Actual old thread: >>2945584

>>2951482
Once you start winding your own, its like a point of no return.

>>2951482
https://electronicprojectsforfun.wordpress.com/rf-module-gallery/the-rf-splitter-and-combiner-gallery/a-resistive-tap-attenuator-for-rf-measurements/

why is this called a 40 db attenuator? isn't this just a resistor divider that reduces the input signal by 0.0199 (2460 ohms in series with 50 ohms) which is close to 20 db attenuation?

File: IMG_20251018_102609.jpg (1.19 MB, 2100x1796)

1.19 MB JPG

Calling out the phaseGODS in here
See pic. Top read is clock, bottom is same clock present in audio output. Need to substract such clock from signal, how the fuck do i turn that square into whatever wave is that in bottom so i can phase cancel it and effectively substract it from output?

>>2951499
1. Filter out harmonics
2. Feed into comparator
3. Profit idk

File: IMG_2406.png (508 KB, 1179x2556)

508 KB PNG

>>2951499
what youre looking at is the sum of a wave and its 2f harmonic. this should be concerning to you since square waves do not contain their 2f harmonics. its being generated elsewhere. my bet is an amplifier being thrown in and out of saturation somewhere. if you post a circuit it might make it more clear whats happening.

>>2951494
Values for attenuators are always given for power, not voltage.

>have couple year old laptop with DC jack
>charger shits the bed one day
>Want to change it to usbc using a trigger board so it's not my only device with a bespoke charger anymore
>Cut off the jack from the charger, hook it to a bench psu and insert it
>Works
>Run 2 bodge wires (pins and body of jack are all on the ground plane aside from the positive) and test with bench psu
>Doesn't work
>After a few minutes of fiddle fucking plug the jack pigtail in with no power hooked to it
>Immediately starts accepting a charge (external power still hooked to bodge wires)
>The fuck
>Turns out the last little bit where it feels like the jack snaps in is what does it
Why would being hooked directly to the pins on the back of the jack which should bypass anything like that not work? I've confirmed that all parts of the jack other than the positive are all connected to the ground plane and do indeed show voltage when hooked up and measured, so I don't understand why some internal mechanism would still be able to interfere or what that mechanism might even be.

>>2951511
Maybe it's a switched jack.

File: IMG_20251018_113206.jpg (691 KB, 1080x1164)

691 KB JPG

>>2951500
The parasite signal is dynamic *meltingemoji*
>>2951503
Ty for the science. See pic, its a BBD delay, so, yes there will be audio distortion/filtering artifacts in the parasite wave. Thing is, the MN3007 from that circuit has two outputs and you pan them to filter out clock noise, however im not using that MN but one with a single output and the clock noise is loud. Im NOT lowpassing the output, i might at the very end but i need first to weaken the clock with phase tricks. Now have in mind that first scope is taken literally from the very audio output, so theres filtering from output pre, ill next post what it looks like taken right out the BBD

File: IMG_20251018_113905.jpg (2.28 MB, 2610x2532)

2.28 MB JPG

>>2951514

File: file.png (40 KB, 1066x444)

40 KB PNG

>>2951514
>MN3007
im not familiar with analog delay lines, but is the clock in question driving the sampling? if so, theres a much larger problem here: youre only able to capture frequencies at half your sampling rate. if your clock is 25 kHz like it is in the first scope pic, you should filter out anything above 12.5 kHz on the output, which will include the clock interference.
if you want to capture frequencies above 12.5 kHz, youll need to increase the clock frequency.
if that decreases the maximum delay to below what you require, youll need to daisy chain multiple BBDs.
and if youre still early in the design process, i would strongly consider using a microcontroller for this. pic rel goes for $6.50 on mouser.

>>2951508
thanks. it's so confusing keeping db vs dbm vs dbw etc. straight.

https://yageogroup.com/products/Capacitors/part/C2220C511KCGAC%7BBULK%7D

what do "current" and "voltage" mean here? as an example, at 10 MHz it shows 1.85A and 57.87V. is this some sort of a derating curve, and so at 10 MHz the absolute maximums are 1.85A and 57.87V? or is something completely different?

(i'm trying to understand how to select capacitors in a resonant circuit, so voltage and current limits are very important.)

>>2951531
if i had to guess, it seems like thats the experimental data they used for all the other data: voltage across cap vs current through cap. why they chose 176.78 V, i couldnt tell you. its also possible that the data is simulated, considering they have data for 176 V @ 10 GHz, which seems odd. but i definitely dont think its a derating curve.

>>2951514
Second BBD with an out-of-phase input clock, running in parallel? You could also use some FETs or an analogue switch IC to make an additional discrete BBD stage on the very end, allowing you to sum the two output signals like in the MN3007. Seems kinda useless to have a BBD IC without the two outputs, maybe an appnote will describe how to use it properly.

If the clock frequency will remain constant, tuning a multipole LPF will be a good idea, you won’t lose any signal information because of Nyquist’s sampling theorem.
If the clock frequency is variable, then maybe you can use some synchronous switched capacitor filter, likely with some integer factor between the BBD clock and the filter clock. I’ve never used switched capacitor filters though.
An edge-triggered sample+hold circuit may even work fine, only ever sample the output signal at the tiny rising edge of the clock (or better yet, of a clock 90 degrees out of phase via twisted ring-counter or PLL). That way you can ensure regardless of the clock’s contribution, you only ever sample it at the exact same place each time, and so turning the clock’s contribution into DC. It may attenuate your signal more than filtration, but you don’t have to worry about getting enough filter poles to fit between f and f/2.

>>2951499
>20µs grid
so that's a 25 kHz clock? and the signal below is way above that? just stick a low-pass filter on there. also you probably need a higher clock signal

From end of last thread I wrote:

Would hotswapping a DAC while power is still connected via the microcontroller (3.3v) be enough to kill its internal reference voltage? I think I did that...Vref is suddenly 0 and can't get it working again :(

someone asked "what DAC"...it's an AD5668.

>>2951583
what do you mean by hotswapping? desoldering it and soldering a new DAC in its place while power is on?

>>2951531
The current rating for capacitors is usually because of power dissipation limits. The current limit in the plots looks like it exactly follows the ESR.
Not really sure about the voltage, https://www.wima.de/en/service/knowledge-base/special-technical-subjects/ says it's because of reduced dielectric strength of the ceramic at higher frequencies, but I definitely see the opposite trend in the paper-oil capacitors I usually work with.

>>2951583
I’m guessing it’s on one of those arduino breadboard modules? If it got voltage from one of its other pins for longer than Vcc then maybe. Especially the analogue inputs, as it may not have internal protection diodes on those.

>>2951586
I have it on a DIP adapter. I wanted to make room for something else, so I took it out of the breadboard (forgetting that the Arduino, which was feeding it 3.3v and SYNC, DIN, CLK, was still powered via USB) and when I put it back, the internal ref was 0. But I think it's not totally dead, just maybe the internal reference. I need to investigate more later

>>2951611
I see. well Anon, this is why we buy spares

File: file.png (111 KB, 866x1069)

111 KB PNG

any particular reason why a photodiode with a smaller capacitance would have a slower frequency response than one with a larger capacitance?
maybe smaller capacitance = wider junction, so the carriers just take longer to traverse it? i guess that makes sense, but i'd be curious to know what wavelength is the sweet spot for high-speed photo response, at least for current technology.

im pretty retarded with digital logic. Anyone willing to point me in the right direction for the following problem? Ideally i want to build this compact and without a microcontroller.
I have a Digital input, a Digital Output and a Stop Impuls.
When the Input is on, the Output is supposed to be on. When the Stop impuls triggers the Output is supposed to shut off, not turning back on again until the next rising edge on the Digital input

File: file.png (12 KB, 1343x415)

12 KB PNG

>>2951737

>>2951741
yea like that, but that leaves out the rising edge part. With In continuosly on and stop just being an impuls, the thing would turn off for just the duration of stop

>>2951512
Yeah, I was wrong and 1 of the pins isn't grounded, measuring between the pins I get 150k ohms and when I ground that pin then it starts accepting power, so I guess I'll just jump it to ground or if I really care and want to get fancy figure out how to hook it to the trigger board so it grounds when a cord is inserted

>>2951748
you can buy latches that are edge-triggered ("synchronous") on both inputs.

File: file.png (364 KB, 944x1119)

364 KB PNG

>>2951748
>>2951751
my apologies, apparently edge-triggered latches are called "flip-flops".
for pic rel, tie D high, tie CLOCK to the input, and tie CLEAR to stop. CLEAR is asynchronous but if stop is really just a pulse then it shouldnt matter.
also, if stop is some sort of analog signal, i would highly recommend using a comparator to convert it to standard logic voltages. also im just realizing CLEAR is active low so use the comparator to invert stop while youre at it.

File: 354543543.jpg (73 KB, 1327x964)

73 KB JPG

>>2951753
thanks, thats exactly what im looking for.
>tie D high, tie CLOCK to the input
i never would have thought of that
>using a comparator
yea that was always the plan

File: Screenshot (2481).png (36 KB, 337x518)

36 KB PNG

Does anyone know what kind of component this is?

>>2951806
a choke?

File: 43.png (275 KB, 1297x916)

275 KB PNG

I listened to the anons advocating against stripboard so here I am learning KiCad for the first time
I'm not sure if I have the fortitude for this

>>2951806
Definitely a common-mode suppression choke. I saw a diagram that suggests the zig-zag means it’s a type-2 choke, whatever that means. Probably a different type of core material, or different shape.

>>2951809
I think there’s a way of autorouting on KiCAD with an addon, but I’ve never bothered. For a THT board I’d go into the settings for this file (one of the buttons at the top left) and change the default trace width to 0.5-1mm, instead of the tiny default.

If you select parts in the schematic viewer, they’ll be selected in the board layout too, so you can easily grab and move parts into sections before laying out. Then follow the rats nest.

>>2951812
does kicad have an autoplacer? would be neat if it did. minimize ratsnest overlap etc

>>2951822

It has both auto-placement and auto-routing. Whether or not those are useful is another matter. I don't use them, personally. In truth, I've never made any real effort to learn how to use the autorouter, and I'm too autismal about component placement to just let KiCad put them wherever it thinks is good.

I also don't do PCBs often enough to have grown out of finding drawing traces manually slightly cathartic, so...

>>2951809
the only thing more overwhelming than being faced with all of the lined up components at the start of routing is when you get about half way through and realise that you need to completely start over again but with the nagging doubt that your new layout idea won't work and you will need to reverse all the changes anyway.
good luck and have fun! if you want my advice just try and get as many nets untangled as possible before laying down any tracks at all

>>2951809
to add to what >>2951831 is saying, I often start with ICs, the place related passives around them. once things looks reasonably untangled routing tends to be easy

>>2951809
I haven't messed with PCB design yet but all in all that doesn't seem so bad. You could knock it out in parts as well, in between gooning

>>2951809
Even if you end up making a pcb for it, usually you make one out of perfboard vector board or whatever first.

File: 1737575957706598.jpg (60 KB, 627x680)

60 KB JPG

>>2951482
need to pad my resume with projects and shit for internships
starting small, any suggestions for fun ones that are good to get my feet wet?

File: toilet cam.jpg (50 KB, 686x386)

50 KB JPG

>>2951843
>any suggestions

toilet cam for scientific inquiry of real-world phenomenology

>>2951843
There’s some project lists in the OP repo. The “Don’t ask, roll” one has a mix of varying difficulties and utilities. I wrote this one instead because I’m autistic:
https://github.com/74HC14/ohmOP/blob/main/ProjectCategories.md

There’s also some utility projects that you should consider making if you don’t already own such a product, namely:
>CC/CV bench top PSU, lots of ways to approach this
>split-rail low-noise linear PSU for audio projects
>function generator, with lots of wave shapes
>amplifier for your function generator
There may be some other obvious ones I’m forgetting.

I have a super cheap soldering iron, and I've realized maybe due to my skill issue, using it has led to many botched repair attempts.

I want to know about equipment for microsoldering specifically, which ones are nice or which ones you use or recommend. thanks

like fuck i can't even reliably repair a USB-C cable

>>2951840
not according to some anons in the last thread.

>>2951866
JBC C210 or C115 cartridges are meant to be good for tiny stuff, and there’s pretty good tip variety. JBC stations themselves are expensive, but the Chinese knockoff stations and USB irons are a plenty, knockoff tips are usually fine too. Importantly, there’s a lot of reviews for them out there, mainly on YouTube.

Seems so odd that there isn't better PCB software.

>>2951884
What more do you want? Personally I’d like it if KiCAD’s simulator wasn’t strictly tied to its schematic view, allowing you to have generic op-amps and power sources on the spice file alone. And for it to be easier to manage spice models, and see which components have spice models.

I’d also like to see a mode designed for laying out breadboards and stripboard.
And for all the damn PNP and PMOS transistors to appear the correct orientation by default.

A microwave toolbox might be handy too, for making high-frequency reactive components as parametric shapes in the copper layer.

>>2951873
thanks

>>2951884
there is, its called Altium, but unfortunately it still has the exact same fatal flaw as KiCAD: your refusal to read the manual and actually learn how to use it.

File: 1760918803940365.gif (3.22 MB, 281x690)

3.22 MB GIF

>>2951809
component placement is like 90% of the layout process. make your schematic with layout in mind (shouldn't be too hard considering that you have like 4 ICs), and then place your components right to minimize traces crossing. actually routing the traces should be pretty quick and is best left for the end of the layout process
>>2951840
>>2951871
that was me. nowadays, i go straight to production without any breadboarding. with the simple 4 layer, mcu-to-peripheral types of boards that i put together, following whatever design guides given in your components' datasheets is almost always good enough to get it working.
if anything is fucked up, i'll fix it with jumpers and note whatever fix down for the next board rev. i also dont pay for my boards so i'm a lot more willing to throw money at a board that might not work the first time around.
even in industry, you'll see people going straight to production with complex (multi GHz, HDI, RF, etc) boards since actual board layout is very important and impossible to test with a breadboard/perfboard. even if it's simple enough to breadboard, they'll just simulate it in ltspice or simetrix or whatever
breadboarding might be fine when you're first starting out, but you're gonna grow out of it pretty quick.
>>2951884
kicad is probably one of the best, if not the best EDA suite available out there for simple boards. it's modern, free, easy to learn, the ui is clean, and it's constantly getting better. it's more than suitable for anything you'll be doing at the hobby/diy level
>>2951900
>altium
>better
lol. lmao, even.

>>2951892
>What more do you want?

i want it to be able to do everything for me with no corrections needed after the fact. there is so much manual work required with the modern software. it makes no sense. there's no reason most of the process can't be programmatic.

File: harness.png (87 KB, 300x300)

87 KB PNG

Anybody know where to source GM plugs for vehicles? I would like to make custom T-harness factory plugs for an alarm system. I was able to find DELPHI clone weather proof connectors on amazon/ebay but what I'm looking for is the square 6-pin connectors they use for harnesses. Is it a metra or am I not even in the ballpark. Thx anon

>>2952002
>pic for ants
look for markings, easiest way

two-fold question about my radiant in-ceiling heat that doesn't work in two rooms.

One room has power going to the thermostat and the thermostat breaks continuity properly. I figure this means the heating elements are broken somehow.

The other room has no thermostat, just a bunch of wires that I can't seem to get voltage to. But then I remembered, they wouldn't show the line voltage if there was no load/circuit attached, right? i.e., the thermostat wasn't completing a circuit so no voltage. Complicating matters, I don't know if this box is actually for a thermostat or if there's heat in the room. There are four red wires, one white, and one black. The red wires are in 2 pair that are capped off with wire nuts.

But then, if the heating elements in the first room are broken, how are they presenting a load and showing the line voltage without producing heat?

>>2952082
>how are they presenting a load and showing the line voltage without producing heat?

since you're able to read voltages and test continuity, we can assume you have a multimeter
if so, learn about the resistance function, and use it to test if the coils are actually broken
also, labeled drawings can clarify a situation greatly, so make one
BTW, make sure the breaker is off when testing resistance or else you can hurt your meter

>>2951936
i thought kicad has some kind of python api?

is anyone familiar with engineering of high performance mouse like endgame gear op1 8k? i'm a software engineer who is curious about building my own. do i need an electrical engineering degree?

>>2952138
it's such a huge market because everyone needs a mouse so there's lots of competition and legitimate engineering that goes into so it's tough for a hobbyist to compete. you can't even buy the sensor as a separate component.

>>2952138
Unless you can really get in and customise the sensor ASIC’s firmware, there isn’t much you can do. I know OpenRGB and other third-party software manages to communicate with computer mice, but I couldn’t say if you could flash firmware to change anything not intended by the manufacturer.

So I think you’re limited to:
>custom shell and ergonomics with an existing sensor from a donor mouse
>some minor custom electronics, like different switch types and scroll encoders
>maybe adding an additional microcontroller for more advanced features, like configuring low-backlash switch thresholds from analogue optical/hall sensors when the switch originally just had hardware switches (idk about denounce though), custom battery charging and monitoring, and things like turbo buttons
There’s plenty of videos of people doing the first one, maybe one or two on the second one, never seen anyone trying to emulate the high-end features. Well it should be doable if you add a low-latency (virtual) USB hub between the PC and the sensor IC, then use an additional USB MCU for the switches and handling some special logic. That way you could use the existing sensor as an image sensor, but do those fancy custom-threshold switches on the MCU, and maybe even have a flick-stick-like feature that only moved the X axis. That way you could pick a cheaper donor mouse. The USB hub and MCU could be one in the same, if you pick an MCU that’s capable of USB HID and USB hosting. If you’re doing wireless instead, no clue how those work. Just disconnect the switches from the ASIC and use two wireless links I guess.

File: mpv-shot0001.jpg (250 KB, 1280x720)

250 KB JPG

to the anon a thread or two ago who tried to regulate two rails using just one active surprise, here's Shahriar (The Signal Path) repairing an arcade PSU with a similar design, which has the same regulation issue that many anons pointed out
https://youtube.com/watch?v=z5MRLtg6QYA

>>2952148
*active device
not sure how surprise got in there

File: vactrol.gif (3 KB, 700x200)

3 KB GIF

Surely there's a simpler way to get variable resistance than this:
https://sound-au.com/project200.htm

What's the benefit of Vactrols? Is it just for peak limiting?

File: Screenshot 2025-10-21 at (...).png (3.47 MB, 2228x1468)

3.47 MB PNG

>>2952138
I don't know shit about computer mice, but this looks like a simple circuit. Granted, I don't know if this is the entire circuit. It uses a M483SIDAE MCU and some sensors from what I can tell. Def don't need an engineering degree.

>>2952172
It doesn't get simpler if you just want an isolated 2-terminal voltage-controlled "resistor"

Can't comment on any benefit though. They were used in old audio equipment.
A slightly more modern (1970s) approach was to use an LM13700 if you wanted voltage-controlled gain.

>>2952172
it's simple and it werks, what more do you want?
if it's part of a larger system however, you might ask yourself "what am I actually trying to do?". because if it's say a VCA then there's a billion other ways to go about it. but for the specific case of an isolated controllable resistor which doesn't have to handle too much current and that doesn't need much bandwidth it's fine. there's some downsides of course
>limited power handling
>distortion
>limited bandwidth

>>2952172
Depends on what you mean by simpler.
Servo/stepper motor on a 10-turn potentiometer.

>PZT3904
Somehow I never realized the 3904 is manufactured in sot-223 package in addition to the usual sot-23.