File: 1714564207454.png (1.2 MB, 3000x1500)

1.2 MB PNG

I just got an esp32

File: 1693544457793523.jpg (323 KB, 2606x1954)

323 KB JPG

>on a factory tour
>see big cap and ask what they use it for
>"it's a motor brake, that one's faulty"
>say I like electronics and ask if I can have it as a showpiece since it's broken
>Sure! Do you want a working one too?
That's how I got pic related. I was trying to pinpoint the fault on the right one, but I measured capacitance and ESR with a scope and they're within tolerance of the spec sheet. That leaves leak resistance: how can I measure that?

>random devices just disconnect
>on one day it's stable, on another day it's acting like there's a loose connection
>trying multiple cables, nothing makes it reliable
>can't make cables too long because voltage will drop and the protocol can't deal with higher latencies anyway
>have to restart software reading from USB because reconnecting isn't a common concept
God why is USB so shit. Ethernet just works. Even if you disconnect it for a moment because you tripped over the cables, software can deal with it and continues as if nothing happened.

>>2794213
>reconnecting isn't a common concept
what sort of garbage USB device doesn't reconnect when you run around and kick the cable loose.

>>2794224
It does reconnect on the protocol and OS level. But software on layers above that doesn't normally reconnect. You run a terminal on n USB serial adapter? You have to restart it manually. Same problem if you have a temporary USB network interface. If it goes down, you have to reconfigure it unless you configured it as permanent interface.

>>2794230
>You run a terminal on n USB serial adapter? You have to restart it manually. Same problem if you have a temporary USB network interface. If it goes down, you have to reconfigure it unless you configured it as permanent interface.

Imagine that. We truly live in hell.

File: 14544-SparkFun_Current_Se(...).jpg (62 KB, 455x455)

62 KB JPG

I want to measure 110V AC current using a Hall effect sensor, (to calculate power draw), and I want to read the measurement over I2C for logging.
The ACS723 current sensor seems like a good fit and is available as a breakout board (pic) but because it's an AC current, it will output an AC voltage proportional to the current. And I presumably want a DC signal to feed into an ADC to query via I2C. Can I just put the output of the Hall effect sensor through a bridge rectifier with a smoothing capacitor? My understanding is that still wouldn't give a perfectly smooth signal.
There also seem to be "rms to dc converter" ICs that would give a DC output.
I also thought about using the output voltage to drive a 555 timer chip and then use a pulse counter to effectively "integrate" the signal over time, that's probably overcomplicating things though?

tl;dr options for measuring and logging mains power consumption?

File: p3 kill-a-watt.jpg (19 KB, 500x500)

19 KB JPG

>>2794284

>>2794284
Acs712 has a filter that might be good enough. I've also considered inductive current sensors which often will allow you to avoid cutting the mains wire.

>>2794290
Sure, but I want to automatically log energy usage and make graphs, etc. (I should have been more specific and said I want to measure energy use, not power)

>>2794252
USB, like bluetooth, was designed for things like keyboards and mice. With very short cables. Stateless things. Not that we really needed it.
Even odder, is that from a programming standpoint is that it’s a block oriented protocol… it’s hard to see the “serial” part of it unless you’re bit-bashing USB 1.1 with a 1GHz mcu in real-time mode.
Originally, USB was seen as a simplification to the full rs232 25-pin d connector spec (which was hardly ever used) since it has only 4 wires and can carry power (although only 5 V instead of the enlightened firewire). Until the came up with USB-C with a lot more wires and bullshit that is hardly ever used. This will pave the way for a “new” and simplifed “serial” bus after the USB-C tower of babel collapses.

>>2794284
> AC voltage through a smoothing cap
Yep, that will work.
The key is to measure that voltage through a very high impedance buffer like an op-amp or a j-fet input op-amp. This is what is done in a “sample and hold” circuit. That way any ripple is minimized, because the ripple goes up proportionally with the load.

>>2794190
Obviously, just charge them up to 90 V (or whatever) then wait a while—perhaps a month—and measure them again.
If one is leaky it will be at a lower voltage.

>>2794342
… oh and something fascinating… I was trying to get SIMPlink to work on my bluray player yesterday (don’t get me started on hdmi please) and checked the manual to my TV and it can actually use the USB port as a true, rs232 serial port at 115 kB and accept commands from a terminal (or, I guess PCs nowadays over COM1) to control it and to do diagnostics. I assume this is also for interfaces for the disabled, but it was incredibly interesting that it still works that way and thanks to laws and whatnot, it hasn’t disappeared yet.

>>2794346
Thanks for the tip anon. Actually having done some more reading it seems that measuring AC power is not straightforward at all, and to deal with loads that are not purely resistive, I need to take many samples of both voltage and current at different time points to get instantaneous power consumption measurements, and average them. I was hoping to just measure current and multiply it by mains voltage but that is apparently too naive.

>>2794291
I do wonder how accurate those inductive sensors are, wouldn't it depend on the geometry of how the wire passes through them?

>>2793771
update: replaced caps. didnt have 1200uF so used 1000uF instead. voltages of chips appear to be within spec now (also the main logic chip on the board is getting a lot hotter now), but the VCR is still throwing the same error code. hmmmm. chip might be dead.

>>2794368
Remove the one RAM chip that looks scorched and try to power it on.

>>2794342
The USB wire protocol is a serial protocol, no matter how you look at it. Serial doesn't mean "byte wise".
God there's so much dumb shit in the USB protocol. Like how you terminate a transfer. Fuck ZLPs.

>>2794368
the RAM chip looks like bog standard sdram, you could probably replace it with anything from an old videocard (assuming the timings are at least as low as the one on the dead chip).

>>2794376
bro i really do not think that that is the issue

>>2794387
Does that chip look right to you? lol

>>2794390
its a low-light phone picture. the chip looks completely fine. its not getting super hot, and it definitely isnt getting hotter than the identical chip right next to it. its going to take more evidence than a smudged fingerprint to make me suspect that chip.

File: simpleadccircuit.png (9 KB, 376x261)

9 KB PNG

So I have this simple ADC circuit to measure the positions of 2 potentiometers.

The ADC is an ADS1115 which I'm using for differential readings. I'm specifically checking between A0->A1 and A2->A3. Both A0 and A2 are connected to the 5v rail.

I can get readings just fine, but it doesn't seem to be very precise even after discarding outliers and averaging samples. How can I improve this circuit to eliminate as much noise as possible?

File: precision full-wave rectifier.png (46 KB, 632x421)

46 KB PNG

>>2794284
?Can I just put the output of the Hall effect sensor through a bridge rectifier

nope
a diode loses the lower 0.7V of the AC signal
if using a bridge, you lose 1.4V
that screws up your resolution and accuracy
so you'd need a precision rectifier, with op-amps helping out the hapless diodes
(pic is for audio, you'd need bigger caps for 60Hz)

>>2794397
>How can I improve this circuit

you could use ground as a differential input instead of 5V
coz the 5V could be noisy
ground is never noisy coz it's the reference point for everything else

>>2794415

also, caps always help
on the supply and the inputs

>>2794397
put a small (or large) cap across A0 and A1 / A2 and A3.

>>2794392
to me, it looks like the plastic has melted. it does not look like a smudge. the plastic looks like it has deformed

Why does steinmetz say that the idea of the electron charge is harmful to the understanding of the behaviour of electricity in circuits?

>>2794431
tell you what, ill take another look at it tomorrow at work and post a better picture of the chip. who knows, maybe im high. it if is the chip, thats an issue, because im pretty sure chips dont spontaneously melt like that.

>>2794432
Because electrons move from cathode to anode.

>>2794364
> not straightforward
Sure it is.
Send the sampled voltage into your a/d input.
You’ll get some (as far as you know) random number.
Now measure what it actually is with a multimeter or clamp meter.
Repeat that with different current draws until you’re tired of doing it. (You have to do this anyway to test it)
Now, when you get a value you can look up what the current actually is. If it’s in the middle of two values then interpolate. Remember your linear algebra?
Simple. Even works if it’s not linear.
This is how I computed square roots and other things when I went to high scool. In the back of the textbook there is a table of known values.
Simple.

>>2794433
>post a better picture of the chip

dont bother
chip looks fine
it could be dead
but it looks fine
only evidence to the contrary is one bored dude playing a joke on you

>>2794383
> USB is serial
I used to test/reverse engineer USB stuff for a living. I meant that looking through the decoded data stream and looking at the app-level software, you’d think it was “block” oriented. Not even bytes, but blocks of bytes. The massive inefficiency of software devs sending individual bytes in blocks (unions of structs) was driving me crazy. Just a shit protocol made even worse by idiot devs. I had to quit when they started sending XML through for the good of my health.

>>2794432
its hard to say exactly what that author was thinking when he wrote that, but i will say that classical electrodynamics was finished by Maxwell, and Maxwell died before the discovery of the electron. you really dont need electrons to explain how a lot of circuits work, and trying to use classical electrodynamics to explain the behavior of electrons is ill-advised because electrons are quantum particles. its best to treat electrons with the respect theyre due, and leave them out of simple circuits althogether.
>>2794437
anode to cathode, actually

File: 1631271058930.jpg (94 KB, 1280x720)

94 KB JPG

>>2794443
You're right. He should just throw the board in the trash. Problem solved.
>>2794447
>anode to cathode, actually
Wrong.

>>2794447
I'm still trying to figure out what he means by dielectric field. but he seems to view charge as waves not electrons

>>2794446
Most standard USB protocols don't use anything like this. No unions of structs, no XML. But the messy transfer termination semantics make protocols awkward, I've seen things just pad their packets to full length to avoid dealing with short packets and ZLPs.
Anyway, electrically, USB is a serial protocol. What the fuck are you guys smoking?

File: file.png (131 KB, 746x174)

131 KB PNG

>>2794449
huh, whoops.
>>2794452
>I'm still trying to figure out what he means by dielectric field.
post a passage thats confusing you.

>>2794443
>one bored dude
I am not the person who made the original suggestion. But I confirm the surface of the chip looks slightly melted.

Go look at the image again. There is a perfect rectangular deformation right in the middle the size and location of where the actual IC would be.

>>2794440
I don't think that extrapolation is valid because inductive/capacitive loads will shift the voltage/current relative to each other to different extents.
>>2794412
>a diode loses the lower 0.7V of the AC signal
good point

>>2794284
If you care about measuring real power and not apparent power, you'll want to measure the full AC current waveform at at least a few kS/s and do the proper integration. With a good anti-aliasing filter, of course.

If it's just connected to a mostly resistive load then it doesn't matter. If it's connected to a dumb reactive load then high frequency sampling doesn't matter, just assume the waveforms are sinusoidal. If it could be connected to rectifiers or triacs then you need to sample at higher frequencies.

Also consider a current transformer instead of hall-effect, they should be more than accurate enough for AC. Hall effect is basically only used for DC, because it suffers from offset drift.

>>2794459
IIRC in a galvanic cell it's flipped compared to an electrochemical cell. Not to mention the photovoltaic mode compared to the photoconductive mode of a photodiode.

>>2794472
Could just be shrinkage from the epoxy moulding process. Also epoxy doesn't melt, it softens at high temperatures but considering the lack of discolouration around the board I very much doubt that. IC surface finishes vary wildly depending on what they've contacted and where they came from.

>>2794484
>surface of the chip looks slightly melted.
>rectangular deformation right in the middle

>epoxy doesn't melt, it softens at high temperatures
really cool cool story bro

>>2794484
>IC surface finishes vary wildly

another important point: the nose is what gives it away, not the eyes
burnt chips have a very unique and powerful smell
like a Trump fart, a stink nobody can ever forget

>>2794473
>inductive capacitive loads…
Yeah, you’re sampling and holding the rectified and smoothed DC value so it won’t matter.
How fast do you need to update it? Faster than a multimeter?
You ever see how fast an LCR meter works?

Maybe you could really nerd out an show an representative equivalent schematic with the resistive, capacitive, and inductive components with their exact values on an oled screen updated every 60th of a second. Don’t forget about back-emf, Q factor, and pulse delay and reflection to compute the relative distance of the load components in the transmission line.

>>2794502
Mate cos phi is so elementary even those kill-a-watt meters will display it. If you want to measure the power consumption of a simple transformer or ac motor, or something with a capacitive dropper like an LED bulb or kill-a-watt meter, your results will be significantly wrong.

>>2794502
>smoothed DC value
So you just get the root mean square of the current and voltage independently, and multiply them to get the power? That doesn't seem to take into account the parts of the wave where the power is flowing "backwards" i.e. into the mains

File: IMG_9275.jpg (1.64 MB, 4032x3024)

1.64 MB JPG

Wiring some other lights to my cars foglights but don’t wanna cut any wires on it so I’m trying to tap into the relay. What number is the switched output I should be connecting to?

>>2794526
Probably 87. Maybe 30 or 87A if they're idiots. If you get it wrong it will just turn on at the wrong time, it shouldn't blow anything up.

Look up a wiring diagram for your car if you're unsure.

>>2794521
The back-emf tends to have a reducing effect on the voltage across the smoothing capacitor over time.

You can also measure the current with a resistor attached to thermocouple. That’s really smooth.

The reason we measure the AC voltage in RMS is partly due to analog methods used to measure it back in they day.

>>2794545
>The reason we measure the AC voltage in RMS is partly due to analog methods used to measure it back in they day
RMS specifically is used because it's the current/voltage that's produces the same average power output for a resistive load. Non-ohmic loads don't follow this trend.

You can use analog methods to measure real power instead of just measuring apparent power, that's what those spinning disc non-smart power meters do.

>>2794512
This started simply and now we’re doing metrology grade measurements on real and apparent power with arbitrary loads. My clamp-on meter isn’t even that accurate.

Just buy a kill-a-watt meter.

>>2794412
My fluke multimeter uses a circuit like this.

>>2794547
>Just buy a kill-a-watt meter
The anon wants datalogging, to have a long-term log of what draws how much power and when, probably for something like a fridge. Cos phi isn't just a few percent difference, it can easily be over a 50% difference, it's something you really want to take into account when measuring a reactive load, like a fridge.

You can almost certainly buy something like what he's asking for, but when it comes to datalogging I tend to find that what you buy is a lot more expensive than an arduino with an SD card shield added to the non-datalogging version. He could hack into a kill-a-watt meter with such a microcontroller circuit, but doing it from scratch would be a cleaner build. The capacitive dropper inside a kill-a-watt meter isn't going to be able to provide much current for a microcontroller, it may well not be sufficient.

>My clamp-on meter isn’t even that accurate
Phi is just the phase angle between voltage and current. It's not hard to measure, it would be pretty easy with just a couple of zero-crossing detectors, or just do it in software with your ADC readings that you're getting anyhow, but you need to measure voltage and current at the same time for it. A kill-a-watt does measure both at the same time. A portable clamp meter or multimeter does not. Some benchtop meters might.

>>2794550
>>2794512
https://sigrok.org/wiki/Digitek_DT4000ZC

File: 4996B23B-379D-4392-881E-E(...).jpg (583 KB, 970x915)

583 KB JPG

>>2794550
> fridge
As I said before, you send the smoothed voltage from your current measurement to the a/d converter, and calibrate it with an accurate meter. Done.

If you change the load to a chinese 2 W LED light using a capacitive dropper, then recalibrate it for that.

> datalogging on killawat
I happen to have a residential logging meter on my incoming 100 A biphase power using two current transformers. The thing is pretty complicated and uses multiple microcontrollers. It also somehow measures the voltage using those current transformers with great accuracy, but it remains pretty consistent where I am, within 1% measured agains my fluke.

If I wanted to log on my killawatt, I’d set up a webcam on it, snap a picture every second, check if certain pixels were black, and figure out what digits were displayed and log that, then discard the picture. Done.
If you wanted to be really fancy, Tie off the LCD but those things are weird signalling-wise.

>>2794546
> spinning disc, analogue methods
Not necessary. We had this covered a long time ago, as in picrel. The two bands wound in opposite directions carry the current. There’s an outer coil that isn’t shown.
The above logging technique will work on this thing from 1954 as well.

>>2794558
That doesn't measure current and voltage at the same time. Multimeters usually only have one ADC channel, and it doesn't seem like the DT4000ZC is any different. The data transfer seems to be only in one direction, so I don't think you couldn't swap quickly between voltage and current automatically either. Even if you could, the settling time on DMMs is usually much slower than the speed of the AC waveform. My meter has a dual display that can show hertz and AC voltage at the same time, or current value and min/max value, but still can't do voltage and current at once.

>>2794560
>calibrate it with an accurate meter
That's fine if you're testing one appliance over weeks, but if you want to test like 20 different appliances it would be a pain to calibrate for each one. I don't know why you're making it out like power factor is some sort of precision instrumentation thing to measure when power meters do it for cheap. Using a pair of zero-crossing detection circuits alongside a pair of rectifying and averaging circuits would make for an effective circuit for ohmic/reactive loads, which is where nonideal power factor will be the most impactful. You can also just buy the ICs they use in power meters if you know where to look.

>The thing is pretty complicated and uses multiple microcontrollers
Maybe because it's got two modular circuits, one for each phase? Those power meters are overbuilt both for reliability and for the ability for the power company to retroactively start charging for apparent power by communicating to it without your knowledge, anon doesn't need all that junk. Just an MCU with two ADC channels and some passives.

The camera trick works but is more tedious to parse the data, if you do it often enough you'd be better off tapping into the multiplexed LCD and reverse engineering the protocol once. Then you'd just import the data as a CSV and run a quick script to convert it into numbers.

This is /diy/, it sounds like a neat project.

File: 2024-05-02-002740_1350x98(...).png (3.81 MB, 1350x985)

3.81 MB PNG

>>2794593
>That doesn't measure current and voltage at the same time.

File: 2024-05-02-002907_1079x93(...).png (2.9 MB, 1079x936)

2.9 MB PNG

>>2794593

File: 2024-05-02-002948_1064x87(...).png (2.66 MB, 1064x871)

2.66 MB PNG

>>2794593
>>2794598

File: sample rate tp400zc.gif (55 KB, 644x677)

55 KB GIF

>>2794600
>the magic of 2

unfortunately, you need sample rates with millisecond resolution to get accurate results
250 ms is useless unless you could get the 2 meters to somehow synchronize their readings

File: 20240501_175145.jpg (3.25 MB, 4000x3000)

3.25 MB JPG

Do you work on repairs? I spent a few hours trying to diagnose this gaming keyboard. Fucking black traces on black mask is bulllllshiittttt. I also think that this board is multilayer.

Managed to.figure out Q1 and Q2 wasn't getting a 3v supply which feeds the infrared leds.

>>2794709
>How did you break the keyboard?
Liquid was spilled on it which was the initial problem. After that some cat pissed on it. So had to clean all that off and clean the corrosion up. Works perfectly fine now.

>>2794670
i fixed a keyboard i found at a goodwill about a year ago. the problem was a wrong-value resistor. i think the factory just put the wrong resistor on the board. poor guy got a lemon and just gave it to goodwill. its like a $100 keyboard.

>>2794719
Yeah it is really nifty finding stuff like that this TV screen I am working on a crappy little 12 inch but the backlight circuit died. It need the supply capacitors replaced and a new fuse and it worked like a charm. I use it for test equipment

>>2794718
Look at the board under a UV light if you have one. You can see corrosion under the solder mask if you shine it at the right angle. Neutralizing it with a weak acid or base (depending on the pH of the spilled corrosive liquid) will stop it from creeping further through the PCB.

>>2794727
There isn't a lot. The liquid damage was minimal the issue is the vias. They get damages with very minimal moisture that's what caused it to crack. I just can't find the actual line where it's broken. Bit if I get the chance to open it with a uv light I'll check it out.

>>2794709
Election tourists are the worst.

>>2794593
>Using a pair of zero-crossing detection circuits alongside a pair of rectifying and averaging circuits would make for an effective circuit for ohmic/reactive loads
So the zero crossing detectors would give you the phase offset of the current/voltage, is there math that would let you use the offset to calculate the true power draw from the RMS of the current and voltage, though?

>>2794670
>Fucking black traces on black mask is bulllllshiittttt.
basically all black PCBs are BS. Which makes 99% of all computer mainboards over the past several years BS

File: IMG_1267.jpg (2.82 MB, 4032x3024)

2.82 MB JPG

>>2794431
>>2794443
>>2794472
>>2794484
>>2794485
>>2794488

>>2794598
>>2794620
>>2794560
In the question, he was mainly interested in getting the current and datalogging. He doesn’t need to measure the voltage at all if you keep in simple and forget about all the what-ifs. Sure if you were making a product and plan to sell millions of them, then yeah you might want to compensate for different load types, but for all we know he might be able to get by just knowing when it was “on” or “off” and then multiply that by how much power it uses.

> tap the LCD
It’s both an AC signal and multiplexed. Now we need to get into rectification and smoothing again. I’d be done an hour after the first post with a webcam, but cheap hacks are something I specialize in.

>>2794620
> datastream
Nice. The E.U. Should do something useful make a law mandating this like they did with USB-C. I hate USB-C though, that was not useful and Apple poisoned their implementation.
I spend a lot of time postulating turning multiplexed LCD screens into big, non-multiplexed LED screens so I can work in the dark and I don’t need to find a magnifying glass so see some of the tiny shit on them.

File: IMG_1268.jpg (3.16 MB, 4032x3024)

3.16 MB JPG

>>2794751

>>2794751
It’s amazing what the light and angle does.
I see lots of SOTs under the microscope where I can’t read the markings at all, but if I turn it 180 degrees it’s clear as day.

>>2794745
Why hate black PCBs. Yeah, they’re a bit of a meme, but so is hating them.
I think they’d be useful for making front panels out of PCB.
Also, I see a lot of computer cases with clear panels so you can see inside, and there is even LEDs on the fans and stuff so it looks like a little discotheque diorama inside there! Black PCBs look great in there.
And, if you’re buying a fan, might was well get one with LEDs in it if it’s not too much extra cost eh?

>>2794753
>lead-free BGA chip
Reflow, reball, or replace.

>>2794755
yeah, clear side panels, lots of lights, and...

PCBs that are any color other than black

>Why hate black PCBs.
cuz you cant see the traces you dumb fuck

>>2794284
>calculate power draw), and I want to read the measurement over I2C for logging.
If only there were some kind of pre-made board that did exactly this
https://shop.m5stack.com/products/ac-measure-unit-hlw8032

>>2794752
>He doesn’t need to measure the voltage at all if you keep in simple and forget about all the what-ifs
P=IV will only be valid for purely resistive loads

>>2794741
>is there math that would let you use the offset to calculate the true power draw from the RMS of the current and voltage, though?
Yes, though this assumes the load is reactive or ohmic:
>phi = 2π * t_offset * f_mains
>P_apparent = Vrms * Irms
>P_real = P_apparent * cosine( phi )
If you assume that the mains frequency is constant, you could do the entire t_offset -> cos(phi) transform with a single lookup table. Or maybe a polynomial, if that's more your speed. Calculating the mains frequency from your timer implies division, but I'd probably use a lookup table for that also, division with a variable denominator is slow.

Caring about non-sinusoidal current waveforms is less important, but if that does matter then that's when you'd need kHz sampling rates with dual ADC channels sampling both waveforms to integrate the instantaneous power.

>>2794771
Cool product, the HLW8032 chip they're using is 44c each on LCSC. I can't tell if it's just a cos(phi) calculator or if it does a continuous integral though, if it's the former then I guess it's cheaper to ADC sample the waveforms directly than to put them through rectifying circuits.

>>2794818
>Yes, though this assumes the load is reactive or ohmic:
Why is it not valid for capacitive loads?
You are talking about the need to rapidly convert the offset time to a phase angle, but doesn't that remain constant? If a load is stably drawing power then voltage and current RMS, and the time delta between the zero crossings should all remain constant, right?

>>2794769
> can’t see the traces
Why do you need to see the traces?
I guess you’re the guy that can fix bad traces in the middle of a 16-layer motherboard with multiple ground planes.
It’s surprising that a thin layer of black paint completely negates your entire skillset.
I just place a light at right angles and I can see the traces just fine, but I’ve also encountered multilayer boards, where you were unaware of there existence apparently.

>>2794771
> P =IV
In this case, P = IV × powerfactor.
And powerfactor is some number. This is calibrated away by testing it with something that can measure the real power. You can largely just log the A/D values from the current transformer (or whatever) and calculate it later.

> but what about continuously varying loads swinging wildly with arbitrary high capacitance inductance and resistance?
> but waddabout wild loads that change their Z characteristcs at 1 GHz, how do you account for that?
> watabout 3 phase… nay.. n-phase power?
> whatabout arbitrary waveforms with sub femtosecond rise times

>>2794836
>the middle of a 16-layer motherboard
ahh yes, where the majority of physical damage occurs

you fuckin idiot

>>2794846
Show us this unseeable physical damage on a black pcb. Maybe we can help.
We got a guy that can spot epoxy moulding anomalies and shrinkage based on lousy photo, so this should be easy. Also the IC was black! You wouln’t be able to see it.

>>2794842
but the power factor will be unique for each different load!

>>2794836
Most damage happens on the surface. Also seeing the trace and the direction let's me see where it is headed maybe I'll see where it broke along the way. Instead I have to blindly probe until I head a beep because it isn't like they gave me a schematic. Why the hell does it even need to be black. It doesn't give you extra angry pixies it's just a pain in the ass to service. Not trying to intentionally make me job harder.

>>2794853
I’ve only seen black PCBs in videos from people complaining about it.
Dave jones is one. Ironically, that’s the first time I heard about making faceplates out of black PCBs—same guy.
I don’t own or have been sent any black PCBs, I think it was popular with people making those “open” devices where there’s no case (just 4 stand-offs) and possibly a piece of clear acrylic over top of that, also with stand-offs. I don’t like my stuff covered in dust, so I tend to put it in cases, where a black PCB would be absolutely pointless.
But there’s no doubt it looks great if you can see the PCB.

>>2794855
>>2794848
>looks matters more than functionality
youre either gay, female, or retarded

>>2794849
> power factor is different for each load!
Yesss! Exactly… I’m making progress here.

Now, if the guy wants to know what his beer fridge is costing him every year, it’s gonna be fine.

If he’s designing a whole-house monitor for commercial sale, then yeah… more complicated things will be needed.

I have several of these kill-a-watts and clones, but I never really used them long term. Some of them tabulate the kWh energy use over time and compute the $ amount. The problem is my power floats so it’s radically different from month to month. So usually I just throw it on an unknown device for a month, and figure out what it costs me per year.
Most appliances come with cost rating around here (energuide) so I just trust that.

>>2794858
It does to the people that buy them. If you hate black PCBs so much, why do you insist on buying things that use them? Furthermore, if you are knowingly in possession of things with black PCBs you should take extra care to protect them from physical damage.

File: 115F2AA7-85E5-4502-995D-9(...).jpg (448 KB, 693x788)

448 KB JPG

>>2794858
Ahh, you’re this guy. I can see the traces and the texture of the fiberglass mesh inside the epoxy PCB. It’s even easier to see the traces on this board than with most other boards.
The E.U. should mandate black PCBs everywhere for ease of maintenance and repairability.
> gaming keyboard
Ahh. That explains everything.

>>2794867
No that's not me. I'm the one who fixed the keyboard. This one was a gift from a friend.

>>2794867
> It’s even easier to see the traces on this board than with most other boards.
made up bullshit

> gaming keyboard
not me, but way to show your retardation

>>2794869
When you look down on the keyboard, through the keys, can you see the PCB? If so, that’s likely why they made it black.

At least it’s not gloss black, that would be worse and difficult to take a photo of it due to glare.
On the plus side, the silk screen is very readable.

>>2794828
>Why is it not valid for capacitive loads?
It is valid, capacitive loads are a type of reactive load.
>doesn't that remain constant?
Power factor may change in the 10-0.1s timespan, for example as the load on an electric motor changes. For sinusoidal current, you don't need to measure faster than each mains cycle (50/60Hz), which is trivially slow for a microcontroller anyhow. That kind of 0.1s time resolution isn't useful for knowing the cost of electricity, just for the finer things of inverter/transformer sizing or troubleshooting startup transients.

For non-sinusoidal current you're not measuring just phase angle, but the shape of the current waveform, since things with rectifiers will only draw short spikes of current at the peaks of the sinusoid.

>>2794878
>At least it’s not gloss black, that would be worse and difficult to take a photo of it due to glare.
On the plus side, the silk screen is very readable.
The silkscreen isn't but the traces are not. The photo doesn't really show what it really looked like to me working on it. Either way I was able to see enough to do what I needed. The board matches the metal panel paint but you really can't see it from the top. Not with the keycaps in place and the buttons themselves. I think it is just branding jank mostly.

did you know that you could easily solder a bigger SMD part onto a smaller footprint and vice versa? 0805 on 0603? no problem. 0402 on 0603? or 0604 on 0805? no problem. it is probably possible to fit 0402 onto a 0805 footprint as well, with a little bit of patience. so, always use M-sized footprints, they are not taking up much more space than N or S and there's a lot more leeway that way.

File: fx32.jpg (82 KB, 899x1599)

82 KB JPG

Is it possible for there to be too much current if the Voltage is correct? The current output on this power supply is 1A and it's going into the battery holder of this DOD FX-32 pedal.

I found the schematics but don't understand any of it.V
https://www.freestompboxes.org/viewtopic.php?t=8570

>>2794970
Yes, too much current is a thing.
The reason is that the circuit might depend on the internal resistance of the battery.
Coin-cell driven devices and LED flashlights are notorious for this.
Best to limit the current with a resistor

File: schem.png (326 KB, 1688x1242)

326 KB PNG

>>2794970
>schematics
It's transistors and op-amps, there's nothing there that will die if you accidentally give it 9.5V. Heck, 9V batteries can give out an amp if you really push them for a short period of time, and they're like 9.6V when full anyhow. The whole thing would probably run just fine up at 12V even, though depending on the circuit it might sound different.

That said, your USB-to-9V adapter is probably noisy as shit. I'd add a capacitance multiplier, or even use a USB-to-12V adapter alongside a 9V regulating circuit. I prefer one that uses a TL431 and a pass transistor, but to each this own. A 7809 won't have much ripple rejection in the hundreds of kHz.

>>2794970
>>2794990
fyi appliances that run on a constant voltage will only draw the current they need, assuming nonshit design

>>2794990
ooh, a cd4007 being used for analog stuff

>>2794995
> 4007
Not surprising.
Most effects pedals are failed attempts at filters, tone controls, or preamps and put together by tinkerers without any understanding of electricity, like “circuit bending” — sometimes they go through 100 transistors to find that one with the “magic” quality they’re looking for. Of course, the result is that it turns the output into a square-ish wave every time.
All this sounds a lot better when you’re high as a kite.

At what point would you say someone is ready for a project?
I already finished up studied power transfer/RLC circuits/frequency response and am on Laplace transforms right now, but I still don't know what transistors do and that Sedra Smith book is daunting.

Would you say going through half of S&S is good enough for side projects?

>>2795199
>S&S
is that control theory? youre ready.

>>2795200
No, microelectronic theory going in depth on pn junctions, op amps, and mosfests/bits and their common configs

>>2795201
what does it stand for?

>>2795201
mosfets/bjts* Jesus christ

>>2795201
oh, Sedra Smith lol. i thought it was "signals and systems" which is what my school called control therory.
just start bro, youll learn with your hands.

>>2795199
I knew none of that when i made my first blinker lol just quit being a bitch

>>2795199
I made my first electronics project when I was a little kid, with a few nails, some wire, a board, and a piece of steel strapping, a home made capacitor out of tinfoil and wax paper I got from my mom. To flash a light bulb. LEDs were not invented yet.
I remember my dad giving me a relay about a year after that. I’m like “holy crap… someone already thought of this and did a way better job”

Of course, I did not amount to very much despite such early promise.

>>2795211
Exactly.

You didn’t mention what the project was, practical advice cannot be offered. Are you going to hit us up for project ideas? I just gave you one.

>>2795125
I’d agree with you a lot of the time, but this circuit uses series pass JFETs, and I refuse to believe someone can do that without at least some clue.

CD4007s are the best way to get 4-terminal MOSFETs, they’re great.

>>2795199
To design custom circuits for a given category, you need both an understanding of theory, and experience building and troubleshooting existing circuits. Just one or the other won’t be enough. Get your hands dirty with some simple projects, the OP has a few in its GitHub repo. I made an elektrosluch for my first project, it was fun.

>fix my 3d printer with a new rotary encoder and a bodge wire to repair a trace that i mistakenly cut like a retard
>fix a smps i had laying around
i am unstoppable

File: Screenshot 2024-05-04 091153.jpg (58 KB, 299x608)

58 KB JPG

Does it make sense to get one of these miscellaneous grab-bags of components if I am just starting out? I don't want to have to specifically order every resistor I'm gonna use. Or will I just end up with a bunch of useless crap?
The general plan is to get a breadboard and work through some simple starter projects

>>2795412
those are fine for beginner stuff. chinkshit passives are acceptable for general use provided:

1) you aren't pushing a part to its absolute limit ("50V" chinkshit capacitors might not actually be 50V, "1/4W" chinkshit resistors are likely not capable of dissipating 1/4W without blowing up) or

2) you need guaranteed characteristics that you see on a comparable data sheet which are obviously not characterized on a chinkshit part which has no data sheet (minimum zener current on a chinkshit 1N47xx might not be anywhere near the minimum zener current on a genuine 1N47xx part.)

i don't really trust chinkshit ICs at all, especially things like op amps or linear regulators.

>>2795420
You can always test them.
I think that’s a worthwhile “first activity” when just starting out. You’ll learn a lot just from doing that.
Test sealed electrolytics outside though.

File: 1714852525653.jpg (3.77 MB, 4000x3000)

3.77 MB JPG

Sorry if this isn't the right place, but I feel like I'm losing my fucking mind. I cannot for the life of me find this LED module. There are a million with three screw holes but these two screw versions seemingly don't exist. I've searched all the numbers on the board that I can see and nothing has come of it. Please, please, PLEASE can somebody save me from this hell. It's the bathroom light so currently I'm showering in the dark which is a giant pain in the arse.

>>2795558
It might be proprietary if you can’t find it on alibay or wherever. See if you can find manufacturer markings on it or the fixture, or even contact the people who renovated the place to see if they know. Worst case you just have to replace the fixture, it’s probably easy enough to get one that will fit there seamlessly, though whether it matches with the others in the room is a seperate matter.

Personally I’m quite happy with my Edison screw LED bulbs that die every three years.

File: cheapo ceiling lamps.jpg (147 KB, 1122x645)

147 KB JPG

>>2795558
>LED module

it's silly to be looking for a module
you need to purchase the whole lamp thing, including the housing
then you get a guaranteed fit, along with standard ceiling mounting points
suggest you get one where you can choose the light temperature
experiment to see which one makes you naked butt look best

>>2794153
everytime i try to desolder components on a nice manufactured PCB, it takes forever. the solder never comes out all at the same time, and i have to take 10mins poking each side of the leg and removing little bits of solder as i go and eventually it comes out.

using 300~350degrees on my iron, keep refluxing, even retinning the point before beginning to try help, and I dont know what im doing wrong. What can I do to make this go faster?

File: desoldering needles.jpg (85 KB, 1276x529)

85 KB JPG

>>2795678
>What can I do

watch and follow all 10,000 desoldering videos on youtube
if still failing, invest in some needles

>>2795678
Don't try and desolder components by taking all the solder off the pads. Just liquify the solder and pull the component off with tweezers.

If you're working with ICs use a hot air gun or short all the pads with silver solder then melt and pull off. If you're working with through-hole parts use a vacuum desoldering pump or snip the leads, liquify, and pull the remaining lead out of the hole with needle-nose pliers or tweezers.

>>2795681

also chipquik, which is usu for chips but ok for other parts
https://www.youtube.com/watch?v=7kyaz4Zrd78

>>2795678
Get a Yihua 929D, if not a proper desoldering iron with a vacuum pump and temperature control. They work great, if a tad tedious.

There’s also the strategy of running thick solid core wire next to all the pads so you can reflow them all at once, but it’s eve more of a pain. Those extra wide soldering tips might be good for this.

>>2795681
These aren’t that reliable, often there’s just not enough room to fit them between the pad and the lead.

>>2795678
melt the solder then quickly whack the board. the whole blob will come flying out, and you can just pluck the leg off of the via.

>>2795667
Now you need to move into a new house now if one of the lights goes out? Is that where we’re headed? Never buy these goddammed things.

Now that the chinese are making LEDs and chips, you’re going to have to come to terms with the fact that there’s a bunch of things that you’ll never be able to buy, and chips you’ll never be able to find specs for.

You might be able to replace the driver chip, the bridge rectifier, or most likely that capacitor spilled it’s guts it looks like.

> how do you know it’s chinese?
I can tell by that font they used on the gloss-white PCB.

>>2795740
>Never buy these goddammed things

they're supposed to last a decade, and the lack of replaceable bulbs makes them look modern and thin
you'll never get laid if you have fat fixtures in your house
good news is what often happens is one of the series LEDs burns out so all you gotta do is find it and short it out, and the rest will come on
this assumes you're not a pussy coz it means using tweezers on a live circuit while standing barefoot on a wet floor
i've done this a lot so i've acquired immunity

>>2795725
>>2795683
>>2795683


thats the thing - and im mainly talking through-hole, surface mounts arent an issue - i can never melt all the solder at once. if i turn the temp up, i scorch the board and lift the 'ring pad' or whatever its called

>>2795681
hmm needles might work. i have a sharp cone, and i could use that. my go-to has always been the wedge shape

>>2795678
are you adding new solder before you try and desolder? helps immensely

anyone have a preferred brand for high quality stranded copper wire with soft, flexible silicone that is rated to 600V or 1000V? something like the quality of wire you would see in a nice set of DMM probes. tired of buying chinkshitted garbage wire. i need 14 AWG and 16 AWG right now.

File: 1703806184828304.png (15 KB, 492x252)

15 KB PNG

i'm having a hard time understanding aliasing. an adc samples at fs. the input is pre-filtered as shown. is the aliased content between fs/2 and fs present in the adc output or just the aliased content below fs/2?

I'm looking to drive a series of relays via a microcontroller. I know I could use an individual discrete transistor to drive current through each relay, but ideally I'd like a single multi-pinned component. No relay will ever be energized at the same time as another one, so I was wondering if a mux could work, but also am not sure if general 3v3 logic level muxes would be suitable for handling enough current to drive a relay.
Is there a particular word for an IC that solves my problem, or will I just need to scan through datasheets to just find a mux that will handle the current? Or more likely just be lazy and use BJTs anyway.

File: 1714893391686.jpg (71 KB, 1130x230)

71 KB JPG

>>2795882

File: 1714893537666.jpg (69 KB, 703x518)

69 KB JPG

>>2795891

>>2795877
when you adc you can only read at a certain frequency
the alias never existed, but because your sample frequency, you cant say for certain that it never existed.
so you filter it to remove anything higher than the adc can sample before it hits the adc
and then you can say with confidence that now only the signal below f/2 remains.

in the diagram the tail of the alias is due to the input above fs/2, by filtering at fs/2 you will filter the alias below fs/2.
(in practice filters are not hard edged....)

>>2795882
well logic level stuff isn't going to work if you need something with a transducer driver stage
maybe a stepper motor driver could be abused to do this?
a matrix of sink/source transistors seems more elegant to me

if I short 120v wires with a 250v capacitor, will that trip a breaker, or just charge/discharge the capacitor slowly?

File: 1714894772502.jpg (136 KB, 675x732)

136 KB JPG

>>2795891
>>2795893

also

>>2795901
depends on how big the capacitor is, but it probably won't trip the breaker. you could always add a current limiting resistor. or use a DC boost converter instead of mains.

File: 1690593870315016.jpg (70 KB, 506x432)

70 KB JPG

>>2795901
are you talking about 120vac?

>>2795908
yes

>>2795911
if it's a polarized capacitor it will explode
if not, it won't charge/discharge slowly, it will do so at 60Hz and whether it'll trip a breaker depends on the capacitance

>>2795901
thats not “shorting.”
I=C dV/dt, so the RMS current would be C * 120 * 60 * 2pi. tripping a 20A breaker would take a minimum capacitance of about 450uF, assuming nothing else is connected to the breaker.

File: simpleadccircuit.png (8 KB, 616x194)

8 KB PNG

>>2794397
Alright, I've simplified my circuit so that it looks like this (not showing the irrelevant I2C and A2/A3 pins).

I've attached the differential to ground instead of the 5V rail and I've added both an electrolytic capacitor and smaller ceramic capacitor to smooth out the noise.

File: signal.png (65 KB, 1435x744)

65 KB PNG

>>2796013
And yet the noise was not reduced that much. Here's the values I'm getting from the ADC across A0 and A1.

The range of values varies by ~100 points from min to max. Any ideas on how to reduce that further?

File: signal_nocap.png (64 KB, 1447x718)

64 KB PNG

>>2796014
Here's the signal without the capacitors. Either I'm using them incorrectly or they are not really doing anything at all.

>>2796013
Use 0.1uF (100nF) instead of 10uF.

>>2795767
>i have a sharp cone

i dont think you got the point about desoldering needles
they're hollow
you heat the joint, then shove a needle into it, placing a barrier between the pin and the PCB

regular (non-hollow) needles are only useful in cleaning out holes filled with solder, when wick or solder pump dont work

File: signal_100nF.png (58 KB, 1426x734)

58 KB PNG

>>2796016
Still seems to range between 25600 -> 25700

Would using smaller ones be better? I was under the impression larger ones reduced noise better but slow the signal's responsiveness. I have ceramic capacitors going all the way down to 10pF though.

File: twisted pair.jpg (21 KB, 677x622)

21 KB JPG

>>2796022
>Would using smaller ones be better?
It's good to use small and large value caps together to filter out both low and high frequency noise. Try twisting your i2c cables together, and your power leads too.

>>2796026
Thanks, twisting the cables seems to have helped a bit. The range seems to have been reduced from ~100 values to ~70.

The sample rate of the ADC is ~860 samples per second, and out of 500 samples I am seeing around ~19 of those large peaks. So that means this weird signal is ~32Hz.

I guess the question is how do I change the circuit to remove that 32Hz signal. That's the major one I need to take care of. Or maybe I should do it in software?

File: signal_twisted.png (60 KB, 1437x754)

60 KB PNG

>>2796040
Forgot image.

>>2796040
>>2796041
If you're using a breadboard, transfer the circuit to perfboard and solder everything. Keep your wires as short as possible and make the component layout nice and tight. Keep the circuit away from other electronics or put it in a metal case and use shielded wire.

>>2795877
The input waves that are above fs/2 will be interpreted by your reconstruction/anti-imaging filter (the thing that smooths out the stair-step samples into a smooth signal, be it in software or after a DAC) as being a lower frequency than it was before sampling. In your case, anything in the original signal at f=fs/2+3 will appear instead at f=fs/2-3 instead. The problem being if you look into your samples and see a wave with a frequency of fs/2-3, you don't know whether it was a real signal at fs/2-3 or if it was originally at fs/2+3, it's ambiguous. If you have important information at fs/2-3, it will be combined with any unwanted information at fs/2+3 and it's impossible to seperate them.

Your anti-aliasing filter will have decreased its amplitude somewhat, ideally you want this filter to drop any aliased wave's amplitude to below LSB/2, but if you don't care about what happens in this region then you could choose to ignore it. Like maybe you'll get aliased shit down to fs/2 - 8, but the signal you care about in your passband only goes up to fs/2 - 10. It's also an option to use a digital filter after sampling to cut out what's above the passband much more strongly, again so long as the alias doesn't enter too close to the passband.

>>2796013
Aren't all ADC readings relative to the 0V rail anyhow? If the fluctuation is the issue, average your samples. If you want your sampled wave to be accurate, then get a proper voltage reference. Consider synchronising your samples with the source of the oscillations (e.g. mains frequency). There's also some good low-noise practice to be done, like putting your MCU to sleep when the samples are being taken.

>>2796040
>32Hz
Are you sure it's not 50/60Hz, due to your actual sample rate being different? Well it probably wouldn't be a lower if you calculated that based on its maximum frequency. I'd mess about with a high impedance scope probe (maybe a JFET amp too) to see if you can sense it anywhere else.

>>2795891
>>2795893
>>2795905
Pretty much exactly what I was looking for, thanks!

>>2794153
anyone here ever built a diy head tracker for flight sim? I have a bunch of wii parts and some ir leds.

File: kinect studio.gif (11 KB, 759x153)

11 KB GIF

>>2796163
>head tracker

probably wanna look into kinect studio instead
you can get fully functional hardware for less than $10 used
and a free SDK lets you directly interface with your windows software

File: ResizedImage_2024-05-06_1(...).jpg (757 KB, 1568x1176)

757 KB JPG

It's me, hard drive fag. Just replaced the SMD fuse and noticed that those diodes (forgot what they're called but I think that they are supposed to short the input as soon as a certain voltage is exceeded, which it did because I plugged in a 20v laptop power brick) permanently conduct.
Aparrently the overvoltage also took them out.

File: DSC_2471.jpg (2.89 MB, 4000x3000)

2.89 MB JPG

Could I just replace them with the diodes from this hard drive?

>>2796215
>replace them

if they're just overvoltage protection, then you dont need them, coz you already learned your lesson

>>2796214
TVS diodes? I guess it's a good idea for them to fail short, as opposed to failing open. Probably shoulda checked for a short to ground AFTER the fuse before you ordered the parts, eh?

>>2796215
As the other anon says, they're not necessary for running, but unless you really trust the PSU and whatever surge protector you're presumably running it on, I'd want to keep the some TVS diodes.

If the part numbers are the same then go for it, if the part numbers differ then I'd want to find the datasheets (using an SMD codebook) so I could verify that they're compatible. Worst case you just trace the circuit to see that they're in the same circuit topology.

File: DSC_2472.jpg (2.17 MB, 4000x3000)

2.17 MB JPG

>>2796222
Was fucking hell to desolder. There's a big ground plane that sucked up all the heat and it took around 5mina per diode. The SMD package doesn't fit. For now I'll just use the hard drive without TVS diodes. I certainly won't make the same mistake twice. I don't even know if there isn't something else on the board that got fried so I should verify that it works in the first place.

>>2794153
>buy high quality multimeter probes, very sharp tip
>drop one of the probes on the floor like a retard
>tip bends
>very lightly grind it down with a 6000 grit whetstone
>immediately removes a large portion of the gold plating

why does literally no one make a set of high quality probes that are just 8/32" threads, so we can buy 8/32" needle tips in bulk and swap them out when they wear out?

>>2796230
> gold plating
Don’t worry, they have gold plating down to 1 atom thick, so it’s not like the plating was compromised.
I can see most of my recent 1/4" and 3.5mm jacks with gold plating wear off quickly and noticeably (definitely within a year).
I’ve found that nickel plating suffices.

> replaceable tips
They used to have these, I remember them from the 80s. We’ve regressed!

Generally, they’re going to get dull even without dropping them. If you have a power supply, some nickel, and some acid you can re-plate them with nickel at home easily enough. Useful for other things too.

>>2796225
Put a resistor in series with the power to limit the current if running without TVS diodes.
What’s a hard drive wattage? Like less than 1W?

Thise TVS diodes also limit spikes from various sources, so never insert an ac wall plug to turn on the device due to arcing. Always use a switch, or, even better, some kind of soft-start mechanism.
I don’t think the hard drive connector is “hot pluggable”

>>2796041
Yeah, I’m guessing this is actually 60 Hz hum.
Let me guess, the potentiometer shafts are aluminum and you’re turning them by hand.
You’re supposed to put plastic knobs on them.

You can probably hear this stuff if you have a headphone or amplified computer speakers.

Add the smoothing capacitors as close to the ADC pins as possible. An inductor in series or a big ferrite clamp-on wouldn’t hurt either.

>>2796288
.............A
HYDROCHLORIC
..............I
.............D

File: signal.png (63 KB, 1434x768)

63 KB PNG

>>2796055
> Are you sure it's not 50/60Hz, due to your actual sample rate being different?
>>2796295
> Yeah, I’m guessing this is actually 60 Hz hum.
I'm not so sure. I measured the time it takes for 860 samples and I get a very consistent 1076ms, which is only slightly longer than expected.

Within 500 samples (the total number of samples in each image I posted), that is ~625ms that have passed. If I am seeing 19 peaks in the signal over those 625ms then that would be 32-33Hz, right?

> Let me guess, the potentiometer shafts are aluminum and you’re turning them by hand. You’re supposed to put plastic knobs on them.
I'm turning the plastic knobs on them by hand, yes. Though I get the same signal issue when I bypass the potentiometer completely so I don't think that's the issue.

File: laptoppower.png (114 KB, 2247x1366)

114 KB PNG

>>2796361
Okay, I think it might be my computer's power supply doing something odd. I tried powering the board from my laptop and the signal is gone and I'm measuring a range of ~10 values.

>>2796361
I’ve never encountered any technology that samples at under 1 kHz. I’m not sure anything can be learned from that. Even your laptop mic in can probably do at least 44 kHz. Try that and look at it on audacity, lol!
Then do an FFT on it so we can see the spectrum.
I used to do this to clean up bad VHS and Betamax recordings with annoying background noises in them. VLC can do it too.

>>2796393
> I’ve never encountered any technology that samples at under 1 kHz.
The ADS1115 is an ADC that provides 16-bit samples at ~860 times per second in continuous mode.

> I’m not sure anything can be learned from that. Even your laptop mic in can probably do at least 44 kHz. Try that and look at it on audacity, lol! Then do an FFT on it so we can see the spectrum.
I'm not trying to measure audio or do any sort of spectrum processing. I just want to measure the outputs of some actuators to keep them in sync.

>>2796425
> ADS1115
Yeah, it’s for DC voltages (as the datasheet says, battery voltage monitoring)
Not generally useful for determining the frequency components of noise from a switching power supply that operates in the kHz range, or 10s of kHz most likely.
You can definitely hear 60 and 32 Hz, that’s audio range. Put it through a capacitor in series to remove the DC component.
BTW… and I’m sure you have… considered rotary encoders?

>>2796493
Are you implying it was aliasing a much higher frequency? Definitely possible, I think by changing the sample rate a small amount he'd see a significant difference in the noise frequency. The moral of the story is to use a multipole anti-aliasing filter.

>>2796519
Or random sampling. After a few minutes or, perhaps hours at that sampling speed… should be able to get a pretty complete waveform.
My oscilloscope does this to get insanely high speed periodic waveforms.
It’s literally like watching the wave pattern slowly materialize while shooting single electrons through a double slit.

>>2796493
> BTW… and I’m sure you have… considered rotary encoders?
The linear actuators I got have potentiometers built into them. There's newer ones that use some hall effect sensors that act as encoders, but getting a set of those would be a bit expensive.

>>2796540
Huh, that's pretty neat.

File: file.png (42 KB, 457x428)

42 KB PNG

>19v 65VA trafo = ~25V DC Full Bridge
Is this getting "toasty"?

>>2794153
guys I attempted to get into serious electrician job training and I failed after 3 months as I wasn't quick enough with learning and fell behind the class. Also the teachers were bastards :(

>>2796665
What were you having trouble with?

>>2796671
Rearranging equations was my biggest problem, followed by calculating with octal and hexadecimal numeral systems and remembering some rules for how you calculate common components in circuits. It was very rushed and I wasn't quick enough and lacked the time to learn due to long commute to work and school. Also suffered autism related problems in class. But I still like electronics

>>2796674
Keep studying and trying. The only way to fail is to give up. Seek a mentor for guidance and advice, and practice solving problems on your own time. Try again when you feel that you're ready. You might have to come to this conclusion through denial and error but I swear on my mother's mating name that when you put the petal to the medal you will pass with flying carpets like it’s a peach of cake.

>>2796665
The biggest redpill for me was that it you can't rely on teachers to actually teach you things, or even give you an idea of what you need to learn. It is your duty to teach yourself. As the other anon said, a mentor that you can actually trust to have your education in his/her interest is very beneficial, if you can acquire such a person. But for people, time is money, and the cheapest time is your own.

>>2796674
>Rearranging equations
Do an online course or whatever on algebra. The triangle method is useful. A foundation in mathematics is what a lot of stuff is built onto, and I found getting into a course that implies mathematical knowledge that you don't have is a recipe for failure.
>octal and hexadecimal
Huh? Electricians need to know non-decimal number bases? Isn't that only ever used for computer engineering?
>commute
If you're on a bus or train, get yourself a workflow that you can do extra work on the bus, be it a laptop or just pen and paper. If you're driving, there isn't much you can do, maybe listen to audio lessons, or listen to podcasts or whatever that you'd normally be doing at home. A shitty commute can really kill your motivation, I guess consider living closer to the learning institute if possible. Otherwise, teaching yourself the rough outline of the course, and the foundational mathematics, should get yourself into a place where you can take the course while keeping up.
>autism
Is a superpower, at least outside of group projects. You just have to understand it and learn how to direct your attention properly.

>>2796783
> octal for computer engineering
Even octal isn’t really used for computer engineering, it died out in the 80s. Digital isn’t making many new PDP11s anymore. There’s just some vestigial uses for it nowadays.

>LMP7721 3-Femtoampere Input Bias Current Precision Amplifier

Why use such exotic OA if whatever device built with it could be calibrated to account for the offset current? Just wire it as a unity buffer and attach a voltage source and a 10Meg resistor in series to V+ , and measure Vout. That will be your offset, including input voltage offset as well. Any complications with this approach?

>>2796909
Mass production.
Op amp itself, the IC, and SMD are all products of mass production and getting costs down.
Easier to buy the part than have some kind of laser trimmer on your line.

Ever seen them adjust those TV tuners found in your atari for channel 3? There’s someone that has to take a plastic screwdriver and pry apart the little coils until they have the right properties, and then put wax over it (sometimes there is a little sponge inside the coil to help hold it) to keep it in position.

>>2796674
I don't know why you'd ever need hex numbers, but I wish I learned them early enough, preferably at the same time as the multiplication table. Since we know it since childhood, we never have to calculate 7x6, everyone immediately knows it is 42. Same thing with the hex. Ideally you just look at 0x1000 and you just know it is 4k. I periodically forget and have to re-program my memory. I typical think of each position as 4 (since 16=2^4), and sum them so in this case it is 000 => 4+4+4 = 12 so it is 2^12 = 4k. But like I said the idea is not even to think about that and just know that 0x10000 is 64k etc. Of course in my case the purpose is not to convert any number, but just know the ballpark when figuring out memory requirements etc, or for example for microcontrollers, say if SRAM is 20k, and I know that 4k is 0x1000, 20k = 0x5000. etc.

>>2796918
I guess another reason is, as Dave mentioned in a video about his nanoamp meter, is that so it would just work, without any calibration, since most people don't have precise equipment that is needed for calibration. Which is funny, since that begs a philosophical question: what are you trying to achieve? Why would you even need a precise nanometer if you cannot verify how precise it is? Their metrology threads is pure autism but still a very interesting read.

>>2796678
Thank you anon, that's very inspiring. Now about to begin a new job training in public administration but I might come back to electronics in the future. At a minimum I'll keep working and continue learning about it as a hobby nevertheless because it's such an interesting and practical field.
>>2796783
Thank you aswell for your kind words and advice. I was always told "who writes is who stays" meaning that if I always listen and just write things down I wouldn't have to worry lol but I'll have to swallow that red pill too now. Good idea regarding online courses, I will check out that triangle method and I'll look consider how I can improve my workflow.
>Huh? Electricians need to know non-decimal number bases?
Oh then it was because of the IT part of the training, as the job was a mix of electronics and IT. I wasn't sure if all electronics technicians have to learn this. Exact job title: electronics technician for devices and systems.
>You just have to understand it and learn how to direct your attention properly.
ig you're right. From experience I can focus and learn very well if I am in an adequate sized class with not much output stimuli. During job training I was in a big hall with at least 50 other trainees and when we changed into the school rooms it was too cramped, loud and smelly (body odours)

>>2796930
>new job training in public administration

good ridance
last thing we need in the profession is another lunk who doesnt even count their change in hex
doesnt even own a binary clock

do you know if coin battery holders only differ by diameter regardless of the thickness, 2016/2032 should both fit in the same holder due to the spring action?

>>2797091
2016 might be a tad too thin to make reliable contact, but 2025 works in a standard 2032 socket.

>>2797091
not unless you use a quarter as a shim

>at work our old 3D printer, a Zortrax M200, just died without a clear reason
>I volunteered to try a quick diagnosis, it was a slow week anyway
>black silkscreen everywhere, multiple mysterious daughter boards, all main chips with grinded out markings
>ALL issues on the Zortrax forum were solved privately
>basically no documentation anywhere
>get filled with impotent rage
Even if I were an EE (I'm mech) I doubt I could have done anything in a reasonable timeframe. This is relevant because I was considering a Bambu P1S, as I have little free time and I want a printer that just works out of the box without fiddling, but I read it's an Apple-like kind of product and I don't want to find myself with a brick I can't repair like it happened with the Zortrax. What would you anons recommend?

>>2797150
Did you check to see if any fuses were blown?
Are the power rails good?
Sounds like it’s fixable, unlike a sony TV where all errors are the TCON board with one big chip on it which is unobtainium.

>>2797152
Yeah, power rails were good, no clear sign of a short, no blown caps, and the silkscreen made understanding the topology a bit hard for my untrained eye. Besides, it happened weeks ago.

File: Vos.jpg (20 KB, 451x422)

20 KB JPG

How does this circuit work? I can kinda guess that gain=1000 is 100k divided by 100ohm, but why is 100 ohm resistor connected across the inputs?? never seen this before. shouldn't the opamp try to make V- = V+? but this is not possible since there will be a voltage drop across the resistor! wouldn't that drive the OA crazy??? and why are there two 50k resistors on both ends? why not make it a standard voltage divider 100k/100???

>>2797150
>I just realized this is /ohm/ and not /3dpg/
I'm retarded.

>>2797223
Don't worry, you fit right in.

File: C16191CF-4158-40F7-AE56-6(...).jpg (105 KB, 904x427)

105 KB JPG

>>2797171
That looks like the “offset drift test circuit” and is probably not useful.
They put the thing in an oven, turn it on, and the offset drift due to thermal changes is compensated by a nulling loop running as 330 Hz, so it returns to zero in 3 minutes or so. They recommend enclosing it.
There’s also an “air flow detector” circuit that amplifies the thermal drift, but normally you’d enclose the device to protect it from air currents.
Not a generally useful circuit, so that explains why you don’t understand it. It outputs 0 V.
Also in the datasheet, they made typographic error for tin oxide resistors.

>>2797170
> power rails good…
Yeah, unfortunately I have the same problem. I’ve seen 100s of youtube videos with people fixing all sorts of shit and it’s almost always the power or caps, or fuse, etc. I never get anything that easy.
I think there’s 100,000 unposted videos where they couldn’t fix it.
In your case, without more information, we might as well assume it’s something like the “service engine soon” light in a BMW… a logic bomb or runtime counter that has to be reset by someone with the private key.

>>2797091
Coin cells are evil.
Hook up a AA holder to whatever it was that uses the coin cell, and either glue the thing to the hilder if it’s small, or glue the holder in or on the thing if it’s big.

>>2797327
>Also in the datasheet, they made typographic error for tin oxide resistors.
i hope someone got fired for that blunder.
i once was reading a datasheet for an MCU and one of the address registers was just completely wrong. the only hint was that the address space overlapped the next one if you did the math. only remedy was pulling up a more recent version of the document. very insidious.

File: 7D78A913-DBA7-4B0F-88CA-7(...).jpg (250 KB, 1829x833)

250 KB JPG

>>2797349
Yeah, I found another one the other day.
Some retard intern at TI decided to replace a bunch of ‘-‘ characters with the more wordy and English appropriate word ‘to’ making it incorrect.
TI is usually pretty good, at least it was in the olden days. Everything is gradually turning to shit.
One guy posted a TI schematic with an extra (and wrong) short circuit connection in it. I hope he didn’t burst a vein, that would be pretty annoying.

File: allah is is is is is.jpg (54 KB, 660x366)

54 KB JPG

>>2797359
>retard intern at TI

more likely, Muslims
many introduce mistakes into their work
as it proves only Allah is perfect
although Andrew Tate comes close

>>2797330
>leaks hydroxide on your pcb
nothing personnel

>>2797349
>remedy was pulling up a more recent version of the document.
You should go out of your way to download the latest version by getting important datasheets directly from the manufacturer's website.

>>2794153
I've got a 100L hot water heater that I want to heat using Solar Panels (4 x 250W).
Why solar panels over actual solar water heating? Because I got these panels for very cheap ($10AUD each).
I've also tested that this is feasible in the sense that
>using a 500W element @ 24V, the water heats up to ~80C in ~1 day assuming good sun
>the water maintains 40-50C after two days of no power being applied to the element
Assuming I run the panels in series (~96V), would I be able to get away with a standard thermostat controller? Or would current probably still be too high (high current relays are expensive).
I haven't done the math yet, but I'm assuming that because a heating element is basically just a resistor, I could swap out for an "AC" heating element too?

File: funi.png (66 KB, 557x344)

66 KB PNG

>>2797376
1100L of water = 5550 mol of water
heat capacity of water = 75.385 J/(mol•K)
heat capacity of 100L water = 418.450 kJ/K
If you want to get that water from 20°C to 80°C, that's 60K difference
heat required = 25.107MJ = 7kWh

So 4•250W of constant power will get the water from 20-80C in 7 hours. But you could only get that much continuous power by having your PV panels on gimbals, which you probably don't want to do. If you angle them towards the equator you'll get maybe half of that.
It also assumes maximum power point tracking, which you won't get just by piping the panels into a resistive heater. Pic related is the output of a PV panel, if you pick the wrong size resistance you'll get way less than 250W per panel if it's directly connected. You'll probably want some form of MPPT controller, though I'm not sure if conventional MPPT charge controllers will pipe energy into a resistive load with no battery. It's probably a problem that's been tackled before, with any luck you can just choose the right size resistance and a normal MPPT controller will think everything is fine, maybe using some funny configs and/or a little voltage divider. I advise using a length of nichrome that you can adjust the length of via crocodile clips.

Consider running the water behind the panels before it gets heated resistively, it will both cool the panels (cooler panels are more efficient) and use the waste heat for heating the water further. There are also some really easy diy solar water heating panels out there, Tech Ingredients did a good one.

File: sn32 wiring 2.jpg (189 KB, 1183x841)

189 KB JPG

This is the wiring diagram of a manual lathe with a 3-phase motor. What does the leftmost part do? Lines 17 and 18 connect to 2 poles of the 3-phase motor. I'm guessing it's a brake, since the lathe doesn't have any fancy functions.

File: IMG_20240509_171331__01.jpg (1.73 MB, 3472x4624)

1.73 MB JPG

What does this shrugging symbol indicate?

>>2797613
a square wave signal that modulates the MOSFET gates.

>>2797618
well yes, but I've also seen the reverse signal ...
i.e. _||_
there must be a difference?

>>2797624
The symbol in your image means "active low signal".

>>2797624
Maybe a 90% duty cycle PWM signal.

>>2797625
yes that's exactly it, thank you
so to activate this gate I set the logical value on the connected pin to 0

File: Capture.png (177 KB, 1137x489)

177 KB PNG

i need a +/- 18V bipolar supply. i already have pic related. anyone have experience with using 2 of these wire in series for a quick and dirty bipolar supply? they have floating outputs so they shouldn't instantly explode. i would put a reverse biased diode across each power supply to prevent negative voltages at startup.

>just get a dual output linear power supply and stop being a cheap faggot
i plan to, just dont want to spend $300+ right now.

I suspect that a 0.1% 1mohm resistor could be a waste of money. 0.1% of 1 mohm is 1 uOhm. that's such a ridiculously small resistance, the solder job and the contact between the part and the PCB would have to so perfect i am not sure if that is achievable?? i am all for using 1k-10k 0.01% resistors in precision circuits as voltage dividers and such since those resistances are much harder to affect by any external factors. but as for a 1-10 mohm current shunt for 10-50A range, i'd think 1% or 0.5% should be enough. it is harder to do better than that.

>>2797714
Can you even measure it that accurately?
We adjust shunts (like a 1” long copper wire on the PCB) by putting it a test jig, and drop some solder on it until it’s within spec. Obviously they are all over-valued, it’s a lot harder to remove material using this technique.

>>2797380
>You'll probably want some form of MPPT controller
Thanks heaps, the resistance curves are actually what had been confusing me.
I was trying to work out ways to keep any kind of "power controller" out of the system, but I might have to do something like this. The controllers I've tried (for other projects) have required a battery, but will do some experimentation as you've suggested.
>Consider running the water behind the panels before it gets heated resistively
It might be a bit overkill for this use-case. Short story is that I'm building myself a cabin on-the-cheap and am trying to avoid gas heating. I'm able to source cheap second-hand solar panels pretty easily, so if I have to add a few more panels to the setup, I'm okay with that. Piping under the solar panels might cause problems when I'm not around if the connections blow off and drain all the water (I've got a jerry-rigged kitchen system like this with black piping on the roof - stupidly, the plumbing connections are in the sun and they've blown off a few times).
>There are also some really easy diy solar water heating panels out there, Tech Ingredients did a good one.
I might look into these too. If I can do them for cheap, it might be worth it for me.

>>2797711
What kind of current do you need? Do you want current limiting, or would an overcurrent latch (or fuse/breaker) be fine? It may be an option to just get a centre-tap transformer and make your own for like $30.

>>2797714
It doesn't matter. ±0.1A is probably fine for your applications, you almost certainly don't need to go down to ±1mA. Also before you go all autistic on tolerance, go somewhat autistic on temperature coefficients too. With the heat generated by the power traces and shunt itself, the thermal drift alone is going to be a good 1%.

>>2797723
You can almost certainly make an MPPT power controller using:
>microcontroller
>current and voltage sensing circuits
>a good gate drive circuit (https://circuitden.com/blog/11 or a dedicated IC)
>inductor and transistor arranged to be a buck converter (or inverting boost, it's easier since it uses an N-channel FET)
And there will be youtube projects describing roughly that. If existing projects are designed to charge a battery, just change the code. Making such a circuit that can handle 1kW isn't trivial, but you can basically just parallelise the thing with more FETs and a bigger inductor until it looks like it will work. The inductor is the expensive part.

I'd consider this an intermediate level project, I'd probably attempt it myself, but it would take some time, and quite possibly a revision or two. Not something I'd recommend for an electronics noob, at least without a sufficiently good tutorial that you don't need to deviate from.

>>2797718
Yes and even it was possible to measure it accurately, I am not sure how to achieve <<1uOhm trace resistance. Even 100uOhm is hard. Unless you go with 2oz or thicker. But Kelvin connection certainly helps. I just noticed that Dave uses a 1mohm 0.1% shunt resistor in his uCurrent meter. His main idea was to minimize the burden voltage especially for high currents. In that sense it probably makes sense to reduce the shunt resistance value as much as possible and then use two stages of precision opamps with a gain of 100. Actually I was wrong, he uses a 10mOhm resistor which makes more sense but still it is a 10uOm tolerance. But the original version had a 0.5% resistor which he later replaced with a 0.1%. Not sure if it's worth it.

>it’s a lot harder to remove material using this technique.
desoldering braid wouldn't work easily?

>>2797728
>What kind of current do you need? Do you want current limiting, or would an overcurrent latch (or fuse/breaker) be fine? It may be an option to just get a centre-tap transformer and make your own for like $30.

like 100 milliamps or less. i'm trying to troubleshoot an op amp circuit. the PCB takes +/- 18V from a flyback and smooths it out with a linear regulator so the switching noise from a bench switcher shouldn't be an issue. i just need bipolar.

>>2797728
>Not something I'd recommend for an electronics noob
It's probably a bit out of my league at the moment.
My knowledge of electronics is basically constrained to simple DC circuits and I still don't really have a proper conceptual understanding of electricity actually works (I do want to brush up when I get more free time).
Based on what you're saying though, if I wanted to keep things dead-simple (and not too inefficient), would my best bet be:
>higher resistance heating element (e.g. targeting 500W) so as to avoid ~zero power from insufficient sun >>2797380
>higher voltage so that I can use a cheaper/low-amp relay for thermostat control
I figure if I can get something going now that semi-works, I can probably improve the setup later on with a proper power controller.

File: Typical-PV-Current-Voltag(...).png (40 KB, 740x474)

40 KB PNG

>>2797731
>like 100 milliamps or less
Even cheaper, ~40VA transformers are pretty common.
>the PCB takes +/- 18V from a flyback and smooths it out with a linear regulator
If your final voltage you're aiming for is ±15V or smaller, you could use your 30V supply with an op-amp to buffer the middle of it.

>>2797737
>higher voltage
Generally yes, though if you're looking for solar controlling hardware that can handle the higher voltage you'd find more options at 24V than 96V. For example, it's probably easier to find a 24V 30A charge controller than a 96V 10A controller, but both will be expensive, neither are investments I'd make without being sure I could configure them to work without a battery.
>high resistance
Well if the maximum power point in full sun is 36V at 6.9A per panel, then you'd want to pick something around 5.18Ω per panel. If you put the panels in series, then that's 20.7Ω total. A 240V 2.4kW heating element is 24Ω, which isn't too far off. The further you deviate from this resistance at full sun, the less power you get. The further the sun angle changes, the less power you get. You'll probably find that a resistance that's somewhat higher than the 20.7Ω will give you more power across the whole day, so 24Ω is definitely a good place to start if you're not going to try and use an MPPT controller.

For a 12V nominal conventional silicon solar panel, maximum power is at ~18V, which I expect is a factor that holds for 24V and 96V systems too. The maximum power voltage for a given illumination doesn't really change much, it stays around that 18/36V level (pic related). Read the back of the panels (or otherwise a datasheet), as these are the numbers I assumed for getting 20.7Ω.

>>2797766
>it's probably easier to find a 24V 30A charge controller than a 96V 10A controller
This one doesn't concern me too much - it's easy for me to wire up in parallel later if need be.
>so 24Ω is definitely a good place to start if you're not going to try and use an MPPT controller.
Thanks heaps for all your help, this is great info to start tinkering with.
I'm still a while away from setting this up (need to complete the cabin), but will report back to /diy/ somewhere on how it all goes.
I know this isn't the most elegant or efficient way of doing things but, given the cheapness of second-hand panels, I think it's worth toying with.
If it fails, I can repurpose the panels for other things.

I want to make a car radio, by which I mean something I can plug an aux cable into and adjust the volume, I don't want radio.
I'm thinking that this just requires an amplifier circuit but I really know nothing.
Anyone know of something similar that I might modify to do what I want or have any advice?

File: output.webm (743 KB, 720x1280)

743 KB WEBM

Making a key with display for my project

File: Screenshot_20240510_14150(...).jpg (615 KB, 1080x1646)

615 KB JPG

>>2797957
the setup in the car makes the biggest difference. If all your speakers are run from a standalone amp, you could just adapt your headphone out to RCA and set your gain accordingly, change volume on the player. Otherwise you need a ~50w 4/2 channel amp module and some method of splicing it into your stereo harness, picrel.
I would desolder the screw terminals and make a pigtail out of the old radio adapter.
For all the effort I would go the amp route but i don't know what the rest of the system is supposed to be

>>2798017
Cool

>>2798017
ive always wished e-ink buttons were a thing. OLED/AMOLED buttons are cool, but i feel like theres a market for a button that could digitally relabel itself and consume virtually zero power.

>>2798017
Reminds me of that guy who potted a tiny OLED and MCU in a lego piece and played Doom on it. Very neat. What's the project?

File: 5FE1246F-9658-405D-82EE-8(...).jpg (552 KB, 1320x719)

552 KB JPG

>>2798180
> e-ink display buttons
Yeah, sounds cheap. And small, too.

I’m sure they would be epic looking.

L)>>2795412
Yes I think this is a good starter kit. These are all passive components.

As >>2795420 said, ICs (integrated circuits) are not passives, and the budget Chinese ICs (especially beware of op amps where it really matters) behaviour may differ from their specification/datasheet in ways that make them unusable. FWIW I do buy Chinese ICs and have been mostly happy.

>>2798241
(((they))) keep e-ink prices artificially high and keep innovation slow, a device with couple of such keys would double the device's price

>>2794190
charge them up to 9v and put your tongue to the leads, you'll be able to identify the faulty one

File: file.png (353 KB, 585x551)

353 KB PNG

>8800mah
these things are bullshit right? they are like half the size of AA batteries and claiming mah up to 8800mah.

How much capacity should I expect from these 16340 li-on batteries while searching for a legit one?

>>2798409
A half-size AA is a 14250 and has 300 mAh typically. Your batter is slightly larger, the dimensions are in those numbers, so you can scale the capacity with the volume.

There are no new revolutionary battery chemistries out there, in the works, or even physically possible. Only minor variations, with different tradeofs. Even lithium ion has about half the recharge cycles when compared with, say, NiCd, and Li-Ion is more dangerous, expensive, more sensitive to damage from cold, destroyed by deep discharge, etc.

I want to amplify the output of a thermocouple, how do I do it? The thermocouples I have (unknown type, chinkshit) measure around 0 volts at room temperature, while body temperature outputs around 400uV. Assuming near-linearity I'll have to amplify in the millivolt range, but it's two orders of magnitude from the minimum input voltage the opamps I have require from datasheet. Is it just a matter of biasing the input above 0.3V and then leverage a 100:1 gain or is there a better topology?
I tried doing a 1000:1 non-inverting with a RC4558, but the output is nonsense, so I assume the input voltage is just too low.

>>2798393
Sounds just like ink jet printers! Why are “they” trying to corner the market on everything related to “ink” — I don’t get it.
e-Ink displays seem similar to LCD, and LCD requires sweet bugger all to maintain their displays… like less than battery self-discharge rates of power. And LCD is cheap as dirt.

File: A536DCBA-F9FC-4344-959C-C(...).jpg (462 KB, 1702x1080)

462 KB JPG

>>2798419
> minimum voltage
I think you’re looking at the offset voltage. It doesn’t exactly mean that.

>>2798423
>e-Ink displays seem similar to LCD
the difference is e-ink screens are patented as fuck and a lot of things are proprietary

File: tegaki.png (7 KB, 400x400)

7 KB PNG

>>2798419
Either get an op-amp that can sense down to 0V (RRIO or maybe just ground-sensing), or use an intermediate voltage reference to push the thermocouple above the 0V rail. I'd do the latter, as the thermocouple voltage can go both above and below 0V. In my image I've shown it as a differential amplifier referencing ground with respect to a bipolar power supply, but you could just as easily use a second op-amp to buffer a voltage divider to be halfway between your rails, or even use a dedicated voltage reference IC/diode. So long as the reference can push and pull current. You could replace the single resistor to ground with a pair of resistors to the rails (of twice the resistance), but then you wouldn't have a reference to use for future measurement.

You could also use a dedicated thermocouple amplifier IC, or even a thermocouple digitiser IC that takes care of the cold-side temperature compensation for you. Thermocouples measure a differential in temperature, hence why the voltage at room temperature was 0V. All practical thermocouple measurement circuits have some method of measuring the temperature at the termination of the thermocouple wires, often a thermistor but maybe a diode or digital temperature sensor IC. If you get a linear enough sensor circuit and get the gains right, you could use an analogue summing amplifier to combine the two temperature measurements to get the absolute temperature of the thermocouple. But most people would digitise both and do any math and calibration and nonlinearity compensation digitally. Analogue signals aren't that useful, unless you're building an analogue PID loop, or displaying temperature on a moving coil meter.

As the other anon alludes to, the most important feature for your op-amp is low input offset voltage. IIRC CMOS op-amps are best for this, but definitely do your own research. The differential amplifier circuit I drew should cancel out input bias current, unlike a conventional inv/noninv amplifier.

>>2798481
>unless you're building an analogue PID loop
Bingo! I'll try with your circuit, thanks.

File: Untitled.jpg (134 KB, 800x800)

134 KB JPG

how do you shop for power conditioners/EMI filters? tripp-lite isobar is supposed to be the gold standard, but it's not affordable in my country. many of the CW4L2-10A-S or CW4L2-10A-T based power strips have anecdotal reviews saying they improve the sound quality of hi-fi systems. they can be had for as little as $20. then there are DIY solutions like pic related which seems simple enough to assemble, but how would i know if it performs as good as tripp-lite isobar?

1> The first stage is the RFI filter DC circuit. Through two pieces 50A bridge stacks and 2 pieces 35V2200UF capacitors, Filter the DC component of the AC power supply to suppress buzzing caused by saturation of the DC component transformer or increased static current.

2> The second stage is differential mode filtering, two differential inductors, differential mode filtering of AC power supply, to suppress medium and high frequency interference.

3> The third and the fourth stages are common mode filtering, through the X2 capacitor and common mode inductor, combined into the common mode filter.

Due to it is use of Four stages filtering, so the filtering effect is more better. In view of the problem of rapid transient group pulse interference with repetition rate of several kilohertz in some users, it can play an inhibitory role.

This module is used in conjunction with a decoder or power amplifier to filter the DC component of the power supply, eliminate and suppress the hum of the transformer, the high-frequency noise in the power supply, purify the power supply, and can significantly improve the sound quality of the power amplifier, enhance the resolution and make the background quiet.

https://www.aliexpress.com/item/1005005422329779.html

File: Untitled.jpg (132 KB, 750x1000)

132 KB JPG

btw i'm all about the noise filtering and not surge protection. i already have a belkin surge protector which i could modify like pic related.

this power strip has surge protection built in:
https://www.aliexpress.com/item/1005005422368771.html
with EMI filter but without surge protection:
https://www.aliexpress.com/item/1005005119066552.html

>DC blockers don't really improve sound unless you had a lot of DC. However if you have DC that's bad, you sure as hell want to get rid of, it's loud when it goes to the transformers.

>They use electrolytic capacitors (in a way that doubles their voltage rating using two at a time), but they are slightly more complicated than that. Here's a google image search with several variance in schematic and pics of working ones.

>By and large enoise is the worst enemy, but if you got bad DC by all means a DC blocker is going to be your good friend.

>Maybe... A lot of replacement light sources like CFLs and LEDs, many dimmer circuits, and a lot of the SMPS (switch-mode power supply) units in almost all electrical/electronic stuff these days kick back charge into the line and cause distortion that leads to what can look like DC to a transformer. We live in a rural'ish area and power delivery is not great, but I was surprised at how much noise there was at the wall outlets of a number of places down in the "big city" as well as around home. That said I am not sure I ever saw any impact at the output of the power supply after rectification and filtering, but in several places adding a DC blocking filter did fix some transformer hum. I've forgotten the details, but IIRC the magnitude of common-mode offset needed to induce hum in a toroidal transformer was a lot less than I expected (led to me doing a little more research to find out why).

>>2798604
I don’t shop for them, I take them out of microwave ovens and reputable appliances. I have a lifetime supply. I even have some from back in the 70s and 80s where electronic equipment (including a lot of computers) had these metal hermetically sealed units, many of them had that little IBM PC style plug coming out of it. You know the kind I’m talking about?

Also take the linear power supply pill. A bigass 12V transformer will take out a lot of shit, plus isolated. In fact, audiophiles have 1:1 isolation transformers from, say, hammond that do exactly that.

Also note that those MOVs are tiny. Replace with bigger ones. Even on your power bar, do you realize the MOVs wear out when they absorb power spikes eh? Anyway, you can easily build your own. And hermetically seal it with some sheet metal.

> aren’t the capacitors going to be bad in the 70s units?
No, usually caps go bad due to under speccing them, hot environments, and leaks. These are hermetically sealed so caps likely fine. Also it was before the influx of “consumable” caps from china.

I think that modern devices should have a plug-in slot like a fuse, 9v backup clock battery, or some D cell-sized holders for your cheap Chinese capacitors so they’re easy to replace. MOVs, too.

>>2798471
> patented as fuck.
All we have to do is get the chinese to start caring about making them. They use the patent filings as a goddamn instruction manual.
I’ sure they don’t even give one fuck about patent IP, so it’s win-win. Now we need the government to fuck off and not hold up my temu shipment since it might have a bullshit gucci logo with an incorrect serial number, oh god forbid.

>>2798620
>DC blockers
Never heard of these before. Doesn't having any net DC current imply you've got loads that run on a half wave rectifier? And even if you do, why would such a load be after a transformer? If it's a load in parallel with the transformer, I can't imagine it would be significant unless your wiring is real shitty and the rectified appliance draws a lot of power. Maybe some heaters do this for a half-power mode, idk. Don't use shitty appliances.

>>2798624
I think those are Y caps, not MOVs. MOVs aren't for power filtration at all, they're for surge protection, so I imagine they're meant to be seperate.

>metal hermetically sealed units
The old ones explode and get gunge all over the inside of your enclosure. Marco Reps showed that in a video or two of his. X and X2 caps have always been consumable parts, unless you somehow have a perfect sine mains lacking any spikes, which would defeat the purpose of having X caps in the first place.

>>2798632
my house is from the 70s, outlets aren't grounded except in the kitchen and bathrooms. i don't know the technical details but everything that's connected to the mains makes a slight humming noise if you care to listen for it when it's quiet at night. buzzing noise is a common complaint on corsair PSUs but it's on everything like my NEC monitor or LED lightbulbs. it's quiet enough that i'm not sure if everyone has it and the average person just doesn't notice or care. it's a transformer buzz that is heard regardless of the power draw, a separate issue from ground loop or the type of coil whine that becomes louder in higher loads like gaming.

>>2798632
the transformers mentioned in >>2798620 >>2798622 are inside the device like the PSU in a computer or a hi-fi amplifier or a guiter amplifier and they would make a humming or buzzing noise

in a PSU the average person wouldn't notice it unless the PSU has 0 rpm fan mode and there's no other fan or HDD running and you would have to have your ear pretty close to it in a quiet room

Can someone recommend me a circuit that only consists of a basic inductor and capacitor and no other elements. I already know the LC filters, LC resonators and LC matching networks so something else?

>>2798688
LC balun, wilkinson divider, directional coupler

>>2798644
Computer PSUs are SMPSs, they shouldn't hum but they might have some coil whine. They rectify the current going into them anyhow so any net DC shouldn't impact them whatsoever.
Conventional iron AC transformers like in old-style amplifiers are prone to hum, and I can kinda see why a net DC current being put through them might make that worse, but I can't see how that net DC current would get there in the first place. The transformer itself only couples AC, and after it there will almost certainly be a full wave rectifier. Then back at the power pole, you also have an AC coupling transformer. As I said in my previous post, I can maybe see it happening if you have a half-wave rectified high-power load on the same phase somewhere, but even that's a stretch.

File: blog_ax860i_technical_det(...).jpg (72 KB, 597x496)

72 KB JPG

>>2798707
the reason i started looking into this stuff was
>Since it's making the noise when idle, it's safe to assume that the noise is coming from the PFC choke coil, which means it could still be from mains noise.
>Have you considered an EMI/RFI filter, like a Tripp-Lite ISOBAR?
https://forum.corsair.com/forums/topic/165666-corsair-hx1000i-very-loud-coil-whine-buzzing/

however on my AX860i it is clearly not coming from the PFC inductor (#3 in pic related)
https://forum.corsair.com/blog/icue-controllers/ax860i-technical-details/

from listening to the fully assembled PSU (so i could not reach everything very precisely) with an empty toilet paper roll i suspected #16 (LLC resonant inductor) during standby, and both #16 and #10 (LLC main transformer) when running in self test mode (as if the computer is running). i applied epoxy resin with an oral syringe mainly inside #16 because it was easier to reach with the coil visibily exposed and it was able to hold a lot of resin without spilling, but i also applied some in #10. the noise was significantly reduced but didn't disappear completely. i think this will be adequate especially with sound deadening foam and optimal placement of the computer under my desk or whatever. but if the buzzing could be eliminated via EMI filter or DC blocker for <$20, also on my LCD monitor, that would be sweet.

another mention of DC blocker
>The hum some people are getting may be due to excessive DC on their AC lines. Some transformers react this way to DC. I would try a DC blocker in such a case.
https://audiophilestyle.com/forums/topic/21206-group-buy-for-full-atx-linear-psu-from-teradak/page/22/

File: Screenshot.png (560 KB, 748x782)

560 KB PNG

this mentions that AX860i doesn't have a full EMI filter built in
https://www.techpowerup.com/review/corsair-ax860i/4.html

File: Screenshot.png (664 KB, 982x804)

664 KB PNG

https://www.techpowerup.com/review/corsair-ax1200i/5.html

File: s-l1600.jpg (157 KB, 1600x1600)

157 KB JPG

copper crimp ferrules like pic rel are much better than aluminium ones. prove me wrong

>Hi
>My friend got a chance to test many SMPS PSU and he told me that his choice is corsair AX 1200i . It was better than Ax 860i and other SMPS PSU such as Seasonic brand
>Recently he tested new model AX 1500i and he concluded that that is the best.
>Before I jump to Teradax I am just a little bit reluctant about Corsair Ax 1200i or 1500i or Teradex.
>I believe that it is unfair to compare SMPS to this Teradak but anyhow sharing ideas is a good choice
https://audiophilestyle.com/forums/topic/21206-group-buy-for-full-atx-linear-psu-from-teradak/page/22/

>Contrary to the AX1200i, this 1500 W monster doesn't includes a server-grade EMI filter. The AC receptacle only hosts a couple Y caps. However, the rest of the EMI filter is on the main PCB, and while there is a ton of glue to make the unit coil-whine proof, we managed to identify three CM chokes, two pairs of Y caps, five X caps (one sits right behind the bridge rectifiers), and an MOV. Overall, the EMI/transient filter is more than complete.
https://www.techpowerup.com/review/corsair-ax1500i/4.html

What properties do capacitors lack (besides energy density) to replace batteries? Something about the internal resistance, and a faster discharge rate?

>>2798828
>What properties do capacitors lack (besides energy density) to replace batteries?
Energy density.
In most other metrics (like charge/discharge speed, nr of cycles, temp range) caps are superior.

>>2798832
don't caps discharge pretty much with a linear rate, as opposed to batteries that can stay at close to 100% for a long time?

>Update:
>My Arcam SR250 does have a minor mechanical transformer hum with my ear to the vents. The CMX2 does successfully eliminate this mechanical buzz!
>The buzz is below the noise floor of my UMIK-1, so I can hear it with my ears -- but not with the mic.

>For my situation, I bought a new 4K TV and I guess the electronics in it (they are more PC than TV now a days), or maybe it was some new LED ceiling lights I put in, created a hum in my APC H15 Power Conditioner that was really noticible. The APC has an isolation tranformer in it for Voltage Regulation and it was singing a song! Thing is, it would come and go throughout the day.
>So after a whole bunch of moving things around and trying all the isolation tricks, I finally decided to give this DC Offset filter a go. Solved my problem!

>I purchased the Emotiva CMX2 last year to help reduce a hum from a tube headphone amp and it fixed the issue. Fast forward to yesterday my Arcam AVR-850 now has a hum coming from inside the AVR-850. No issues with noise coming from the speakers. I disconnected everything from the back of the AVR-850, noise was still present. I moved the AVR-850 to another electrical outlet and the noise (hum) was still present. Connected the AVR-850 to the CMX2 and it removed the hum and the problem was fixed. I have had the AVR-850 since 2017 and I was thinking great it’s time to either find a local electrician to diagnose the noise / fix or buy another AVR.
https://www.audiosciencereview.com/forum/index.php?threads/emotiva-cmx2-ac-line-filter-with-dc-offset-eliminator-review-and-measurements.45071/

File: cap discharge.png (10 KB, 511x486)

10 KB PNG

>>2798842
>linear rate

linear enough for Straylia

I am near the end of my project but too pussy to move forward

>>2798887

I think he means in reference to voltage relative to charge, in which, case, yes, the voltage present at a capacitors terminals is directly proportional to its charge.

>>2798828
>What properties do capacitors lack (besides energy density) to replace batteries?

Energy density. That's literally it. They're straight-up better in every way to an electrochemical cell. Effectively unlimited cycle life and extreme power density.

Well, I guess I can think of two other potential problems. The fact that voltage is proportional to charge can complicate the design of your power supply. At the simplest, you can simply drop a linear regulator in and the battery will happily supply it until it's nearly dead. Depending on your exact application, you may even be able to just power everything directly off the battery. Capacitors would pretty much require some sort of switching power supply, due to the voltage/charge issue. A linear regulator would waste a huge portion of the capacitor's charge, either by being set at the high end and not working once it discharged past the first 20-30%, or at the low end and dumping 70-80% of the capacitor's initial charge as heat.

The other problem is that a catastrophic failure of a battery tends to turn it into a fireball. A catastrophic failure of a capacitor tends to turn it into a bomb.

>>2798893
>too pussy to move forward

you've proven to yourself that you COULD finish the project
so, you've already won
actually finishing it would be an act of pure egotism and arrogance
time to start something new and egg-citing

is this correct general to ask
>how to learn how to make stuff like that, i know how to program https://youtu.be/41qIQzLS304?t=53

>>2798913
>is this correct

no, we dont do projects
we argue over esoteric minutiae
you want the arduino general here >>2774325

>>2798842
Yeah, the pack voltage drops quicker with capacitors than with batteries (depending on chemistry, look up lithium ion and lifepo4 discharge curves). But that only becomes a problem if you’re dumping the full voltage through your load at maximum duty-cycle in the first place. In the case of an EV, that would mean full-throttle. Driving around normally you wouldn’t notice much difference, only when you put your foot down would you notice it. Even then you can use field weakening to get up to higher speeds than the pack voltage would conventionally allow for. It’s mainly the torque that will be diminished.

>>2798913
Idk bro I’m not clicking a YouTube link on cellular internet. Did you read the OP?

File: 8ba12d73c6808a441b7533c90(...).png (443 KB, 636x900)

443 KB PNG

should a seperate analog / digital power supply also have seperate ground planes?

Could switching noise somehow get into the analog plane?

>>2799105
>should a seperate analog / digital power supply also have seperate ground planes?
Probably, especially if there's high current (spikes) going anywhere. A common method is to use a star ground, or just have a single point where the digital and analogue grounds are connected. You could also connect them with a choke to reduce interference, or even connect them with an appropriate value resistor if there's no supply current that needs to flow between the analogue and digital grounds. You'd still need it to be able to handle signal current though.

There are seperate methods for mitigating common-mode and normal-mode noise. Having seperate Vcc rails, plenty of capacitance, and star grounding should kill any normal-mode noise, common-mode is harder but grounding things properly with low impedance planes is what I'd use. Watch some tutorials on how to lay out grounding for circuits, there are some unintuitive things to learn out there. Like having ground traces/paths that follow your signal traces. You can always add more vias to your board, they don't charge for that at the fab house.

>>2799107
Hmmmm. This guy says splitting grounds is a bad idea since it leads to EMI problems:
https://youtu.be/vALt6Sd9vlY
So maybe it will suffice if I just keep a good distance between the analog and digital parts.

>>2799119
EMI is largely not a problem unless you’re trying to get a product tested and certified for sale in a socialist country.

If you are DIYing, don’t worry about it.
In fact, there’s so many variables to “creating an antenna” that this advice they are dishing out may as well be reversed to “always split grounds” or you run into EMI problems.

>>2799154
Hmmmmm I'm quite undecided on weather I should split the GNDs on my audio circuit.

I'm also undecided on how to best switch audio sources. Initially I thought of using a multiplexer but those add distortion to the audio line, so maybe a relay is the cleanest way after all.

Can I use the GND path of the source to switch it?

>>2799172
> switch audio sources
They make mux and “soft” switching ICs exactly for this purpose, they’re just mosfets with a resultant esistive load.
> relay
You are going to blow somehing with the pop n the line with a relay. Only place for a relay in audio is in slow-start current unlimiters in power amplifiers with big caps.

>>2798625
Pfizer's Viagra patent was thrown out (2012 SCC 60) because it describes billions of molecules without singling out the one that is the actual invention. Patent authors deliberately leave out key details. Pfizer lost the case, but the patent was already about to expire, so they'll do it again. If they would retroactively cancel a patent and give the excess profits back, patents would be worth reading.

>>2799178
Which IC would you suggest?

I have HEF4066BE, with the issue that exceeding -0.5v max at the input, it continuously conducts.

Also, it doesn't mention TDH in the datasheet, but if the CD4066 is the same IC, it mentions around 0.1% THD, which is a bit suboptimal, adding that much distortion to my quite low distortion audio chip - TDA1575.

File: my audio switchboard.jpg (91 KB, 800x250)

91 KB JPG

>>2799277
>I have HEF4066BE

forget it
the whole family of 4016,4066,4051/52/53 etc cmos switches are crap and have huge crosstalk between channels
either find something modern or use the only guaranteed fool-proof solution: giant relays in steel boxes

>>2799215
why did we trust these people to roll out billions of compulsory experimental vaccines again?

>>2798758
stop listening to audiokikes

File: Analog Mux ICs.png (340 KB, 1642x1304)

340 KB PNG

>>2799119
Then at least use star grounding.

>>2799277
The DX4xx and DG6xx series analog switch ICs are pretty good, some have seperate supplies for the digital and analogue sections so you don't need to worry about negative voltages. Pic related.

Using relays should be doable before your amplifier, so long as you ensure there's no DC offset to the signals and you have a low-pass filter (f=20kHz) to get rid of the worst pops it should be ok. Using an additional circuit to attenuate the signal during switches be even smoother, maybe a series resistor and a JFET to ground, but it depends on your signal amplitude.

>>2799119
just keep everything on one big ground plane, why tf are you splitting supplies

>>2799306
https://www.youtube.com/watch?v=ZoU5N0RD_Is
this guy found out it was his electrical stove
also i think people in general are just clueless, you might not be aware of the issue if you haven't listened carefully specifically for it when it's dead silent at night
i will try the EMI filter with DC blocker

File: relay.jpg (1.05 MB, 2100x2800)

1.05 MB JPG

>>2799280
What if you just used the cheapest common 12V relays?

File: Tektronix 500Mhz oscillos(...).jpg (67 KB, 693x377)

67 KB JPG

>>2799374
>common

why be common when you can be extraordinary?
if frankenstein relays arent your thing you can still be extraordinary by choosing super miniature relays like those found in modern scopes

I'd like to power a weather station with a solar cell which sounds like a contradiction. The station needs to be in the shade yet the cell obviously has to be in the direct sun as much as possible. What do?

>>2799392
>station in shade
>wire
>cell in sun

>>2799398
too messy
wireless power would be awesome

>>2799325
Because noise travels on your power rail, and you can get voltage ripple even on a ground plane if you don’t use a star ground.

>>2799392
It won’t take that much power if it’s designed to run on AAs for a year, use amorphous solar cells. They’re found on calculators and caravan trickle chargers, and work even in very low light levels, at the expense of having a low maximum output.

>>2799305
It’s not that viagra didn’t work, although I think it was originally a heart medication, it’s that they were not quite legal or ethical in their business dealings, as often is the case in many companies.
Patents are driven by the business side, as an added bonus, they don’t even know enough about the patented process to by “lying”

>>2799277
> which IC?
I don’t remember exactly, there are lots of them. I took them out of old stereos over the years. Also soft volume controls are very common:

So, use your relays, but before switching any of them, use the TDA1524 volume control to automatically lower the volume of the amplifier to nearly zero. That way there will be no pop when the relay’s contacts open and close.
This is pretty standard practice.
Most modern stereos do this at power on, too, and use soft muxes that implement the same feature when you switch inputs. My yamaha amp I got for free out of a dumpster has these relays when I switch inputs.
I’d probably use big relays so there is a satisfying kerchunk noise when switching them. My amp makes a “kertink” which sounds weak.

Relays on oscilloscopes are also important because of the high speeds (RF) involved. Relays are also commonly used in RF switches.

You’re not gonna notice the (maybe) .1% distortion from the cd4066 on a little home-grade amp.

can someone explain me how to make an efficient charge pump voltage doubler, without sending massive current spikes from a square wave into the capacitor(s)?

>>2799443
How much current are we talking. All you need is a sine-wave oscillator and you can do that with one transistor easily enough. Often, they need a coil to oscillate, and that coil can also act as the boost inductor. Kind of like the so-called “joule thief” circuit.

File: 8B4229DB-74B0-4179-8401-2(...).jpg (2.38 MB, 3840x2160)

2.38 MB JPG

I have a printer (big boy) that has this earth wire running from a rail that just hangs loose. The rail it’s connected to is already connected to the earth wire in the power cord so I have no clue why this would just dangle here. Any clues?

File: Bildschirmfoto_2024-05-14(...).jpg (37 KB, 607x533)

37 KB JPG

>>2799310
Thank you Anon!
Should I put the OP amp VCC pins to POWER or SIGNAL GND?

File: Bildschirmfoto_2024-05-14(...).jpg (48 KB, 810x459)

48 KB JPG

>>2799310
Should I put a switching converter in there at all?

File: Bildschirmfoto_2024-05-14(...).jpg (41 KB, 722x464)

41 KB JPG

>>2799310

>>2799452
I have a few projects that would use it. One is just a charge pump for keeping high-side MOSFET gates charged up, probably around 10mA consumption or less. For that I already tested using a 10Ω resistor in series with the cap and it seems to work fine. The other though is to run 1A of LEDs, which may or may not be a suitable application for a charge pump.
>sine
>inductor
Kinda defeats the point of using a charge pump, if I need an inductor anyway I'll just make a boost converter, with the feedback and extra space that requires.

>>2799477
OP-amps should usually be on the signal side, definitely on the signal side so long as you're using them as amplifiers and not comparators. Though they have decent PSRR so it's not as critical as an ADC.

>>2799481
>switching converter
Depends on the project, if the inefficiency and heat generation of a 7805 isn't a problem I'd use one just for simplicity. Not that there's a difference if you have pre-made modules. The issues with switching converters are numerous but not insurmountable:
>noise back to its input: use large caps on the input of the converter and on the grounding star point
>radiated magnetic noise: minimise loop areas in your circuit, especially around high-impedance analogue signals, consider iron shielding cans around the converter and/or delicate signals
>noise out its output: not an issue if it's powering digital shit, this is bad for ADCs that lack seperate supplies
>radiated electric noise: use grounded guard traces between switching supply nodes/rails and high impedance analogue signal nodes, consider metal can shielding
>mystery noise: try ferrite sleeves, optical coupling, idk

>>2799482
I did not know KiCAD had these. I've been using SolderJumper_2_Bridged all this time.

File: Bildschirmfoto_2024-05-14(...).jpg (28 KB, 509x517)

28 KB JPG

>>2799491
I'm already exceeding the max power rating of my transformer by a few Watts...

Getting those 200mA that I suppose my Arduino+LCD down from 25 to 5V with a linear regulator will devour around 5W.

On the other side I don't want to risk adding noise to my audio rail. Pic related seems to have isolated coils. Is this any good?

I saw the dg211bdyz quite cheap on aliexpress. The datasheet says they can go down to -2V at the input. So I think I'll go with that , if you don't have any objections.

ib4 reddit formatting