/diy/ - Do It Yourself

Hi, I'd like to measure the power that is consumed by a load connected to AC mains. I was thinking about using a current transformer to measure load current and line voltage (line voltage measured with a resistor in series with current transformer) and an ADC connected to a microcontroller. The issue I came across is that the line voltage is mostly a sinusoidal wave, while the load current can be really fucked up. How would I go to measure the power when the current has all sorts of frequencies, including some high frequency in the case of some loads such as switch mode power supplies. In other words, how would I design a reliable current sensor that can handle such loads and measure the real power used by the load? Bonus points if it can also measure reactive power.

FROM THE PREVIOUS THREAD:
>>2526860
>Gotta multiply instantaneous current by instantaneous voltage, and make some sort of weighted running average. Probably using DSP, since analogue multiplication is hard. Also consider a transformer or isolated ADC+divider for mains voltage, to prevent the majority of your digital circuitry from being mains referenced. A differential amplifier measuring live-neutral and referencing it to ground also works. Naturally you'll need anti-aliasing filters to obey the nyquist sampling theorem on both ADC inputs. Figuring out the power factor and reactive and apparent powers would just be a software problem.

>As for the sensor itself, hall or shunt would be my desire, as current draws aren't necessarily symmetrical around 0V (see half-wave rectifiers

Can you go in more detail about the following?
>you'll need anti-aliasing filters to obey the nyquist sampling theorem on both ADC inputs
If I am only interested in power (and not recreating the signal), is the nyquist rate important at all? Can't I just sample without an anti-aliasing filter, then multiply voltage*current? Furthermore, if I'm going with a current transformer, how do I account for phase shift?

Reposting request from previous thread: >>2523575

tl;dr I'm looking for good books/manuals on SMPS design and on transformer design and manufacturing.

>>2526929
>Can't I just sample without an anti-aliasing filter, then multiply voltage*current
No, because any frequencies above the sample frequency divided by two are going to show up as aliases at lower frequencies. Worst case you've got an SMPS putting out high frequency current noise so you just get a bunch of aliases wildly throwing off the power factor calculations. At least dampening them with a low pass filter doesn't give false frequency spikes. Arguably you might be fine with a pretty low frequency for your ADC, since parallel caps and series chokes are everywhere, though it depends on the purpose of your power measurement. If you're trying to guess your power bill, then I'd want to use at least as high a rate as a common smart-meter.
Some delta-sigma ADCs come with internal anti-aliasing filters anyhow, and just need a simple RC before them.

>how do I account for phase shift
Phase shift due to the transformer itself? I'd assume it to be negligible, but I'm not familiar with their operation. Watch out for frequency-dependant action with a current transformer, you may find that high-frequencies get through at a different amplitude than the 50/60Hz baseline.

>>2526934
If you're wanting to get into custom transformers for isolated SMPSs, you'll need the equipment to characterise them. Not just their turns ratio, but their voltage ratio as a function of frequency, their saturation characteristic, the coupling coefficient, etc. Getting cores with air gaps is also kinda difficult looking, so personally I'd stick to forward converters instead of trying to make a flyback transformer. Well actually I'd recommend just buying transformers from suppliers like digi-key, but if you're willing to make a coil-winding machine or are autistic enough to do it by hand I won't stop you.

No clue about literature on the topic, but chances are TI and Maxim and such have some good appnotes to look through on high-power SMPS design.

Also why not a van de graff?

>>2526476
>Pick an LDO
i don't think there any any ldos that work with a load that draws only couple micro amps, while at same time the ldo using like 1 micro amp to power it self
the voltage reference as power source is an interesting idea though

>>2526943
>No, because any frequencies above the sample frequency divided by two are going to show up as aliases at lower frequencies
Oh true, I forgot about that. Thanks for explaining it clearly.
>Some delta-sigma ADCs come with internal anti-aliasing filters anyhow, and just need a simple RC before them.
I'm not very familiar with ADCs so I'm not sure if the ADC I'm using has a built in AAF or not, probably should look in the datasheet, but if it doesn't, what type of filter should I use? Is an RC filter enough or do I need some sort of active filter? In the latter, any advice in what I should look at? Maybe an OP amp fitler? Preferably keeping cost minimal, so no ICs worth 30-50$.
> If you're trying to guess your power bill, then I'd want to use at least as high a rate as a common smart-meter.
Smart, thanks for the idea.

is this correct?

is this correct?

>>2526956
almost, I found R2/R1 not R2/R3

>>2526947
Have a look at the ADP160. Rated at like 560nA.

>>2526954
>but if it doesn't, what type of filter should I use?
Depends on your desired SNR and bandwidth, and the amplitude of the expected high frequency noise. Ultimately you want the attenuation at and above your Nyquist frequency to be such that any signal there gets stamped into nothingness. If you want 8-bits of resolution then the attenuation will be a lot less than if you want a full 24 bits of resolution. If you increase the Nyquist frequency significantly past your desired bandwidth, then you'll have more room for attenuation. Say if your desired bandwidth is 4kHz, and you want 60dB of attenuation, then a 5kHz nyquist frequency means 60dB/0.1dec = 600dB/decade (30 pole filter lmao). If you set the nyquist frequency to 100kHz then that means 60dB/1.4dec = 43dB/decade (2 pole filter). N bits of effective resolution means 20*log(2^N)dB of SNR (8bit = 48dB), but there's nothing stopping you from choosing a desired effective resolution lower than your ADC's resolution, which reduces the required filter attenuation.

t. retard who's never built an anti-aliasing filter before

>>2526957
>>2526962
gratz

>>2526957
oh you changed it, yes that's what I found as well.

>>2526963
>>2526964
yeah sorry I meant R2/R1, have it on paper and copied to mspaint incorrectly
thanks for checking

>>2526954
>do I need some sort of active filter
Usually a good idea if you're after more than 1 pole (20dB/decade) of attenuation. Sallen-keys are my go-to, but there's elliptic and other strange topologies that can give much steeper rolloff out there you may want to look into. I've heard stacking a bunch of filters together can negatively influence stability, and all those stacking phase shifts can be kinda funky. Honestly anything other than a sallen-key, igmf, or active twin-t is wizard shit to me.

>>2526963
>ADP160
>Maximum output current: 150 mA
damn thats a LOT way more than i expected, i thought those can do like few ma at most.
So i was thinking about hooking the reference to power the micro and then all the other stuff on the board i will connect to normal LDO that is sitting behing a fet and the micro will turn the fet on when it needs to get high current to power everything
but fuck if the reference can handle 150ma then i can basically just hook everything up to it directly

>>2526963
you know how lm317 and ams1117 are basically thes famous linear regulators, what is the voltage reference variants of that? i am going to be ordering from chinks so i need something that is mass used so it's likely chinks have it and it's not a fake

need a 3.3v nmosfet for switching from microcontroller. Must be through hole for breadboard applications (no SMD). Must also be cheap (i.e. buyable from aliexpress for pennies) and expect to conduct currents up to 500 mA or 1A preferably. To-92 package style (or to-220) Any recommendations?

I want to do current measurement for a battery in both directions. A low side shunt with a differential amplifier can do this. I'm going to use a rail to rail opamp
Do i need to worry about the ground being different when the battery is charging. Could the possible negative voltage (with respect to battery negative aka ground) hurt the opamp input?

>>2526979
You can use a normal MOSFET and implement a charge pump to get a higher voltage from 3.3V
This may or may not work with high frequency switching though

>>2526982
i'm not familiar with charge pumps, Do you mean some kind of mini boost converter or a voltage doubler? Any chance to get a small circuit diagram? Like assume the i/o pin outputs 3.3.V that drives a mosflet load.

>>2526980
save yourself all the bullshit and do a high side monitor. using any rail to rail opamp in a differential config should work. theres even purpose built current sensing opamps with high gain built in so you can use really small value resistors which wont affect your positive rail in a meaningful way

>>2527001
If the shunt is on the high side, I would have the same problem while charging. The only difference would be that instead of negative voltages I would be seeing higher voltages than the battery (which is the opamp positive rail)

>>2527011
your using a differential setup youre seein voltage across a shunt, not the voltage of the rail. and you should put everything after your shunt. use like a 0.1R or less resistor and you legit wont see a difference thats meaningful. ive never had a problem doing this

>>2526866
Can anyone recommend an online class that will teach me enough about electronics to design a circuit with an arm chip in it and a solar panel and battery?

>A SSR with back to back thyristors
Does it matter where you connect the power supply and load on the output?

>>2527043
>online class
Use your search engine of choice to find projects that others have already completed, and tweak their designs to meet your requirements. Unless you actually like wasting time and losing money.

>>2526866
really dumb question, but how do i make the gate voltage fall to 0 V minimum rather than ~0.5 V minimum? and why is this happening?

>>2527049
Money is not an issue. I want to take a class to accelerate learning in less time.

>>2526943
>Also why not a van de graff?
Because I wouldn't learn anything about SMPS and transformers that way, which is also part of the reason, beside safety, I don't chain a DC PSU to a ZVS circuit to a rewired MOT to a voltage multiplier. And because it's not compact enough to be practical. Also power efficiency.

>>2527073
Engineering is about self-learning. You can't buy yourself a way out of this one. Just download books from libgen and start studying like one of those south korean kids trying to get a job at samsung.

>>2527060
>why is this happening?
The collector-emitter voltage of Q1 cannot fall below 0,5V (in this particular case). Khirchoff's law for the gate circuit would be: Vgs=Vr2 + Vce(of Q1).
If you want a lower collector-emitter saturation voltage then use a S8550 or another PNP transistor with a very low collector-emitter saturation voltage. There will still be some voltage drop (from the current trough R2 and like 0,1-0,2V across CE of Q1). If you really need 0V then use a BS170/BS250 combination for the push-pull driver you have, the MOSFET's are small signal small current (200mA max) but their minimum channel resistance is around 5 ohms so this should get you closer to 0V, using lower resistance MOSFET's is even better.

>>2527108
Also use a MOSFET for Q3, it's collector-emitter voltage will only fall a bit below 0,2V, the maximum collector current would be 12mA (1k resistor and 12V supply) so this means you should give Q3 1mA (or more) into the base to make it's collector-emitter voltage fall as much as possible. But if you want to get it even lower use a BS170 (drain-source resistance of 5ohms, which forms a resistive divider with the 1k) which will give you a 0,04 drain-source voltage (1k/1k+0,005k x 12= 11,96V => 0,04 drain source voltage).

>>2526987
You could use a BS170 and put multiple of them in parallel, their max current is 200mA and their drain-source resistance is 5ohms, a low resistance (0,1ohm) source resistor would help equalise the current between them.

Anon from the misplaced post on /g/ here (yes I am a retard). Can a 555 timer and a high-voltage H-bridge create a PWM dimmer for a neon bulb? Is this even possible for gas discharge bulbs?

>>2527136
>BS170
From the looks of it, it's very similar to the 2n7000 which i have a butt loads of. I suppose I can try to test it out and see if I can use the 2n7000, I thought 3.3v might be too low for those ones.

>>2526178
>Pic related, though in your case the resistor on the far left would already exist in your circuit. Just the main square of three resistors, the TL431, and the PNP transistor will replace the zener. Would also give a much more stable output value.
Alright, thank you. I'll try doing this whenever I have the time once the holidays are over.

>>2526968
No it's an LDO, not a reference.

>>2526970
TL431. Though the chink ones are often TL432s.

>>2526979
In SMT there's the AO3400/3401. Maybe you can find something similar in TO-92, but probably just go for a breakout board. Logic-level-FETs are kinda sparse in THT packages, besides relics like the 2N7000.
>aliexpress
If it's obscure you may end up with fakes, not worth the risk imo.

>>2526980
Literally a non-issue if you use a ground rail for the differential amplifier that's halfway between your low voltage rails. Say, 2.5V if you're powering the op-amp off 5V. Read this: >>2523562

>>2527142
It's possible, but a waste of parts. A series 1M potentiometer would probably work fine without overheating it. They run at 1mA when striking and less while running. Just ensure there's a minimum resistance.

Starting to get into soldering and filling out the cart. Is there anything else I should consider?

>>2527215
A solder pump. The braids will sooner suck your soul out than any solder.

>>2527215
I like tin sucker over braids
fume extractor is a meme, waste of money for hobby soldering
get also rosin/flux core solder, 60/40 instead of pure tin, it goes on lower temps and you might avoid overheating some components, pure tin is a meme for greenfags
37e for fucking flux ? you are doing drugs ? get some hard rosin rock kilo for a dime, syringe is a meme unless you do some puny SMD
rather invest into good solder station with replaceable tips and good selfserviceability
34e for wipes ? wtf ? you dont have wet rag ? put some paper/thin plywood where you solder and you good, no need to clean it
wtf is even tip thinner ?
good cutters are indeed a must, maybe put saved money there
having third hand is sometimes good, but one can live without it
liter of isopropyl alcohol might also come in handy
angled tweezers and angled pliers (picrel) are handy sometimes

protip you can make your own flux syringe by dissolving some rock rosin in isopropyl

>>2527215
I'll get the sucker over that then, they don't seem that expensive
>>2527278
Can't buy lead solder in Europe so I'm stuck with lead free solder. If I want to get resin core SAC305 then it's like 90 Euro and I can only seem to buy 500g rolls of it. My options are very limited with solder wire it seems, I've spent the last 5 days looking at this and I just want to buy the stuff and be done with it.
Idk about the flux, Just saw people saying the pen is easier to work with and that's what was recommended.
I actually have like no rags so just buying a shit ton of them doesnt seem too bad
Pretty sure its just a misspelling in that screen, it's just tinner
Already have helping hands so that's good
snips/pliers and isopropyl alcohol I'll just get in local shops then
Will have to try the diy syringe, seems pretty easy to make then
I'd like to mention I got an ADS200 for free so I dont think I can get that much better of a station and is definitely overkill for me starting off

>>2527278
I'll get the sucker over that then, they don't seem that expensive
>>2527278
Can't buy lead solder in Europe so I'm stuck with lead free solder. If I want to get resin core SAC305 then it's like 90 Euro and I can only seem to buy 500g rolls of it. My options are very limited with solder wire it seems, I've spent the last 5 days looking at this and I just want to buy the stuff and be done with it.
Idk about the flux, Just saw people saying the pen is easier to work with and that's what was recommended.
I actually have like no rags so just buying a shit ton of them doesnt seem too bad
Pretty sure its just a misspelling in that screen, it's just tinner
Already have helping hands so that's good
snips/pliers and isopropyl alcohol I'll just get in local shops then
Will have to try the diy syringe, seems pretty easy to make then
I'd like to mention I got an ADS200 for free so I dont think I can get that much better of a station and is definitely overkill for me starting off

>>2527301
ah right, tinning, never done that much, when tip is dirty a bake it for a bit in rock rosin, the acids eat the impurities and then it's fine
you should be able to also tin it just by melting normal solder on it, no science really, might be good if you intend to store tip for longer periods of time, to prevent oxidation

you are right about the lead solder, even shop where I used to buy it all the time is empty now, but there is plenty others, usually small eshops selling old batches, maybe look into other countries in EU, might be worth the delivery, I found some 200g rosin core for 20e with bit of search. I need to stock up for two lifetimes at least before tinjews rob me blind.

for the fluxes, its more about what you want to do, Pen/Syringe might be better for SMD/Hot air/Baking for precision but not so much for prototype drilled boards, but you can manage with anything you have really, you will get touch of it after few hours of fume sessions

>>2527215
Isn't that a bit much? When I started (and I'm still on a beginner level), I only had a very basic iron and leaded solder.

>nanodla crashes pulseview on mac
is it strange that running obscure software is the reason i bought a raspi?

is paralleling FETs as simple as pic related? want to build a test device for measuring inductor saturation. realistically it's going to be used at <10 amps but i'd like to build it to hypothetically go up to 50 amps or so.

>>2527381
Yeah nothing wrong with that. Consider adding a diode (and series resistor) across the 10Ω resistor to have a faster turn-off time than turn-on time.

Not sure if this is the right place but does anyone have experience making pcbs with 0.5mm bga using jlcpcb? I'm trying to route an nrf52840 but I can't make sense of jlc's capabilities page. Is it possible to have traces or vias like in my picture using jlc's capabilities for a 2 layer board? What about using 4 layer board capabilities?

Datasheet for nrf52840: https://infocenter.nordicsemi.com/pdf/nRF52840_PS_v1.1.pdf

>>2527412
thanks, forgot about that trick.

IC brand tube power supply caps looked bulgy so I took 'em out. Turns out they all test okay in capacitance, don't have an ESR meter though. 22uf 500V caps from 1991 and they may have still been okay.

Already bought new replacement caps anyway though. Guess that was $15 I didn't need to spend.

>>2527215
Get a solder pump, wicks suck ass, they more often end up getting soldered to the board and removing solder, unless you douse it in 10 gallons of flux. Also for your solder wire, make sure it has flux in it. And make sure not to get a solder that is too thick. Get your cotton swabs from a regular store to save money. Finally, you might want to get some prototype soldering boards.

>>2527525
I have 33uf mains rated caps branded with "WEST GERMANY" that still work

>>2527338
Wasn't too sure what was the way to do tip maintenance but those seem like solid ideas too
If you're talking about lead solder with resin core then yeah I guess you can buy that easily as a business. It's just the lead-free solder with rosin cores are either sold in 500g rolls and cost ~90 euro or I have to buy in bulk. If you're talking about lead free idk what sites youre looking at, haven't seen anything that cheap and I went to like page 50 of the search results.
Not sure what I want to do just yet, no harm is having various fluxes really
>>2527339
I guess so, I go a bit overboard on things and like to have all the essentials I may need. Buy it all from the get go instead of figuring out that I actually need xyz at various periods and waiting to get them delivered
>>2527543
Well solder pump it is
As stated above, having trouble trying to get rosin core solder that isnt like 100 euro per roll. Does brand really matter or are they all pretty much the same?
Local stores don't have wooden cotton swabs, only plastic ones which idk if that really matters, just saw mentioned that it's better to get wood ones.

Thanks for the replies, seems like I'm not missing anything really and just replacing current stuff

>>2527692
>Does brand really matter or are they all pretty much the same?
Honestly, it's more about the materials the solder is made of. Lead is lead. Different brands won't change that. Solder with rosin is solder with rosin, very little differences across brands. Sure some might be of cheaper q2uality (for instance, no rosin core, or different alloy than lead), but as long as it says lead solder with rosin core, you should be fine. I know it's hard to find lead solder in europe, maybe try to get it online? I can tell you right away that lead is a must or else you'll have a hard time soldering stuff. The reason has to do with temperature, the non-lead crap requires usually a higher temp which is hard for low end soldering iron to achieve on top of increasing the risks of burning your components.

>>2527154
The datasheet says it's minimum GateSource threshold voltage is 1V and the typical is 2,1V. 3.3V should be good.

>>2527215
Ignore the fume extractor, get a 12V PC fan instead. Also get some 0,8mm solder for the thicker leads or joints.

>>2527813
And get a tin of colophony, always dip your soldering iron's tip into it before soldering.

>>2527800
>I know it's hard to find lead solder in europe
Idk here in Romania I can get it from any local electronics store or online with no problem. The only problem is for businesses because of the enviromental regulations.

>>2527142
Would be better to use a triac dimmer circuit than a 555 and H bridge. Or you could use a 100k potentiometer and put it in series with the neon bulb and a 100k 0,5W resistor.

>>2527800
>europe
You can order it online. I prefer lead free, because it works well, even if I probably need a higher temperature. Also you can't mix leaded and lead free solder.

>>2527692
>Well solder pump it is

dont do it.
de-soldering can be a big hassle, so you need all kinds of strategies for success.
so you def need both pump and wick.
and sometimes you'll need hollow needles (and regular stainless needles) of diff sizes.
occasionally you'll need more heat than your iron can generate, so you need two of them, or hot air.
e.g. to replace a usb port, a pump wont work, wick will kinda work, and hot air will def work.

I have a question regarding TOF range measurements.
Nowerdays optical TOF range measurement is cheap enogh for consumer devices and achieves at least cm level accuracy.
Could the same be done with RF and an active target to be used as local positioning system for robots?
>Device 1 sends out a signal
>Device 2 recieves the signal and sends a response
>Device 1 recieves the response and can calculate the distance to device 2
Clock drift between the devices shouldnt matter to much since all device 2 does is sending out a response once it gets the signal.

There must be a problem with my thinking since any existing system I can find gets down to meters at best. Why can't it achieve the same resolution as optical TOF?

>>2527824
you sound like someone that only does solder blobs on a perfboard

>>2527800
>https://eu.mouser.com/c/tools-supplies/soldering/solder/?description%2Ffunction=Leaded%20solder&product=Solder
>hard

>>2527820
100k pot might be too small to get it dim enough

>>2527841
don't forget desoldering shims

>>2527858
I think it should be doable to get 10cm accuracy if the receiver retransmits at a different frequency. No identifying data contained in the radio ping, as that would just make it take longer to ensure that you've got the right one, you want minimal clock cycles between device 2 receiving and transmitting such that clock error doesn't make the delay significant. Might even be able to do it without an MCU in there, but whether the resultant delay would be more or less repeatable isn't something I could tell you.

Key IC is a time-to-digital converter.

>>2527858
>RF MHz to GHz range
>light is at 500 THz
Translate that to wavelength and you'll get your answer. When your devices are only a couple of wavelengths apart your resolution will be low. 1 cm is 30 GHz, and to achieve cm level resolution something like 70-100 GHz is required. That's what automotive radars use and get ~5 cm resolution. You can't cheat physics here.

>>2527908
Why is wavelength important when I'm looking at propagation time?

>>2527915
It still determines your temporal resolution. If let's say you send a (sine) pulse with very long wavelength - where does it start? If you want to measure at cm resolutions this means you need to sample everything at very high frequency. And even if you manage to sample that fast you won't be able to accurately determine when does the pulse start, because it will rise very slowly compared to how fast you need to sample to get 1 cm resolution. Add a bit of noise to the mix and it will get very difficult to determine start of the signal.

>>2527946
Well 5GHz is 6cm wavelength, isn't that good enough?

>>2527965
Which you need to sample at at least 10 GHz, most likely multiples of that if you want to get anywhere close to 6 cm. This is not a hobby grade project and would most likely require custom asic hardware.

Would this work?
I'm trying to make a device that can connect over usb to a computer or to other devices over I2C using usb-c ports. My thinking is that in order to use usb and I2C in the same usb-c ports I need to use muxes to choose whether I want to connect to the usb or I2C pins on my microcontroller. And I can detect if the device is connected to a computer or I2C device by reading the CC pins to switch the muxes accordingly. High CC pins meaning a host computer is connected and low CC pins meaning an I2C device is connected.

>>2527980
>sample
Why would you do it digitally? Just use an RF transistor after the band-pass filter to detect it and trigger the time-to-digital converter on device 1, or to trigger the second transmitter in the case of device 2.

>>2528001
I think it would probably work, since both USB and I2C are bidirectional standards. I'd just ensure there's pullup resistors on both sides of the link.

>>2528012
>pullup resistors on both sides of the link
1.5k pullups on both sides for DBUS+ right?

im using this relay with the arduino to toggle this mini water pump but the pump doesnt toggle.
https://www.amazon.ca/Gikfun-Submersible-Fountain-Aquarium-EK1893C/dp/B07BKXJXK1/ref=sr_1_15?crid=2ZZHJG6UK080R&keywords=water+pump+5v&qid=1671850150&sprefix=water+pump+5v%2Caps%2C218&sr=8-15
for some reason, the pump is constantly on even though im toggling the relay input pin every 2 seconds. I can see the red led light toggle but the pump is on the whole time.

>>2528133
the connection looks like pic rel but the power supply is
https://www.amazon.ca/Breadboard-Supply-Module-Voltage-6-5-12V/dp/B08W1KSWY9/ref=sr_1_9?crid=2ZJ40MY3MTDTJ&keywords=power+breadboard&qid=1671850449&sprefix=power+breadboard%2Caps%2C229&sr=8-9
with the power set to 5 volts.

i'm slightly drunk and observing that all the cheap sot23-6 buck switchers want an electrolytic output cap (>4u7). this is unacceptable. are there any examples to the contrary? i only need 12->3v3 at 300mA for an ESP.

>>2528059
I meant on the I2C pins on either side of the muxes. So on U4's pins 3 and 4, and on U?'s pins 8 and 7.

>>2528133
Did it work before? Take the pump away, can you hear the relay clicking when you touch the input pin to +5V or GND? It's possible that an inductive spike from the pump arced in the relay and melted something under tension, or spot-welded the contacts together.

>>2528140
If its ESR that can't be too low, then a tantalum cap should be fine. Otherwise, a ceramic with ~1Ω series resistor should be fine.

>>2528145
never used a tantalum and NEVER will. it's not the ESR, it's needing a bunch of 4u7 ceramics to hit the 22uF+ they think they need for stability.

>>2528146
Just use ceramics then. They go up past 10µF easily.

So what is the best practice to program microcontrollers on my pcb if I have more than one?

>>2528163
You mean communication between MCUs? Get some practice writing to the registers of communication hardware (SPI, UART, I2C, whatever makes most sense for you), especially how they interact with interrupts or "has been written to" flags. It's also an option to just present a flag or two by toggling single pins (e.g. an abort line) between MCUs, or to output whole words by writing to an entire I/O output port at once. Arguably faster than SPI, depending on what you're doing, unlikely to ever be useful.

Beginner here. Given a cat5 cable that carries signals on pins 1, 2, 7, and 8 (as in pic related), is it reasonable to use n-channel MOSFETs connected to a common MCU pin to activate/ deactivate the bridge between two RJ45 jacks? Does this idea work/ make sense? If not, what would you recommend to control this communication bridge between RJ1 and RJ2?

>>2528182
Not communicating but programming. If I have one I just make a serial/USB connection on a pcb to connect to my pc. What if I have several MCUs? Do I make a programmer for each one on pcb? It's kinda stupid, no?

are the LEDs on this kind of CoB LED chip in series or parallel (or is it chip dependent)

>>2528192
Not that anon, but you should provision the use of a TAG Connect cable on your prototypes for ICSP - one footprint per MCU. That way, you need only connect the right programmer to the right footprint and upload your binary. You can also route debug UART traces through your TAG cable too. This is just how I do it.

>>2528192
>a serial/USB connection
Use an ICSP header for each MCU, and have an external USB programmer. For AVRs that would be a USBasp clone, or a serialUPDI setup. For STM32s it would be an STlink v2. Direct USB programming can be useful for a dev-board, but ultimately a waste of space on the PCB, and also a waste of flash memory if it needs a bootloader.

>>2528197
Depends. You can usually figure out how much is in series by the rated voltage. You can then see how many LEDs total by putting a really low current through them and counting the bright points.

>>2528205
>not making your programming pads as non standard and non user friendly as possible
Now I'm gonna go coat my boards in conform to keep the plebs away

>>2528187
bumping. anyone?

>>2528206
>>2528206
Are you fucking kidding me? Im not spending 60€ for a cable wtf.

>>2528221
Oh wait, it's not just 60€. It's 60€ per MCU manufacturer.

>>2528221
If space and cost don't matter, just put a 2x3 0.1" header on the board like a normal person.

>>2528211
why cant it be active all the time? seems like it could just be crossover setup. what are you attempting to transmit over the cable?

>>2528187
Gotta put two of them antiseries, because you know there's an integral diode in each FET. And even then you need to ensure that the gate voltage is high enough compared to the source voltage even when the source is at its most positive, so you may need a voltage converter (beware noise) or a floating gate driver. Lastly you may find that the junction capacitance to be problematic for fast signals. Especially so for the body diode of the FET, but at least you can put a schottky diode across each FET to solve that.

Also I'm not sure about how ethernet signals look, but if they go below 0V, you need to ensure that your gate voltage can follow them.

Altium Vs KiCAD vs EasyEDA for a beginner?
Price doesn't matter, I'll pirate anything no problem.

>>2528248
EasyEDA is kinda shit. I use KiCAD and it's pretty good, there's a fair bit of community support for it and some nice addons. I think Altium has a wider suite of features, but they're not really features a beginner would find much use in. I say Altium is worth considering if you want to get employed by some company doing circuit design, while KiCAD if you're more likely to make your own business or join a small business that can't afford Altium. It's the small companies that the corpo lawyers target after all. If neither then I'd try both and see what you like better, or at least watch (up to date) tutorials of both.

Side note, EasyEDA might be worth dabbling with if you want to try JLCPCB's assembly service, I hear it's a bit finnicky to get it working with KiCAD.

el wire makes everything better

>>2528187
>>2528211
>what would you recommend

a relay (or two) of course.
coz it doesnt mess with the signal in any way.
(well, it untwists part of the signal path, but any solution will do that -- not really critical on CAT5 )
BTW, the normal ethernet signal lines are not 1,2,7,8 but 1,2,3,6.

>>2528012
>Just use an RF transistor after the band-pass filter to detect it and trigger
If only it were that simple... If you want precision you'll just have to do it digitally. Analog is too unpredictable. You can get around having to sample at GHz by using clever modulation schemes and mixing down to MHz to then use digital processing to extract range data. But again, once you hit GHz frequencies, you can forget about lumped element models because everything behaves in unconventional ways. I wouldn't even attempt to make something like this, because its best left to people who have access to MMIC chip making capabilities. Instead of dreaming about RF why not make it with ultrasound? It makes everything 100000 times easier because speed of sound is slow. Then you can use transistors and breadboards to trigger timers no problem.

>>2528261
>sending GHz impedance matched signals through a mechanical relay
ishygddt

>>2528268
Basic “relay” idea for switching high frequency stuff is not uncommon. Cheap, but can be bulky. Check out TSP #92.

Also you can get high speed HDMI mux/switching chips out of discarded/thrift store TVs and whatnot.

Circuits to measure

>>2528267
Original question asker here, I am aware of the ultrasound solution. My question was less practical and more aimed at broadening my understanding. That has been sufficienty answered. Thanks.

>>2528288
Marked as solved. Locked.

>>2528289
Stop necrobumping old threads.
________
I7 6800k @ 4.20 GHz | RTX2080 | Windows 10 LTSC

“Two things are infinite: the universe and human stupidity; and I'm not sure about the universe.” - A. Einstein

>>2528295
>total posts: 1

Merry Christmas to my Western friends
Here's a small adjustable 555 boost converter as gift
https://falstad.com/circuit/circuitjs.html?ctz=CQAgjCAMB0l3BWK0BsAmBAWA7HAnAmHgMwrHHiYhKYUICmAtGGAFBgo0AcIakPGFL34ghkVgCVwYNOGxCWsxVBWZI1FepgJWAFxC1ZgkMQRGEY8MjCQbGRWjx4UmJ1RhlMabDmwJbkHjY2OBwIAAm9ABmAIYArgA2uqwATiZmvGrpSmg86jZwktngucVcWqrqSBXaqQbE5kJmAhYqaHysAObF3rKmfXyarADuPVlg8plaI3JCvbyt8+KjxvPGhlAz6w3UMvWy4lG7shsIeCc7O2DQ1bBoRa5CpDwbz5oGVe+1CQbOJig8M59AGaWBwNBEJxQ6HQwHQVxcbDEPDlPBgTCYFgMRh5GZA-6A84GDibUb4jGEk5-cTdcl-fHNIbhahEt4AnhvWSRWKJZIAY1EXA5IPZBkweWQ8FsomgpjwkGIXEcQXQ4ooMDgbG6oopgsBpWmaS4LgJIGNVEeKgKNLNJst5uoBs2AuU4vyeze+SYPEYxFl5ADgaDktsXWkJ3K4eOQ1Grsj6ItkeWlET+QxpvEAuYpTU6mzgL2+RuTBsIYR5QaXHR2DAQq4eBDWpA+ejLaw7k2aRbuebMgE2AOoXgdRb8xbHAq1vEUnHKDzfZA1MqGhqNxmzCekETJi3pIMHNIvaeh+TG5Mh7PPdPJJ73bC183VDP5DEM3FU2b77QWVPX6yjC-Ac9wA-tZFHIDxD4EJGWMOcWiEKgADEACF2B8M0wkZeQXhNCAp2QBB6ynKV4AgCR6AAZwASwo3QYgAOz5egZnKdQ3gdT0WLCNkhQzLiqlKVjqHTG0hPbDCqkmaYoJAbDfiEAd1EtJDUKNMJMBNITPXPYcAHsDCUSonGESUMDaAyTFYIA

>>2528145
Yeah I can hear the clicking sound every two seconds.
>It's possible that an inductive spike from the pump arced in the relay and melted something under tension, or spot-welded the contacts together
hmm should I try another relay then?

Why do you need motor controller if you have a microcontroller on board which can do the job?

>>2528443
Software is buggy and hard to get right. Unless you're developing a commercial solution cost savings aren't worth the pain of debugging.

>>2528454
Isn't it extra fail point?

Can someone help me understand this datasheet? What do the current leakage values actually mean? What are they plus or minus relative to?

>>2528495
It means that value will be somewhere between -1 and 1. It can be 0, it can be -0.9 it can be 0.2... Leakage means current that will flow into control inputs - it's like having a current source of 1uA at that pin that constantly tries to push/pull current. You need to consider this if you're dealing with big input resistances connected to that pin. Even if it's only 1uA it can generate errors because current has to flow. At 100k for example you would get 0.1V of extra input because of leakage.

Why this is bad for a hobbyist, who is not planning to work as an electrical engineer?
https://youtu.be/H_FFytMCXaA

>>2528389
Then chances are the relay is fine but the load is wired wrong (or faulty). Hook it up without the relay to see if it comes on. And use a freewheel diode across the pump.

Hi everyone, I know little to nothing about electronics but I'm trying to get into it. Here I have a very simple circuit; when you press 1 (see pic) on the trace it makes the led light up and it makes component 5 play a tune. 3 and 4 are the batteries.

My questions are:
- why does it need two batteries? Was it just the manufacturers choice because they thought that for whatever reason they could save money this way or is there another reason?
-What is the blob on 2?
-what is 5 exactly? An active buzzer?
-how does it know what notes to play in the little tune? They must be hardcoded somewhere, right? Is that in the buzzer or in the circuit? could the note sequence be changed using an Arduino or something?
-is it also possible to remove the buzzer that is there right now and place my own buzzer that plays whatever note sequence I tell it to?
-assuming I had a multimeter, where would I need to place the pins to know what's going on in the circuit?

Sorry if they're dumb questions and I'm not sure if they all make sense. Thanks anyway

>>2528606
>why does it need two batteries
Looks like they're low-voltage (1.5V) alkaline cells, like LR44s. 1.5V is kinda low for a lot of electronics, so I can see why putting two in series might be a desirable option. That said, some low-power controllers are definitely able to run at such low voltages, and there's also the option of using a 3V lithium coin cell instead. I imagine it's just a cost thing.
>What is the blob
Silicon chip stuck to the board with thin bond-wires from it to the surrounding traces, and a blob of thermoset resin over it to protect the silicon and bond-wires. It's a cheaper method than soldering a plastic-encapsulated chip for very large quantities.
>what is 5 exactly
Looks like a piezo disc. To play sound you need to send them an AC waveform to them. Electrically they behave like a capacitor, the disc is deformed by applying a voltage across it. In the case of this circuit, one side is pulled to 3V and the other to 0V, then the other way around, to get a square wave with a peak-to-peak amplitude of 6V.
>how does it know what notes to play in the little tune
The controller under blob 2 would have internal read-only-memory containing the notes in some simple format. Chances are there's an oscillator in the chip for timing the notes and for basing the frequencies off. Read a microcontroller datasheet (e.g. ATmega328P) to see how you can set one of its timers to play a tone if you're interested. Considering the cost, it may not be a microcontroller chip under the blob but an entirely purpose-built chip, but regardless I imagine all programming in there is read-only as it's cheaper that way.

cont.

>>2528606
>>2528623
cont.
>is it also possible to remove the buzzer that is there right now and place my own buzzer that plays whatever note sequence I tell it to
For the above reasons, you'd keep the piezo disc and remove the blob to replace with your own microcontroller. The Arduino IDE has a function called "tone()" that uses the aforementioned timer hardware to play a note. Though due to RAM limitations, for more than a short series of notes I'd use program memory to store it instead. In addition, I'm not too sure how you'd get two pins operating as a synchronous push-pull waveform output, you'd likely have to write your own replacement for the tone function in order to use two timer pins with the same comparator value but with one inverted.
Pic related is my own shitty code from 2 years ago for an rgbduino (the one with the anime girl on it) to play megalovania when you plug it in. I used python to generate the array of note values from a midi file.
FYI, I'd refrain from using unmodified Arduino hardware as the built-in LEDs make low-power operation pretty tough; you'd eat through the batteries kinda quickly. I'd also flash the fuses to make it run a lot slower.
>where would I need to place the pins to know what's going on in the circuit
So putting the VDC probes across the power input of the blob/LED should be 0V when the button isn't pressed, and ~3V when it is pressed. Then putting the VAC probes across the piezo disc itself should show a some Hz and some approximately average voltage when a note is playing, depending on whether your meter can measure 200-2kHz square waves. An oscilloscope is a better tool for measuring this.

>>2528634
thank you very much I really appreciate it!

>>2526866
is it worth putting an optocoupler between a MCU and a FET gate driver, even without an isolated power supply? i guess i mean to say, what is the risk of a FET exploding and feeding my bus voltage back into the MCU and blowing shit up?

Why would I need RTOS on a microcontroller?

>>2528675
I don't think gate-channel shorts are common. You also need an optocoupler sufficiently fast. Common PC817s aren't good past 20kHz or so.

>>2528752
It makes it easy to "do multiple things at once" without making spaghetti code and it also allows you to define response times to guarantee something will execute in a certain time-frame.

>>2528752
when one or two ISRs isn't enough for real-time response I guess

>>2527887
Yeah you gotta play around a bit with the pot value to get it just right.

Just graduated and for the first time in years thinking about doing a fun project. Mad how studying electrical engineering just made me stop doing any hobby projects. I was always playing around with Arduinos but quit as soon as I started studying. Anyone else experienced this?

>>2528536
Nothing wrong. Besides, much of IC design in becoming increasingly aided by computing tools. It used to be just for validation, but I foresee a point where you throw in specs you need, costs, power etc and run an optimization algorithm. I see it more and more.

>>2528752
You don't need it until you know you need it.

>>2528838
Yes, once you go full engineering you can't go back to being a dumb retard playing with arduinos. It just sucks the fun out of everything because you already know what is going to happen. What helps somewhat is ramping up complexity of the projects, but this also means spending more on components. You can either become a HAM boomer and start erecting huge antennas in your back yard, become an audiophile and just spend your time on trying to lower imaginary distortions and try making sound warmer by using some ancient USSR tubes or start building modular synths. I just can't think of anything interesting that I would want to build. I'm even having a hard time finishing my master's thesis because last 2 years just sucked what tiny bit of joy I had left in me. Fucking boomer professors and their covid rules. Muh lab requires vax, because you must use software that I can't share with you because of licensing agreements. Not to mention how making projects for courses is worthless in their eyes, I spent a lot of time on some stupid project just to have then same grade as retards who barely made anything. And then of course prof didn't even look at what I made and gave me a shitty grade, treating me like a slacker because I don't want to memorize retarded facts from his book. Then as a cherry on top feminists started invading and girls became protected species or something when they "do work" in various societies and because of that you're seen as lazy asshole because you're not participating in one. It's like uni is for everything except engineering. And they then ask us if we want to become slaves and do a PhD while they rake in research funds. It's just horrible, I'm going to start gardening instead...

when designing shit in kicad i can never decide if i should use mostly wires or just mostly labels with no wires

>>2528866
>use mostly wires or just mostly labels

please dont make stooooopid choices.
when you use a lot of labels, you're forcing the reader to hunt around for the corresponding label.
so following one connection takes 20 seconds instead of 0.2 seconds.

>>2528866
I think labels should only be used for generic things, like power rails, data bus, or connections between big blocks, or any other case where connecting lines would just add unnecessary clutter. Remember that the ultimate goal of a technical drawing is maximizing clarity and readability.

i have this shift register, https://www.ti.com/lit/ds/symlink/sn74hc595.pdf my question is if i set an output to 1 it puts out say 3V, thats fine, but if i set output to 0 what does it output? is it connected to GROUD or left floating?
i want it to control mosfets and i need to know so i that way i know i have to add pull down resistor on gates or not.

>>2528928
>? is it connected to GROUD or left floating?

It pulls the line low; per pic related it will be less than 0.1 volts driving something like a mosfet.

nice, so no pull downs, thanks

>>2526866
what's the difference between an LED driver IC and a current mode PWM IC like UC3843? i understand that LEDs are constant current devices, but can't you do this same exact function with an UC3843?

>>2529007
"current-mode" means current-mode switching, not constant current output. It's not like you can't use a regular switching controller for constant-current output, but generally that only works for low currents. High currents will use a low value shunt resistor, so the voltage across it will be in the range of mV, so you need an amplifier before sending that to the feedback input. Presumably switching LED driver ICs have a built-in amplifier.

>>2528804
>>2528782
>>2528851
I have an esp32 with camera, which checks video frames for presence of certain color and if it finds the pattern it activates the servos. Do I need to use RTOS functions for this?

>>2529010
>Do I need to use RTOS functions for this?
No, but it will be a lot easier.

>>2529019
Could you please exain the logic how it could work with rtos?

might be a dumb question but if I have a circuit that can be powered from either usb or some other external 5v power source can I just connect them together with diodes? would this mess up the usb somehow?

>>2529022
You can create separate tasks for various things you need to do.
>a driver task that extract data from camera
>servo task that moves servos based on data from driver task
>user interface task for serial CLI interface for control and debugging
>other things you can just implement in new task without getting lost in spaghetti
Just apply common sense.

>>2529026
Is it easy to communicate asynchronous tasks?

>>2529028
Yes, that's why os is there for. You can use queues or mailboxes. It's already done for you, just a function call to push data to queue and read it in a safe manner.
https://www.freertos.org/Embedded-RTOS-Queues.html

>>2529029
Thank you anon.

>>2529008
okay that makes sense. i happened to find some weird project where a guy used a uc3843 to drive a big powerful LED. https://oshwlab.com/zipdox/led-boost-converter-driver-ina155ua he has an ina155 amplifying the FET's switching current which makes sense. i have no clue, however, what's going on with the lm358 op amp? it's feeding into the 2.5V "VFB" error amplifier but i don't understand the purpose of this.

Sorry for the shit video first of all, but I have this LED halo desk lamp that seems to be providing too much power to its LEDs. I already replaced a bunch of them but they too are immediately shitting the bed. I don't know much about LED drivers but I believe they need to receive the appropriate voltage in order to not draw too much current, correct? What would be the easiest way to make the LED driver not kill my LEDs? an inline resistor? The LEDs are wired 4S20P for 80 LEDs total, if that matters. That is to say 4 "arrays" of 20 parallel LEDs each, wired together in series

>>2529023
It would be fine, assuming what's on the output can work with the reduced voltage caused by the diodes.
An alternative would be to use 2 P MOSFETs to switch the unused output off, like in picrel

>>2529023
It would be fine, assuming what's on the output can work with the reduced voltage caused by the diodes.
An alternative would be to use 2 P MOSFETs to switch the unused output off, like in picrel

>>2529019
>multithreaded programming
>easy

how does one build a high current (>5A peak) high speed FET gate driver with BJTs? this is just for autistic dinking around in LTspice, i know the practical solution is to use a $1 gate driver IC. pic related wastes a shit ton of power on R1 yetstill shows pretty awful current amplification. this FET should be able to take like 4A peak (fed by an ideal voltage source at 12V) while i'm seeing only roughly +2.5A and -3.0A peak in simulation. what am i doing wrong?

>>2529148
Experiment with Sziklai pairs vs Darlington pairs. Use Schottky diodes from base to collector to improve transistor speed.

Next try to make a bootstrapped floating high-side gate driver with BJTs.

>>2529037
If they're in parallel without balancing resistors, then replacing some could mess with the matching and result in premature death. I'd consider rewiring them to have balancing resistors, or otherwise to have a constant-current going to each series string.

>>2529055
This method is kinda dodgy when you have both 5V from one source and 5V from your USB connected together. It's possible to feed the 5V back into your USB port.

>>2529148
In addition to >>2529182, you'll likely want to use different BJTs for the 1st and 2nd stages. The 1st stage can use lower-current higher-hFE transistors, while the 2nd stage is more limited by current.
>20Ω resistor across 12V rails
That's gotta be a 10W resistor remember.

>>2527215
Get yourself pic related and a Hakko 599B. Tip Thinners are a waste of money.

>>2529271
I like this thing. Made of metal, not like that chink shit. Very effective.

>>2529266
>This method is kinda dodgy when you have both 5V from one source and 5V from your USB connected together
How exactly? When you have both sources connected both MOSFETs' gates are pulled high and they turn off. There should be no output.

>>2529377
Oh you're right, the logic is inverting and the FETs are in the opposite direction to the PFET circuits I usually see. But it's not very desirable that the output is disabled when both sources are connected. A differential-output RRIO comparator would be a nice solution, if you don't care about increasing quiescent current.

>>2529392
>it's not very desirable that the output is disabled when both sources are connected
I agree but I'd say it's good enough and a fairly safe design. Anon was likely looking for something simple, which is why he suggested 2 diodes originally. This is simple and removes the voltage drop of the diodes without being too hard for a newbie to understand and implement.

fuck you this is fucking 2022 not middle ages anymore and this is a new part how the fuck is this still allowed?

>>2529401
Use a 1Ω/10µF MLCC low pass filter?

>>2529401
Don't believe the manufacturers. Get the part on your bench and qualify it yourself.

>>2529401
You will consume the electrolytics and you will be happy.

>>2529401
>anon is too poor to afford a set of electrolytic caps
Bruh moment

>>2529477
if you are making products in billions of units the cost quickly adds up

>>2529484
I made 1 billion pez dispensers in 5th grade and none of them had capacitors. You don't even need capacitors in anything ever. Just throw them away and pretend like they don't exist. Then you can spend the money you saved on hookers and blow.

>>2529485
>chinese engineering

>>2529266
If they're in parallel without balancing resistors, then replacing some could mess with the matching and result in premature death
They are not, there's four circuit boards onto which 20 LEDs are "surface mounted" in series. Those boards do each have 2 diodes on them, quite what they do I have no idea. Didn't check.

>>2529511
forgot to greentext your quote like a retard and only just now noticed, soz

smd resistors have labels
why not smd capacitors?

>>2529484
Are you making billions of home made circuits? No? Then shut the fuck up, a set of half decent caps is under 20$ and will likely last you for years to come.

>>2529076
"Threads" are doing separate things, it would be more accurate to call them processes or tasks. It's not solving the same problem with multiple threads, it's solving separate problems with isolated computational units that communicate via various IPC mechanisms. It's not as hard to to this compared to solving same complex problem simultaneously on different threads. Only problems that can give you headaches are accessing same hardware resources in multiple tasks. But that's usually avoided by having dedicated driver tasks. So yes, it's relatively easy.

>>2529608
Oy vey! Stop noticing things.

>>2529511
Wait so they're all in series?

>>2529055
good stuff but the orientation gave me cancer

>>2529608
first time using something other than electrolytic huh?

are tayda electronics branded 1/4w metal film resistors and dupont cables any good? i'm getting sick of using this ebay/amazon chink trash that i have. the resistors have leads so fucking thin that they bend going into the breadboard, and the dupont cables have this gummy bullshit all over the end that's probably fucking up my $9 breadboards.

If i take say 22uf ceramic cap and charge it to 3V how long will it take for it leak charge and rop to 2V? approximately
no load is connected

>>2529658
4 circuit boards
each circuit board contains 20 LEDs wired in parallel
those 4 circuit boards are wired together in series

Okay big brain time.
So after thinking for a while on how to power my micro that sleeps most of the time but needs bursts of 200ma i came up with this.
I bought ldo that can supply 500ma and only has 170ua quiescent current. That is very low but since the sleeping circuit only consumes 3ua that is a MASSIVE waste, the ldo is wasting 56x more power than rest of circuit.
I looked at some voltage references and stuff as a power option solution but then i came up with this:
Press button, P fet turns on, ldo charges caps and supplies micro with power, micro then sets the P fet gate to low and i can let go off the button and the circuit stays on.
The micro does any high current operations it needs to do, then sets P fet gate to HIGH, which shuts down the LDO, making its current draw 0.
The micro goes to sleep for 4 seconds (circuit only uses 10ua in deep sleep) power is supplied by the caps. After 4 sec micro wakes up, pulls P fet gate LOW, ldo turns on, recharges caps, micro does more work, goes to sleep and so on.
This way i get all the benefits of using an LDO while wasting at most few nano amps to ceramic caps leakage currents.

What do you think? Seems to work on paper. I ran the math and the numbers check out.

>>2529720
>I ran the math and the numbers check out

>>2529680
yes. i don't know why you'd trust chinkbay or scamazon in the first place.

>>2527215
Save yourself the money and get paper towels instead from your local Tesco and these since you appear to live in the UK:

https://groceries.asda.com/product/cotton-wool-buds/johnsons-baby-cotton-buds/1000108538653

Can someone explain what I'm doing wrong here? I'm trying to detect battery voltage using a voltage divider that I can enable/disable. The circuit simulator doesn't show the right voltage at the divider when it's enabled.

>>2529683
are you asking how long a cap will take to discharge on its own, not connected to anything?

>>2529808
is your top resistor supposed to be 1mega or 1mili?

if its supposed to be 1 mili then you probably have the incorrect fet being used, falstad can be gay with the orientation of the fet youre using.

>>2529808

yeah youre using the wrong fet also, wtf are you trying to accomplish using 1 milliohm on your n fet??? or the other one??

>>2529694
Are you replacing all 20 LEDs in a lot at once? If not then balancing is still an issue. Also there better be some sort of current limitation for the whole lot at least.

>>2529720
Not sure about the math, but you need a pull-up resistor on the PFET gate (maybe another NFET so that pull-up doesn’t kill the MCU) and an anti series diode across the LDO. P-FET power latches are somewhat common circuits.

>>2529854
>>2529864
whoops forgot to capitalize it. They're all supposed to be 1M. I copied this from stackexchange.
https://electronics.stackexchange.com/a/404285

>>2529906
yea it works now thanks. It was because of the typo

I need a little bit of help with finding a replacement part used in this schematic: https://github.com/mck1117/wideband/blob/master/board_module/export/module_schematic.pdf
The BTS3028SDR part is mostly out of stock or ridiculously priced. But BTS3046SDL seems to be similar enough to not cause trouble as far as I can think of.
The heater described as heater_pwm uses around 12V and 8W so that seems to be comfortably within spec. Only the on-state resistances that change are confusing me, this should not have a negative effect on anything right? Looking it up only tells me it increases power waste and I feel like the change in resistance makes for negligible power waste anyway.
I'd like to know if my guess is correct and I can use it as a drop in replacement or if it'll burn my house down.

>>2529925
The BTS3046 has double the drain-source resistance. If you know the heater current then you should be able to calculate whether the power dissipation will be a problem (P = R*I^2). I'd consider 1W a comfortable maximum without any heat-sinking, maybe up to 2W.

Also that's an interesting part, it's a gate driver and mosfet integrated into one package.

>>2529869
>resistor on the PFET gat
i never used series resistors between micros and fet gates before and so far i never had any issues, the capacitance of the fet gate is tiny so i doubt it poses any threat
>anti series diode across the LDO
what do you mean by that? so the caps are back feeding the ldo when its off? that would be a problem since diode wastes too much voltage the ldo is fixed outputing 3V and i need at least 2.8v

How would you solve transients messing up with electronics?
So I have some machines that combine 3 phase motors actuated by contactors and VFDs and a printer for each machine that prints labels for the products it handles.
The problem is the print head keeps dying and those things are really expensive.
We've solved the problem connecting the printers to a UPS, but this is viable long term solution.
I've been thinking about either a transient suppressor circuit on the AC line leading to the printer or some TVS diodes between the printer and the print head, to filter out any voltage peaks.
Any input is appreciated.

>>2529994
>i never used series resistors between micros and fet gates
Not a series resistor, a pull-up resistor. The button pulls it low, but what pulls it high?
Also FYI the FET in your pic is backwards, consider the parallel diode's direction.
>what do you mean by that?
Again, it's not series, this one is anti-parallel. Pic related. LDOs can't handle more than a tiny reverse voltage difference between output and input. If you'll ever see a situation where the input goes open while the output caps are still charged, you can kill your regulator.

>>2530002
Line filters on all appliances, at least on noisy or noise-prone appliances. Ferrite suppression clips clamped at the ends of all long signal wires. Ensure proper grounding of bare metal parts. Consider soft-start/stop circuits on all high-power appliances (especially reactive loads like motors).

How to solder? The fucker stays on the tip of the iron and refuses to stick onto wires. Wtf?

>>2530003
>The button pulls it low, but what pulls it high
the micro
it switches the pin between high or low to turn the fet on or off and keep them in that state also when the micro is sleeping, the button is only to provide the first boost to turn it on
>Also FYI the FET in your pic is backwards,
true, whoops

>Again, it's not series, this one is anti-parallel. Pic related. LDOs can't handle more than a tiny reverse voltage difference between output and input. If you'll ever see a situation where the input goes open while the output caps are still charged, you can kill your regulator.
don't most linear regs have some sort of protection for this?
because one thing i do all the time with my boards where i use ams1117 ldo (with no flyback diode), is that when i program the chip the programmer supplies voltage which then also sits at the output of the ldo while the input of the ldo is not connected to any power and i haven't had a single case of the ldo dying yet
Also wouldn't including the diode cause a problem where Cout would then be used a power supply for the LDO which would basically instantly drain COUT dry and thus screwed up my sleeping power supply design?

>>2530021
hmm i looked at the pdf and they do use a diode there
interesting
i have been using it without one for years without a single problem

>>2530023
to be more clear, with 5V power source connected and Vin being not connected what is the current at the question mark?

>>2530023
>>2530028
PMIC

>>2530042
that seems like an overkill for my case, plus it seems like it has a lot of extra functionality i don't necessarily need and they need a whole ass interface to set them up etc
can you recommend a specific one i can look at?

Newfag here.
Trying to make a simple two motor switch controlled toy car.

For some reason when I press both switches simultaneously it doesn't work.
Am I short-circuiting this somehow? I thought maybe I needed to up the voltage and added another battery holder but it solved nothing.

>>2530101
Update: it only doesn't work if I press the same direction on both, otherwise it works as expected

Mims III, Getting Started in Electronics

Anyone knows where I could get 2020 ed? Libgen only has 2003

>>2530016
Are your wires aluminum?

>>2530132
imagine being this indian

Guys, im modifying this wallplug switch to be controlled by a relay, to this I should just attach the brown wired to the relay right?

Now, why is the blue wire attached to the switch at all?

>>2530016
Tried flux? You might have burned it all before you were able to solder. Wires might also be too cold. You need whatever you're trying to solder to heat up. Solder doesn't want to flow or stick to cold things. Soldering iron might not have enough heat if it's a crappy one.

>>2530176
Are you trying to kill yourself?

>>2530180
Why

>>2530028
I think it would just be 0. There would be no current path since the ldo should shut off because Vout>Vin

>>2530185
Because you're messing with mains power without having a clue.

So I get that earth is a literal rod on the ground but why can we consider that a reference point? If I measure the potential between the live wire in my wall and a ground rod there should be a 120v potential. Doesn’t that mean the physical earth has to be conductive to form a closed circuit? Where does it connect to if it is?

>>2530204
What am I doing wrong then? I just don't know why the manufacturer wired the blue wire to the switch when its already making contact with the wall pins, would be better if you helped the other person actually not getting killed than to just scream non conclusive stuff.

>>2530212
I don't even know what exactly I'm looking at. Like what the fuck is that broken wire with the metal thing? If you do this yourself without being a licensed electrician, you will be held accountable if something goes wrong, and you get someone killed or your house burns down. Just saying.

>>2530176
Are you from Africa? South of the Sohairy Desert?

>>2530145
Yes probably, can't tell. It's car wires for my audio.

>>2530178
I'm using this solid orange thing as a flux, whatever I had in this crappy chinesium soldering kit.

>>2530216
This you looking at

>>broken wire with metal thing
Its the actual parts that plugs in to the wall socket

>>2530218
My plug is better than yours

>>2530221
Your buttplug is hazardous.

>>2530221
I assumed it was built into the wall. Is this just a normal adapter with a switch, and you're trying to replace the switch with a relay? Try to draw up a circuit of how it was weird.

>>2530219
It's probably a factor of everything.

If the wires aren't copper colored then they're not copper. Make sure you're using the flux correctly. Look up how to use solid rosin flux. What I would do is twist the wires together then hold them on the rosin and put the iron on top of the wires to melt some rosin onto them. Then put some solder on the iron and solder as normal. And using more flux than necessary never hurts. The solder is probably unknown chinese metal which doesn't help but it should be usable enough. If it's a chinese kit then it's probably one of those wall powered irons with a little temperature dial on it. The temperature probably isn't accurate so just turn it up so that the solder melts within one or two seconds of touching the iron. If it has a pointed conical tip then use the thickest part/side of the tip. You want as much surface area as possible for heat transfer.

If your wire is aluminum then you're going to need aluminum specific flux. The aluminum oxide on it makes it almost impossible for solder to stick. There are some hacks to solder aluminum online but I haven't tried them.

If all else fails you could try mechanical connections like wire nuts or crimp splices.

>>2530021
>the micro
The micro can only output 3.3V, which is 1.7V away from the source voltage of the FET. You may find that to be a bit low, hence why usually people usually add a transistor and pullup. Also consider using a supercap.

When I simulate the circuit I end up with a lot of current going back into the LDO after the FET turns off. Before anything I'd test your LDO to see if it suffers the same problem, diode or otherwise.

>170ua quiescent current
>That is very low
Mate the ADP2138 linear regulator has a quiescent current of 1µA. You bought the wrong LDO.

>>2530264
>Is this just a normal adapter with a switch, and you're trying to replace the switch with a relay?
Yes

See pic, why is the blue wire like this? Why is it going to the switch?

why are those resistors and the l1 around the antenna connector?

>>2530289
That's so it's in compliance with SIberian electrical wiring standards.

>>2530291
>SIberian
Siberian? Where can i read more on that? How does it help to have an extra wire for no reason?

>>2530290
Those are not resistors, but inductors and capacitors most likely there for impedance matching purposes and/or filtering, because they form what's called a pi network, in this case of low pass type.

>>2530294
In case of arc flashover fatality, the neutral wire is supposed to act as ground.

>>2530297
will it work without them?
I just want to connect a simple external antenna

>>2530289
I can't make sense of it. The blue wire would be neutral, brown is the hot wire (or phase). Maybe the intention is that if the switch is off, to connect neutral to the hot wire of the consumer to "ground" it. But a CEE plug doesn't guarantee which wire is neutral or hot. Maybe it's OK to connect hot from the mains to both hot and neutral for the appliance, because appliance aren't supposed to connect neutral to ground (false ground). Thank god I'm not an electrician.

>>2530298
Should i connect the blue wire going to the switch to the NO on the relay then?

>>2530301
neither i, the other anon provided a reason, there's got to be one, what it costs for them to have this in milions of these is just too much.

>>2530299
It most likely will, but depending on your antenna performance will be reduced. You need to short L1 - solder pads together.

>>2530304
>Should i connect the blue wire going to the switch to the NO on the relay then?
Yes.

>>2530312
I mean, NC.

Why isn't it common practice in projects with relays?

>>2530289
>why is the blue wire like this?

the light inside the switch needs constant power, so it's connected across the power lines directly.

>>2530313

>>2530273
>ADP2138
>1ua
i can't solder ball arrays, it needs to have legs

next time i'm ordering parts i will order some better ldos but for now it's the best i have.. i think i will just totally toss the fet idea and just deal with the 150ua of wasted current.. it's still pretty tiny and i can just swap the regulator out later when i buy better one, it's better than ghetto riging it with some capacitor and fet fuckery
Getting ldo that has smaller idle current than about 10ua is probably pointless anyway since something like 2000 mah lipo has about 50ua self discharge current

>>2530329
this looks pretty good only 2ua idle current, input voltages up to 16V
peak current only 250ma that is kind of shit, but it is just within the range i need, my radio peaks at about 150ma. It even doesn't force me to use those gay tantalum caps on the output.
But jesus that thing ain't cheap almost $1 a pop
With thinks i can buy a roll of 100 ams1117 for 3 bucks ;_;

>>2530317
That's not it, the light is only on when the switch is closed.

>>2530321
Load will be on NO, so it closes when vdc is at high, and from what you're telling me, i should connect the blue wire to NC but i've never seem anything like it, i might just fry some chicken like that.

>>2530334
>thinks
*chinks

on a scale from 99 to 100 how likely is it that this is a fake and it's just 1117s with different name printed on them?

What metal are the legs of THT components made of? Probably something tin-coated?

>>2530337
also let me spam some more, does this mean i don't have to put external diode from vout to vin?

>>2530341
tinned copper-plated steel

>>2530343
some chink pin headers i have are magnetic and some are not, which ones are the shitty ones?

>>2530346
The magnetic pins are inferior. Copper costs too much so they compensate with whatever they sweep off the factory floor.

>>2530336
When the switch is closed, the resistance between the connected power lines would be near zero. The light couldn't work. Can you give us some specs on the switch?

>>2530351
It's just a regular 250V 16A switch.

I just tested with a multimeter, the switch only lets current through in the on mode, and that's the only mode where the light came on.

>>2530360
And what does the third lead do?

>>2530364
The one connected to the blue wire? It's never on.

>>2530365
Then maybe it's the lamp? Please try to know what you're doing.

>>2530368
>Then maybe it's the lamp?
You mean the load? It's on when the switch is also on, as it should be, can't really find a reason why the blue wire is connected to the switch.

>>2530372
Your fucking switch apparently has a lamp. The lamp needs a power supply, at least if it's on when the switch is closed.

>>2530377
>>2530336
>That's not it, the light is only on when the switch is closed.

As is, on, with current passing through

>>2530378
Look at this simple switch:
https://www.digikey.com/en/products/detail/adam-tech/SW-R2-1A-F-LR13-2/15284420
There is a datasheet with schematics. Is this like your switch? As you can see, there's no way to use the lamp without connecting the third lead.

>>2530398
Oh i get it, that might be it then, the only thing the switch says is "HW-4-1", i couldn't find anything online.

Why do you need a power supply unit, if a chip like LDO (or whatever the correct name is) can convert the voltage?

>>2530405
If a LDO is sufficient, it IS your power supply unit.

>>2530406
How do I know if it is sufficient? And if not, what is my second and third option?

This capcha is killing me.

>>2530290
Impedance matching components making up a pi network. You can maybe get away with a little hand-wound air-core inductor and some ceramic disc capacitors, if you read the datasheet and figure out what values they should be. You can calculate backwards from those values to figure out what the impedance ratio is, assuming the antenna and RF IC act purely resistive. With that you could figure out how significant the signal loss would be. Leaving the caps open and the inductor short would mean the signal strength could easily be 10 times weaker, so it may or may not be ok for short-range testing.
Your antenna should also be properly length-matched.

>>2530329
Oops, meant AP2138. Comes in SOT-23.
https://datasheet.octopart.com/AP2139AK-1.5TRG1-Diodes-Inc.-datasheet-137404753.pdf
>Getting ldo that has smaller idle current than about 10ua is probably pointless anyway since something like 2000 mah lipo has about 50ua self discharge current
Yeah I'd say it's only worth it if you're doing some cool shit with energy harvesting and supercap energy storage.

>>2530337
I had alibay AMS1117s fail on me for no reason a week or two ago.

>>2530341
Tin-plated copper preferably. Cheap magnetic ones are just tin-plated iron. I think I've seen some nickel leads before too, they were too flexible to be copper or iron.

>>2530342
I'd say so.

>>2530405
LDOs are inefficient, so for more than 100mA-1A or so, people use switching regulators instead. Furthermore, they can only handle so much input voltage, so for higher-voltage systems switching regulators are also preferred. Then if you want to step down from mains voltage, a switching converter (or AC transformer) power supply is basically a requirement.

>>2530413
Depends on how much power you're going to use.

pain locked loop

>>2530564
plls are cool though

Uuuuh...
1000uF 450V cap has ESR of 120m and ESL or like 20n
Let's say I get 10 of them, just to be sure. That is 800J at 400V which is more than enough.
That would be $40.
Next thing is some sort of high current switch that is also fast. Problem is that all parts are expensive, and I have peaks of 30 kA flowing though the circuit. What would you suggest? SCR? IGBT? IGBTs are slow, but idk, maybe technology have changed a lot... Mechanical contact would just weld itself and I don't think it would be fast enough for the purpose

On the plus side, one 18650 has about 30 kJ of energy, and it can charge main capacitor bank about 30 times, maybe, depends if I make DC-DC converter good enough. Problem is charge time, because powerful inverter would be size of a computer power supply, which is a problem in this device, as every cubic inch matters, and it is already too big. Charging from household socket is possible and would be fast enough, but it limits portability of the device.

What would you suggest me to do? Raise voltage even higher? Borrow IGBT modules from local train?

Other thing is about coil winding and mechanical construction. Does lead stick to steel, or it need a bit of tin in it in order to work?

>>2530652
Actually, i think I should think in terms of power output of each capacitor, and as I figured out, all I need is not much, about 3.8 MW of power for just 1/14000 of a second. With overall (in)efficiency of whole device I think 5.3 MW would be much more reasonable.
But apparently one capacitor can output just 0.315MW or 0.15 MW, im lost in my calcs.

>>2530652
I think SCRs are the way to go, but I'm not exactly familiar with them. Technically an SCR and an IGBT both use BJT junctions to switch their current, so all else being equal I think you'll be able to get more current via an SCR since it's got the cascading base current feature. But it really depends on how the parts are actually implemented, so read some datasheets for modern parts you find using parametric searches to see if they can handle the peak current. A bit of a hunt through octopart and all the SCRs I'm seeing seem to be designed for mains or DC use (pulse lengths of 10ms) and often don't have high ratings for pulses, can't respond particularly fast, or simply lack ratings for anything other than slow-speed operations. You'll need to look around for guides from manufacturers who do actually make pulsed-power thyristors. Littlefuse may be one such manufacturer.

Also consider a spark-gap as a switch.

>Problem is charge time
If you're wanting sub-second reload times you're looking at kW levels of power, but a custom TL494 circuit with a FETs and half-bridge driver IC should still be usable for that kind of level (synchronous rectification of course). The magnetics are going to be the hardest part though, I'd opt for a forward converter topology since that way you don't need an air-gap in your transformer. You may want to consider a ZVS circuit or other quasi-resonant converter topology, if size is a big deal. High frequency means more compact.

>Does lead stick to steel, or it need a bit of tin in it in order to work?
Lead is probably ok, but lead-tin solder is probably easier to use by all metrics.

completely new to this, but I was playing with this circuit and changing different values of resistors for the R2 resistor to see when the led would turn on.
when I inserted a 47 ohm resistor for R2 the led stayed off, but when I inserted a 100 ohm one it turned on. The led is a blue led rated at 3-3.3v 20mA.

what would be the way to mathematically select the right resistor value for the blue led to turn on?

and does the led have a voltage drop or resistance? is the NPN transistor a load? why am I only seeing 1 mA of current flow to the led even though my bench supply is maxing out the current limit of 20mA?

>>2530728
>completely new to this
Yes, I can see.
>The led is a blue led rated at 3-3.3v 20mA.
Here is a first problem. In your circuit you're powering everything with 3V supply (V1). This might be just barely enough to turn on LED.
>when I inserted a 47 ohm resistor for R2 the led stayed off, but when I inserted a 100 ohm one it turned on.
From perspective of R2 transistor looks like diode from base to ground. Because it's silicon, it has a voltage drop of about 0.7V when it's conducting. This means when R2 is 47 ohm, current through it is equal to I = V/R = (3V - 0.7V)/47 ohm = 49 mA, which is way more that your limit of 20 mA which causes voltage on your power supply to sag well below 3V. With 100 ohm R2, I = 23 mA which causes voltage to sag less and leaves just enough voltage to power the led.
>what would be the way to mathematically select the right resistor value for the blue led to turn on?
You first need to check how much gain does your transistor have. According to the datasheet minimum gain is 80 - this refers to how much base current is amplified. If we want to use transistor as a switch we would need to operate in saturation region. This means that less current is flowing through it than it could if supply permitted. If we want 20 mA going through LED, we would need to push transistor to amplify it to something like 200 mA, but it wouldn't reach that because it will saturate before it could pull that much current. For V1 = 9V and Vbe = 0.6V and gain of 80, to get 200 mA capacity R1 would need to be R = (9V - 0.6V)/(200mA/80) = 33600 ohm. You can safely use even something like 10k, 33.6k is maximum value that would work as intended. This will turn transistor into a switch which would have 0.2V across collector and emitter. Now you just need to select R1 to limit current through LED. For 20 mA, R = (V1 - Vled - Vsat)/I_led = (9V - 3V - 0.2V)/20mA = 290 ohm.
>and does the led have a voltage drop or resistance?
It has both. Read wiki/books about LEDs.

>>2530728
>and does the led have a voltage drop or resistance? is the NPN transistor a load?

also, how do I know the voltage drop across a resistor?

>>2530753
Reading literature below will explain things way better than anyone trying to do it with text only here. You just unfortunately have to grind through boring parts first.
>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

So I've bought an IR laser diod with 200 mW power. How to power it on and not blind myself? Do I need laser driver or connect it like simple LED?

>>2530766
You're dealing with class 3B laser. Looking directly at beam could fuck up your eyes in an immediate manner. Don't power it on unless you're wearing appropriate protection.

Anyone else made a SMPS controlled with a microcontroller (atmega8a). I plan on making a buck converter using a IRF9540 P-Channel MOSFET.
I'm just thinking what kind of optimisations can I do with the buck converter. Variable duty cycle for starter, incredibly fast control loop, variable frequency for (reasons??), simple PID or maybe more complex control scheme etc etc

Just wanted to know if someone had similar experience and how did everything work out

>>2530770
I made a boost converter with PI control and it worked fine for what I needed it, but I also burned 3 MOSFETS and a instrumentation amplifier because I forgot to clamp output voltage during turn-on. Adding complexity is asking for trouble. Here a mistake in software means magic smoke will escape from something, fet gets stuck in ON state or everything just craps out because of noise from switching. It's great for learning and you can add various digital filters to do whatever you want, impose voltage/current limiting, automatically detect fault conditions...

>>2530751
appreciate the explanation, this helps a lot. currently reading through Practical Electronics for Inventors, but clearly not enough yet

>>2530774
Sounds interesting anon, yes I too intend to do it for learning purposes but if I manage to learn enough to build useful stuff that's just a cherry on the top

>>2530768
Mom please.

Five minutes ago I received package with another laser, this time 30 mW because I was thinking 200 mW might be too much for my needs. Anyways, I've ordered glasses with proper wavelength as well.

This new laser is in a steel case with focusing ring, and there comes two wires from it. So do I need a driver or not?

>>2530780
>So do I need a driver or not?
Most likely driver is already integrated in case, because diodes usually have 3 pins. See if you can take it apart, otherwise power it with a power supply that has current limiting.

>>2530786
>driver is already integrated in case
You are right, looks like a small pcb with couple of resistors and capacitors, and a voltage regulator. There's also a variable resistor for screwdriver, I assume I should choose input voltage on it?

As for the power supply I have this standard breadboard one with 5 and 3.3 volts output. You think it will be fine?

>>2530793
Yes, it should be fine. Just measure current consumption if you can and then adjust resistor until you get about 60 mW power consumption (12 mA at 5V), assuming diode is 50% efficient.

>>2528854
Agreed with the projects becoming less interesting but can't relate to the feminism and vax stuff. You must be american from the sound of it because here people stopped caring about covid stuff on unis since februari. Furthermore the whole departement is just dudes and some chinese/muslims women so never have to deal with feminism stuff.

Don't lose hope though, just graduated and my friends who have graduated seem to be really enjoying their work and I look forward to it as well.

When I wrote my master thesis and had to do measurements, there was almost no high end equipment besides an old VNA and osciloscope. Also no lab technician. Had to build the entire setup myself from wood with my dad lol. Now I am thinking about working for the company where I did my internship and they have everything that you can dream of, way less frustrating.

Also don't do a PhD right out of uni. Certain companies pay you to do one. Also what are you writing your thesis on?

Where do you buy your capacitors ? Should I order some batches of different capacities or get them on need ?

>>2530828
farnell, rs, digikey, mouser are good options

>>2530828
Depends on the caps in question: for generic ceramic and electrolytic amazon sells packs with a bunch of caps for each most common value for quite cheao,; same for resistors. If you need something specific and/or quality then >>2530848

>need part for $0.2
>$60 shipping cost
How do you not go crazy?

>>2530864
I just put stuff I fancy in the cart and buy it all every couple of months

How do I fix sulfation in my battery? Every website gives a different answer from buying their chinky device to pulse high voltage to the terminals, to letting the battery "charge" for 24 hours at 3V per cell

>>2530857
Yeah but generic SMD ceramic caps don't have a voltage rating. Will that 10µF work for 12V? Who knows!

>>2530132
>Mims III, Getting Started in Electronics 2020 ed
The book is too new anon. Libgen and zbook's onion site don't have it.

>>2526866
Anyone here had to replace a apple 30 pin connector before? the one on my old ipod has gone, and need to replace it. Main problem I'm having is sourcing a replacement, nowhere seems to stock them, except one seller on ebay and aliexpress, and fuck that considering the effort involved in this. I don't want to replace it then have it break a few weeks later due to chinkshit ports.
Is it even possible to get a few original or good quality ports without paying stupid amounts of money? The only seller on ebay charges £9 per port, which considering I'd like a couple spare seems really fucking expensive.

>>2531036
Dude just replace it with an USB mini or something. Look up which of the pins correspond to the power and data then connect to those.

>>2530152
lol

>>2530132
I have one that I got for free (i aint paying for that shit) and, unless you're 12, you really aren't missing out.

You can find a Grobs Electronics edition in libgen ez and they are far more comprehensive.

Idk what edition they have but you're not missing anything with newer editions. They typically just include some access code to prove to youre school that you shat out the $$ for it. They literally jew every penny out of you, possible.

If youre trying to get an absolutely definitive book that completely explains everything from top down/inside out, then I recommend buying old books. Like really old books. Like 1930's - 40's books. A lot of those books were written with with actual lab books that were very detailed.

>>2531036
just buy them on aliexpress its literally exactly the same thing as whats in your ipod now.
you don't think they were made in california did you?
they just run the same machine they use to make livcence parts and sell them backroom for side hustle. nobody is building a factory from scratch to clone a 20 years old proprietary connector to sell on internet flea market

>>2526866
I would like to make a heating cable but can't find the right material for the cable. Here's this for reference: https://www.hydrofarm.com/p/jump-start-soil-heating-cable-12/jshc12

((120 V)^2) / 32 W = 450 Ω
450 Ω / 12 ft = 37.5 Ω/ft

8 AWG copper is 0.00064 Ω/ft.
8 AWG nichrome is 0.047 Ω/ft.

Bigger gauge wire will be even less resistant. What material are they using? Do I need to put an actual resistor every 3 inches?

Realistically I want probably ~1.5 kW 12 ft cable meaning:

((120 V)^2) / 1500 W = 9.6 Ω
9.6 Ω / 12 ft = 0.8 Ω/ft

So nichrome gets close but not close enough unless I make the cable 240 ft long :/ please help.

>>2531212
8awg is massive mate. Use 32awg nichrome.

>>2531235
You're right. Dunno why I was thinking 8 AWG was good. But I need the cable to be thicker than 32 AWG, what should I wrap the nichrome in to give the cable some mass?

>>2531238
>I need the cable to be thicker than 32 AWG

absolutely, #32 is crazy thin and fragile.
you'd want #28 at worst.

>what should I wrap the nichrome in

i dont think mass is your problem. it's corrosion.
put an electric wire in the ground and it'll rust to shit in a month.
but put a current thru it, and it starts to rust in 5 minutes.
i'd use 1/4-inch polyethylene hose, widely available at hardware stores.
good enough for boiling water.

Why is it so hard to find ice blue 5mm leds?

>>2530857
>>2530848
ok thanks.

To learn about repair technics, I'm atm trying to restore an old motherboard.
I will make my order list based on those picrel's brands

shieeeet
the eternal dilemma

>>2531419
Black and white are ass to reverse engineer, you can't see the traces through it. Don't be a prick to the next guy who has to work on your board, though putting a link to the cad files in silkscreen would be fine I guess. Green will be faster, if not cheaper.

>>2531419
also 8 layer pcb for $2.. bugmen are insane and and need to be stopped, they are going to completely bankrupt western fabs

>>2531423
green is the worst looking one
but in 98% of cases i always use black because it just looks cool as fuck with the golden pads and white silkscreen on top

>>2531425
They don't do blind or buried vias though, nor do they do microvias or vias underneath solder pads. They also only do single-sided aluminium boards, meaning if you want to put SMD power transistors on an aluminium board, you almost certainly can't put the gate drivers on the same board.

>>2531427
I'm a yellow guy, the copper looks golden through it, and I use tin plating so there isn't a funky contrast between it and real gold. I bet blue could look great with gold plating though. And yellow or orange silkscreen, but the bugmen only do white (or black?).
Green can look decent for retro designs, but yeah I only choose it for prototyping of things I don't care about the aesthetics of.

Mitxela's curvy trace kicad plugin is a blessing upon us all.

wot?
i don't understand i ordered a bunch of pcbs with vias that are 0.2mm hole and 0.4mm size and never got charged for it

>>2531496
anyone else experienced this at JLC? can i just reply to them to skip the test?

>>2531497
i replied them to skip the test i hope they won't try to fuck me there is no way i am paying 30 bucks extra on my 2 dollar order

>>2531499
come on Ping Pong stop blueballing me and send it into production already

>>2531513
this is so weird
i always make 2layer pcbs with 0.2/0.4mm vias and never once got extra charge and yet this page says their 2sided pcb via limit is 0.5mm
Limit on 4 layer pcb which mine is is actually LOWER then that, 0.45, mine are 0.4.
Well i guess it's my fuckup for not reading the page properly but still why then i never got charged before? confusing.

I'm building a scale model and will try lighting for the first time
I've been researching how to build a simple LED chaser circuit. There would be twelve leds that would blink in series over and over. but as you can see, it's very very small (each opening is less than 2mm.
Do any of you have advice on how to approach this? I've never built a circuit before, only done electronics repair and component replacing, should I get a breadboard?

So, I have a 60v moped battery, and a 60 to 12v dc dc converter so I can use some car socket appliances with it.
I got for free a working UPS with a fucked battery, what would it take to rig it to my setup so I can have like 200w of ac? Do in need to add a battery and a battery charger? or would just feeding the 12v line to where the connectors were be enough?
The dc-dc converter should produce more than enough watts for the ups, unless some of the components specs are complete bullshit.

>>2531540
light pipes

>>2531540

>>2531540
Pretty sure there are SMD LEDs small enough to do that on a board.

>>2531684
do i have to make a special pcb for this? im new and trying to learn

>>2531691
If you go for SMD LEDs, yes. If the centre-to-centre spacing between openings happens to be 2.54mm then you may be able to use those 2mm square LEDs in perf-board. Going for light-pipes from 3mm LEDs on perf-board would work, but you may find that the round pipes themselves don’t look wonderful up against the square openings.

Whats everyones favorite cable to cable pluggable connectors? Signal level mostly. Ive used Metri-Packs in the past, but Im looking for something with screw terminals.

>>2531701
>round pipes themselves don’t look wonderful up against the square openings.
thats another issue i have, no idea how to cut square openings into plastic that size. it may very well be round openings with a tiny drill bit

>>2531728
Butt each pipe against its own diffuser sheet, taken from an old LCD panel.

What your favorite size hammond box?

>>2531722
DB9. You can buy them everywhere, they can be cable or panel mounted, you can make them waterproof, they’re easy to terminate both to wire or directly on the edge of a 2-sided PCB, and they’re solid as shit. The only downside is that they’re already used by RS232, a standard that doesn’t carry any fixed voltage rails aside from ground. People mistaking them for VGA is also kinda ass. Arguably larger sizes are better because they don’t have that overlap. You can customise a connector by cutting off a pin and filling the corresponding socket with epoxy, that way you can’t improperly plug anything together if you have a bunch of different D-sub pinouts in one location.

Din connectors and microphone connectors are also decent if you don’t need as many pins, but they’re not quite as available. The microphone connectors do screw together via one surrounding captive nut, and I suspect waterproofing would be straightforward.

I don’t know any more modern connectors.

>>2531728
Matching light-pipes directly to drill diameters is an easy way to get something that looks decent. Though the resulting specs of light may look a bit small. While more involved, stepping them back a few mm and using a diffuser across a larger area would probably work well, but would involve difficult fixturing. I suspect butting the light pipe up against the diffuser directly might not look any good, but I may be wrong.

I have an old tape deck from the 80s im working on, I can read 120V at the input but zero V after the step down. Am i just measuring it wrong? Putting any order of pos.neg probes to the 5pins after the step down gives me 0VAC

>>2531942
thermal fuse inside the transformer blew
he needed the money ohhhhhhhhhhh

>>2531948
ah yes the thermal fuse.
good news is that i have some in my stock
bad news is that idk if its worth the time to fix something that isnt worth 10 bucks. i got it off the curb with 5 other pieces of varying functionality.

>>2531952
Keep it just in case you get a wild hair up your ass and need to do something to distract your mind. Or smash it with a hammer and set it on fire for hipsternet hate points.

>>2531956
aaaa where is the thermal fuse. im all the way down to the windings and i still dont see it. what have i got myself into

>>2531977
Should be attached to the primary posts (white) under the green plastic wrap.

>>2531984
its pretty solid windings under that green on the primary side, can i just solder the fuse across the top like this?

>>2531985
Try it and see. Test the coil resistance too while you have it disconnected, both primary and secondary.

>>2531985
the fuse is in series with the winding, not in parallel. is the winding reading open-circuit?

>>2531990
>>2531989
putting the MM in resistance mode reads OL across the primary pins and OL across any of the secondary pins no matter what combination.
this transformer is fucked and thus so is whole unit aside from being parted out now

>>2531993
Maybe there's a thermal fuse on each winding? Be kinda strange though, especially for all of them to die. Sure your resistance measurement is working? It's also possible that overheating caused the solder binding points to melt, or maybe they cracked.

i ditched it and moved on to the Sony unit. This one has a corroded out crystal at the VFD unit that fell off as soon as I touched it, its possible the two ceramic caps are corroded too, but im having a hard time finding any info on this crystal.
if i could fix this and get the VFD display working again the unit is 100%

>>2532049
Maybe 11MHz? None of the other numbers sound like a crystal frequency. The capacitor values should be available in the crystal datasheet, so whatever you replace it with shouldn't be a problem.

Chances are you can try a bunch of crystals and see what one causes it to work the best, I doubt anything wrong will happen from too low or high a frequency. Though a more tactful approach would be to look on digikey for crystal markings from a variety of manufacturers to see if anything rings a bell.

>>2532050
hmm thats a good idea. the closest thing I could find without data sheet is a footnote which suggests it could be 27.12mhz

> 27.12 MHz quartz starts in less than 800 s. For example, quartz like TAS-3225A, TAS-7 or KSS2F with appropriate layout.

>>2532051
https://www.datasheetarchive.com/kss2f%20crystal-datasheet.html

>>2531882
military connectors are the shit

my signal generator kit is finally here
just look at all those babbys first solder job tht components.. disgusting but also nostalgic
too bad all these use the knockoff chip that shits itself on higher freqs but i mean it's so cheap buying the original one is just not worth it

>>2532053
Yes that’s the same place I was at. Kss2f only shows up once as the first context link

>>2532167

>>2532215
well fuck me. thank you friend i owe you

>>2532236
Merry Christmas and happy new year and all that stuff.

im sure it would be fine to use a 7.37mhz instead of 7.2 right? since they all have some % of hertz variance allowed? 7.2s are only sold in bulk and i dont need 50 of them, not do i want to wait 3 months from china

https://www.digikey.com/en/products/detail/ecs-inc/ECS-73-20-1X/827470

https://www.ebay.com/itm/195447910502?mkcid=16&mkevt=1&mkrid=711-127632-2357-0&ssspo=uixesydftnc&sssrc=2047675&ssuid=kTZzj1llSQ6&widget_ver=artemis&media=COPY

>>2532257
>fine to use a 7.37mhz instead of 7.2 right?
It should be okay as long as the replacement buffer caps are the proper value.

just noticed that the samsung MLCC product pages have this very useful graph. wish more manufacturers would do this.

>>2532163
I have one of those pending to do, which is the original chip?

>>2532333
all the chink ones are fake, they sell some of the old original ones on ebay and shit but they cost more that the whole kit

>>2532257
The PLL uses fixed dividers to get a desired output frequency. Using a crystal that's 2% off will mean that the speed of some synchronous part of the system will be 2% off also. Idk if that corresponds to a tape spool speed or radio frequency stuff or whatever

>>2532320
Is that capacitance dependant?

nice sine wave there bro
GOD DAMN IT CHINKED AGAIN!

>>2532454
nevermind i am retarded and clicked the wrong timing setting, it's actually a gorgeous triangle shaped sine

this thing outputs the weirdest fucking sines..
have you ever seen a thick colored in sine wave? now you have

>>2532461
Does that mean there's a lot of noise/variation?

>>2532461
I can't understand if you're complaining about the scope or the signal generator. Whatever is making that signal clearly sucks if it's meant to be a pure sine, and as for the scope you've activated min-max holdouts.

>>2532488
>as for the scope you've activated min-max holdouts.

i dont even wanna know what that is, but that wave definitely breaks the laws of physics.
at any one point in time, the DAC can only read one value, not 2 or three.
and some of those lines seem to go from right to left, i.e. backwards in time.

Is there an free online database of all schematics in a standardized format, sorted by categories with tags? Would you like to have one if the answer is no?

>>2532493
>laws of physics
Scopes usually trigger on specific values (depending how you configure it), like on a rising edge. The min-max hold would just show what happens after the trigger over a long time.

>>2532497
>standardized format
Is there even one? One that can be edited. All CAD programs do their own shit.

>>2532497
>Would you like to have one if the answer is no?

i'd like a full case of them for me and my colleagues.
in lieu of payment, am willing to spend one full hour of downloading from the Pirate Bay.

>>2532505
We need one more format. This time open-source gnu friendly, something like openSFF (open schematics file format).

>>2532507
>We need one more format.

absolutely.

>>2532519
It's true, though. New formats are invented because someone feels the need for it. Most of them will be ignored or stay obscure. Some of them will add to the proliferation of formats and add another one, as the comic shows. And sometimes, there is a new format that is good and popular enough to replace all the older formats.

>>2532497
Not that I know of, but that audio circuits guy comes kinda close on his own. They’re all image format though. I’d love a database of netlists that you could do parametric searches through. Like:
>p-chan pin D connected to n-chan pin D and no Schottky diodes
I would definitely help make such a thing. Not sure if the part list and the net list would be seperate, if so the part list also works as a BOM. Ideally the format is such that it can easily be ported to spice and CAD. But converting a net list to a schematic means placing the parts in a sensible way, not sure if that’s easy or a travelling salesman-tier problem. But at least if it’s automated it will be possible to have a few different options. Like to what extent labels are used instead of spaghetti. Manually doing it might be a pain in the ass, same for manually transcribing images.

>>2532524
Format darwinism. Witness the death of megafauna (DB25) as the miniature life-forms (USB C) rediversify into the ecological niches.

I want to use a current transformer. To keep things safe, the current transformer should never be open-circuited across its terminals to prevent high voltages. I was thinking of slapping two zener across the current transformer to make sure that if it's accidentally open circuited, the voltages get shorted through the two zeners. Setup is in pic related. Is that safe or is there a better way of doing things. I'd like not to solder the burden resistor since I might want to change it for different test cases.

>>2532674
I think it would be fine, I'd do something similar for sure, not that I've ever used a current transformer. You could also use a dedicated TVS diode, since they have both zeners in one package.

sometimes i see circuits where reversed polarity protection is just a diode like this, the fuck is that shit? all that will do is it will just short the power supply and blow up

There's no laser/optics specific thread, so I'll ask here.

I try to make my own infrared laser measuring tool (kinda). I've got an IR laser diode with 980nm wavelength, got a camera with removed IR cut filter and got an IR pass filter for 950nm and higher. The results is on the left, I still see bright sky even though its overcast and filtered everything before 950nm. So it is pretty much impossible to distinguish between bright roof and bright laser point. If I look for lasers beyond sunlight spectrum >2000nm the prices are in thousands of dollars.
So my questions:
1. How come cheap lidars (~900nm) and laser tape measurment tools (~600nm) work if the laser is indistinguishable in daylight?
2. What I do wrong?

>>2532839
>filtered everything before 950nm
First thought: the pass filter is shit. Where'd you buy it?

>>2532843
Filter does the job, it filters everything before 950 nm, I can't see anything through it with naked eye. Got it form a local german manufacturer of filters from Nuremberg.

>>2532847
Did you go through the camera's settings? Maybe there's some image processing happening in real-time.

>>2532849
It is in manual mode, I've changed exposure gains and it kinda works if I make everything darker until I put laser mark on that roof and it disappears with sunlight.

My only thought is lidars and other laser measurement equipments use highly calibrated narrow bandpass filters, which cut the spectrum not only below certain value, but also above. But they cost a lot of money, cheapest around 160€ for a small piece of glass. There should be something else. Laser tapes cost like 30 bucks.

>>2532852
You could try stacking another filter if you have enough room, or aiming the camera at an IR reflector which is aimed at the target. I really don't know enough about optics to give further advice.

>>2532858
> You could try stacking another filter if you have enough room
I would but I can't find an IR cut filter for >1000nm.

>>2532785
cheers

>>2532493
RTFM you dumb fuck. Those lines are just crosshatching to indicate the area within which the values fall in. You shouldn't use an advanced tool without understanding how it interfaces with the user and how it can display data.

>>2532882
which camera are you using?
regular ccd or cmos sensors aren't really capable above 1000nm, you should check your sensor data. Maybe it is already shit in the 980nm area, but way more sensitive around 950 and that's why you got those results.
Have you tried with a polarization filter in both the laser and the camera?

>>2532794
this "works" as a bare minimum and is garbage tier

>>2532839
1. crazy dsp in hardware. lidars are also sometimes a grid pattern of multiple points. example is an iphone front camera face recognition.

2. nothing. sunlight is extremely strong in the infrared spectrum. the only way to get around it is to have a brighter emission source. or emit some sort of special pattern that you recognise in the camera. like flashing specific sequence.

I can buy a 2n7000 from aliexpress fairly cheap (around 1 cent per transistor). Does anyone know if there is a pmos variant around the same cost? I can't afford to buy a bunch of pmos transistors for 50 cents a piece

>>2532794
What you're missing from your circuit is the fuse. Once the power supply is shorted, the fuse blows up, protecting your load from significant negative voltages (above Vdiode). But you are correct that if there is no fuse, this is just going to short the power supply through the diode and blow up. Then the load just gets the full negative voltage. It's a terrible polarity protection circuit, but is useful in a pinch. Look into other polarity protection circuits for better alternatives.

>>2532922
>>2532935
I think there's no particular problem, but a combination of smaller issues.
It's dark outside, I tried the same experiment and my camera still can't see the laser on the neighbors roof 50 meters away even though my laser diode has 30 mW power and should be plenty powerful to strike in one km range. So I figured those might be the problems:
1. focusing ring in cheap laser housing is shit and cant focus properly in distance
2. laser 30 mW might be too weak?
3. camera cmos sensor is too weak for higher wavelengths (simple dev board camera ov2640)
4. can't flash specific sequence because for this a perfectly stable fps and big fps numbers are required
5. my IR filters filter just a lower part of the spectrum.

I really would like to finish the project, so here are my solutions to above problems:
1 - buy another laser housing, for example a laser dot sight and switch the diod for mine
2 - I have another laser diode with 200 mW power but without casing and a driver, need to make a driver and will be able to try it
3 - no idea lol, how can I check spectrum of the camera sensor? I assume even expensive cameras might have unsuitable spectrum
4 - working on it, currently waiting for another board to try and get more fps, but I highly doubt i will go this way with flashing sequense, it's too complicated
5 - I've ordered calibrated narrow bandpass filters from china for reasonable price, but they come in europe in month or so. What does polarization do in my case?

What do you think about my essay?

>>2526866
i picked up a bunch of cell phones from my ancestral home that i was clearing out. found 2 Nokias (among many others), one of them a 6630.
It works, but the display is bust. MOre specifically the ribbon cable that connects the display to the mainboard is bust.
I can't find a display on Aliexpress that would work with this and i have no leads now.
now there are 2 options:
1) keep looking for a new display
2) run a jumper cable to the display.
how hard do you think the second option is? i have a soldering iron, patience and the pinout diagram. I don't have a microscope which i presumably need to do this.
how should i go about jumping the connection, directly to the display? is it even possible?

>>2532922
Ok lmfao, I had the datasheet but didnt check its spectrum. Now I see part of the problem.

>>2532953
https://www.ebay.com/itm/300748956848

>>2532949
1 yeah maybe the focus is shit
2 is that a dvd laser? I would say so
3 you already did and its bad. Also are you using a color sensor? get one that's black and white if you ever want to work in the infrared. Pigments may filter the infrared.
4 no idea
5 if you polarize the initial beam and then you filter for only polarized lights it may help you to see only the laser. It helps a ton to eliminated reflexed light from the sun or whatever. But I am not so sure as you will lost a ton on light and you have a really bad source to start off. Also i don't quite remember if laser sources are already polarized or what?
can any other anon comment on it?

I won't answer for a day or two, just to let you know. Happy new year

>>2532609
>Format darwinism. Witness the death of megafauna (DB25) as the miniature life-forms (USB C) rediversify into the ecological niches.
That's such a beautiful metaphor and nobody else will get it because biology is considered hard to understand for some reason.

>>2533032
>dinos die off
>chickens take over
>price of chicken wings goes through the roof
What's hard to understand? P.S. Muricans at 73 trillion chicken wings in 2019.

>>2533034
*ate
*74 trillion

>>2532852
Band-pass filter is how automotive laser detectors work. You can make them cheaply through this method:
https://youtu.be/iwj78pR46zM

The other way is to modulate the laser at some high frequency. That way you can put a high-pass or band-pass filter after the photodiode, blocking all DC (ambient light) and low frequencies (mains light bulbs). I suspect this is what laser tape measures do, but it isn’t really an option if you’re using a camera semsor with integrated ADC. Maybe you can do it with DSP, but it still assumes that you can modulate the laser in the first place.

>>2532942
SI2301, BSS84.
>inb4 through-hole
There just aren't any, use LCSC's parametric search to see for yourself. Take the SMT-pill.

>>2533043
Well the frequency I will need is over 300 THz, I don't think anything can generate it.

>>2526866
I've got the PS4 HDMI problem. I've purchased replacement ports. I have a decent soldering station and bought a good reflow station. I have some leaded solder, flux, and wick. I've spent many hours trying to get it repaired but I can't D':

The pads are all intact but I can't get a little bit of solder on them to stick for use in reflow. Videos I've watched just use flux then go across with their iron and solder and it sticks well to the pads.

What do I do? I could buy solder paste in a syringe and maybe hot tweasers or a hot needle. Would that help?

>>2533141
Could be that your iron isn't powerful enough or the tip is too small. Maybe your solder is made from goats and the flux is really liquefied roadkill. The motherboard has a large ground plane which sinks heat from your iron faster than it can heat the solder to a molten state. One way to mitigate this is by pre-heating the board from the bottom side while simultaneously desoldering the HDMI port with your iron from above. Another way is to destroy the port bit by bit, and desolder the remaining legs once the port is out of the way.

Happy new year.

>>2533099
Are you talking about modulating the laser or modulating the Rugate filter's electric field?
For modulating the laser, you can just pick some frequency like 1kHz, FFT the data, and just take the amplitude of the 1kHz spike. Your camera sensor probably can't handle 1kHz though, you should pick something significantly below the nyquist frequency of your system. The laser itself's frequency doesn't enter into it, all that matters is you're receiving its signal at a known frequency, so you can ignore the steady-state ambient light and just focus on that one modulation frequency.
For the Rugate filter, the function generator frequency is proportional to, but nowhere near the same as, the light frequency. He uses it to make visible light filters after all.

>>2533141
buy chipquik

>>2533099
>>2533169
another way of looking at it is:
>turn the laser on and off 10 times per second
>sample image when on
>sample image when off
>subtract values
>find point of most contrast
>measure amplitude
What's your metric for measuring distance anyhow? Brightness?

Or consider using a fast pulsed laser for a significant brightness spike to feed into a fast-photodiode that signals a time-to-digital converter. Something like a ruby laser or Nd:YAG, or just overdriving a laser diode if they're rated for it. If you can figure out how to Q-switch a laser diode that would be cool too.

>>2532949
2. you can probably open thr casing of the laser and turn the pot to increase the power. most laser modules are a 50mW laser thats been turned down to reduce the output power
3. ov2640 is a dogshit cmos for anything specialised. you will need to get a different cmos which likely wont be on a devboard.
4. same as above. my suggestion is get a monochrome cmos with high sensitivity in IR. unfortunately most of those fall <900nm wavelength. also for high fps you need a very fast processor to grab each frame from the camera. if you need any sort of resolution this will make it even harder

>>2533043
>inb4 through-hole
Well kinda important if you're using it on bread boards. How would you use an SMD in your circuit if you're not building the PCB? are there SMD mounters or something?

For THT being "deprecated", they're still selling a lot of these components for sure.

>>2533259
>are there SMD mounters or something?

you can buy/build adapter PCBs, but why bother when you have solder and wire.
if a transistor/table/person is missing legs, you can just add 'em surgically. easy.

>>2533259
>are there SMD mounters or something?

you can buy/build adapter PCBs, but why bother when you have solder and wire.
if a transistor/table/person is missing legs, you can just add 'em surgically.
it's SofaKing easy.