging dock for my Dualsense, it just needs usb c power
I also recently started using a pi pico 2 W as a dongle for it on my PC (https://github.com/awalol/DS5Dongle), and I would like to be able to "integrate" it with the charging dock, so I could use both from a single usb cable going to the pc.
obviously the simple answer is a usb hub, but I wanted to know if I can "tap into" the usb-c connection of the charging dock, since that only uses usb-c for power and not data.
if it requires adding a few components and not just wiring, that would be fine too.
thanks in advance

>>2991351
Just split it off of the charger PCB, or use a type C splitter cable with a type C to micro USB adapter on the Pico. There are other options too.

File: eboard.png (721.5 KB)

making an extruder board for my 3D printer.

>>2991389
Cool stuff. What’s the motivation behind a custom board?

File: eboard_fit.jpg (1.2 MB)

>>2991402
I want to replace the OEM main board with an off the shelf part and making custom CAN tool head board is the least janky way to do it.

Just found rbdimmer.com, they sell a couple of TRIAC dimmer modules that are controlled via UART or I2C, which makes it nice and easy to drive a load without worrying about zero-crossing synchronicity. But the 8A module has awful LV/HV isolation so I’d only ever go for the 4A. Especially good for non-real-time control from a computer or raspberry pi or whatever.

>>2991404
What off the shelf part do you need it to work with? Some sort of bed level sensor or accelerometer?

File: 6ufxbdehforz5xzmqa7hobo3e1xp.jpg (92.3 KB)

>>2991406
I have a Manta m5p with a CM4 I want to install, the SOC the OEM used is barely adequate for the application. the CM4 might have enough UART's to connect everything but I also want to add a few things to the tool head, plus the printer is littered with Chinese STM32 knockoff parts that makes running main line klipper problematic, if the tool head board works out then I'll make a custom CAN load cell board also.

File: 20260517_231636.jpg (2.6 MB)

Alright, very cheap and questionably safe AliExpress hot plate is en route, do any anons have recommendations for solder paste for SMT boards? I don't have a stencil made, so I plan to just dispense from a syringe.
A buddy recommended the low-melting Sn42/Bi57.6/Ag0.4 but I'm sure there are other factors to consider besides just melting point.
If my calculations are correct, then everything here except the Pi will fit on my ~100 x 80 mm board.

>>2991412
Oh the printer didn’t originally have a toolhead board. How does your board compare to other toolhead boards?

File: IMG202605151307024 (2).jpg (1.6 MB)

Ignition control board

1/3

File: IMG202605151402548 (2).jpg (2.8 MB)

>>2991428
Ignition control board

2/3

>>2991426
I'm using the schematic for the EBB36 as a guide so it's an EBB36 in a form factor to fit my printer.

https://github.com/bigtreetech/EBB

File: IMG202605151405025 (2).jpg (1.4 MB)

>>2991429
Ignition control board

3/3

Soviet ICs were used

>>2991428
Wow, EPROM in a real project is cool. What kind of work does it do, just timing control?

File: image_2026-05-18_084944333.png (1.1 MB)

I really want to buy an oscilloscope but don't want to spend 200-300$ on something I don't know how many times, how often I would use
I'm getting into electronics, I want to learn embedded and circuit design as a hobby and not necessarily a job; ESP32 programming, analog audio circuits and processing, repairing things that break at home maybe too
do I hunt for a locally sourced 30eur old analog oscilloscope, or get a digital one? I think I'd get more use out of something I can connect to a PC, to also be able to look at the data more comfortably, collect it and run algorithms and analysis and whatever I'd like, but I don't know if a USB oscilloscope is a good idea, and if a cheap one has the necessary safety features, I don't want to kill both the scope and my PC
also, I'm in EU

>>2991444
The downside with analog oscilloscopes is they can only show you repeating signals. Perfectly fine for audio use, but for microcontroller stuff that's pretty limiting. Some later/fancier CRT scopes have an ADC and the ability to swap into digital storage mode, mine does this but only stores a hundred or two samples, and these tend to be a bit more expensive. Also my CRT scope is way out of calibration, and it's only drifting more with time.
You could get an alibay USB logic analyser as well as an analog scope, that will cover 90% of use-cases. The analog scope will show you discrete voltage levels on repeating audio signals and such,, while the logic analyser can sample non-repeating digital communications, and even decode their protocols for you. I've got a NanoDLA, it's pretty good. There's a cheaper one that also worked fine for me 99% of the time, but it could be a little strange to get connected, and apparently on Windows there's the possibility of a driver or firmware update bricking them.
But that remaining 10% is non-repeating analog events, like inrush current surges and testing components to failure.

A USB scope is definitely an option, I'd browse the EEVblog forum to see reviews of different models. If you're worried about damaging your computer with one, you can buy USB isolator dongles for $10 or so that should prevent computer damage. Same goes for the logic analyser.
My biggest problem with USB scopes, the few times I've tried to use them, is the user interface on the computer itself. Dragging the mouse around on the screen to move dials and knobs, it can be cumbersome. And rando chinese software is never too reliable or full-featured. I'd want to find one that works with 3rd party open source software, like PulseView, they tend to be better.

I got a handheld battery scope with 2 channels (1 channel is a toy). I picked the cheapest 2-channel one from an EEVblog forum post comparing them. It's decent, but the UI is a bit cumbersome still.

>>2991444
>I think I'd get more use out of something I can connect to a PC
Since you're a noob, you don't want that. All these cheap PC scopes have a shared ground so there's a super-high chance you will make a mistake probing and you will fry your scope and even worse, your PC's motherboard.

Get a cheap digital scope. If you buy one of those old CRT-based scopes, they will be so heavy and complicated you will never use it.

>>2991428
>>2991429
>>2991431
nice. what is it for tho?
ps: should have used cyrillic for the labels & everything.

>>2989599
what are the /ohm/ approved soldering irons/stations? i have a shitty no brand iron i got from a hardware store near my house and the tips are too big and it doesn't get hot enough for micro soldering so i need some recommendations.

>>2991518
i have a quicko t12 (or knockoff, can't remember) that works fine for me. i've used it for small stuff as well as power electronics boards with heavy copper pours (t12 spade tip.)

more people are recommending jbc based designs these days. geboon makes well rated soldering stations, like the tc22.

>>2991325
JK flip-flop. Should have synchronous J and K inputs, as well as asynchronous reset and maybe set. Tie J high and feed clk into CLK, so the output goes hogh from a positive edge on CLK. Then pass clk through an inverter, OR it with rst, and feed both into the reset input of the flip-flop. If it’s active low, NOR it, or otherwise De Morgan the logic. I’d want to test this to see if the clock pulse at t=22 would result in the output being on for a split-second, but I suspect it wouldn’t.

>>2991518
If you plan on doing micro-soldering (idk, 0402 and smaller), then the JBC-like stations compatible with C115 or C210 handpiece are what I’d be looking at. If you need lots of power for copper lugs or XT90s or D2PAKs, going for a station that can also drive C245s at 100-200W might be worth considering.

There’s quite a lot of YouTube reviews of these sorts of stations, there’s even an open source design I saw a build-log and review of recently, though I’d wait on that one for the next design.
The Miniware TS1M looks pretty good to me, though the guy who reviewed the open-source station noted that many Chinese stations have unacceptable tip current leakage for sensitive electronics, not sure if a tiny doodad like the TS1M can have good isolation.

I use a KSGET T12 station and an Si012-Pro with the same tips, but I am looking at upgrading to C245.

>>2991546
>K flip-flop. Should have synchronous J and K inputs, as well as asynchronous reset and maybe set. Tie J high and feed clk into CLK, so the output goes hogh from a positive edge on CLK. Then pass clk through an inverter, OR it with rst, and feed both into the reset input of the flip-flop. If it’s active low, NOR it, or otherwise De Morgan the logic. I’d want to test this to see if the clock pulse at t=22 would result in the output being on for a split-second, but I suspect it wouldn’t.

tried this in falstad but it's not working as expected.

https://www.falstad.com/circuit/circuitjs.html?ctz=DwYwlgTgBAZgvAIgIwKgFwM6IAwDpsEECsqYIiRuALEQJwAcSRBAzC0lVQExWogBGFbKgAOghM1QA3CBVQBbTBQCmAWiQoAfACgoUYACkA0gDEANmBHmA9iKgAPRElpcoPbG-r03VYbEQAbAgA9Dp6wADuDk4uPh5IAR7uqPAIwqG6+lGOyIlxUAlJ2PQpOCFh+gBK0QgBVG4A7K4srlxNpcgNqBGpXCmyyAQKAIb2UohcuCRQ-GDDWAiTJBnh2Yj0Hr7xSK5bHekVwACSNRq7BAUNAVB7-mmoGGY4U6gAFmByK-oA8tA5SFcbglGrsEvtyplIqdAVQWlANjcWuCvsBqjkEbDmrskXcAd1ev1EFQulB5KNxosXjM5gslqgACYAVzQAE8AMIskBmZTPFAotaLLwgqBsa5tPp3A6QgXioHXUVy5GHAAypzyXGKBTyAKCktETghqxqCpYRGaLGuVDBev5xotIrNIvt4qVkLMADsagjZd7fOD0DyEGzlUZDfoPV7NnCMXV-WhA5UAMoAFTDwAjOQV7idYuKccD3wAqqnbTkuELs1m-TbDgKq5sLska5CALI1Lb5K1i6upKVGnJ1YXe9rN-vrJJNeENvy9tMC4euZytEezlHWKDKd1OXhQDAiJy61L2XX7+41YOhirhOwUvwYcgSMPXqAUlgPB94PmHYLWQ7rzeICSe4Hh0x76meORJqmV76DeZS7g+ywwcAcEIG+CG8mmP5-huW7ICUu6ntaR4nvBORFtBmTPre77PIQ2BXFwtAsLQNDsFwSFUbBL6IOh96YSi2HaMAwTgBAOhAA

>>2991325
>>2991546
>>2991642
https://www.falstad.com/circuit/circuitjs.html?ctz=DwYwlgTgBAZgvAIgIwKgFwM6IAwDpsEECsqYIiALLgBxLUBMAbAJwPNLasOogBGiRbKgAO-BINQA3CANQBbTAICmAWiQoAfACgoUYABEAYgBswwkwHthUAB6IkAdkZRHz+tmounqeAiFRJJFkAem1dYABZW0R6B3ooegpsKABmbAoEpJ8cBFCdPQB5aDsEFPjE5PVyrNgcvPCAd2jkJGrKigyK7L9csL0mkqQOzPbOj26hev7mobHPVwTx2p6p4AHEMoTEqGT6RIne-LXmyqcdrYoD1YBJZpUUxk9Y+JV6aie4g6gMYxxcEigAAswCE+sAAIIAOwAJndZgkiM57uVEVcwcVEK93gk4lBXh4cfQvjJkIRDuEQHd3B8XtTUg8vuRkKoHKRJH8CChviT-JJoZRcMwhRREWUPPRmPQiKzVgBzKkElIPPF0lJEInLSZg9YIfE0vHw55oo46tTDCp4h5PGq+LUmu5WkaWx6E42NB0u554ijUXafTXk6YlZEI5ybKWMN16ABKd3VnhUvrxRGGif8vkcqAavg1MBJHH8cgAhjZ2Qh6P9ULwwEWsOXK6tTT7kmnk288b6o8dg82Da1vUm013TWKOy2I2Ou-Lg6O1EjR+m6mDKTP3GPLWu54z7Cy2RzsFyMDypPy-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-X9Tsu-TjOBmrbPfBzuDa2CutB-rlAAgT4ck5Hhvx6zrtJ8zCtlxr6eZ0c2fAME4AQNoQA

okay, i think i got it. this was harder than i expected, man i am shitty at digital logic. and i wish falstad had an easy way of letting me define an arbitrary waveform for the CLK/RST relationship.

>>2991444
Decent used ones hold their value. You will learn a ton with a scope that you wouldn't otherwise. The only reason they're not a must for beginners is the expense. Any scope is better than no scope, but you'll get the most for your money with a Rigol or a used Tek lunchbox. Analog scopes are very old now and can quickly turn into repair projects. A lot of the regurgitated advice to go analog for price/performance is decades out of date.

File: Screenshot 2026-05-19 172606.png (379.2 KB)

Amazon recommended me this, any good for electronics assembly? Seems cheap enough to take a gamble

>>2991715
Donegan Optivisor

>>2991715
it's fine. it works well
>>2991716
>$128
come on now.

>>2991740
Good tools are worth it. You don't know what you're missing until you use something better than Amazon/Ali/Temu offerings.

>>2991765
>You don't know what you're missing

beer, hookers and weed

>>2991279
Snub a dub

File: Screenshot 2026-05-21 181226.png (388.4 KB)

Do you anons think I can use a hot plate to recover a SMT IC surrounded by through hole components? Obviously the board won't sit flat on the hot plate. I wonder if I can get away with cutting off all the solder joints underneath, maybe even sanding the board a bit to get it flat (the board is a dud so it doesn't need to be kept intact)

>>2992004
>use a hot plate

use chipquik
https://www.youtube.com/watch?v=7kyaz4Zrd78

>>2992004
Get a length of heavy gauge solid copper wire and bend it into shape so that it makes contact with all the pins at once, then wrap the other end around your soldering iron tip a few times.

>>2992004
>Obviously the board won't sit flat on the hot plate
You can pile it up on sand or salt or whatever, it's what people do in a frying pan. Copper or brass or other metal shavings would be better, but sand works with enough time. You'd be reflowing every damn component though, I'd want to take those aluminium electrolytics off first so they don't boil.

The solid copper wire method >>2992012 is probably what I'd try, but it's still a pain, and somewhat prone to delaminating pads if you're insufficiently careful.

File: Screenshot 2026-05-21 214945.png (1.9 MB)

>>2992010
I'd actually never heard of that but it seems worth a look, thanks. There is some knock-off version on Amazon called "fast chip for quik removal of ICs"...
>>2992012
>>2992017
Wire method sounds fiddly but worth a try (and I have wire, but not sand). It's the chip I want to save, not the board.

Given that I'm posting anyway here's a progress pic, I am hand soldering everything for my first board, but I might try the second using the hot plate. I need to make 4-6 of them. Actually I got to take the new hot plate for a test drive earlier than I expected, because the first thing I did was solder a TL074 upside down. Hot plate allowed me to reorient it.
I never thought I would take the surfacemountpill, but I can see the advantages.

>>2992035
>There is some knock-off version on Amazon called "fast chip for quik removal of ICs"...
You could always buy a chunk of bismuth for $10 or whatever and mix it with leaded solder until the melting point is nice and low. Indium and gallium are expensive, but a lead-tin-bismuth desoldering alloy should be decent and much cheaper.

File: 1772521268999110.jpg (480.4 KB)

My dad asked me to fix his trimmer and he wants it repaired because he can't find this model anymore. It doesn't hold the charge (it's like 10+ years old). I managed to open it up and this battery pack is completely dead.
How do I replace it? What can I use in its place. Seems to be 2xAAA batteries. Where can I buy something like this? Thanks!

>>2992173
It’s just two AAA rechargeable batteries in series. If you have room, you could replace it with a battery holder for a couple of off-the-shelf rechargeables. If not, you could either assemble your own pack from seperate cells (ideally ones with solder tags), or buy an existing one. Searching eBay or wherever should show plenty of 2.4V 2xAAA batteries, I know lots of cordless phones use similar batteries. If you do end up buying such a battery, it would be a simple matter of cutting the wires and splicing them together. Butt crimp terminals, heat shrink solder splices, or solder and heat-shrink will be your main options. I’d want to solder them with an iron, but it depends on what you’ve got and how much room there is. Alternatively, if there’s a connector on the other side of that battery, you should be able to identify it and find a replacement with that same connector, no splicing needed.

>>2992173
They're just rechargeable NiMH AAAs. I recommend Eneloop (Panasonic). You'll have to solder (or spot weld if you have one) three joints: the positive and negative wires from the pigtail connector, and a series bridge on the other end of the pack.

>>2992176
>>2992177
Thanks!!
Apparently, this type of a pack is used for wireless phones. They have a diff connector though.
I don't have a spot welder and I'm weary of soldering batteries so maybe spending the money on a pre-made pack is the best option in this case.

>>2992180
If you're careful you can push the pins out of the old connector housing and swap it with the new battery's housing. Just make sure the polarity is oriented correctly.

>>2992180
Post image of connector.

>>2992180
You could also forego the connector and solder the wires directly to the PCB.

File: 1770199208888185.jpg (482.2 KB)

>>2992181
>>2992182
I can't reach the connector. I can't see it... it's somewhere deep inside the device. The thing is tho, I can't even open it any further. There are no more screws left. Everything else seems to be clipped or welded. The external screws allowed me to remove the blades and the battery compartment had a cover + screw.
Battery seems to be soldered to these wires so I don't think I have to open it (and probably break it) any further.
>>2992183
>You could also forego the connector and solder the wires directly to the PCB.
... or do what they did and just solder the wires.
Now I just gotta find this battery...

File: anon learns engine dev sucks.gif (946.6 KB)

can someone explain what i would need to get to make my relay switch A: not knock out my esp8266 from sucking all the amps out of it, and B: not sending any pulses through the rest of the electric grid? I have AI help me with some of this shit becuase I'm a retarded newfag but every time i look around online it feels like I'm being asked one of two things: be horribly ripped off for components because local part stores arent in my area, B: get a decent deal on multiple whatevers and theoretically never ever use the rest and haul them around forever.

I have no idea how to select and buy parts. I feel like I'm too stupid for stuff like this sometimes. building a gizmo to turn a fan on depending on indoor and outdoor temps.

>>2992229
>explain

- replace wimpy power supply by brawny alternative
- add big ass cap across supply to smooth out spikes
- use 2 diff supplies (with grounds tied together) one for sensitive parts, another for brute parts
- isolate relay power from rest of circuit by adding diode in series, and a big cap in parallel
- replace relay with power transistor, TRIAC, MOSFET, or solid-state relay

File: Capture.png (318.2 KB)

>>2992010
you can get knockoff chiquik on aliexpress for a fraction of the price.

>>2992229
You have to drive the relay coil with a transistor.

Turning mains voltage off and on with a relay is not a good beginner project.

File: Screenshot 2026-05-23 at 11.55.05_AM.png (182.8 KB)

Respectfully seeking review on this PCB design. I’m making a low powered “dumb watch” (MCU runs at 2MHz, i2c is slow, and total power is like 7µa) and this is the first PCB I’ve ever made. Idea is that a coin cell battery clip is attached to the bottom of the board (negative terminal sits directly on exposed pad which is part of a larger GND fill). I believe this constrains all power and signal to the top layer, lest signal routes make contact with the negative battery terminal on the bottom, but it does make the signal routing pretty messy. Curious how to improve this; the i2c ‘bus’ was particularly hard to route but I tried my best and am hoping this works.

>>2992010
> chipquik
I fuckin love low-temp mystery metals and contamination in my joints. When components get hot, i love it when they fall off the board onto the hv power control pcb. I love the way the mystery metals oxidize into this grey nonconductive powder and cause cracking with thermal cycles.
> why not use mercury?
Yeah, at least I know what the metal is there.

>>2992259
I haven’t worn a watch in 27 years. No need.

File: 72249F50-F131-4565-8147-A222D7CD9493.jpg (164.1 KB)

>>2992229
Look into snubber networks for relays.
You can isolate the whole thing with an optoisolator circuit on a separate power supply (or just a led and photodiode/transistor and a piece of heatshrink)

File: BATT.png (83.3 KB)

>>2992185
man its so low current that you can just twist the wires properly and heat shrink them
also I'm sure someone sells that exact configuration on nimh

don't use lithium, you can use 2 double A or AAA NiMH batteries in a spring loaded housing if it fits
make sure the housing puts the batteries in series, because a singular cell is 1.2, 2.4 is when two is in series
you can tell this like, 2 separate tabs and one connected tab, these usually come with cables coming out of them too

>>2992229
Yeah I’d use an optotriac switching a proper triac, but that would require a properly isolated circuit layout, not something you’d want to do on protoboard.

If you’re using a mechanical relay, you might be able to swap one relay for another that has a higher coil resistance, and so draws less power. I assume you’re switching it with a transistor and a freewheeling diode? Post schematic and current parts BOM.

>>2992259
Looks fine. You may want to add a little solder blob for the negative battery contact to touch. The ground plane and low frequency makes any shitty routing quite forgiving. Don’t be afraid to order a board and test. What MCU?

>>2992259
you don't need a ground pour, put 3 or 4 pads for GND
with such a small and low speed mcu the quality of the gnd barely matters I guess
if you post the project folder I can try an alternative route

File: IMG_0423.jpg (3.9 MB)

>>2992273
It’s comfy to leave my house without a phone but I still like to know the time.

anyone who doesn't at least own an f-91W is not a real DIYer

>>2992288
GND fill is mostly just for making routing easier, and don’t really see a downside to keeping it. Thanks for the help.
https://mega.nz/folder/unRCzIjK#ODTxdGjxlZ3eZhUtUBLmVQ

>>2992292
I can't even right click pick the bottom gnd fill zone, what gives?

File: Screenshot 2026-05-23 at 4.48.52_PM.png (367.5 KB)

>>2992296
Not an expert in this, but is it the “selection filter” in PCBNew? I always toggle with the GUI depending on what I’m working on, but when I shared the file the zones part is unchecked as shown in the kicad_prl file.

File: watch.png (889.8 KB)

>>2992298
I didn't know such a feature existed
anyways what do you think? 3v3 fill on top, GND on bottom

File: improved.png (1.2 MB)

>>2992304
Thanks for helping me, anon. The 3v3 fill on top is brilliant and definitely cleans things up. Do you see any issue with the green vias shorting to the negative terminal of the battery (red circle)? I thought that even signal traces would be an issue on that bottom layer within the red circle, since solder mask is the only electrical insulation.

>>2992318
Unless you want to add an extra layer of lacquer, I’d want to avoid vias in there. If it was a static device that’s one thing, but something in motion with knocks and vibrations doesn’t seem like a robust design.

Anyone have opinions on those lillygo LoRa watch devkits?

File: watchhh.png (596.0 KB)

>>2992318
you can put silkscreen, the white stuff, as an extra protection layer on top of them, but I assumed you were going to add slight amount of solder to that GND pad that touches the button cell? that would put a 0.1mm gap between the trace? you can add kapton tape there as well

is there a picture of the battery holder you gonna solder?

File: 20260523_213506.jpg (191.1 KB)

this could have been a really sleek low profile board were it not for those gigantic capacitors :(

>>2992327
you can mount them sideways, or go for polymer smd caps?

File: Screenshot 2026-05-23 215055.png (15.3 KB)

>>2992329
on reflection I suppose I could have used axial capacitors, I already had those ones in stock, though.
Surface mount electrolytics sounds hard, although I will be trying my first hot plate soldered board soon which might make it more feasible than using an iron.

>>2992331
I meant the niobium capacitors

File: limbo.jpg (6.8 KB)

This image just makes me think
>NOT ELECTRONICS

File: Screenshot 2026-05-23 215923.png (26.2 KB)

>>2992333
huh, I'd never heard of those. Thanks anon, interesting

>>2992336
you know they also make flat solid polymer aluminum caps
https://www.jmbom.com/blog/capacitor/sp-cap-conductive-polymer-aluminum-electrolytic-capacitors/734
look at all this marketing

>>2992339
well gosh it seems like I have been living in the past
truth be told I have very little idea what I am doing and it's a miracle I've gotten this far

File: IMG_0424.jpg (25.0 KB)

>>2992287
MCU is an STM32L072 in LQFP48 pkg. Low power usage is the biggest consideration and I’m trying to get 2+ years of battery life on a 180mAh CR2020 battery.
>>2992321
Never heard of that dev kit. TI made one too, but it’s been discontinued. Not really a fan of smart watches; already have too much junk that I need to remember to charge.
>>2992325
Yes, I’m planning on putting solder on the pad, but I’d rather have reduced failure risk on the bottom layer than prettier tracing on the top layer. They also make little springs for negative contacts, but I’ll try the solder dots first.

>>2992349
At some point around the 1 year lifespan mark I’d be wanting to try incorporating energy harvesting. Amorphous solar cells are the obvious way, but a full wave rectenna might be able to work with the right electrostatic design. A little sprung magnet bouncing about inside a solenoid could charge while you’re walking too. Worst case you could have a wide power harvesting coil and a bedside magnetic transducer to trickle charge a device within a metre of it.

Gonna have a backlight? A screen sleep option based on orientation or brightness?

File: x1.jpg (149.2 KB)

Japanese capacitor company released some Chinese cartoons to promote their tech
https://www.yuden.co.jp/en/product/special/feremo_comics/

>>2992334
It's a tank with oil spilling out

>>2992272
>I fuckin love low-temp mystery metals and contamination in my joints.
remove it with braid? also what components are regularly getting over 138C?

>>2992327
I forgot to add a couple of resistors in the schematic and the board is unusable as a result
it's been nice knowing you, anons

>>2992436
You can't cut a few traces and bodge them in?

File: ௵.png (8.1 KB)

>>2992477
nah, the trace I'd need to cut is actually underneath an IC. And I need to make 4 of them, so they need to be right.
time to wait for the replacement boards to be shipped I guess

>>2992436
>it's been nice knowing you
>>2992494
>time to wait

oh, got impression from prev post you were gonna do the only honorable thing, which is seppuku
but you've decided living in shame is better
it isn't

>>2992499
I figure the least I can do is leave the corrected PCBs in a box for somebody to discover in the future and continue the work I was unable to

>>2992494
So you remove the ic, flip it upside down, and dead-bug skywire it after cutting the trace. Kids today.

>>2992393
When that shit mixes with leaded solder it drops, probably below 80C, and also becomes extremely weak. But it should be pretty easy to flush it away with a solder wick and excess solder.

>>2992494
Bah, I shipped 4 boards with horrible bodges on them. Specifically, wires going to a voltage regulator because the included AMS1117’s 15V rating was too marginal for automotive use, and bent leads of a P-channel MOSFET because it was backwards.
WHY THE FUCK DO P CHANNEL AND PNP TRANSISTORS SPAWN IN UPSIDE DOWN?

>50+ parts on perfboard, a couple hundred connections
kill me, the fuck was I thinking
Next time I'm ordering a circuit board, even if I have to populate it myself it's so much better than perfboard. Still gonna finish it at this point but man this is so ass

File: images.steamusercontent.gif (1.0 MB)

>>2992233
>>2992236
>>2992274
>>2992287
you fags are vastly overestimating my skill level.

>>2992690
Are you at least using vero-board?

>>2992732
I said post schematic you root

File: (((sticks))).jpg (39.4 KB)

>>2992734
I dont have one of those, but its also not necissary, imagine an esp8266 with 3 temperature sensors on the pins and two going to the relay which is this monster.
https://www.amazon.com/Iot-Relay-Enclosed-High-Power-Raspberry/dp/B00WV7GMA2
this thing is probably meant for actual arduinos but I didnt realize it would suck my esp dry like a 6 year old squeezes a caprisun because I dont know anything about electronics. i just dont want to buy off the shelf crap that doesn't do exactly what I want (take the temperature in multiple locations and do math to turn on a fan). my actual small electronics skill level is "i know what a trace is" and not really any further.

>>2992736
Fair enough. Sounds like the relay box is drawing too many amps, but the question stands: is it just drawing more current than a GPIO pin can supply, or is it drawing more current than the ESP32’s power source can supply? If it’s the former, not much you can do besides adding a transistor module or whatever to switch it on and off. If it’s the latter, upgrade your power supply. What is powering the esp?

>>2992736
>>2992760
If you go to digital-loggers.com/iot2faqs.html you’ll see that the signal current should be just 0.2mA, which the ESP is more than capable of providing. I suspect either you’ve wired something wrong (ground reference maybe? inb4 capacitive dropper) or the iot relay is faulty.

>>2992761
i dont know what you mean by ground reference but the relay pin ground is on the ground next to one of the 3.v pins on the very edge of the card. the relay pin is on d1 which i thought is safe enough, am I wrong there? flipping the relay also causes one of my monitors to turn off and on thats not even connected to the power splitter in any way, and I've tried powering it through the downstream splitter or on the always on portion of the relay with no differences in outcome.

>>2992762
That wiring seems fine. Yeah I’m thinking it’s faulty. Ditch the ESP and turn the relay on with a 9V battery or whatever to see if it still causes power surges, or trips a breaker or whatever.

File: 71kEm6NqgRL._AC_SL1500_.jpg (182.7 KB)

>>2992763
i used a 9 v battery and no backflow happened and the thing smoothly switches on i guess. i dunno if that means my esp8266 is just dogshit, i can produce the exact model if you suspect it

i also used slopgpt to do the coding but it seems to just be a standard Arduino relay command.

>>2992764
yeah after putting everything back together the way i had it before the relay still causes the esp to stop taking any readings from the temp sensors until a power cycle. the wierd problem where my monitor would flicker seems gone for some reason though. maybe it would come back if i plugged everything back into the strip?

>>2992764
>>2992765
Ok, now try powering it from another 3.3V source, maybe two AA batteries in series will be close enough. If that works, try powering it from the GND and 3.3V pins of the ESP directly.
Again, what's the power source for the ESP? A USB power brick?

I'm guessing the monitor flickering is from some load feeding noise back onto the grid from switching a funky load, might want to look into an inline filter. But for now continue testing with nothing plugged in.

File: perfboard.png (2.6 MB)

>>2992733
Nope. Pretty much just this shit, but a little bigger. It's will be fairly densely packed when it's done.
Already have a bunch from years ago and all the parts on hand so I figured I might as well use them. I should really stock better prototyping board.

>>2992690
learn how to etch your own PCBs. i don't know why the reddit types (that's a lie, i know exactly why) are so opposed to etching. it's perfect if you only need 1 layer, or 2 layers if you're a masochist.
>Y-YEAH BUT WHY WOULD YOU WANT TO DO THAT WHEN YOU CAN PAY JLCPCB $10 AND WAIT 3 WEEKS TO GET 5 BOARDS WHEN YOU ONLY NEED 1 BOARD NOW!!!!!!

is there a good way to switch 230V AC with a microcontroller? got about 12 components to control, and they should all be <1A, except the heater
would a triac directly on the MCU IO pin work, or would i need to fuck around with optocouplers/a dozen parts for each thing i want to control?
dishwasher timer assembly got stuck in drainage mode forever and replacements are very expensive, so i'm considering programming one

File: rbdimmershill.png (478.6 KB)

>>2992782
If you have a reliable and repeatable process without too much labour, yeah. But if you don't have a laser printer, it's kinda shitty. I never had much luck with inkjet transparencies, the contrast ratio was never good enough for crisp traces and idiot-proof developing. Also my shitty photoresist film had thin spots in it. I prefer just hand-painting traces with etch-resistant lacquer instead.
If anything, I'd recommend a CNC router or laser.

>>2992786
If you don't care about continuous control (e.g. PID) and just want to turn it on and off no more than once a second, probably just a triac, optotriac, and a couple of passives. If you want to do something like PID for pulse-skipping or full-on phase-angle control, you'll need to measure zero-crossing and trigger the optotriac synchronously to the mains sine wave.

You can buy modules that do the hard parts for you, see:
rbdimmer.com
Which I already found and shilled earlier: >>2991405
They do boards with your optotriac and triac on them and just need jumper wires to a micro. They also do boards with their own pre-programmed chip that talks I2C or UART and outputs a proper synchronised signal, if you need that. It's cheaper on their website than on aliexpress for me.
You could buy someone else's triac module, though these guys (and Kemo) are the only people I know that do a digitally controlled idiot-proof synchronous triac control.

File: webcapture.jpg (108.8 KB)

I lack understanding regarding the workings of the baxandall bass/treble control.

how do I properly display the controls digitally if the pots are linear?
Right now I am just mapping the dB scale linearly like this:
dBf = (float)(bitValue - 128) * maxDb / 127.0;

the range of the controls is -15 to +15dB.
bitValue is the 8 bit value sent to the digital pots.

is this correct or should I use the amplitude gain function somehow:
dBf = 20.0 * log10(gain);

>>2992782
If I did circuit boards regularly I 100% would consider it.
CNC milling boards is also appealing, but the lower precision is pretty big downside and I'd never recoup the cost unless I needed it for other things.

>>2992787
granted i've only done it a few times, but i used transparency on my shitty $50 inkjet and it worked just fine for SOIC-8 and 0402 packages.

>>2992805
Having the decibels proportional to the potentiometer position is probably the way to go. That means the actual gain equation is nonlinear.

Also consider using fract and accum instead of float.

>>2992844
so the 20*log equation?
>Also consider using fract and accum instead of float.
interesting. can you link me some video/guide on how to replace float with this?
thank you, anon.

>>2992922
Yeah. Though if you’re multiplying your signal by a gain, it will be signal_out = signal_in*10^(dBf/20). Exponentials and logarithms like that are computationally expensive, so I’d probably just use a lookup table instead, assuming you’re doing it live on a microcontroller.

As for the Fract and Accum, they’re fixed-point data types that are optional in the C programming language, they’re implemented in the GCC compiler. Tried to use them in my arduino IDE ESP32 project, but whatever compiler that uses doesn’t support them. All I could read about the data-type was just occasional forum and stack-exchange posts. Again, fixed-point is kinda autistic and superfluous unless you’re pressed for processor power/time.

>>2992925
optimization is not that important for this instance. I'll be happy if it works. Even with the on the fly calculations it's still fast enough not to have any noticable difference on the sluggish hitachi LCD.

Now displaying dB for the audio pot works - 6dB is at 50%.
but with the baxandall controls it's at 9dB when mid point - shelving when pot is at 50%:
void showDecibel(byte x, byte y, bool showSign, bool showUnit, bool oneDecimal,
int bitValue, float minGain, float maxGain) {
lcd.setCursor(x, y);

// Eingangsbereich absichern
bitValue = constrain(bitValue, 0, 255);

// Pegel berechnen
float gain = minGain + ((float)bitValue / 255.0f) * (maxGain - minGain);

// Schutz gegen log10(0)
// if (gain < 0.001f)
// gain = 0.001f;

float dBf = 20.0f * log10(gain);

// Auf gewünschten Bereich begrenzen
//dBf = constrain(dBf, (float)minDb, (float)maxDb);

float absVal = fabs(dBf);

// Vorzeichen
if (showSign) {
if (dBf >= 0)
lcd.print(F("+"));
else
lcd.print(F("-"));
}

// Zahl ausgeben
if (oneDecimal) {
if (absVal < 10.0f)
lcd.print(F("0"));

lcd.print(absVal, 1);
} else {
int value = round(absVal);

if (value < 10)
lcd.print(F("0"));

lcd.print(value);
}

// Einheit
if (showUnit)
lcd.print(F("dB"));
}

showDecibel(8, 3, true, true, false, treble, 0.1778f, 5.623f);

>>2992931
Ah, so you're not doing DSP, rather you're just calculating and displaying the gain of an analogue circuit using the second wiper of a 2-gang pot?

>it's at 9dB when mid point
dB is a multiplicative measurement, it may be you just have the wrong offset? If it's a log pot it doesn't matter, even if the mid-point isn't mid-resistance, the calculation should still be correct. Are you measuring the gain difference of the tone control with an oscilloscope? It's possible you're measuring an area still on the 20dB/dec slope and not seeing the full gain change.

Not sure what you mean by "shelving". Post images if you can, diagrams and/or waveforms.

>>2992932
I'm just controlling a baxandall with 2 digi pots.
I also have a baxandall logarithmic volume control circuit with another pot.
50% at the volume pot should be -6dB volume attenuation.
For the baxandall it's different.
>Not sure what you mean by "shelving".
there should be no amplification/attenuation of the bass/treble when the pot is at 50%.
Right now I'm just showing a simple percent scale.
https://www.guitarscience.net/tsc/baxandall_2.htm#RIN=600&R1=22k&R2=22k&R3=22k&R4=10k&R5=10k&RB=100k&RT=100k&RF=600&CB=47n&CT=1n&RB_pot=Linear&RT_pot=Linear

File: 1745403585586029.jpg (27.0 KB)

>>2992777
no longer needed. I begged ai for some help about the issue, after narrowing some constraints like me being a retard who cant just buy parts easily because i dont really know what i am doing, it came up with a pretty decent idea. run the pins power from the open gpio D and Rx. once I do that, I can hard cycle them in software and save the thing from having to restart, thanks for your general assistance regardless. also i dont know why but the monitor no longer flickers, though again, that could be that i have less stuff plugged in right now.

File: baxandall_2_pic.png (12.4 KB)

>>2992932
turns out using a reversed log scale is not represantative of the baxandall setting.
Problem it's not exactly logarithmic but also not linear, so since I already wasted too much time on this miniscule issue I'm just gonna display dB linearly from -15 to +15dB, even if it does not correspond to reality.

File: challenged.png (456.5 KB)

When a piece of equipment running it's own embedded OS says "Power - 3.3V - 5V (operational voltage 3.3V)", does this mean just any USB battery pack with a jack adapter will work from amazon?

I'm electrically challenged and not trying to fry expensive shit.

>>2993205
It means that you use a 5V power bank or a 5V USB charger, and there is circuitry on the PCB to convert that 5V DC to 3.3V DC.

>>2993205
>does this mean just any USB battery pack with a jack adapter will work from amazon?
Yes. Make sure it's a straight-through jack adapter, there are some out there that step up to different voltages.

>>2993212
>>2993214
thank you guys, ordering the right stuff now

>>2993205
>any USB battery pack with a jack adapter will work from amazon?

no
you also have to take into consideration current rating
your supply must meet or exceed required current
and just to be super sure, check polarity of the gadget
most everything now is positive-on-inside but you never know until you check
ratings and polarity are sometimes found as a sticker on the bottom of the unit

File: AEgIS.png (1.8 MB)

>>2989690
>Where you'll end up?
Celestial Arch Priest of the order of Saturn inscribing summoning circles to bind Laplace's demon and harness its power for your photon and antimatter imager.

File: fuse clip.png (428.6 KB)

How do you model these fuse clips in KiCAD? You need two of them but the standard fuse symbol in KiCAD can represent just one of them.

>>2993365
Make a footprint with two of them, i.e. four holes. The rail pain is that the DRC will complain if you don’t connect both pins together, don’t think there’s a way to define pins as internally connected yet.

>>2993222
spooky, hopefully the one I got works. USB A 5V 2.0 amp.

I'll take a multimeter and test both first before I try I guess.

Just fixed a car key by throwing away the ripped battery basket and soldering two wires instead which I then promptly taped to the CR2032 battery it needs.
feels good.

File: 1778383834870255.jpg (6.2 KB)

Where can I get practical experience in electronics?

I graduated as an electronics engineer, but I have temporarily moved to a region where there are no relevant job postings apart from basic jobs in the power/energy sector. I am even willing to design electronic circuits for free if it helps me gain skills or build a portfolio, but I haven't found anything like that.

Currently I just design electronic circuits for myself, but honestly it feels different from what might actually be needed in the industry (eg. Kicad vs Altium problem). It's sad that in this field there are no equivalents of remote unpaid internships like there are in IT.

>>2993510
Step up your projects, aim for something fancy. Like a GaN BLDC driver, or some ultra low noise audio equipment, or some UHF radio equipment. Doesn’t matter if you don’t know enough to build it, you learn on the way.

>>2993510
>practical experience
Start out with repairs. Then move on to modifications or conversions of existing products. Once you get an idea for your own project, look around junkyards, thrift stores, flea markets, and garage sales for prototypes. Avoid Amazon if you want to save money. Avoid Ali/Temu if you value your time (and possibly life lmao).

>>2993510
>I graduated as an electronics engineer

accept your obsolescence or move to Shenzhen and compete with 130,000 new EE graduates each year
your best selling point is that you can at least write proper English documentation

can I substitute a 2N6045 N-channel mosfet with a 13N10 NPN? it's a simple shortwave amplifier. the maximum values seem even higher on the 13N10.

>>2993645
Would have to see a schematic. I’d suggest probably not, because gate drive and bias circuitry differs a lot between them. If it’s a linear amplifier, getting a MOSFET into the linear region is quite different from getting a BJT there.

>>2993700
yeah I see, thx. It's a linear amplifier from rm italy. schematics are all over the net. but I had a feeling that wouldn't work.

>>2993645
TIP102

>>2993708
You mean the RT35? Doesn’t look like a linear amplifier with that common-source FET, more like a class-C switching amp, but I don’t know shit about RF amplifier topologies.

>>2993720
noted

>>2993776
no, it's the kl-60

>>2993645
2N6045 is a darlington, not a MOSFET.

File: Capture.png (96.2 KB)

>>2989599
i'm trying to understand the inner workings of peak current mode controllers. i managed to get a single ended controller working, now i'm trying to model a push-pull output controller.

i copied the UC3846 block diagram here. it gives me a convergence error after a few picoseconds. why isn't this working? i know that my "t flip flop" equivalent works fine (in the boxed outline) and that running this in open loop also works fine, but obviously this creates an unregulated output.

>>2993830
Maybe it has something to do with unconnected inputs of your logic devices?
>https://www.allaboutcircuits.com/technical-articles/an-introduction-to-using-logic-gates-in-ltspice/
SPICE is also not the ideal tool to simulate power supplies. Take a look at SIMPLIS, it is much faster as it uses piecewise linear models instead of complex devices like in SPICE.

>>2993871
thanks. i did try this but it gives the same non-convergence error.

truthfully i'm having trouble understanding the logic at the block diagram level. i think the block diagrams are over-simplifying something for ease of presentation which is giving these errors... but i can't picture what "should" be happening to understand why.

File: Capture.png (115.9 KB)

>>2993830
>>2993871
>>2993875
okay, i feel like a massive retard. if i check "start external DC supply voltages at 0V" it appears to work. not sure if it's working correctly, but it at least regulates to 12V.

not sure why the model absolutely refuses to start up without this option enabled.

File: 1753477635160108.jpg (154.5 KB)

So i want to build a little stepper controller and checked out what the established designa are doing. The TMC2240(stepper driver IC) has an advanced reference designs which looks like pic rel, but everyone else uses decouplers, yet this guy uses ESD/TVS diodes. My question is why, what's the advantage?
Chatty only gives me a laudatio diodes and capacitors being fundamental different things..

File: crocodileclip.jpg (100.8 KB)

Does anyone know of an alligator clip that has a screw in addition to a spring, so that it can be tightened to keep it in place? It's probably by a different name as nothing is coming up on a search.

>>2993909
TVS diodes will clamp any overvoltage spikes. I’ve heard of stepper motors generating enough voltage to damage the control electronics, though I’ve hand-driven my 3D printer axes by hand and never had any issues. By decouplers, do you mean capacitors across the output or from the output to ground? Seems like putting a capacitive load on the output of a motor driver is a bad idea, even if it’s got chokes beforehand.

What’s the project?

File: 1764250127745187.jpg (462.6 KB)

>>2993934
Decoups would be sitting in the same place. pic rel the reference design, what most western OSH controller boards seem to do. Chinks usually leave them out altogether.
Right now the goal is basically just make stepper move and get a feeling for future designs, got a load of Nema17s from a university dumpster, a few of them cost 100€+ anew, so those i would like to treat rather well. The TVS diode design is from a 3d printer board "prunt" that boasts being able to turn/adjust the motors without blowing everything up, but then we're back to A as, as far as i got it, a decoup does the same job, as long as i don't turn the stepper actively like a Dynamo and even then the TMC comes with an overcharge protection.

>>2993936
If the TMC can handle power going backwards, you’d just need a single TVS diode or MOV or whatever across the main voltage rails, if anything at all. Though with clamping diodes right on the motor, you have the opportunity to add fuses to prevent the TVS diodes from failing, but that’s likely overkill. If it’s a prototype, I figure why not add footprints for optional components, you can solder them in later if you need them.

>>2993937
>If it’s a prototype, I figure why not add footprints for optional components, you can solder them in later if you need them.
I guess that would've been the obvious step. Thanks, nonetheless.
>Though with clamping diodes right on the motor, you have the opportunity to add fuses to prevent the TVS diodes from failing, but that’s likely overkill
I'm not sure i get this sentence, i want to add diodes and caps together and put a fuse where?

>>2993938
If you have the diodes as in >>2993909, you could put a fuse on pin 1 and a fuse on pin 2 of the stepper connection (one for each motor phase). That way, if the stepper is generating excess voltage for a prolonged period of time, it will blow one or both phase fuses. Without the fuses, the TVS diodes would have to be capable of dumping all the heat produced, which may result in them failing open-circuit, and so sending the overvoltage condition down to the driver and the rest of your circuitry. But this really isn't something I'd worry about unless you specifically believe you're likely to find prolonged backdriving of your steppers, and at speeds higher than they could reach at their supply voltage (motor voltage/emf is proportional to speed).

I'd probably just add caps, add a footprint for chokes that's easy to short out with a jumper wire or solder blob, and add a footprint for the TVS diodes too. And maybe add a ferrite sleeve around the wire going to the stepper motor. Through-hole footprints take up more room, but are probably easier to tweak. That's what I do for my first prototypes. I'd probably add a beefy TVS (or at least its footprint) from GND to VDD, to handle spikes from your PSU or from your motor or from any other peripheral you'll have.

>>2993940
Yeah, i probably stick to the multifootprint and check how close i get to kaputt everything. Thanks for the explanations.

File: file.png (73.7 KB)

I calculated 2,685A peak inductor current for 80% efficiency, 600kHz, 3.3uH, 5V in and 14V, 500mA out.
this seems to be well below the 3.7A switching current or am I missing something here?
https://www.ti.com/lit/ds/symlink/tlv61048.pdf?HQS=dis-dk-null-digikeymode-dsf-pf-null-wwe&ts=1780295300234&ref_url=https%253A%252F%252Fwww.ti.com%252Fgeneral%252Fdocs%252Fsuppproductinfo.tsp%253FdistId%253D10%2526gotoUrl%253Dhttps%253A%252F%252Fwww.ti.com%252Flit%252Fgpn%252Ftlv61048

File: 1768030914170262.png (470.1 KB)

>breadboard circuit
>everything works
>solder everything on a perfboard
>nothing works
The virtual ground is getting pulled down to the negative rail by God knows what and I'm so fucking depressed I can hardly deal with it.

File: 1752584476746361.jpg (638.0 KB)

In case you're wondering why the traces look that retarded, pic related is why (the green wires are the aforementioned virtual ground).

>>2993959
>>2993961
Get a multimeter and start probing for shorts or bad connections. Also make sure devices are oriented correctly (i.e. pin 1 is where it is supposed to be).

>>2993942
I get 2.586 A with your numbers, but otherwise this should be fine. 3.7 A is just the value where internal current limiter kicks in. You need to make sure you stay below that and everything will be fine.

>>2993959
Put a current limited 0.5V or less across the shorter nets (e.g. 100mA), and use your multimeter to find where the voltages on the two nets converge.

>>2993968
>>2993995
I ended up isolating the signal generator subcircuit by desoldering connections. I basically had the virtual ground isolated, just the resistors+electrolytic caps voltage divider, and a single ceramic cap going to negative and that fucking node still didn't stay put at the expected 4.5V. Then I accidentally touched an empty pad and it read back like 2V and change. I knew about leakage capacitance and the perfboard pads should cause it, but that's excessive, also because other pads didn't have this issue. In other words the board is FUBAR.

>>2993998
If it’s not a low-impedance short, then give your board a good scrub with IPA and an old toothbrush. Flux residue can absolutely cause that kind of behaviour. But if it’s under 10 ohms, it’s got to be a solder or component short.

File: file.png (70.0 KB)

>>2993976
unfortunately the efficiency graph doesn't go beyond 300mA, even though it seems somewhat linear at the 5v graph. it's propably a gamble if this will work reliably.
I searched the established manufacturers for hours till I found out about Silergy. seems the chinks seem to have exceeded at the switching reg business.
The SY21222 sheet shows exactly the graph that I want and is also a real bargin at 50ct/pcs.
https://www.mouser.de/datasheet/3/1517/1/SY21222_1MHz_2A_Current_Limit_Step_Up_Regulator_Datasheet_Rev.1.0.pdf

File: E3F668FB-17B2-44DF-9000-B9C9EB90EA01.jpg (418.9 KB)

Ground loop isolator. Now I can’t hear the problem it was meant to solve, so improvements will be put aside until I can recreate the buzz.
I see some ground loop isolators have series capacitors on their inputs, I can’t imagine why this would matter since transformers block DC anyhow. To stop DC biases from roasting the transformers?

File: xlequals2pifl.gif (14.6 KB)

>>2994070
>since transformers block DC anyhow

in this formula, if frequency is zero (i.e DC) then the impedance of the coil is very low
or almost a shot circuit
which could easily cause problems for a given source, such as one with phantom power

>>2994085
Yeah phantom power is the only real case I can think of this being a problem. But I see this circuit being used for 3.5mm isolators, where the only phantom power you'd ever see would be the <3V ~1k impedance for a miniature electret capsule. Those are usually more than tolerant of short circuits, at least in my experience. The transformers are still hundreds of ohms in DC anyhow.

File: optoi2c.png (31.0 KB)

I want to transfer data through a transparent plastic watertight container, and plan on doing so via optical isolation with discrete IR LED and phototransistor. Pic is for SDA, SCL just needs a single opto. The sensor in question is an SHT35, and the datasheet claims a maximum rise/fall time for its SDA and SLC lines of 300ns. Unfortunately for me, my phototransistors have a rise and fall time of 3us, and my LEDs are unspecified, I'll probably look into using a faster photodiode alongside a speedy small signal transistor. Anyone have any better ideas?

I'm already using a wireless power coil pair for transmitting power through the wall, with a magsafe magnet array so I can easily remove and replace the transmitter+sensor.