[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip / qa] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/diy/ - Do It Yourself


Thread archived.
You cannot reply anymore.


[Advertise on 4chan]


File: skeleton circuit war.jpg (2.42 MB, 3001x2158)
2.42 MB
2.42 MB JPG
Thread hit by shortage:>>2496402

>I'm new to electronics. Where to get started?
It is an art/science of applying principles to requirements.
Find problem, learn principles, design and verify solution, build, test, post results, repeat.

>Incredibly comprehensive list of electronics resources:
https://github.com/kitspace/awesome-electronics
Additional resources below:

>Project ideas:
https://adafruit.com
https://instructables.com/tag/type-id/category-technology/
https://makezine.com/category/electronics/
https://hackaday.io

>Don't ask, roll:
https://github.com/Rocheez/4chan-electronics-challenges/blob/master/list-of-challenges.png

>Archive of Popular Electronics magazines (1954-2003):
https://worldradiohistory.com/Popular-Electronics-Guide.htm
>Microchip Tips and Tricks PDF:
https://ww1.microchip.com/downloads/en/devicedoc/01146b.pdf
>Li+/LiPo batteries required reading:
https://www.elteconline.com/download/pdf/SAFT-RIC-LI-ION-Safety-Recommendations.pdf

>Books:
https://libgen.rs/

>Principles (by increasing skill level):
Mims III, Getting Started in Electronics
Geier, How to Diagnose & Fix Everything Electronic
Kybett & Boysen, All New Electronics Self-Teaching Guide
Scherz & Monk, Practical Electronics for Inventors (arguably has minor issues with mains grounding)
Horowitz and Hill, The Art of Electronics

>Recommended Design/verification tools:
KiCAD 6+
Circuitmaker
Logisim Evolution

>Recommended Components/equipment:
Octopart
eBay/AliExpress sellers, for component assortments/sample kits (caveat emptor)
Local independent electronics distributors
ladyada.net/library/procure/hobbyist.html

>More related YouTube channels:
mjlorton
jkgamm041
EcProjects
Photonvids
sdgelectronics
paceworldwide

>microcontroller specific problems?
>>>/diy/mcg
>I have junk, what do?
Shitcan it
>consumer product support or PC building?
>>>/g/
>household/premises wiring?
More rules-driven than engineering, try /qtddtot/ or sparky general first
>antigravity and/or overunity?
Go away
>>
>>2504416
You can push pretty high instantaneous force ratings out of a solenoid (e.g. 10 times the rated nominal) just by overvolting them, though they'll overheat if you don't cut back on the current. You may want to do this just using a PWM circuit on a constant high voltage rail, or have two voltage rails and two transistors, or boost-charging a capacitor that you discharge into it. It's also an option to use another inductor like an ignition coil to produce a spike of voltage to get current flowing in the actuator really quickly, if you're careful. Definitely put a thermistor in/on the windings, for safety. Push-pull puts another requirement, that you can reverse the thing's direction. A higher voltage rail with 4 N-channel FETs in an H-bridge is what I'd do, but that's because I have a stockpile of FET gate drivers, quality FETs, and now dev-boards to solder them to.

If on the other hand you're after high continuous-force, there's not much you can do. Maybe try and find some meme silver wire for lower heat production for a given geometry, add forced airflow, add water vapour injection, etc.

As for stopping at a particular point, if you need accuracy then you'll need some sort of method other than timing. Be that measuring the magnetic field of the plunger (and mathematicing away the magnetic field of the solenoid), measuring the current through the solenoid and using some sort of differential equation wizardry, using emission-reflection optical/ultrasonic intensity measurement, or an optical/magnetic linear encoder strip. If you can get optical to work it's probably going to be the cheapest and most compact, modulate it at high frequencies to prevent ambient light being an issue.
Then it's just a matter of throwing the system into reverse. If you REALLY overvolt it for that instantaneous force, you could absolutely get it to stop on a dime, provided a properly tuned microcontroller PID loop and a low-latency feedback circuit.
>>
>>2504398
Those open frame motors are ideal for running 24/7 for years, which is why they're used in things like fans and clocks.
If you need a particular RPM and want to use 110V / mains power with no conditioning, I'd use the motor and gear or belt it down to what you need. Also if you need cheap; you can find those motors in any of a number of scrapped devices like bathroom vent fans.
I'd only use DC and PWM if you need / have DC power and need a fine level of control on the RPMs.
What are you building? That would allow better recc.
>>
Replacing mouse switches a few times cause a lot of strain or damage to the board if you do it carefully? Like the act of heating the soldering and putting a new one.
>>
>>2504578
It depends on the quality of the board, but in general, if you are careful with the heat and don't try to use lead free solder it should be fine. I have replaced my Razer mouse switches >5 times in the past 10 or so years
>>
>>2504587
thanks I'll try I hope I don't moms spaghetti this board :( I don't have a lot of experience
>>
>design boards with M3 mounting holes because M4 would be silly
>boards arrive
>screws don't fit
>measure screw
>3.5mm
what the fuck
guess who's gonna be doing some drilling
also need to find some washers that fit or try to drill some out
>>
>>2504736
>anon learns existence of M3.5 bolts
Some mechanical engineer somewhere is sleeping in a very comfy manner, knowing that he fucked over a sparky today.
>>
Got an old chinkpad x120e, fan burned out. Got a replacement off ebay, but now it's behaving weirdly. Going on and off constantly at about 1hz. Reset bios, no avail. No bios setting for fan anyway.

Took a meter to it, power pin fluctuates between ~3 volts and ~0.9v, relative to ground. Yellow pwm sense line reads fairly constant 3.2v.

Do you think I just receive a bad fan and the system is running practically open loop, or is something more sinister afoot?
>>
>>2504849
Are there any other fans to swap the plugs around?
I'd guess the fan has a faulty tacho output and isn't giving the right signal for the fan controller to trust, or maybe it's drawing too much current and the fan controller keeps cutting off. If you can look near the fan plug maybe there's a fan controlling IC with a datasheet you can check, but maybe all that stuff is done in firmware. You should at least be able to look for a current sense resistor, though I don't think it's too likely.
>>
>>2504484
i figured /ohm/ would know more about control loop theory than /mcg/ so here we go...

i'm trying to implement a digital PID loop with a micro controller except i'm pretty ignorant on theory and practice. i know about the nyquist theorem and i've read that 10x oversampling is more typical in digital control loops, for purposes of loop stability. does "10x oversampling" just mean that i sample the process voltage 10 times and then average the results together before feeding them into the PID's error calculation -- or am i fundamentally misunderstanding something? if this is correct, how is it beneficial to do this?
>>
>>2504868
No, sadly it's a netbook, and I don't have any other fans with tiny jst connections. I do have the cable from the dead original fan though, maybe I can solder it to a generic 3 pin 5v fan to test?
>>
>>2504872
I think it refers to a 10-sample running average (i.e. square wave convolution), as opposed to just splitting the samples up into groups of 10. You could use a non-square wave for the convolution instead though, like a nice curvy window function, which might give more desirable characteristics, idk. Such a 10-sample average also means a delay of at least 5 samples, so ensure the ADC sample rate is high enough that such a delay isn't an issue. You'd need a FIFO data structure, not sure how to do that with an MCU, I think you'll want a FIFO for the integrated signal anyhow.

t. knows nothing about DSP and PID algorithms on an MCU

>>2504873
>maybe I can solder it to a generic 3 pin 5v fan to test
You can try that. If it has the same issue then I guess that means it's an issue with the tacho signal.
>>
>>2504872
im not sure thats the correct interpretation. the 10x oversampling means more along the lines of being able to accurate capture a signal of N hz frequency requires 10xN sampling frequency/rate. in terms of PID, thats sort of independent from the sampling rate. you design your PID controller to be able to handle such a high sample rate (your ID terms) but generally you would put your samples through filter to make it more stable before you actually feed it into the PID. that is the average filter youre talking about. If you want it more responsive you should fft the data and put it through a window function. like the other anon said. or even a hardware filter before it hits the adc.

someone correct me if im wrong, thats my understanding.
>>
>>2505057
>like the other anon said
Well I was actually talking about a rolling average convolution with a window function, as opposed to a window function before a STFT. Though for all I know, the mathematical result of that could be equivalent, FTs and convolutions go hand in hand, apparently. But STFTs kinda require a shit-ton of RAM, so I wouldn't use them unless you absolutely have to. I think PID makes more sense in the time domain anyhow, though making a Z-transform-based PID algorithm might actually be useful.
>or even a hardware filter before it hits the adc
You always need an anti-aliasing filter anyhow. Some ADCs have internal anti-aliasing filters, like nice TI-branded stereo audio ADCs, which also do their own downsampling now that I think about it. Might just be part-and-parcel of the whole delta-sigma process.
>>
>>2505068
i think theres some confusion about DSP versus control theory here. op said he wants to do PID on a mcu. basically as long as he smooths the input signal, the PID can take care of the rest. the sample rate in a PID only defines the initial responsiveness of the system. but that can be altered with the I term. anyway, anon needs to figure out what P, I, and D coefficients he needs in his calculation. and then he can run that through the negative feedback system to implement his idea. i cant guess as to what that might look like unless he says the exact use case
>>
>>2505074
Yeah, a basic trial-and-error PID loop needs basically no understanding of control theory. You can figure out things like heat capacity compared to power and use that to inform your constants, same for measuring hysteresis or latency, but just tweaking the values in trial-and-error should work well enough for most cases. Knowing where the poles and zeroes of his system are basically entirely superfluous, even if you’re writing an algorithm to auto-tune the PID constants.

Also what the hell these MOSFETs are turned on even with no gate voltage.
>>
File: amp.png (13 KB, 1068x767)
13 KB
13 KB PNG
Can someone help me see what I'm doing wrong?
The thermistor is around 920 ohms at room temperature .
The voltage divider holds point A at 1.3V so the inverting input must be held the same.
So 1.3V across the 1K3 resistor is 1mA which goes through the thermistor dropping 0.92 volts and the 4k75 resistor dropping 4.75V so point B is approx 7V and point C is approx 2.25V.
I've verified these values in circuit and they match up.
So then, the 2nd op amp should try to keep the inputs matching so point D should be 2.25V too.
Then 2.25V dropped across the 3K4 resistor would give 0.617mA flowing through it
The drop of 7.75V across the 10K2 resistor would mean 0.759mA flowing through it
The difference of .142mA must therefore be flowing through the 25K5 resistor giving a voltage drop of 3.621.
The op amp is single supply so it can't go less than zero so it can't output -1.37V
Also when I actually measure the circuit, the op amp output is 0.6V and point D is actually a tiny bit higher than 2.25V at 2.32V.
I assume I have to take the resistors going to the alarm comparator and then ground into account but I'm not sure how
>>
Best cheap linear dc power supply? Need it for work and my boss is a penny pincher.
>>
>>2505074
>i cant guess as to what that might look like unless he says the exact use case

i'm trying to build a """high performance""" MOSFET-based electronic load using a digital PID control system. i'm starting with CC mode but ideally i'd also be able to build out CV, CP, and CR modes.

i have an stm32g474 nucleo board which has several onboard ADCs which can sample at 4 MSPS and several onboard DACs which can sample at 15 MSPS. my thoughts were to buffer the sense resistor with an op amp, apply some sort of analog low pass filter (not sure of bandwidth), and then sample with the ADC @ maximum sampling rate with 10x hardware averaging. then run through the PID loop, push results to DAC, etc.
>>
>>2505106
I don't even know what is the point of your circuit. Are you trying to trigger an "alarm" when temperature crosses certain threshold? What do you intend to measure?
>1mA which goes through the thermistor
Unless you're making a water level/flow detector then this is going to cause your thermistor to heat up which will mess up your temperature reading. You only want to have a couple of microamps running through it and having it in a feedback path is also unusual.
>>
>>2504739
it's retarded m1 m2 m3 m4 m5 m6
thats it no other sizes of machine screw need to exist ever
nobody ever fucking goes
>oh maaaan i have m5 screw but i wish i could only have m5.12 screw it would be perfect for my project!
>>
>>2505169
It's not my circuit.
It's part of a driver board for an IGBT stack.
The thermistor is embedded in the igbt
The buffered reading goes out to the controller and the alarm signal trips at a certain threshold
>>
>>2504349
take the Manhattan-pill
>>
>>2505189
>The thermistor is embedded in the igbt
Ok, self heating can safely be neglected then. Measuring 0.6V most likely confirms that output opamp is saturating to GND and that voltage at D is different because of that. If the thermistor is of NTC type resistance will decrease with increase of temperature. This means that voltage at point C will decrease, which will decrease drop over 25k resistor. Output of omapm is of the form: Vout = Rntc*1mA * some_gain - offset_adjust. Circuit is focused on temperatures above some value that is then scaled for whatever is consuming the temperature signal. Divider at the output has no effect on output and is just used to scale down the output signal for comparator. IGBT datasheet must contain thermistor parameters and looking at those values other resistor values will probably start making sense.
>>
>>2505189
>IGBT stack
pervert
>>
>>2505211
Ah, so the only reason the 0.6V is there is because the op amp can't drive itself right down to the negative supply?
I'd never considered that.

IIRC the thermistor is 1k at 0C and decreases most likely non linearly.
Given that the igbt will most likely be operating at 60C or so usually it makes sense that it doesn't measure as low as the 15C where I am.

I'll throw a 1k pot across it on Monday and see how it acts
>>
>>2505329
>Ah, so the only reason the 0.6V is there is because the op amp can't drive itself right down to the negative supply?
It depends on the opamp. Some (rail-to-rail) can go very close to ground, others need some space.
>>
>>2505138
Just get a switching one and put a capacitance multiplier after the output, and take the feedback line from the output of that.
>>
>>2505355
I wouldn't, the output impedance of those audio outputs is pretty low. If you just want to swap between two channels then a DPDT switch would work. If you want to combine two audio streams then you'd want a pair of summing amplifiers with half-decent op-amps.
>>
File: o no.png (110 KB, 1230x972)
110 KB
110 KB PNG
>ordered PCBs
>MOSFET symbols were the wrong way around
WHAT THE FUCK HAVE I DONE
and the local shop only sells GDS P-chans
>>
>>2505399
thank you.
>>
File: panic.jpg (1.41 MB, 2524x2219)
1.41 MB
1.41 MB JPG
>>2505417
screw it, this will have to do.
>>
>>2505456
This is art.
>>
File: noodles.gif (474 KB, 460x460)
474 KB
474 KB GIF
>>2505456
>>
>>2505456
>dmm probe brushes against mosfet tab while measuring mcu pin
>kills mcu dead
>now have to desolder mcu from underneath mosfets without disturbing their folded pins and somehow removing their 3dp plastic bracket
at least i have like 15 more of these micros but damn
>>
>>2505149
sounds like a solid plan anon. youll just need your PID to be responsive. having it lag might not be good for something that needs high performance.
>>
File: desoldered-resoldered.jpg (2.11 MB, 3052x2472)
2.11 MB
2.11 MB JPG
>>2505537
Hey this "solder a solid core wire across all pins on one side" desoldering technique actually works pretty well. Might even be able to get it working across both sides of a SOIC like this, maybe even all 4 sides of a QFP.
Locking forceps / haemostats are also great for holding onto components while you desolder them.
>>
how we doing tonight anons
>>
>>2505575
Bretty good fellow ohmosexual
It's 4 am and I'm hitting the sack
>>
>>2504873
Wow, easier said than done, i don't think I own a single little fan that isn't 12v or 24v
>>
>>2505569
Just drag solder across all of the pins like the rest of us
>>
>>2505670
Try and hook 5V up to a 12V fan, they sometimes work ok. Just wiring the fan power pins up to an LED+resistor might be enough to see if it behaves erratically when the tacho pin is left unconnected.

>>2505691
You mean get a full bead across it? Too hard to bridge them all at once, which is the same problem as the perfboard blobbers from earlier, but not quite as bad since the pitch is closer. More doable with a SOIC-8, but for a SOIC-14 or higher it's a pain. Even with a T-12 K knife tip. The thermal conductivity of copper is more than 10 times greater than that of lead, and even with the copper wire there it took me like 5-10 seconds to get that side to reflow (at 320C).
>>
File: 1657872933834.jpg (1.82 MB, 3024x4032)
1.82 MB
1.82 MB JPG
>>2504484
What soldering station would you suggest to a beginner?
>>
File: Hotplate v104.gif (103 KB, 1136x800)
103 KB
103 KB GIF
What do you think about my hotplate design?
>>
>sweet bit of kit
recently obtained a "46336007h-fpnx" "tau 2" thermal camera core from the scrap heap, is there any cheap premade interface boards that output standard NTSC video? or will i have to fart around with the 0.5mm pin pitch connecter and talk to it over uart (3.3v TTL RS232) manually?
>>
>>2505765
T12 station. The one that uses an external PSU, because the PSUs have suspect quality and subpar design (HV clearance is noticeably lacking).
A Pinecil or TS101 are similar in a standalone iron format, much better for portability but arguably worse for UI.

There are some T245-tipped stations and controller kits that are probably quite competitive in price, I imagine we’ll see something like a TS-T245 higher-power portable iron eventually.
>>
>>2505986
I’ve got a similar hotplate myself, I’ve designed a mains triac controller board (the one with snarky writing in the OP pic) without care for form-factor, I’ll shove it underneath the hot plate in a case, probably use a layer of aluminium foil or something on top for insulation. I’ll ”koptan” tape a thermocouple to the hotplate and do PID control, and just control it via USB-UART. Including a display and rotary encoder would be doable, but I was running out of pins and decided it would be more user-friendly to write a python script to speak UART anyhow. Writing a GUI that shows the ADSR reflow profile on a little OLED is not something I consider easy.

Not sure if 3D printed plastic will be robust enough, especially PLA as it deforms over time. I may go full JWST and have multiple layers of spaced out aluminised Mylar between the housing and the hotplate, though it’s likely that the PCB standoffs will conduct the most problematic heat.

>>2505992
0.5mm is better than a micro USB port. Consider hot-air or reflow plate or oven if it’s too much. I’ve heard those thermal cameras can be real shitty about not giving documentation (military tech), but if you can get a datasheet that describes its communication protocol then the rest shouldn’t be too tough with a DSP or half-decent MCU. It’s possible a generic composite video encoder will work too, maybe, but it’s much less likely you get one with the right connector. Better off looking for a breakout board for that connector.
>>
>>2505992
Hirose 50-pin mezzanine connector
>>
>>2505992
If it's this camera, theres an interface board made by manufacturer. It's pretty fucking expensive but shouldn't be hard to make one as it's just a connector soldered to pcb. If you can't make pcb's, get a bunch of those connectors and solder small copper wires to corresponding legs. When you get one connector correctly done, solder the wires to vero board and pour some epoxy over it.

Camera: https://www.flircameras.com/flir-tau-2-336-7-5mm-thermal-imaging-camera-core.htm

Interface pcb: https://www.flircameras.com/flir-tau-pcb-wearsaver-with-solder-pads-421-0047-00.htm
>>
>>2506025
It's not shown there but there's a 40mm fan on the base. Hopefully that'll avoid the pla from melting by having fresh air flow inside the case.
>>
>>2506113
I'd still give at least a cm or two of distance. Looks to me like there's direct contact between the case and the corners of the plate, but I may be wrong.
>>
>>2504484
>needs PCBs made
>local company says they can take orders from private customers
>tfw they will probably just order from pcbway and charge extra for "quality control"
been stuck in the thought of etching it myself for 4 weeks.... GAARRGH!!!
>>
>>2506116
There's a couple mm, but yeah. Probably a smart move.
>>
anybody make a Wienbridge oscilator run at very low frequencies? like I want to run one at 1hz and below but online the frequencies are all over 100hz
>>
>>2506118
If you have a CNC laser machine (or CNC router) then home fabrication is largely hassle-free after the initial software hurdles. If you have an SLA 3D printer then you can use UVtools to easily expose photoresist, which is decently easy, but for 2-sided boards gets kinda tough to not over/underdevelop, and is still not very convenient for solder mask. Toner transfer is arguably better. Printer transparencies are maximum hassle.
Some sort of projection LCD/DLP exposure setup could project light onto boards that are face-up, removing the issue of having a plastic layer atop them for solder masking, and allowing use of diluted liquid solder mask as an etch resist so you don't have to worry about development issues, but laser ablation is going to be far easier still.

via riveting is ass.

>>2506123
If you want a sine wave, just connect a 2-pole (sallen-key) filter to an inverting schmitt comparator. Or just increase the capacitances on the Wien-bridge until it works at the desired frequency. Either way you'll hit the rails without an AGC or pumping your THD with a diode soft clipper. I hear incandescent bulbs are surprisingly good for this.
>>
>>2506128
>If you have a CNC laser machine
Thanks! I have nothing and am just getting started on a hobby level. It would be no problem to order it except I'm autistic about doing it myself ("for self-reliance") in which case I'd go with the printer method or even just pulling enameled wire in a 3D printed (PLA) frame.
There are no issues with scalability of fabrication as it's for personal projects. It's just that the more local you go here in Europe the more obvious it gets we just don't do PCBs ourselves anymore. Every national company stating some collaboration with Chindia. There seems to be NO way whatsoever to support local fabrication to ensure environmental responsibility and workers rights...
t. eurogay
>>
>>2506143
>pulling enameled wire in a 3D printed (PLA) frame.
You could also consider 3D printing a stencil to spray-paint through, giving a crude etch-resist. I bolted a 0.5W laser to my 3D printer but the software end is a pain in the ass. Also might not be powerful enough, we'll see.
>>
>>2506152
>stencil
>spray paint
>laser
Can you give a quick run down on the process? I'm imagining a lot of different things at once now.
>>
>>2506152
>>2506165
Never mind I found it! Looks cool and easy to get going with. Thanks for the heads up!
>>
>>2506165
The idea behind both methods is spray-paint protects parts of the board from etching. It's so chemically resistant you don't have to be fussy about softening during a slow etch.
With the laser ablation method, you coat the entire board in the stuff and burn away where you want it to etch.
With a stencil, you physically block the paint from being applied to the parts you want to be etched away. The issue with a 3D printed stencil is you also need to print the supports that hold loose parts together, with some amount of clearance for the paint to get underneath. This adding of supports may or may not have an existing toolchain that will automate it for you. Also the whole paint running thing.
A vinyl cutter could make a decent stencil also, with a different set of problems.
>>
File: fdmpcb_detail.jpg (167 KB, 794x464)
167 KB
167 KB JPG
>>2506176
Oh here's someone making such a stencil, though by the sounds of the article they're using it to expose a photoresist.
https://hackaday.com/2022/07/26/a-new-way-to-produce-pcbs-with-your-3d-printer/

There's also methods that add a permanent marker to a 3D printer to draw over locations that need to be protected from etching, and methods that cover the entire board with permanent marker ink and simply scratch away the unwanted ink. I've had mixed results using permanent ink as an etch resist though, seems to not last slow etches.

For the laser stuff, Marco Reps has touched on using lasers to do PCBs a few times before, though he tends to use fancy fibre-lasers with incredibly fine focal points, which results in an incredibly fine PCB pitch. Much finer than my shitty diode laser.
>>
Cool thread guys! I often learn lots lurking these.

I'm hoping someone could help me identify this chip. My searching hasn't led me to the answer... I think it's a diode, perhaps surge protection of some kind. It's situated at the top of a row of 12 MOSFETS. It got blown with the MOSFETS when some doofus other than myself connected grid ac to the ac output of this hybrid solar inverter.

Thanks in advance for any insights
>>
>>2506365
Is it a chinesium inverter? Don't recognise that manufacturer logo. If you can measure the size of the diode it might be possible to figure out what package it is. If there's another of the same diode on the board then measure its voltage drop to figure out if it's a Schottky.
>>
File: board.jpg (1.32 MB, 2079x3355)
1.32 MB
1.32 MB JPG
How hard it would be to make a sensorless BLDC motor controller using just discrete logic ICs?... With field oriented control :D

I've made this hall emulator (basically LM339 comparing virtual middle point with phase voltage) but it worked like shit because ebike controller PWM was coming back through motor into halls and i couldn't really figure the fuck is going on since I don't have an oscope.
>>
>>2506379
Sensorless? Gonna be tough but with some fast op-amps and clipping circuits you can probably measure back-EMF reasonably easily, then use comparators for commutation. Problem is, the back-EMF voltage scales with speed, so you'll likely want some way of scaling the op-amp-gain or the comparator reference voltage as a function of speed as well.
>With field oriented control
Maybe if you count using an EEPROM as a combinatorial logic array to fill with an 8-bit sine table, then use a 4046 running off a motor pulse output itself to feed a counter that steps through the EEPROM addresses. Think you could run the PLL via the last bit of the counter, but I may be wrong.
FOC without some sort of absolute encoder is nigh-impossible though, hence why everyone uses current sensors on the FETs. Though you likely could interpolate from digital hall sensors or back-EMF.
Personally I'm going to try to shove analogue hall sensors inside my motor for FOC, though I'm not particularly hopeful about how well that will work. They all saturate way too easily, it's as if there somehow isn't any market for analogue hall sensors that can measure up to a tesla or two. I'll be trying to angle them at like 89° so there's only a barest hint of magnetic field that's normal to their sensor axis.

>phenolic PCB substrate
hope you like delaminations
>>
File: 1557149151253.jpg (55 KB, 540x699)
55 KB
55 KB JPG
Dear /ohm/,
My name is Ron Jambo. I'm being held prisoner in the jungles of little Cambodia (Rhode Island). I have one xilinx xc2c128 and some bamboo.

What would you do with the CPLD if you were in Ron Jambo's moccasins, Anon?
>>
>>2506396
I caca dau VC,
>>
WTF is wrong with my post, it's not spam
>>
>>2506405
Yeah you'll want it to handle that kind of high voltage if it's a TVS. Not that such a tiny TVS on a mains inverter can handle much before giving up the ghost, as you've discovered. Some searching in the TVS section on LCSC shows me a lot of parts rated 120V/133V/193V, which means a reverse standoff voltage of 120V, a minimum breakdown voltage of 133V, and a maximum clamp voltage of 193V, which would also be too low for US mains. Same for some Littlefuse diodes I found on Octopart, which at least look to have about the right size package:
https://datasheet.octopart.com/SMDJ120A-Littelfuse-datasheet-130255154.pdf
It's an SMC, though it may well be taller than your one. Still the wrong markings.
This onsemi part comes up when I search GGP:
https://datasheet.octopart.com/SMCJ70CA-ON-Semiconductor-datasheet-85604422.pdf
But the GGP one is different from the 170V one which is different again from the mention of 133. GGP is the part marking though, and these ones are also SMC.

If it's before the transformer stage and not after it I could understand it being a lower voltage, 70V or so could make sense for a 48V solar setup.
>>
File: pinout.png (24 KB, 471x561)
24 KB
24 KB PNG
>>2506025
>I’ve heard those thermal cameras can be real shitty about not giving documentation (military tech)
pinout documentation is on their products page, and in a separate document, they also give you info about the serial communication protocol and tables of the commands and their hex values, i know these are neeche products so the chineze wouldent know to make a simple board to boot the core up, and set it to NTSC output mode
>>
File: 3d.png (65 KB, 1656x1256)
65 KB
65 KB PNG
How is this layout? Inputs too close to power rails? This is just the top side.
>>
>>2506405
The URL may be blacklisted.
>>
>>2506438
>capital K for kΩ but lower case r
>U instead of μ
Also I can’t see any connections to the ground plane. Ground stitching to prevent snaky current paths is a good idea. The mounting holes look too close to components to get a good washer on.

Also show a multi-layer view with higher contrast instead of the 3D one. KiCAD’s Eeschema view shows both front and back layers at once now to some extent now that you can see through solder mask. If you are using KiCAD, go into the addon manager and get Mitxela’s trace melting script.

Also post schematic. Looks like a class-B amplifier with op-amp feedback, but not going for a V-be multiplier may result in a bit of crossover distortion due to the op-amp’s limited slew-rate. Better not be connecting it to a switching power supply.
>>
how come no one talks about fpga?
>>
>>2506486
Far too esoteric to be used by us dabblers. They’re more expensive and less useful than an upper-end 32-bit MCU unless you’re doing some really high bit-rate shit like live video encoding. More at home on the microcontroller thread anyhow.
Oh and the software is fuckawful.

t. never touched an fpga
>>
File: 8c.png (31 KB, 1656x1256)
31 KB
31 KB PNG
>>2506465
It's a current amplifier made in eagle and it'll be connected to a 317/337 regulated supply. Mounting will be with adhesive standoffs through from the bottom.
I'm not sure how ground planes are supposed to work but it only connects to the bottom left jumper. Everything is star ground on the bottom. I wanted to make the top ground plane like a shield and not use it for carrying current.
Did I do it wrong?
>>
>>2506487
>Oh and the software is fuckawful.
okay so my experience so far is the normal.
thanks anon
>>
How do I read the capacitance of the capacitors from this?
>>
>>2506507

ch-27p-j = 27pico
100m-m = 100 micro
f-r01M-Z = not sure; best guess is .01 micro
suffixes, KMJZ are tolerances; google those
e.g. F=1%, G=2% J=5% K=10%
>>
>>2506543
thanks, do you also know whats up with 2r2m-m?
Simply 2uf and tolerances?
>>
>>2506549

seems R is a decimal point, so 2r2m-m would be 2.2uF
>>
>>2505986
How did you make this gif? Is there a thread for CAD software?
>>
https://a.aliexpress.com/_mP61Dcy

Can I use this thing to power something and charge at the same time?
>>
Should I get a Pace ADS200 or a Hakko FX-951?
Or are there other soldering stations i should consider. In Europe if that makes a difference.
>>
>>2506562
Get one from chink express
>>
>>2506562
The Pace model is more robust, but both stations are good. If you solder all day long, get the Pace.
>>
>>2506128
>If you want a sine wave, just connect a 2-pole (sallen-key) filter to an inverting schmitt comparator. Or just increase the capacitances on the Wien-bridge until it works at the desired frequency. Either way you'll hit the rails without an AGC or pumping your THD with a diode soft clipper. I hear incandescent bulbs are surprisingly good for this.

I don't know what all those acronymsn mean, I am still kind of a noob. I don't think I need that much stability, but for more context I an looking to build a variation of the circuit mentioned in this paper "Experimental Study of Synchronization of Coupled Electrical Self-Oscillators and Comparison to the Sakaguchi-Kuramoto mode" to drive some LEDs that start off out of phase and then slowly lock in phase.
>>
>>2506118
just order from the chinks
>>
I hope this is the right thread.
Is there a way to open socket like in picrel without damaging it?
>>
>>2506622
Probably not. Why would you open it?
>>
>>2506557
I just took png. screencaps in fusion 360 and stacked them in gimp to make a gif.
Btw, does anyone know what temperature sensor should I use? 3D printer thermistors only seem to go up to 250ºC. Maybe a thermocuople?
>>
>>2506662
k-type thermocouple
>>
>>2505765
for an absolute beginner? not a station at all grab a ~30w fixed iron with SMALL tip and learn the basics first.
certainly not a station with sensor tips, you will fuck up and those tips are expensive.
americans blab on about hakko all the time idk never used one. depends on where you are what kind of irons and tips are commonly available.
>>
>>2506660
>Why would you open it?
It's an old walkman. Only one channel of headphones works, but second one connects when fiddled with. It's not the matter of dirty conections since I cleanded it with alcohol. Seems like that circled part must have wear off and bend, so maybe if I open the whole socket and bend it back whole thing will work again.
I really don't want to replace that socket since I'm not that good with soldering, and I like that walkman.
>>
>>2506486
if you went to the trouble of learning an hdl, if they provided you with the software suit for free (sometimes its not), if you knew how to use it properly (takes years) if all perifpheral software modules you needed had open licence (they don't) if the specs you needed were clearly given on each part and chosing one wasn't a labour in itself,
at the end of the day, anything you can do with an fpga you can do cheaper (and with better clock rates) in discreet logic. the only reason to use them is size and ease of post manufacture bugfix
>>
>>2506675
>the only reason to use them is size and ease of post manufacture bugfix
and cloning of extinct chips for drop-in replacements.
>>
>>2506675
You're exaggerating. You can program Lattice FPGAs (and some others) with fully open source tools. Learning Verilog isn't all that hard. Most times, the reason to use a FPGA is because you want to create the peripherals yourself in the first place, or a MCU would do.
>>
File: mat.jpg (184 KB, 1500x1177)
184 KB
184 KB JPG
I was thinking about buying one of these mats I see everybody soldering stuff on, how come I see people in the comments saying they're dangerous and will absolutely certainly with 100% accuracy fry anything electronic you put on them with ESD? Are they safe or not?
>>
>>2506687
I have that one and haven't fried anything yet. I doubt it would be that popular if it 100% fry everything. Put your sensitive in a separate place if you wish
>>
>>2506687
Isn't that ESD-safe (conductive silicone)?
>>
File: static charge.png (21 KB, 1121x116)
21 KB
21 KB PNG
>>2506690
>if it 100% fry everything
I was just parroting the alarmist comments a little there. Thanks for the first-hand account. Does yours ever build up static charge?
>>2506693
It says it's esd safe in the description, but I don't know if I would believe the chinese to not cheap out on the addictives needed to make silicone actually safe (by itself I know for sure it isn't).
>>
>>2506488
Oh a star ground. I’ve seen some star grounds that are effectively just a full ground plane but with small gaps going everywhere. Lower trace impedance I guess. That said, for something low-frequency and low-power like this, I doubt you’d get much out of a star ground, they’re more useful for high-power circuits where the voltage drop across the ground trace is significant, or circuits with analogue and digital going on where you want to isolate sudden current draw edges. Also consider guard traces around sensitive inputs.
As for the plane on the other side, it’s probably fine so long as it’s connected to ground somewhere.
>>
Do they sell transfer paper at Staples that's worth a fuck for PCB etching? I don't feel like waiting 3 days for presensitized boards to get delivered
>>
>>2506770
I've heard people using glossy magazine paper
>>
>>2506754
Thanks for the feedback on grounding. I read some papers on it and came away even more confused. I just try to imagine where the return current is gonna go.
I couldn't resist and I ordered the board. I'll check it for oscillation with a scope the best I can.
>>
>>2506672
I call it a signal lug / signal pin
sometimes the plugs can be out of spec, make sure you try a different cable/headphones. Also deoxit really does work better than iso.
>>
File: a.jpg (4 KB, 225x225)
4 KB
4 KB JPG
What should I use to extend the height of a kailh silent switch? The red part where you "click". I don't have a lot of tools. I used tape, temporarily, but it makes it even mushier.
>>
>>2506973
There are extenders for smd switches. If you need it quickly, make a cone from tape on top if and fill with epoxy (araldite or such)
>>
>>2506973
a piece of q-tip stick or toothpick?
>>
>>2506973
parts from a disposable pen?
>>
>>2506977
nice idea it would make it fit perfectly and I can sand it to fix any extra height
>>2506981
>>2506985
thanks I'll see what fits
>>
>>2506988
Just use epoxy putty then. Less mess and easier to work with.
>>
>>2504484
ROLLIN
>>
Sorry if this is a retarded question, I'm new to electronics. I recently tried making a rectifier circuit, but when I checked its output signal on my oscilloscope, the peaks of the AC signal I was supplying it seemed to be getting inexplicably clipped. The diode is in series with a resistor, so I have no idea why this would happen. Anyone know what's happening? I can post my oscilloscope output if it helps. Thanks.
>>
>>2507069
Post schematic.
>>
>>2506390
>hope you like delaminations
Better than silicosis. I hate to with with fiberglass
>>
I want to design my own driver board because I cant seem to find ones using the LDD MeanWell drivers. How should i go about that? I know verry little about acutal EE.
>>
>>2505106
>>2505211
Turns out it's actually got a positive temperature coefficient
>>
>>2507352
Driver board for what, LEDs? Not too tough, but you’ll have to study basic electronics, feedback loops, and the specifics of switching power supply design and LED driving.
>>
>no switch debouncing
>no issues whatsoever
huh
>>
File: Untitled.png (109 KB, 603x375)
109 KB
109 KB PNG
>>2506668
Aight. This might be a stupid question, but, since thermocouples are based on joining two different metals, could you cut these and solder them to a 2 pin jst connector, or would that affect the measurement?
>>
>>2507978
>could you cut these and solder them to a 2 pin jst connector
Yes.
>>
>>2507987
Thanks. I think I'm gonna make my own instrumentation amplifier with LM358 to go with it, since special ones like the AD8495 cost like 10$ a piece.
>>
>>2505106
So the sweet spot where the op amp comes out of saturation is when its inputs are at about 2.35V so the thermistor should be at 1050 ohms
>>
>>2507978
You can’t solder chromel and alumel. Best off just using screw or spring terminals.
Also not only do you need to measure the differential voltage, but you need to measure the cold-side temperature. The differential voltage just tells you the difference in temperature, which itself isn’t fully linear and probably needs a lookup table.

An IC like the MAX6675 will do it all for you. Looking at its specs it has a 330k input impedance. If you do try and create such a circuit with an op-amp, I’d try to get one with minimal offset voltage. 358s are a bit shitty, apparently JFET-input op-amps are even worse for offset voltage, so you want good CMOS op-amps. Could even build a chopper amplifier. You also need a decent ADC, I’d check out the specs of any op-amp and ADC you plan on using to ensure their noise and unideal behaviour doesn’t impede getting a reliable measurement.
>>
>>2508128
>You can’t solder chromel and alumel.
I think Anon was talking about cutting the plug off and replacing it with a JST connector.
>>
>>2508135
If you're doing that properly you crimp, not solder.
>>
File: 1581031252129.jpg (257 KB, 960x870)
257 KB
257 KB JPG
>>2508180
>If you're doing that properly
doing things properly is for fags
>>
File: Capture.png (242 KB, 1143x565)
242 KB
242 KB PNG
>>2504484
can these chinkshit signal generators read .csv files, ideally from a USB thumb drive connected to the back?
>>
>>2508244
or if that particular model can't, are there any budget small form factor generators with similar specs that can take .csv? i don't feel like spending $300-500 on a "proper" signal/awg, especially since i'm running low on storage space in my cuckpartment.
>>
>>2508244
what kind of frequency range do you need? it's not that difficult to make your own if you just want audio frequency ranges by using an audio dac. for faster stuff you can still get decent speedy dacs (like for composite video) for reasonably cheap. then just amplifiers with a few different gain settings, maybe some circuit protection stuff. need an adc too if you want to do stuff like modulating waves, not that i've ever wanted to do so. if you want it to measure signals for making bode plots (very handy) then you'd need a decently high-specced mcu.

the hard part is the ui, but you could just have it constantly plugged in via usb in order to use a computer window as the ui instead
>>
>>2508258
right now it's pwm at maybe 50 khz. but i would like something that can do 25 mhz square wave minimum for... who knows what. it sounds like an interesting DIY project but it's one of those things where i'd rather avoid the headache of DIY.
>>
>>2508273
buy an si5351 module from aliexpress for $2
>>
File: file.png (66 KB, 654x996)
66 KB
66 KB PNG
This is a dumb asf question but im hacking together some trash and one thing has a button like this, with the existing configuration where its 2 sets of linked pins. I need to get it so 1 button closes 2 different circuits. Is there a component or an easy way to do this while still using this form of button? I can't replace the button itself as it sits in the housing, but i can wire other stuff up to it.
>>
File: double up a switch.gif (22 KB, 1037x502)
22 KB
22 KB GIF
>>2508393
>1 button closes 2 different circuits.

if you're working on low-voltage DC, you can use a couple of transistors whose base is driven by your switch.
google ''low-side switch'' or ''high-side switch'' as needed.
a more general solution is to use a SPDT or DPDT relay.
>>
File: 4066_pinout.png (98 KB, 2128x3230)
98 KB
98 KB PNG
>>2508393
You can also do >>2508413 with a 4066
>>
>>2508423
He says it's so the button "closes 2 different circuits". Using the term circuit instead of signal makes me think he's turning loads on, so I'm guessing using some MOSFETs would be better suited than bilateral switches. That said, the only reason I can think of to use two switches instead of one is if he's switching to different nodes, so maybe one of the switching elements needs to be floating or tied to a different power rail.
>>
i find this funny
https://www.youtube.com/watch?v=HaUzUwNBFcc
>>
>>2505691
>>2505707
>>
>>2508461
Mate I'm talking about desoldering techniques. Not getting solder on all pins at once, but getting a solid bead of solder built up across and between all pins at once.
>>
File: chipquik.jpg (45 KB, 601x428)
45 KB
45 KB JPG
>>2508478

have a look at this if you havent yet https://youtu.be/FTQqjggeklo

my preferred technique is to cut the pins off with an x-acto, then sweep them off with a solder blob.
>>
I want to start doing challenges from the op. Do I purchase the components individually for each product or are there components I should but in bulk because I'll need them later? I already have a set of resistors from 100-100kΩ.
>>
File: alibaby kits.jpg (148 KB, 1079x850)
148 KB
148 KB JPG
>>2508538
>individually for each product or are there components I should but in bulk

you do both.
kits for common components.
individual buys for very specific functions.
most importantly, dont forget to serenade your transistors to sleep every night.
>>
>>2508521
Hey that's actually real neat.
Was somewhat afraid to cut the pins off as I don't have a good craft knife, so would have tried to chomp them off with flush-cutters instead. Do craft knives actually cut through the leads fine? Probably wouldn't be too tough not to cut the PCB traces beneath.

>>2508538
Small amounts of assorted passives (1k, 1.2k, 1.5k, etc.), then larger amounts of the common values (1k, 10k, 100k, etc.) is what I prefer to do, though kits don't usually come like that.

Also buy MOSFET gate drivers. Everyone should have mosfet gate drivers. I got bad ones from alibay, but if you buy ones that are already chinese (like the EG2131) you might have better luck. You just have to translate the datasheet yourself.
>>
File: file.png (19 KB, 224x205)
19 KB
19 KB PNG
Hello. What are these resistors for at the output of this amplifier circuit?
Are they nessecary?
https://electronicscheme.net/2x12w-stereo-amplifier-with-tda1521tda1521q/
>>
>>2508553
pls respond
>>
>>2508550
>Do craft knives actually cut through the leads fine?

real easy.
press tip against pin, press down.
if you're skillful, tip wont even touch PCB.
but if it does, no harm coz it's just a tiny point of contact
once you see how easy they cut, you're tempted to run across a whole row but that might cut traces.

>>2508553
>Are they nessecary?

no, but put it in anyway coz you never know if your particular chip will oscillate w/o it or not.
>>
>>2508553
Snubber network. It's for dampening high-frequencies, usually to improve amplifier stability. Might also decrease audible crossover distortion, maybe.
>>
>>2508560
>>2508558
>snubber network
Have heard that somewhere before, will look into it, thanks.
>>
I wanted to make my own test leads and wanted to use the same type of wire (stranded, compact) just dunno the type of insulation.
>>
File: file.png (134 KB, 456x335)
134 KB
134 KB PNG
>>2508560
Another question: Do I need that much stabilization at the supply input?
>>
>>2508572
Silicon would be best as it's heat resistant and very flexible, but that could get quite expensive.
>>
File: silicone rubber cable.jpg (7 KB, 225x225)
7 KB
7 KB JPG
>>2508572
>dunno the type of insulation.

easy.
silicone rubber wire, coz it straightens itself out (no kinks) and rarely gets tangled.

>>2508573
>that much stabilization

you need enough that you cant hear (or see on scope) any 60Hz noise.
rule of thumb is 2000uF per amp of supply current.
>>
>>2508572
For a multimeter? They’re usually silicone insulated, but with somewhat small amounts of copper and rather thick insulation, comparatively speaking. If you buy some wire that’s rated for rather high voltages it might be close. As for strands, I believe there’s no disadvantage to having large numbers of small strands. More flexible, less prone to work-hardening. For chemical ingress resistance I’d want tin-plated (or gold or maybe nickel) oxygen-free copper. Any marine-grade wire should fit those specs.

If you plan on passing current through it (e.g. 10A setting on meter) then naturally you should ensure the wire can handle that kind of RI^2 losses.

The harder part is getting the strain relief. Test leads get tugged about quite a lot, so something that has a really solid grasp of the insulation is a must. I suspect some sort of compression-like fitting that grips all around the wire is required, since the focused points of contact you’d get from a less symmetric crimp would tend to dig into the silicone and potentially rip through it. Sealing the wire in via silicone adhesive may actually be best. With a thick insulation you probably won’t have to worry about pulling the cable at an angle too badly, but a nice strain-break cone wouldn’t be a terrible idea just to ease the radius. Honestly I think some casting silicone in a 3D printed mould would be the way to go, for both the bananas and the probes.

>>2508574
Still cheaper than PTFE insulation.
>>
>>2508575
What insulation is it listed as in the National Electric code? Seems like it's similar to SO cord insulation.
>>
>>2508577
>The harder part is getting the strain relief. Test leads get tugged about quite a lot, so something that has a really solid grasp of the insulation is a must. I suspect some sort of compression-like fitting that grips all around the wire is required, since the focused points of contact you’d get from a less symmetric crimp would tend to dig into the silicone and potentially rip through it. Sealing the wire in via silicone adhesive may actually be best. With a thick insulation you probably won’t have to worry about pulling the cable at an angle too badly, but a nice strain-break cone wouldn’t be a terrible idea just to ease the radius. Honestly I think some casting silicone in a 3D printed mould would be the way to go, for both the bananas and the probes.

This would be used mostly for anything low voltage, like to have some long test leads for automotive electrical troubleshooting, I have some wires shorting to ground (negative) and want it long enough to test both ends from back to front of the car
>>
File: TDA1521 ENDSTUFE.jpg (780 KB, 1890x1181)
780 KB
780 KB JPG
>>2508575
I used more stabilization than in the circuit diagram cause I have some premade amplifiers where you could hear the 50hz humming noise of the transformer very clearly due to insuffiicient stabilization.
I also used bigger caps for the output. 1000uF still seems pretty small. Does that make sense? It's a 2* 12W amplifier.
>>
>>2508580
Well going for high-voltage rated wire just means that the insulation is nice and sturdy. Vulcanised rubber wire is really nice stuff, sturdier than silicone, sometimes found around kettle or toaster or oven power leads. Though honestly if you don’t care about temperature resistance or kinks then PVC might be fine. Maybe some sort of braiding if you want abrasion resistance.

If you’re tracing a short you might want Kelvin sensing, milliohm-meters aren’t hard to make.

>>2508583
Ultimately the PSRR if the amp is what determines the noise, and how close you are to clipping. If you get the total peak-to-peak ripple to maybe 5% of your rail-to-rail capacity you should be able to trust the PSRR of your amp to handle the rest. Do the math with the capacitor calc equation, or just sim it. PSRR should be specified in your amp’s datasheet as something like 60-100dB.
>>
>>2508128
>MAX6675
It seems like a really good IC, since I was originally planning on using the 10bit ADC on the atmel chips. Thanks for the tip!
>>
>>2508584
In the end I'm not that space restricted that it matters. Anyways, does this thing need a preamplifier? https://pdf1.alldatasheetde.com/datasheet-pdf/view/19206/PHILIPS/TDA1521.html
>>
File: file.png (10 KB, 522x379)
10 KB
10 KB PNG
>>2508583
Are baxandall equalizer circuits like pic related that can be built using a generic OP Amp suited for Hi Fi audio applications?
>>
What's up with various configurations of dual diodes in a single case? Sounds like they could be common cathode, common anode, anode to cathode, or not connected at all? Why 4 different configurations? And if I want to build a bridge rectifier with two of them, I'd need either #3 or #4?
>>
>>2504484
I'm designing something for my company. It runs on a coin cell and we would like a long battery life on it. It also has a small 2.4GHz radio on board so instantaneous current consumption will be beyond what the coin cell is capable of.
What's a good way to model a coin cell in circuit sim. I have the entire circuit including decoupling, etc in ltspice but I can't think of a good way to simulate the coin cell. A voltage source with a resistor in series is a simple way but I doubt it is that simple
Not really looking for anyone to spoonfeed me the answer but pointing me towards it is appreciated
>>
I want to put NTC in series with incandescent lamps to make them fade on and last longer. Lamps are 60w 100w or 40w. Can I just use a 1k ntc? there is no voltage rating or current on the ntc
>>
>>2508754
im on 220v btw
>>
>>2508677
0.3μA input bias current suggests no buffer needed. The amp itself has a maximum of 30dB of gain (also 30V/V coincidentally), so if you need more than that then yes another gain stage could be handy.

>>2508692
Not if that’s meant to drive a speaker, otherwise sure. Just watch that your THD won’t increase at the kind of output current you’ll be using, some precision op-amps are like that.

>>2508709
If you want diodes to ground or to VCC (for clamping purposes) then having one of those packages is a bit cheaper than two discrete diodes. Can also put the diodes in parallel for higher current capability.

>>2508754
100W at 240V = 0.42A, 576Ω (when hot). A 10k NTC would need to get its hot resistance significantly below 576Ω, which might be possible but I’d lean towards 1-2kΩ.
Also there’s a positive feedback thermal runaway effect while the NTC’s resistance is greater than the effective resistance of the bulb, where increasing temp means increasing the power dissipated by the NTC. So long as you don’t get stuck in that zone due to the limits of the NTC’s resistance range you should be fine, but keep an eye out.

Also due to the PTC of the bulb itself it may be possible to end up with oscillations.
>>
>>2508753
You could pick a coin cell from a good brand (ideally the one you guys plan on using) and try to match its discharge curve (from the datasheet) using a model based on LTspice's non-linear capacitor.
See: https://ltwiki.org/?title=Modelling_a_Ni-MH_battery_with_hints_on_Li-Ion_battery_modeling
>>
>>2508778
>Also due to the PTC of the bulb itself it may be possible to end up with oscillations.
that would be kinda cool.
>>
>>2508692
https://www.sound-au.com/articles/eq.htm
that guys website has everything
>>
>>2508754
No, just get a dimmer.
>>
Any good experiments for a power transistor? I have an extra BD437 and want to make some kind of switch for a LED or a relay. Kinda want to try some latching circuits or something.
>>
>>2508808
class-A amplifier
>>
how do you pronounce "18650"?
>>
>>2508896
eighteen six-fiddy
>>
>>2508804
>just replace every switch in the house
no thanks
>>
>>2508896
"one eight sixty-five zero"
>>
>>2508896
18mm 65m 0
>>
爆炸火
>>
I have a shitty radio that I don't care about that uses AA batteries.
if I bought some protected 18650 batteries, and then bought a circuit that prevented overdischarge like this:
https://www.aliexpress.com/item/1005003988318853.html
the another circuit that allowed the battery to lower it's voltage to 1.5 (since the 18650 battery has 3.7 volts), like this:
https://www.aliexpress.com/item/1005004130212100.html
reference here
https://www.matts-electronics.com/wp-content/uploads/2018/06/MINI-360.pdf
would that work?
I am just starting and know very little, but I assume I can just wire this up (battery -> protector/charger -> buck converter -> radio)
>>
File: igbt_chopper.png (37 KB, 984x522)
37 KB
37 KB PNG
>>2504484
Why isn't my IGBT chopper, chopping?
Switching circuit is typical 555 timer, putting out a square wave with 50% duty cycle at the 12 volts Vcc from a battery, circuit frequency is about 8 kHz, output current about 250 mA
The IGBT model is IHW40N120R5, breakdown voltage of 1200V, max collector current 40A, gate emitter threshold voltage is 5.1-6.4 V at 1.0 mA collector current. Happy to provide more info
I'm not sure what I'm doing wrong here, to my knowledge the common emitter configuration allows for voltage-amplified switching, however when I connect my oscilloscope leads as shown I just see a low voltage DC signal, even at 1000V (I assume some sort of leakage current?)
Why am I not seeing the load switching on my oscilloscope?
>What I'm trying to do
This is a test for a beefed up system, I happened to have some 1200v IGBTs laying around and wanted to make sure I knew what I was doing before ordering more. My ultimate goal is to switch a voltage of about 3000-5000V across a load with an equivalent resistance of about 500,000 Ohms (about 5-6 mA expected current) at a frequency of 5-10 kHz. I say expected because the load is actually an electron beam which I need to switch at kHz frequencies for research reasons
I'd probably want to work my way up in current but max would be a few hundred milliamps.
I am thinking about switching to using high-voltage MOSFETS but IGBTs at those voltages would be cheaper and more readily available.
>>
>>2508928
If the thing runs on 2 AAs, then it probably runs at 2-3V, and if it runs on 3 AAs then it probably runs at 3-4.5V, at which point you don’t even need a converter since lithium ions should be run from 4.2-3.2V. Not sure what those links are for, personally I’d use a TP4056 board with included DW01 protection chip. If you need a voltage converter, research efficiency and quiescent current draw. Make sure any voltage converter goes after the power switch.

>>2508945
Ooh, an electron gun! Sounds fun, hope you’re making a free electron laser.
As for your problem, I can only suggest idiot-check solutions. Like probing the 555 output and gate-emitter voltage, replacing the transistor, double-checking the pinout, measuring the resistor, and testing the PSU. Also hope your scope is rated for 1kV, because that could explain unusual readings.
>>
>>2508945
hope your hvdc supply wasn't sharing a ground with your scope
>>
>>2509097
>hope your hvdc supply wasn't sharing a ground with your scope

Dont be be ridiculous, dude. I'm not some Yuropoor fag. I can afford individual grounds for each of my instruments. With en suite bathrooms to boot.
>>
>>2509101
yeah, but with 1000v isolation? my bathroom is only good for 400v.
>>
>>2508778
>Not if that’s meant to drive a speaker
No, I'm the guy you quoted before that. So, yes I'll build an equalizer with some extra gain. The output of my phone is not quite enough it seems.
>>2508801
Just what I was looking for thanks!
>>
>>2509124
>The output of my phone is not quite enough it seems
Phone outputs, when the volume is turned up, can usually output enough for a standard line-level. But maybe you need something higher, at which point I'm not sure if you'd call it a preamp.
>>
I'm in school trying to understand this lab experiment with pulsed RL circuits, I have an inductor and resistor in series and I'm probing between the voltage source and the inductor, Channel 1 and getting a square wave and with the 2nd probe I'm probing between the inductor and resistor on Channel 2 I'm getting a ripple wave, almost looks like a sawtooth, is that because the inductor is using that voltage? That's called inductance right? When voltage is induced in the coil, no? Please explain as if I'm literally retarded (which I am)
>>
File: BodePlot.gif (29 KB, 625x438)
29 KB
29 KB GIF
>>2509140
>V = L * dI/dt
If I is constant, V is zero. If V is constant, I is a smooth ramp. Of course that only holds so long as V is actually constant, i.e. so long as your inductor is dropping much more voltage than your resistor. In reality, the current through the inductor is limited, so as you decrease the frequency you'll start to see the curvy edges of the wave.

Going more in-depth, all waves can be represented as a sum of sine waves:
>A * e^(j*(omega * t + phi))
Where A is amplitude, omega is angular frequency 2*pi*frequency, t is time, phi is a fixed phase offset, j is the imaginary unit sqrt(-1), and e is the fundamental exponential constant. If we define an inductor's impedance/reactance X as a function of angular frequency as being the ratio of voltage and current, we get:
>X = V / I
>X = L * dI/dt / I
>I(omega) = I_amp * e^(j*(omega * t + phi))
>X(omega) = L * d/dt(I_amp * e^(j*(omega * t + phi))) / (I_amp * e^(j*(omega * t + phi)))
>X(omega) = L * j * omega * I_amp * e^(j*(omega * t + phi)) / (I_amp * e^(j*(omega * t + phi)))
>X(omega) = L * j * omega
Note that this is purely imaginary, there is a 90° phase difference between current and voltage, just as we would expect from the defining differential equation. We usually define this reactance in terms of the imaginary angular frequency s = j * omega:
>X(s) = s * L
On the other hand, a capacitor (defined I = C * dV/dt) comes out as X(s) = 1/(s * C)

So by using that equation for reactance you can construct a voltage divider like you normally would:
>Vout = Vin * R2 / (R1 + R2)
>Vout(s) = Vin * R / (sL + R)
Note that when sL is much smaller than R, the output will approximate Vout = Vin, but when sL is much larger (at higher frequencies s) the gain will drop steadily. With this equation you can plot a graph of the impedances and phases of a wave as it passes through this filter, called a bode plot, pic related. By considering each harmonic of a square wave, you could recreate your sawtooth-like wave.
>>
>>2509159
>so as you decrease the frequency you'll start to see the curvy edges of the wave.

the voltage increases and the waveform lengthens as you lower the frequency, ya?

How would I figure out the minimum frequency of the square wave that will provide enough time for the current to reach it's max value?

You can give me bread crumbs since I'm having trouble really conceptualizing this stuff.
>>
>>2509164
>the voltage increases and the waveform lengthens as you lower the frequency, ya?
Yes.
>How would I figure out the minimum frequency of the square wave that will provide enough time for the current to reach it's max value?
Well, you could do like I've done at least twice, and differentiate it to find the peak value. But as you might expect from that image I posted, or by looking at the equation, there is no peak. At high frequencies, the wave is approximately a low-amplitude triangle-wave. At lower frequencies, the curved facets of the waves are segments of an exponential decay that falls out of feeding a step function into "V = L * dI/dt" and getting a natural log out, which you could calculate yourself if you want. As you decrease the frequency the amplitude only approaches that of the input square wave (and so does the shape of the wave). In terms of impedances, the inductive impedance/reactance is directly proportional to frequency. At DC (freq = 0) the input and output amplitudes are the same, or at least they would be if the real inductor didn't have some nonzero resistance, and at DC the inductive reactance is zero.

If you want a bit more freedom in messing about with circuits, try LTspice or maybe falstad / circuitjs. LTspice is a bit unintuitive, but if you watch a quick tutorial on it you should be able to manipulate it quickly and efficiently in short order. You can easily plot a bode plot with an ac analysis (e.g: .ac dec 6000 20 20k) or plot a time-domain waveform with a transient analysis (e.g: .tran 0 100m), and iterate on it quickly.
>>
File: file.png (106 KB, 680x835)
106 KB
106 KB PNG
What is the difference between an Asymetrical and Symetrical power supply and why does the former perform so much worse for this particular linear amp IC?
>>
>>2509184
>>
>>2509184
Being able to keep input voltages away from the rails helps keep the internals properly biased. Straying too close to the negative rail could well result in the distortion getting worse. When the input signal is going into a singe-supply amp like that, it gets AC-coupled and then pulled by resistor to the ground rail. The signal at that point will just be going on either side of ground, so for half the time it will be negative. Very few transistor differential amplifier circuits can handle voltages below 0V without suffering from distortion, I think ground sensing amps like the LM358 can get a tiny bit below the rail, otherwise it’s P-channel JFETs. Which don’t exist.
>>
File: IHW40N120R5.png (63 KB, 877x622)
63 KB
63 KB PNG
>>2508945
>12 V battery
Your IHW40N120R5 is toast.
>>
>>2509191
Didn't get it sry.
>>
File: gate_emitter_voltage.png (115 KB, 999x647)
115 KB
115 KB PNG
>>2509005
Well it's not a FEL, I can't say exactly what I'm building is even gonna work so let's just say I'm trying just build a kHz-switched electron beam source for now :)
To answer your questions
>probing the 555
Already done and the voltage and output current looked fine.
>replacing the transistor
I've already used like 4 out of the bag of 10 I have, all had the same issue
>double-check the pinout
I'm pretty sure the pinout is right on the 555 timer otherwise I wouldn't be seeing the square wave trace on the scope. The IGBT only has three pins and I'm pretty sure I'm not messing that up
>measuring the resistor
Also already done
>test the PSU
works fine according to my multimeter, the dial on the front is a bit miscalibrated but otherwise it puts out 1000V if it's turned to 1000V
I can triple-check everything again today but I'm not sure it's something obvious
>>2509005
>>2509097
>high voltage scope
The scope is rated for 300V. I have a 2kV 100:1 probe hooked up for all my high voltage measurements. I'm not an idiot, however the 2 kV probe could be pure chinesium. The same port on the same scope I've also used for low voltage measurements on the same channel and everything looks fine.
>>2509198
Are you sure that's how it works?
Isn't the max gate-emtter voltage on the IHW40N120R5 the +/- 20 volts as listed (25 for a transient?)
12V really doesn't seem like a lot of volts for an IGBT Vge
>>
>>2508573
it is a meme pushed by big cap companies, don't fall for it
>>
>>2509198
That’s the threshold voltage you tosser.

>>2509212
Can you measure an emitter-collector diode on the IGBTs you’ve put in the circuit, then check the IGBTs you’ve not plugged in yet?
They’re from a reputable source, right?
>>
>>2509446
That might be a good idea
I ordered them from Mouser, not sure about the manufacturer
>>
I'm going to place my first pcb order ever. Can I score the board with a knife and break it apart if it's 0.6mm thick?
>>
>>2509517
that's thin. you can break 'em with a vice
>>
>>2509517
Depends, are you cutting through copper or not? How sharp is the blade? You can just put in V-scores and breakaway tabs:
https://jlcpcb.com/capabilities/Capabilities
>>
File: diypsusketch.png (410 KB, 1354x1245)
410 KB
410 KB PNG
Got parts and getting ready to make a case to cobble together my first PSU for my first "actual" electronics project.

Here's what I wanna do, but I don't know if it's allowed.
I'm pretty sure I'm making parralel connections there for my DC fan and split between the 5V USB step down and banana plugs (both detected by the voltmeter)
Did I do that right? They'll sip the 12v and 5v line and just sip whatever amps they need?

Other questions:
1. What's the most practical way to add a standby LED light for the whole thing? The display on the RD6006 and two voltmeters will be for those individual ports, but an LED for the whole thing to know at least it's getting power in. Probably not worth it but I hope a neat and easy feature to add

Am I retarded?
>>
>>2509522
>>2509524
I'm just ordering a 10x10cm board from jlc with multiple 2x4cm designs on it. There's nothing connecting each individual piece. Was just wondering if it's just easier to score with an exacto knife and snap or should I still try breakaway tabs?
I'm using easyeda - v-cuts only seem to be available when panelizing?
>>
>>2509546
Yeah might require panellising. I just ordered multiple seperate boards in the same order, since the shipping cost is like 10 times more than the actual board cost. Plus board cost goes up with area, so I doubt the difference would be worth the pain of having to break them. I'd just add edge cuts around the outside aside from small middle point breakaway tabs, assuming easyeda lets you do so. Kike-cad masterrace.
>>
>>2509546
They're not very easy to score with an xacto. That's why I just break 'em in a vice.
>>
File: acrylic scoring tool.jpg (124 KB, 600x600)
124 KB
124 KB JPG
>>2509531
>Am I retarded?

could be. that drawing was def made by a patient at an asylum.

>>2509546
>just easier to score with an exacto knife

used to be you could get an acrylic scoring tool for $2.
seems inflation has hit that too, and now they're like 4x that.
but they do work fine on PCBs.
>>
>>2505537
>killed another MCU on the same board
>no idea how
just fucking end it
>>
>>2509546
just put rows of 1mm holes between each section. then you can just snap them apart with your fingers.
>>
>>2509718
>snap them apart with your fingers
https://www.youtube.com/watch?v=23rnAs9BHkk
>>
File: 1663404583270925.jpg (27 KB, 400x400)
27 KB
27 KB JPG
>>2509714
>could be. that drawing was def made by a patient at an asylum.
yes but does it make circuit sense
nothing should blow up or melt right
>>
>>2509826
You should probably use an isolated buck converter to power both volt meters.
>>
>>2509827
I have plans on using only one of the 5V lines at a time, either nanner plugs or the usb thing and figured it'd be okay.
I'd be much appreciative if you could point out the flaws in that logic though. regular cable guy wanting to make little things move as a hobby
>>
>>2509828
inaccurate readings/noise

https://www.digikey.com/en/articles/use-isolated-dc-dc-converters-with-embedded-transformers-to-ease-assembly
>>
>>2509835
Thank you, because I can imagine I wouldn't want that. I've got like 5 of them anywho.
>>
I'm adding a solar power to my RC plane to maximize flight time. I already have cells, buck converter, and the diode that will prevent back flow from the battery (this plane is very small so I don't have space for a proper charge controller or MPPT). The plane will be run on two 18650 batteries.
Solar output will be wired in parallel to the battery pack.
My concern is that with the solar in parallel to the battery, I won't know the battery voltage because I will always be getting 7.5V from the regulator. I want to use a free pin on my flight controller to control a relay that can turn off the solar power so I can see the battery voltage, but don't know anything about them besides "small power on one pin can allow big power from other source on other pin". I was thinking about using an auto relay. It needs to be as small as possible for weight savings. What relay would you recommend for this?
Cruising current: 4A (won't pass through relay
Max current: 10A (won't pass through the relay)
Battery volltage range: 6.2-8.4V
Maximum solar output: 2.2A (will pass through the relay)
Solar voltage from buck converter: 7.5V
Pin high: 3.3V
>>
>>2504484
you guys may like this one:
>>>/sci/15011366
https://techxplore.com/news/2022-11-sweet-microchip-patterns-curvy-surfaces.html
imagine printing CPUs at home with this technique...
>>
>>2509947
I think he solar voltage will mirror the battery voltage, not the other way around. Should be fine without a switch, but if you need one use a mosfet instead of a relay.
>>
>>2504484
What's the best brand of helping hands?
>>
>>2510018
rubbery mat.
>>
>>2510018
Panavice. Also blu-tak.
>>
>>2510018
QuadHands
>>
So I decided to dive into electronics and bought myself an esp32 with camera. It does not have its own usb connector, but I got a serial2usb board for that. I blinked couple of LEDs and started testing the camera. There's an example sketch for camera and it outputs the result onto a web-server. I've checked ytube and everyone does exactly that - watch the camera feed from the server or take pictures on SD card. So my question is, how the fuck can I get what the camera output directly on my pc via usb? Should I build myself a fancy serial monitor which can decode images from the text feed?
>>
>>2509949
No. It wouldn’t be precise enough to allow alignment of multiple exposure/doping masks, and even if it were you’d still need the same sort of semiconductor lithography tech in order to make those masks in the first place. In no way does this make semiconductor manufacturing easier or more accessible.

If you want to make silicon at home, you’re best off making a wafer exposure unit using a DLP projector and some quality first-surface mirrors, and making a vacuum deposition chamber. Could probably do that for a few thousand dollars, and allow you to make maybe a few hundred transistors. Well short of a modern CPU, but you might be able to get a 4004.
>>
>>2510052
The USB feed is just UART, not something that could be modified to be expressed as a USB camera. You could bit-bang USB through some equivalent to V-USB, get an EDP32-S2 with native USB hardware, or my favourite, just buy a camera with USB output. Cheap replacement laptop webcams fit this requirement.
>>
>>2510057
>The USB feed is just UART, not something that could be modified to be expressed as a USB camera
I don't understand what you mean. I don't need plug and play etc. I just want a video feed from camera in a window just like I have text feed in a serial monitor window. How can I get the text but not data (which is also pretty much text)?
>You could bit-bang USB through some equivalent to V-USB
What should I read or watch to understand this? Sorry I'm new.
>get an EDP32-S2 with native USB hardware
I think native and non-native change nothing in this case? I have a USB connection just fine.
>just buy a camera with USB output
But I don't need a USB camera, it has to be standalone.
>>
>>2510057
Okay I get another question. How can I monitor the serial output on my pc outside Arduino IDE?
>>
>>2510057
Or how does a software like Ardupilot deliver live fpv footage and all telemetry directly on PC?
>>
>>2510052
>watch camera feed
This needs a lot of bitrate. You can't do this over a shitty UART. Even USB 2.0 isn't good for cameras, they all transfer compressed video at low framerates.
>>
>>2510082
115200 baud via uart serial is ca. 11 MB/s. How is that not enough?
>>
>>2510086
More like 11 KB/s.
>>
>>2510086
Also, even 11 MB/s is not much for video. If you have 640x480 pixels resolution and 1.5 bytes per pixel, you would need 13 MB/s to achieve 15 FPS.
>>
>>2510089
>>2510086
But how do you get a live capture from any standalone device with camera like a drone connected to ardupilot? I wanted to make myself something like this and specifically bought a board with camera. Why you make me feel bad anon?
>>
>>2510091
You have to be more specific. But unfortunately, you won't get very far trying to use a 11 KB/s connection for video. Not impossible, but likely not even close to good enough.
>>
File: IMG_20221126_024110.png (178 KB, 1200x474)
178 KB
178 KB PNG
>>2510092
>You have to be more specific
Look I've made a scheme.
>>
>>2510063
>How can I get the text but not data
Well UART is just a relatively low-speed protocol. Get maybe 2 Mb/s max, which is pretty damn shitty for video. Compare that to USB 2, which can give 480Mb/s, which is barely enough for a passable webcam.
>What should I read or watch to understand this? Sorry I'm new.
A quick google search suggests nobody has actually implemented bit-banging USB in an ESP32. As I say, the ESP32-S2 has this feature natively, and is probably powerful enough to directly process images to send through USB. As for video, not sure. Might want to look into a Cypress FX3 programmable USB 3.0 controller, or equivalent, assuming there isn't an ASIC that does what you're after.
>I have a USB connection just fine
You have a USB-UART connection. As I say, maximum baud rate of 2Mb/s. Plus you'd need to write your own driver to receive the serial data and format it into an image.

>>2510066
It's really easy to do with python, which IIRC has its own methods for opening bitmap files so you may be able to display images through received data. Not that it would be fast.

>>2510069
Most FPV video is analogue, not digital. The analogue video receiver would then feed what's basically a composite video capture card, something you can buy off the shelf. The technology behind digital FPV is a lot more complex due to the low-latency compression algorithms, something I'd expect them to be using FPGAs or dedicated video codec ICs for.

>>2510086
Can you count the decimal places? You're out by a factor of 100, it's 110kb/s. And that's a lower-case b meaning "bits", not "bytes", so you're out by another factor of 8.

>>2510094
>those shitty 433MHz modules
AHAHAHAHA
Those will probably be more of a bottleneck than UART. And also likely enough an affront to the FCC, not to mention your neighbours' garage door openers.
>>
>>2510101
So basically you mean I'm pretty much fucked and there's no good way I can buy a camera module, hook it to a development board and watch its feed?

>AHAHAHAHA
>Those will probably be more of a bottleneck than UART
I found those receivers in google images, at the moment my camera is hooked straight to the board.
>>
>>2510105
Well it's not impossible, but considering the kind of bitrate even a low resolution camera will output, you'll need a pretty fast MCU to even just pass that data on without any processing. The dev-board is a negative, not a positive. If you do need processing, you should try your best to use an ASIC instead of a microcontroller.
If you don't need it to be wireless, just use a USB camera, or maybe a composite video camera with a USB converter. If you do need it to be wireless, then buy a composite video camera and wire it up to a composite video transmitter.

At least say the specs of the camera.
>>
>>2510109
>wireless
Wouldn't that mean Wi-Fi for high enough transfer rates? Or even 5G over a mobile provider.
>>
>>2510111
Hadn't thought of that. Yeah you could use your computer with a wifi network card for receiving pretty easily, but getting something to transmit that video feed at high bitrate is still not trivial in the slightest. Also don't expect long range.

Ask the remote control general why they all still use analogue video.
>>
>>2510116
I thought the new meta in fpv rc was the digital dji setup?
>>
>>2510118
Yes and no. Higher definition, higher range while still looking good, lower absolute maximum range, probably still higher latency. More expensive when you inevitably crash. If you're flying for fun and not for cinematography, analogue is still probably the way to go. At least that's what it was like 2 years ago, probably hasn't changed significantly since then.

Maybe someone out there has done a teardown on a DJI system to see what ICs they use. I'm betting there's either a high-spec ASIC or an FPGA in between the camera and the RF transmitter. Not cheap.
>>
>>2509716
>78L05 has the opposite pinout to 7805
WHAT THE FUUUUU
>>
Why does a Teensy 4/4.1 board cost €50 in Germany? Esp32 is for example €7.
>>
How do I make a timer for a nine volt battery that leads to an electric match?
>>
>>2510253
>timer for a nine volt battery

sounds like you're trying to build an IED
you need...
- casio watch F91-W
- interface circuit (and power transistor) to convert 3V alarm signal into a 9V-powered switch
- a light bulb (with the glass broken but element intact) to light the fuse
>>
>>2510139
>78L05 has the opposite pinout to 7805

just spin it 360 degrees and plop it in.
>>
>>2510255
No I have a very autistic schedule and I want an automatic cigar lighter via a propane
>>
>>2510255
Thanks though
>>
File: 1554680299639.png (165 KB, 803x445)
165 KB
165 KB PNG
>>2510253
Is this for a menorah? for celebrating honkukkah?
>>
>>2510139
As a software developer I find this scary. Make a tiny mistake, and your hardware is fried. I especially hate when I don't know whether I accidentally fried a part or whether it doesn't work because I made a different mistake.
>>
>>2510368
That's the part of the fun. Now imagine that but with $40k FPGA.
>>
File: 1664048827214964.jpg (15 KB, 480x360)
15 KB
15 KB JPG
I bought a KSGER soldering station for my first soldering station.
I then read about the grounding issue, but easily fixed with some soldering.
I then bought a Pinecil v2 since I have a 65W PD power bank so solder the KSGER
I haven't gotten around to doing the grounding fix for the KSGER and just use the pinecil
What kind of retardation do I have?
>>
>>2510253
MCU + RTC (or GPS module) if you want accuracy from days down to minutes. Or just take apart an old watch like IED-anon suggested.

>>2510256
Well it's more that it fed 12V into my microcontroller. It was bodge-wired in place anyhow, if it was in the PCB design from the first place I'd have noticed which way around it should go.

>>2510368
At least in this case it was an easy issue to trace. A continuity measurement across the voltage rails said 6Ω total, and the only different part was the 78L05. The scariest and most demotivating thing for me is to have stuff going wrong, and to have absolutely no clue WHY they're going wrong. If a bunch of shit goes wrong but I'm slowly working my way through the cause and effect of each then that's just part of the engineering process, and I'm learning something from it. But when I run into dead ends, it feels like the skies are conspiring to prevent the actualisation of my circuit. And anything that kills your motivation is particularly awful when you're self-employed.

>>2510454
Use the Pinecil to install a barrel jack and/or XT60 on the back of the T12 station. That way you have two portable irons.
Also arguably you can upgrade the T12 station to use JBC T245 tips, though I'm unsure if the internal MOSFETs can handle 2.5 times more current. It would need a different PSU also, but you should have bought the external PSU one anyhow because of the internal PSU issues.

Another modification is to ditch the PSU altogether, take the controller and put it in a case with a power tool battery holder built in, a heat-resistant slot for the iron itself, and a belt clip. Should make a convenient and long-lasting portable setup for a tradie friend.
>>
>newest eboom video
Teaches you about the importance of NO.
>>
How to blink with LED?
>>
>>2510531
>discrete circuit with capacitors and resistors to generate a square wave
>FPGA with explicit logic to switch a pin on and off
>MCU with a program to toggle a GPIO on a timer
Choose your poison.
>>
>>2510531
555 with astable configuration.
>>
>>2510531
brushed motor in series with LED
>>
>>2510531
Buy a blinking LED.
>>
>>2510531
if you didn't bare metal code that DSP using assembly, you haven't truly blinked an LED
>>
>>2504484
>https://github.com/Rocheez/4chan-electronics-challenges/blob/master/list-of-challenges.png
roll
>>
Yo got another question regarding my retarded linear amp hackjob.
Basically these caps at the input are called filter caps, right? Do they need to have this exact value given in the example circuit on the datasheet? - For now I didn't have that value in polyester caps, so I put 3 100nF in parallel. How do I know which frequencies they filter out?
>>
>>2510808
Are they PSU filtration caps or signal filtration caps? If they're PSU caps then precise capacitance doesn't matter much, but will determine ripple voltage. If they're signal caps then their capacitance will likely form an RC filter, the cutoff frequency and bode plot of which is very easy to google. Chances are it would just be a 20kHz low-pass or 20Hz high-pass.
>>
File: file.png (10 KB, 298x494)
10 KB
10 KB PNG
>>2510808
>>2510811
I forgot to include the image . sry
>>
File: file.png (70 KB, 1312x806)
70 KB
70 KB PNG
>>2510811
>>
>>2510811
>>2510814
Yeah so they're high-pass filter caps. The 220nF + 20kΩ input impedance means a time constant of 0.0044s, or a corner frequency of 227rad/s = 36Hz. Using 300nF instead means a corner frequency of 26.5Hz, which is fine. Personally I'd want to aim for ≤20Hz, but there's bugger-all below 50Hz anyhow so it basically doesn't matter.
>>
>>2510531
The eternal electronics question.
I've tried a blinking led, transistor circuit, 555 chip, arduino, dual opamp. Depends on duty cycle/waveform, physical size, current, adjustability.
>>
>>2510831
>dual opamp
Not dual comparator? Op-amps can act kinda funny when run like a comparator, though it depends on the op-amp.

The worst of all the solutions is a reverse-biased BJT oscillator.
>>
>>2510830
Ok, thanks
>>
>>2504484
>>2506118
>>2506152
>>2506176
Here's another novel idea I won't have the opportunity to try for myself in a foreseeable future
>CNC etching using EDM techniques (electronic discharge machining)
Applied Science have documented the process of building a conductive metal cutter and I can't see why it wouldn't work for PCBs to skip a lot of steps. For bulk manufacturing CNCing every trace won't be competitive but an EDM attachment to a 3D printer and the proper tracing software to avoid making non-conductive islands of copper would be required.
>inb4 "just CNC the traces with a cutting bit"
the copper sheet is to fragile IIRC
>>
I'm making a drone bomber to drop bombets on russians. Anyone have a good algorithm for calculation of engagement (or whatever is correct English term for timing of drop)?
>>
File: 1537303792366.jpg (63 KB, 680x521)
63 KB
63 KB JPG
>>2510886
>>>/lgbt/
>>
>>2510887
>>>/pol/
>>
>>2510867
just order from China already, like OP tells you
>>
File: 1535362269867.png (338 KB, 657x527)
338 KB
338 KB PNG
>>2510897
>>>/mlp/
>>
>>2510887
>>2510897
>>2510959
The fuck is wrong with you fagets, give me algorithm already. I bet here is half of anons play with RC and drones.
>>
>>2510974
>>>/qa/
>>
>>2510974
It's probably the wrong thread anyway. Maybe the /mcg/ or /rcg/ thread is better?
>>
>>2510974
>>>/out/
>>
File: Backlight-Cables.jpg (59 KB, 528x350)
59 KB
59 KB JPG
hi can you help me please, i am trying to find the name/type of the connector in the picture on the top right corner.
so far i have only found the JST SYP-2 which looks similar but does not match.
thanks
>>
>>2511002
BHS connector
>>
File: JST BHSR-02VS.jpg (21 KB, 266x282)
21 KB
21 KB JPG
>>2511006
thanks
>>
What is the best method to soldier SMD components at home?
>>
>>2510834
chip is a TL072, used to make an oscillator. gives off a fade rather than a blink but it worked excellent for my purposes
>>
>>2510867
I can’t think of any reasons to do that over CNC routing, which needs just as much thought put into levelling, but can also do drilling operations and soldermask. Or using a laser to ablate away etch resist, something that’s far easier to mount to a 3D printer.

>>2511031
Iron with small screwdriver tip. Hot air is also pretty accessible, so is a reflow plate, both of which are better for soldering tricky components like QFNs.
>>
hey peeps, I was wondering what you guys use as a overhead work light on your bench? The one I am using I am finding more annoying than helpful, and may repurpose it for something else.

I am not sure how to ask this, but I will do my best.

For reference this is the current one I am using.

>https://www.amazon.com/gp/product/B07R22DMCR

the magnifying light is not a requirement, I have rarely used it and find it useless.

I am looking for something I can clamp to my desk, I use the longer side of this desk, to work on stuff. Not all of, I have some space to do my college work as well.

>https://www.amazon.com/gp/product/B076Q6GNYH

But I am wanting to buy two of them, so I can put one in the middle, and at the end of the desk, the one I linked earlier creaks so badly when moving it, and it can't even move it enough to have it over me if I am working on something more towards the middle.

Would be cool if there is one I can move the light horizontally without it losing vertical height, but I am just looking for something that is better overall than what I have.

I may go budget route, and use two of these hard freight LED I have, do some cad modeling to put themon these flexible joints like from adam savage's video.
>https://www.thingiverse.com/thing:3938011

Thanks in advance for any recommendations.
>>
>>2510018
>doesn't know about blu tac putty

I remember buying helping hands from radioshack, and pcb vise from amazon, and rarely use them, like this anon said here >>2510024
I find myself soldering on ESD mat I have, if anything I just use helping hands to hold a wire for some cases, and blu tac putty.
>>
>>2510454
>I then bought a Pinecil v2

Any issues with it? I have thinking about buying one to put in my backpack to solder some shit while its slow at work, which has been alot lately because of covid.
>>
I have a magnifying desk lamp similar to pic related which I use quite a lot, but many if the LEDs have started burning up so I'm planning on replacing them. From what I could cross-reference on Google they look like concave 5 mm LEDs, but will any regular set of those just work? How do I know they're receiving the correct amount of power? I know LEDs are weird because negative resistance or something (althought I never delved too deep into exactly what's up with that), but what do I need to know to make sure I have the right LEDs for the job?
>>
>>2511141
if its leds keep burning like that then I would guess it is already not supplying correct amount of current for those leds
>>
can you solder silver wires with regular soldering stuff?
>>
>>2511170
Yeah. Use eutectic solder and a lower temp setting on your iron.
>>
A mouse I bought stops moving the cursor in some situations like when I move it. Tried different pc, cleaning, new mousepad, level of distance settings, everything you can imagine.
Do you think it's the sensor? Is it easy to replace it? I can't see any pixart seller on Ali tho
>>
>>2511208
correction, when I lift it and put it back in normal movements
>>
>>2511208
>>2511209
Wired mouse? Replace the cable.
Wireless? Reflow the battery contact solder joints.
>>
>>2511211
and check the little power switch for high resistance.
>>
>>2511211
I did tried a different paracord cable and unfortunately didn't help. Everything else keeps working like buttons and you can even see the cursor is still moving in the right direction but barely (a few pixels).
>>2511212
Oh I don't know shit about electronics ha
>>
>>2511214
Then the optics are dirty or the sensor is faulty.
>>
>>2511215
It's possible the LED emitter is bad, but less likely.
>>
>>2511215
thanks
>>2511217
is that a different component or comes bundled with the sensor? I know how to de/solder so I'll just buy a new one and see if I can somehow fix this mouse but I doubt it
>>
>>2511218
>is that a different component or comes bundled with the sensor?
A laser mouse has the laser and sensor in one unit. In a regular LED mouse, they components are separate.
LED - https://www.youtube.com/watch?v=NrjV0srfFnE
Laser - https://www.youtube.com/watch?v=sAklrT_HL0A
>>
File: kapton and copper.jpg (1.08 MB, 2742x2258)
1.08 MB
1.08 MB JPG
Did someone say "diy flex pcb"? Not sure how standard solder mask will hold up under bending though, or if this copper foil will work-harden too easily.
>>
File: 1669601286644.jpg (60 KB, 720x1254)
60 KB
60 KB JPG
I've come into a really large supply of components from a garage sale, but it's all undocumented.
I have part numbers for a lot of it, but could I use one of these components testers to identify them without having to Google a data sheet on everything?
>>
>>2511264
Not a bad idea for initial categorisation, but you should download datasheets for all parts you own anyhow.
>>
File: baseus-gan2pro-120w.jpg (70 KB, 604x601)
70 KB
70 KB JPG
>>2511104
Not him but I have one and it works pretty good. You definitely want a 100w+ USB c supply though. I'm powering mine off a Baseus 120w supply. At 20v you're getting pretty close to peak performance out of it. (without going 24v bench supply)
I do get a little bit of coil whine from the supply when it's heating the iron up at first.
>>
>>2511304
here's the power chart from the pinecil wiki for your reference
>>
>>2511307
What's stopping it from running on 28V from the barrel jack?
>>
>>2511347
idk. officially it only supports up to 24v. The schematics are on the pinecil wiki if you want to dive into it
https://wiki.pine64.org/index.php?title=Pinecil#Pinecil_V2_2
https://files.pine64.org/doc/Pinecil/Pinecil_schematic_v2.0_20220608.pdf
>>
File: niggy.png (94 KB, 1314x464)
94 KB
94 KB PNG
>>2511355
>shitty dual P-channel chinky MOSFET package
>driven by 2:1 voltage divider from VCC
>5.1k effective series gate resistance resulting in 2mA gate drive
do these guys not know what gate drive ICs are or something? it's not as shitty as the TS100, but it's still somewhat awful. also not too sure why they're using two FETs in anti-series, is backwards current flow a potential issue somehow?

if they used a proper gate drive circuit and an overkill FET it might be able to run at like 48VDC, with a sufficiently conservative PID. or be powerful enough to run a T245 tip if you swap out the mounting hardware.
>>
Why come the hakko 888d is like $110 but the iron itself is $80?
>>
>>2511382
>but the iron itself is $80?

product pricing is only tangentially based on technical aspects (or even logic).
mostly it's a psychological science used by marketing professionals.
they've investigated diff strategies and found this one maximizes profits.
>>
>>2511378
That second 5.1k (R9) is pretty much redundant. I'm guessing they didn't want 0.6V at the gate for some reason so they used (VBUS_TYPEC / 2) + 0.6V.

>also not too sure why they're using two FETs in anti-series, is backwards current flow a potential issue somehow?
Nah its not a problem since the FETs are still on the "high" side but I too wonder why they did that. If they were in parallel to increase current capacity it would've made sense but not like this
>>
>>2511355
>>2511378
>>2511407
/ohm/cil when?
>>
>>2511407
R9 isn't redundant, it keeps the gate voltage between Vcc and Vcc/2 (0 to -Vcc/2 Vgs). If it weren't there, the gate would get to 0V (-24Vgs), which kills most FETs.

The antiseries thing might be for thermocouple measurement, for tip sensing, or maybe for internal antistatic protection reasons.

>>2511408
It would be a T245 (or even beefier) hand-held unit that can take up to 12S/50V for 1kW peak of soldering output. Heats up in 0.1 seconds. Can be programmed to follow reflow curves at the touch of a button. Has a sous-vide mode where you immerse the tip in a bucket of water with your steak, also can be used directly as a meat thermometer. XT60 socket and 3.5mm headphone jack on the back. Reprogrammable with a mini-toslink transceiver in the base of the headphone jack, can also play eurobeat to make you solder faster. Readout for internal MOSFET temperature is larger than the tip temperature because it's more important. High-frequency internal HV generator for easy arc starting. Measures hand moistness via electrodes on the handle. Vibrate motor for feedback, also ribbed silicone handle. Mercury tilt switch to deliberately violate RoHS.
>>
>>2511144
Well I can replace them again in the future if need be, but I'd just like to get them all working again for now. For all I know the company that made these just used a bad batch of LEDs or something. Still, how do I know the replacements I'm buying will be the correct ones?
>>
>>2511411
That makes sense about R9. I assumed VBUS would be something like 5V or 9V or 12V so the thought didn't cross my mind.
>>
>>2511430
Shoulda used a zener instead of R8, much more versatile. This way your FETs aren't going to be driven properly if you run it off 12V.
>>
>>2511437
>Shoulda used a zener instead of R8, much more versatile. This way your FETs aren't going to be driven properly if you run it off 12V.
Absolutely no doubt there, zeners are the more "standard" way to drive FETs in this type of configuration. I guess they were just trying to save a bit of money there
>>
I want to build a portable ~50Ah 12V lifepo4 battery and work my way up to something more substantial that can provide 230V via inverter and 400Ah+.
I've been using rechargeable batteries of all sorts for a long time and have never had anything go wrong other than simply a dead battery. I know very basic circuitry and should overspec for safety reasons. Safety is also why I want lifepo4. I want it to have fuses, a kill switch and general on/off button, connectors where I can to avoid having exposed wiring etc.
I'm not buying individual cells but a ready-made quality 50Ah for my first one. What's the worst that can happen apart from accidentally shorting?
>>
>>2511307
Thanks anon, I've heard of mixed things but I think it was about the old version, I actually have a power supply I took off my ender 3 and swapped it for a different one, I can probably use that to power the pinecil, thanks for your input.
>>
>>2511472
np, I just recently bought mine off Amazon, they also have an authenticity checker on the pinecil site.
https://pinecil.pine64.org/
Don't bother with the aliexpress ones since they want the same price as legit ones for the clones.
>>
>>2511307
>3A over these tiny tiny USB-C contacts
>>
>>2511477
you should read about the new epr usb standard
they're gonna be pushing 240w at up to 48v down usb c cables
>>
>>2511479
In 10 years it'll just be 120V AC. USB is the cryptocurrency of power delivery.
>>
>>2511480
come on we would rectify it first then send it down a usb c cable
although their primary reason for stopping at 48v is safety
>>
>>2511482
>come on we would rectify it first
lmao
They will use recycled Bell wire for future USB cables. The fancy ones will be a maximum of 3 feet in length, and no insulation aside from Flex Seal™. Every cable will behave differently and they'll cost as little as $49.99 unless you buy in bulk from aliexpress and wait for 6 months.
>>
Edison won.
>>
>>2511485
only took decades of advancements in power electronics to make it work.
>>
Guys, can i embed this ibto my phone? I tried using NFC tools but i get no reading both from the tag and the receiver, isn't this supposed to be NFC?
>>
>>2511545
That's an RFID tag.
>>
>>2511555
How can I use this with my phone? If possible
>>
>>2511562
What exactly are you wanting to do? Read up on RFID vs NFC and the different types of tags, as not all of them operate on the same frequency.
>>
if i have a chink instrumentation amplifier and the gain resistor on it does nothing and the amp it self doesn't work as expected (the amplification works but is very noisy) do you think i got chinked and it's actually just opamp with fake name on it?
>>
>>2511564
There are 13.56MHz tags as well anon
>>
>>2511104
I've used it three times at 85w. I'm not a real electronics guy but it's melted the solder. However I did buy the $25 tips because they send you a tiny needle-point thing that's not going to heat anything I'd use. That's where it stops being -as- cheap but I like it.
I use a 65W power bank so I'm not even maxing the power.
>>
>>2511766
>I've used it three times at 85w
65
pardon the retardation
>>
File: wheresolderpoints.png (1.1 MB, 542x860)
1.1 MB
1.1 MB PNG
I got these from China but thought they'd be bigger
where the actual hell would I solder on here?
>>
>>2511865
post underside
hope there's some larger smd pads on it
>>
>>2511865
SMT - surface-mount technology
SMD - surface-mount device

They differ from THT (through-hole technology) in that they are soldered flush against pads on the PCB with solder paste and hot air or infrared light. You can solder them with a regular iron if you have the right tip and magnification.
>>
>>2511872
>>2511873
lel smd is not what I want but no big it was like $1.50 for the whole lot. I wondered how they were so cheap.
thanks anons
>>
File: 1556419096830.jpg (94 KB, 720x615)
94 KB
94 KB JPG
>>2511875
no problem
>>
>>2511875
Do you not look at the datasheets when buying parts?
>>
File: horse porn.jpg (63 KB, 750x752)
63 KB
63 KB JPG
>>2511879
>Do you not look at the datasheets
I would if they had more horse porn in them.
>>
just starting out, is it a bad idea to make a 21v battery buck boosted to 36v to a brushless motor controller (ECS).
I have a brushless motor (shimano ebike) that has a proprietary ECS that I plan on replacing it with a chinese one, but the 36v BMS's are more expensive than a cheap one I found here (I only need 10A):
https://www.aliexpress.com/item/1005004205819976.html
>>
>>2511915
If it were 5A or below I'd consider it, but it's really better to just use a higher voltage battery. A 10A continuous capable boost converter is almost certainly going to cost more than a BMS, but maybe you can find otherwise. Aliexpress power products can be a bit dubious on the cooling side and quote the datasheet value assuming proper heat-sinking.
>>
>>2504849
>>2504873
>>2505670

Unsolicited update: found a working 5v pwm fan, plugged it in, works fine, problem is a DOA replacement cpu fan, fuck the ebay chink who sold it to me, lol
>>
>>2512184
RIP
>>
>>2511915
fuck BMS's are expensive. I am pretty sure the ones that advertise balancing for $20 are just lying.
but apparently I could just get a dumb battery protector for very cheap for the purpose of charging and protection, and then buy (a somewhat expensive) Active Balancer which will just balance the batteries out (not sure if it would work while charging), so I guess I could just periodically balance the batteries to fix issues with unbalanced batteries.
The question I want to ask is, if batteries in parallel can become unbalanced, why can't batteries in series also become unbalanced?
If I used a active balancer maybe I could manually separate the batteries in series (I haven't thought about how I would do that yet) and balance them out manually, would that make the battery life even better? or do the super expensive BMS's have some sort of solution I am unaware of or it's just not an issue.
>>
>>2512200
>bms are expensive
so are batteries
so you buy one to protect and maintain the other
>>
>>2512202
true but i wonder if the price is worth it.
$200 worth of batteries for a $100 bms.
>>
>>2512200
I mixed up parallel and series.
>>
>>2512200
IIRC the expensive part is balanced charging, which you can offload to an external charger like a B6AC. Often you can just ignore balancing while discharging, especially with monolithic lipos, as the matching ensures no more than a couple of percent imbalance. For discharge balancing circuits, I guess you could have both active (transistors to bypass certain batteries) and passive (cell voltage monitoring with cutoff threshold). Check out what existing BMS circuit diagrams look like and see what you do and don't need. Maybe the $20 have sufficient functionality for you, I think they just cut out when any cell drops below a threshold and don't do any charging stuff at all.

>>2512210
Batteries in parallel can't get unbalanced. The issue with balancing cells is the voltage across a cell could get low enough to damage the internal chemistry. But if they're all in parallel they necessarily all have the same voltage. Because that's what being in parallel means.

You can put a 1Ah cell in parallel with a 0.1Ah cell and they're going to behave just fine, no balancing needed. Unless you get into the case where you draw so much current that the 1Ah cell's ESR is such that the voltage after it is low enough to damage the 0.1Ah cell. But for normal packs (i.e. not mismatched batteries made from a handful of laptops) this is a non-issue.
>>
>>2511915
>10A
nope. You need more. Expect 15A battery current, because I know you're fat ass.
>>2512200
>fuck BMS's are expensive. I am pretty sure the ones that advertise balancing for $20 are just lying.
Check the schematic. If you see resistors or somewhat low resistance and transistor - it has balancing.
Google IC datasheet, maybe it does low-current balancing all the time, like TI BQ76952. There you can have 1k resistors or so and still have battery pack withing 0.03V.
Balancing is dumb simple circuit anyway, TL431 with some transistor is all you need.
They even sell just balancing boards.
Sure, this is lowest tier of balancing, but I don't think it is worth it making an active balancing circuit for <300 W-h for energy.
>The question I want to ask is, if batteries in parallel can become unbalanced,
Nope, but if one battery fails, other gets all the pressure and this cell group would be charging faster than others and discharging faster than others.
>>
>>2512312
I think I will probably go with this for the UART or RS485 so I can have an LCD display or go cheaper with bluetooth + android app, it's 2x more than the cheapest 10s BMS I can find (which was $20), but it seems cool.
I don't know why 10s is so much more expensive, I might consider going with a step up converter.
https://www.aliexpress.com/item/1005004650181011.html
>>
>>2512331
why do you need uart on a bms?
>>
>>2512337
you can connect a LCD display that shows stats, like how well the batteries are balanced.
https://www.aliexpress.com/item/1005004715317036.html
>>
>>2512340
Is that necessary though? Just trust the BMS and/or charger to not kill your battery.
>>
>>2512350
true, but it could be useful for diagnostics (it's like an extra multimeter).
Also I gave the wrong url for the BMS (it was only 4s, not 10s).
https://www.aliexpress.com/item/1005004862569135.html
this BMS also allows me to select from 10-17 series, so a motor upgrade to over 36v may be an option (since I'm doing something sketchy with a shimano motor, it might just not work). and something else about the UART and bluetooth app is that you can configure settings for the BMS apparently.
I do worry it would drain my battery, I probably need a off switch.
>>
>>2512358
>off switch
Make sure it can handle the current. In my case such a thing would be far too large for the arc-breaking capability (120A) so I'll likely use some beefy MOSFETs instead.
>>
>>2511545
Tape it
>>
File: 1661375456356.png (18 KB, 601x152)
18 KB
18 KB PNG
So I bought this surge protector and saw the USB is 2.4amps at 5 volts.
That means it's 12 watts right?
I'm pretty disappointed if that's the case. I was hoping it's be more so I could stop using my usb charger.
>>
>>2512432
Is anything faster than that actually useful? Sure if you're in a hurry and need to charge your phone or tablet up from dead, but that's a pretty rare instance and I think it's fine having a seperate charger lying about for that. Or better still, one to go in the 12V socket of your car. Fast charging is worse on the battery anyhow, and you've got to make those things last considering most devices these days have their battery sealed in.
>>
File: aaa.png (760 KB, 760x745)
760 KB
760 KB PNG
great job texas shitstrumets, where the fuck is pin 1??
i ASSUME is top left of that weird lined stripe, but assumption is pretty gay when it means you might fry your pcb or have to resolder several chips
>>
>>2512557
with the stripe to your left:
top left pin = Vcc
bottom right pin = 0V

Should be, anyway. lmao
>>
File: file.png (22 KB, 346x229)
22 KB
22 KB PNG
>>2512561
>top left pin = Vcc
>bottom right pin = 0V
nope that is completely wrong
i just need to locate pin 1 and this gay as TI chip doesn't have a hole like normal, or at least an engraved line, so i have to ASSUME they mean the printed line to point to pin 1 on top left which is very retarded
>>
>>2512562
>nope that is completely wrong
you forgot the part of the post that said
>with the stripe to your left:
which meant orienting the chip so that the stripe is on the left as you look at it. Then the top left pin is Vcc and the bottom right pin is 0V.
>>
>>2512565
>>with the stripe to your left:
why would you even do that? rotate the chip sideways? standard way is to have the chip rotated like in that schematic, so putting it in a different orientation, like wide side up, to then describe the pins is so completely unusual literally nobody does it which is why your description was understood wrong
standards exist for a reason, what you did was almost as bad as not putting a hole next to pin 1
>>
>>2512567
Doing a little trolling today? Wednesdays suck.
>>
>>2512567
Writing-side-up is constant among all SOICs I've seen.
>>
>>2512389
I think I could probably throw away the balancer, and just use a plain low voltage protection module either DIY or prebuilt (I have to look deeper into the specs, many prebuilt modules only support up to 36v and I need it to go up to 42v because 36v is only nominal).
based on this video, but I may need to modify it for higher voltages (I think I'll stick to just unplugging the battery wires from the protector module).
https://www.youtube.com/watch?v=1Fs4SfVSsLk
And then for charging I would use 10 TP2045 modules based on this video.
https://www.youtube.com/watch?v=ZLLQ9Br4GBI
I might also add in a fuse in case of short circuit, I have also seen people using lightbulbs as a short circuit protector (current limiter), but because there are no 36v bulbs I probably will need to use a step up / down converter to a more common voltage.
And maybe I could look into using zener diodes to prevent powering the wrong direction.
Am I missing anything?
>>
File: 1610147466192.jpg (408 KB, 2997x1687)
408 KB
408 KB JPG
Any professional PCB designers here? I can make my own PCBs, but I'd like to know what standards and certification processes are required to put a PCB inside a product, specially here in Europe.
>>
>>2512718
CE - conformity europe or some shit. thats the regulation you eurocucks have to adhere to if you want to legally sell something. part of that is getting it tested an accredited testing body where they put your shit through the ringer. that costs a few thousand bucks, per evaluation in my country. so if you fail the test, you have to do it again after fixing the issue. if you want to tell shit internationally you have to go through the same process for each countries regulation. FCC, CCC, RCM. for burgerland, chinks and kangaroo fuckers are the most common

good luck
>>
File: 9k=.jpg (457 KB, 1135x1600)
457 KB
457 KB JPG
>>2512716
>a bunch of isolated voltage converters and linear converters
What a mess.

Look at the B6AC circuit to see how they work. It's just one series charging supply, plus parallel shunting transistors to keep things balanced.
>>
>>2513050
I thought the reason why he has 10 converters was because he a 100v DC supply from a transformer, and because transformers aren't regulated, it goes down in voltage for each battery.
I was planning on using an spare ATX PSU and just using the 5v rail which has a 16A on it (is the current regulated?, that would concern me).
I shouldn't have included that video, I didn't even watch it, and it's pretty weird.
I am considering going back to the BMS (which includes a balancer) after I realized the video with the protector with the buzzer only works for 8s (I thought it used the total voltage), and I can't find any equivalent module for 10s.
Another option is I could just use have no voltage discharge protection, and judge the recharge time by voltage display, since this is for a ebike, but protection would be ideal.
also the charger you are offering supports 6s for Li Ion, and it's as expensive as the "smart bms" I posted.
>>
>>2505170
It's ok, there is just m2.5 and m3.5 that are not whole nominal values.
>>
File: Z.jpg (451 KB, 2780x2500)
451 KB
451 KB JPG
>>2513057
Charging is the easy part, because you just need a TL431 and a transistor and a handful of resistors on each cell, that's what those cheap charging BMS boards probably have.

>voltage display
Again, depends on how well-balanced the cells are. If you're charging them up balanced (again, balancing charge boards are piss-cheap and dead-simple) then it shouldn't be too big of an issue, though I may want to raise the minimum voltage by a tad for a safety margin.

The B6AC method I posted is to have a differential amplifier circuit on each cell, feed those into a 4051 1-of-8 multiplexer, and then that goes into and ADC input on the MCU. Personally at that point I'd just have one comparator for each cell too, have the outputs all commoned together to an SR latch, then to a collection of series MOSFETs to gate motor current (assuming you can't turn the ESC off with an LV signal). The advantage of using differential amplifiers like that is the actual voltage at any op-amp input is going to be no more than 1/3 times the cell voltage, due to that voltage divisor ratio. So with 36V op-amps you could get up to 100VDC worth of cells, which is like 24S.

The alternative method is to use a DW01 chip or equivalent on each cell, like pic related. Kinda a pain since you need logic level shifting on each DW01 output (that BJT cascading is funky man) but arguably it's simpler and more reliable. Then that goes to the same array of series MOSFETs, as you can see. The only question I'd have for active use, is would it latch? Latching is probably a requirement, not having that feature likely means that high-current draw would pull the cell voltages low via ESR drop, and when the FETs shut off that current you'd get the cell voltage coming back up again. Best case you'd get oscillation as it turns on and off, possibly not very good for the FETs or the ESC. Worst case and you'd get linear regulation by the FETs, which would kill them in short order.
>>
Who should I believe?
>>
>>2513097
If in doubt, trace the schematic. Pin 1 of an MCU being VCC matches with a few tinyAVRs I've got ATM, though I doubt the chinks would use AVRs over PICs or Padauks. Unless they're copying something like a USBasp or transistor/component tester.
On the other hand, pin 1 being ground matches common LDOs like the AMS1117, so who knows.

I should make one of those component testers myself, really simple design, really neat firmware.
>>
>>2504872
It's probable that just reconstructing the waveform isn't sufficient to get a stable PID loop. I think you should run the algorythm (and the sampling) at 10x the nyquist sampling frequency, so 20x the bandwidth of your feedback signal.
>>
File: cells.jpg (135 KB, 936x503)
135 KB
135 KB JPG
So here's a question regarding li-ion cells that I need some help with anons:
I got a headlamp and it takes 18650 cells. I'm going on vacation tomorrow, so I don't have time to order proper new 18650 cells (which I will do once I am back), but I'd like to take my headlamp with me and use it.
I still have a pile of old salvaged 18650 cells. I don't know quite how old they are or what health they have. I now need to decide between two cells which one I want to take with me:
A red sanyo cell, that is the right one in picrel. When healthy it has a capacity of 2100 mAh. I charged it to 4.021V in the headlamp, over night it dropped voltage to 4.005V.
A pink Samsung cell, the left one in picrel. Healthy capacity would be 2600 mAh. Charged it to 4.056V in the headlamp, over night it dropped voltage to 3.949 V.

So which one would you anons take with you? The Samsung one because it has the higher capacity on paper (when healthy), or the sanyo one, because it didn't drop as much voltage? Just gut feeling is enough, I just want to know which one to take on vacation.
>>
File: merged.png (63 KB, 1112x649)
63 KB
63 KB PNG
WHY
>>
>>2513190
>higher software version
>more bloat
>slower performance
Wirth's law.
>>
>>2513194
Why does it matter when none of those new libs are going on the board?
>>
>>2513190
>ardushitto
>>
>>2513240
Alternatives?
>>
>>2513244
I tried the RPI pico. It seems to be better in every aspect. Even the firmware looks much better, and doesn't want to use the shitty arduino IDE.
>>
File: 1586520783669.png (1.22 MB, 400x800)
1.22 MB
1.22 MB PNG
>>2513244
notepad++
makefiles
>>
File: file.png (89 KB, 404x295)
89 KB
89 KB PNG
I'm going to build a solar tracker and when I took a look at some of the sensors available on the market I noticed that a lot of them use simple LEDs. I didn't even know LEDs could sense light before that, so now I'm curious why one would use these instead of photoresistors? Or even "photodiodes" "phototransistors".
Between all these, what would work best for a simple solar tracker made with an Arduino?
>>
>>2513252
Solar tracking with light seeking is a bad idea. You don't need any sensors because movement of the earth is very predictable. All you need is time and latitude.
>>
>>2513262
Yes I know about that but this is going to be for a moving trailer so it wouldn't be that simple. At first I wanted to apply the same principle and use a gyroscope+accelerometer+magnetic field sensor and some algorithm to determine the perfect angle at a given time, position and direction but I just realized that the algorithm part would be way harder than I originally imagined. I thought there was already some formula/algorithm available for such an equation but it seems like I would have to figure out everything myself and this is way beyond my capabilities.
So the light seeking is my best bet, I think.
>>
>>2513265
>I would have to figure out everything myself and this is way beyond my capabilities
Things are always scary when you start doing something you've never done before. But after a few days you'll start understanding things and it won't be as difficult anymore. Start with simple things, like limiting yourself to one axis at a time and only one variable at a time.
>So the light seeking is my best bet, I think.
This will also involve math, at minimum you will have to learn about some minimization algorithm and possibly plane geometry in 3D depending on the type of your sensor.
>>
>>2513185
there is a possibility you live near a vape store, they might have batteries (hopefully it's not only button top protected & high discharge cells).
generally speaking, you could still use a vape battery that has a similar mAH as a high capacity cell, it might have a 20% less life than a normal high capacity battery due to it being specialized in high currents.
you could also take both just in case one fails if it's possible to replace the batteries easily (but it might be easy to get the polarity wrong in the dark).
>>
>>2513275
>like limiting yourself to one axis at a time
The tracker/frame is only going to be one axis anyway so that's okay.
>But after a few days you'll start understanding things and it won't be as difficult anymore.
For sure anon, but the problem is that going this way would require complex mathematical astronomical equations and I really don't see myself figuring this out.
So far the only hobby type people I've seen doing it this way are these guys
https://www.instructables.com/Solar-Tracking-With-Arduino-Intelligent-System/
https://www.youtube.com/watch?v=dYRKqdo6pAs
And they're so proud of their algorithm that they monetize it and don't publicly show how it's done lol.
+ it requires a computer and I can't use one in my project
Honestly anon, if I give you an azimuth, a position and the time of the day, where would you start with the math/algorithm to get the ideal one-axis tilt angle as a result?

Yes the light seeking way will also require some math but from what I've seen it'll be pretty easy, at least compared to the other option.
>>
>>2513282
https://hackaday.com/2010/05/24/sun-tracking-solar-panel/
https://www.youtube.com/watch?v=d-Rz0nYSkLM
>>
>>2513286
If the frame is to remain stationary, add a limit switch to the end of travel so that when the sun sets, the heliostat resets the panel position for sunrise.
>>
>>2513286
>>2513294
Anon, the two examples you linked are using light sensing to track the sun, not an algorithm. That's my point, it's way easier to do it this way, especially if your solar panel will be mounted on a constantly moving object.
The hackaday example even uses a Redrok Energy sensor module which I've looked up before and noticed it uses LEDs instead of LDRs, which made me wonder why and posted >>2513252 lol.

However thanks for bringing up that "The Colony" show, didn't know about it and it seems very interesting, I'm gonna watch it.
>>
>>2513300
Sorry, should've said "different anon" so you didn't confuse me with the ornery anon before. >The Colony
That show is kinda hacky globohomo shit, but there are some cool bits scattered throughout. They did a second installment with a different group, but it sucked pretty bad.
>>
>>2513305
>kinda hacky globohomo shit
kek how so?
>>
>>2513306
Just watch it and you'll see. The show aired during Obama's first term and they were pushing certain narratives. You'll pick up on it (or you won't because you're indoctrinated and assimilated into the cult).
>>
>>2513310
Nah I definitely will. I'm torrenting it rn, thanks.
>>
>>2513084
Doh, I completely forgot about the controller system. Of course it has discharge protection. I was getting worried about how I would wire up the voltmeter to the steering bar without the possibility of short circuiting the full power of the battery by accident.
I don't really understand what you said about the B6AC method working for 24s, but if you have a URL that sums it up that would be cool.
This is the controller I think I would use (I wonder if the brakes are for anti-theft or not).
https://www.aliexpress.com/item/1005001855335856.html
>>
>>2513279
Actually good advice anon, thanks. I don't think I have a vape store nearby and more importantly don't want to buy a vape battery, I'd much rather buy a proper one, but it's a good idea regardless.
I'll take both I guess and just hope for the best, just gotta hope the airline doesn't complain because "unlabelled battery cell in a bag" is the kind of shit I'd imagine them making an issue over.
>>
File: Solar_declination.svg.png (118 KB, 1920x960)
118 KB
118 KB PNG
>>2513282
>complex mathematical astronomical equations
You don't need much astronomy because all the hard work has already been done for you.
https://en.wikipedia.org/wiki/Position_of_the_Sun
You need to calculate pic related for your latitude at some specific day, then orient your panel to the north and tell motors to follow angles.
>>
>>2513364
>to the north
where do you live?
>>
>>2513371
I should have said to orient the panel in default reference position so that motors can operate relative from that point assuming 360° movement capability.
>>
>>2510051
Seconding this. I love being able to solder pcb's with multiple wires with mine.
>>
How can I determine what mosfet will work to replace one that's too small?
I have some shit that has one which switches on a 12v component but it just isn't liking high amps. It gets up to 0.7v droop before the shit shuts off. It's only rated for less than it needs to so there's that.
My plan is to Desolder it and mount a much larger mosfet somewhere remotely but I'm not sure if it's going to be as simple as pick one that does more amps
>>
>>2513792
Slap a heatsink on it first.
>>
>>2513792
What FET do you need to replace? How much current do you need to pass? What are you switching it with?
>>
>>2513810
>>2513810
>>2513810
NEW THREAD
>>
>>2512980
I see. I've been doing some research on the topic, but regulations here in europe are a mess. In my country there's an organization that makes the norms you have to follow, which are basically a copy of european normative. It's a state sponsored scam, it's illegal to sell public legal documents, but they just excuse themselves by saying they are selling a "recommendation", not an actual law, but if you don't follow those rules you'll be in a world of pain to get anything done.



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.