/diy/ - Do It Yourself

foist for deadbug position

>>2519612
>antigravity and/or overunity?
Go away

If i know a speakers ohms but not watts, is there any way to figure it out?

I wish I wasn't incel

>>2519688
Maybe you could become a rapist.

>>2519688
Time to grow up be a man and fuck a hooker

>>2519612
Can I get a schematic for that?

this shit is starting to make me suicidal
After the mic preamp not working properly due to shitty fake chink instru. amplifier, i made a brand new pcb and this time used locally sourced high quality standalone grass fed opamps to make my own instrument amplifier, seen in the bottom of the pic
But it is still not fucking working.
I checked all the wires and they seem fine.
I even built the whole circuit in falstad and it works flawlessly there:
I have the audio output hooked up to a scope and i am getting nothing basically. A flat line (with a bunch of noise of course but that is fine since there is no shielding yet). Only if i blow into the mic hard i get slight wobble in the output voltage.
The capsule is electret (https://micbooster.com/jli-microphones/262-jli2555.html)) i am using the tranny they recommended. (2N4416) I measured the legs and they are correctly getting about 13V and -13V supplied to them, from my preamp circuit
The only chink part left is the 2N4416 attached to the capsule, so i could try to change that as well (i have some J111 njfets on hand), but i doubt that is the culprit

>>2519724
>I even built the whole circuit in falstad and it works flawlessly there:
stupid fucking 4chan won't let me post url, here
https://www.falstad.com/circuit/circuitjs.html?ctz=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-k4AMYgO0O+y-B67BpwYj2eqoVB6bDNHx4azWKANMDsdO2bMu37ulP5yCF2vIPioHjELB94gXFSQRZ1g45+m-MzEkCFgEaugtr4YUh-Bc3DP2tmZ+clv7ILCLpvLpZPNdSlDHpyx0+X132X0e2N72fb2ikXbY9OGvG+Chyh5ds5C7QwxkcYQMjHU1wJxX4HwQ3APVhJDPiER14OVelahKTDUK+d9fjAWCsKoEiclBUZSNuU0KIqOigX9RdcNeA04BpWZmK3KjwDo+0JFg+diBWWFePo3jiH7dlWJUdidwwbIfwYnIpPYg81JAzsrD1XS9L0nhwGQaBhCaRTLEUsxGlrOB60mWScPjOSNxEtMgloRTzWmTytOIHT9IC0FqmMsFCAQSxQQs8KpQnWE-wzJT4u2EkkvmJLMNjVLyPjVAuOxTKBL40ZCuQaYUIcA96NkDTpKcAAlFS4xOcUKBebx-SQb8YAQdgGtY3KchLRgBpTahGGkUbupxYCsSnXlYWAmJsy+NsnGQ-QfOHHIqREjdkAAfTwfbIH25C9WO2ArHCkgEAuuQqOyYd9rYfbiFeg7kBxQDaGAlptWNdbwD88EIySFavkYPpDtwY7Tu0LBXugW7DDgB611ul63qhz6ACs2XUMbvDcXwWBi2zh2m3wsWQ7xxRJRagRpqMGoeAnSrxNx-SCGyIC6pH2EBsR7nmVRwTpjMYaOk6zoRsBoCewx0eesBXuh-bPsBzTqUJgU-l6NXpfhu6rssBX8yPMcnsxw6MHVgW0mocV8ToU0KQSg24fOmAlcVq2Vax26NYdvBz3sMEPKUxhSA9mXjesU3kY7NG-dV6PPqAA

Just a flat line on the scope. My next step is to try to try to scope the trace on the right side of the C2 to see if there is actually any sine even going to the opamp in the first place, since if the tranny on the microphone is producing no sine it can't be amplified of course. And if that is the problem i will try replacing the tranny with the J111, but i dunno man, i have wasted so much time on this project i am about to smash it.

>>2519728
worlds cannot explain how much i hate working with audio signals

>>2519728
altho scoping that trace going into the opamp (next to C2) will be hard since my scope can do at most 50mv per div and the expected sine peak is like 6mv (according to falstad), so i won't be probably able to see shit anyway

can someone recommend me the best bang for my buck soldering station? I've been using a shitty $10 radio shack iron for like 20 years. I want to get my 10 yo kid into electronics and soldering so I figured I'd buy him a station for christmas.

>>2519724
You still haven't given up? The problem is most likely jfet biasing. Without resistor to ground from gate you're relying on whatever crap like flux residue and balls sweat you put on the capsule to get some magical path to ground. Its just fundamentally a horrible design.
>I even built the whole circuit in falstad and it works flawlessly there
This means that your circuit model is probably shit and doesn't represent reality very well, you're missing some key parameter that you're not modeling - AC source is not accurate representation of capsule, not even close. You need to focus more on how capsule works (modeling, expected impedances and output values at different sound levels) and how jfet works (biasing types, operating ranges, tolerance ranges, gain, parasitics). You can either learn about those things (plenty of resources on the internet) or just throw everything in the bin and buy a commercial microphone - now you know that you aren't just paying for component cost, you're also paying for guaranteed specifications, testing and EMC. Manufacturers aren't ripping people off for no reason.

>>2519788
https://hakkousa.com/fx-888d-digital-soldering-station.html
Do not order from Amazon. You will get a used station in an open box.

>>2519794
thanks anon. I was kinda eyefucking one of these, but it's a lot moar than I was planning on paying, plus it takes up a lot of room on my desk, but I could use some of the shit on my 3d prints too.

fucking forgot link like a newfag
https://www.amazon.com/YIHUA-Soldering-Station-Charging-Voltage/dp/B07S4GPM66

>>2519803
>>2519805
That's a decent chink brand, but will you use the hot air enough to warrant the extra size? It's less expensive in a single unit, although larger and less portable. The integrated PSU is handy. Wonder how long it'll last.

>>2519791
>. Without resistor to ground from gate
well the original design doesn't have one and it works fine for the other people who built this project
i will try adding one to see if it does anything but i doubt it. 10k should be fine? between the tranny gate and ground

>>2519818
>well the original design doesn't have one
Exactly. As I said they're relying on flux or cum crusts, dust, humidity and other parasitic paths to ground. There are special microphone JFETs that don't require external resistors because they have integrated diodes that act as resistors. Normal JFETs don't have them and you must provide a reliable path to ground. Some people might be lucky and have enough parasitics to do the job, but that's not reliable engineering.
>10k should be fine?
10k is too little, you need at least 1meg and even that will take away some low end response so something like 10meg-1g would be appropriate, but again that's why you need to know your capsule, how it works and what are its parameters.

I'm a dumb attention whore Imma leak myself.
This is my setup in the mental retard subhuman facility

I have a 2000s era kids Spiro pinball machine I had in the attic I thought would be a good place to start with electronics repair.
It had no sound, but all other functions work. So I went in and tested the speakers with a function generator, and the power supply with my ociliscope. Both are fine.
I hooked up a 4ohm speaker with 4.7 ohms in series, and got some garbled sound for a second, but then nothing again.
I have the board out, and see it's a main processor under a large blob, and then what I think is an amplifier under the small blob in pic rel.
I mapped it out, and scoped the points going to the large blob processor from the small blob. But I don't see any signal that could be audio.

I tried putting 10mV into the amp and seeing if I could hear anything on the speaker. But, once that didn't work I've become pretty stumped.

Any ideas on what I should try next?

>>2519842
It's spanning several tables. I also got a laptop to shitpost on the chans the whole day, a kettle to boil me some tea and beneath the table is a cabinet with high quality knipex etc. tools.
There are big towers with all kinds of comonents and boxes full with miscellanious Microcontroller stuff etc.

>>2519822
So i tried adding 1mohm resistor as it's the biggest i have but sadly no change.
After some more googling i found this cool preamp circuit directly from TI, so i assume it's some expert tier stuff and i will try to build it on a breadboard, since it's so simple, just to see if i get any output from it or not. I'm starting to think if the electret capsule itself isn't dead. I wonder how fragile they are. Since i accidentally dropped it on the ground one time.

>>2519845
>I tried putting 10mV into the amp

hook up some powered computer speakers to either side of R4 and R10 to see if blob is dead.
if you get sound, assume the lil blob is dead and bypass it with an LM386 or two.

>>2519863
>found this cool preamp circuit directly from TI,

that circuit makes sense where yours doesnt.
electrets all have (as far as i know) an internal fet so dont need any external ones.
i presume the external fet was used to create a symmetric ouput in order to feed the instrumentation amp.
which sounds too precious to be reliable.

>>2518794
gainclone is synonymous with chip amp
includes pretty much anything with a single chip that's trying to act like a wire with gain

>>2519863
Considering that they have to withstand sound vibration, capsule would probably survive a light drop. I think I posted that TI application note a month ago, because this is how it should be done (at a basic level), it even explains how to calculate values and theory of operation. Just copy over design and swap values for your capsule. Only thing you need to add is properly biased jfet.
>>2519873
He has a bare capsule without fet.

>>2519871
Sweeeet.
I put the scope on the small blob side of R4 and then R10.
I looked it up to some headphones and hear the music off both

Not sure why there are 2 nearly identical sound outputs from the large blob, or what the rest of these nearby components are doing, but I'll just amplify one of the signals.
Ty

>>2519918
>Not sure why there are 2 nearly identical sound outputs

prob that newfangled fad called stereo.
i doubt it'll catch on, same for that interwebs nonsense.

>>2519612
>>2519656
Hey, put the "xxx edition" in the thread title next time, that way you'll have room for the whole paste

>>2519873
Is a discrete JFET even needed? Why not use a JFET-input (instrumentation) amplifier right next to the capsule?

>>2520004
It only has 1 speaker.
Else, duh.

>>2520016
>Is a discrete JFET even needed?
the charge in the capsule has to remain unchanged, so any resistance that wires going to the amp present would degrade the signal, which means you need to have a tranny sitting directly on top of the capacitor plate basically

>>2520221
Ignoring those tiny little capsules with JFETs literally inside them, the JFETs put on to those large capsules are like 1cm from it. With good shielding I don't see an issue with that. So considering you can make a PCB with an inst-amp on it significantly smaller than the mic, why not just put it right there?

>want to drive a vfd
>want to use 4511
>want to display V and A on two of the digits
>realise that would require seperate drivers on all of the segments, plus a bunch of discrete logic
>decide to just signal those drivers directly without bothering with an external 7-seg decoder
>check out 4000 series for buffers and inverters
>first look at 4009, goes up to 18v and has seperate vcc and vdd
Man, the 4000 series is based. Now I'm split on whether to use a couple of spare 74HC595s, or to select an MCU with enough extra pins, for once. Think I'll go for the latter, since I have an ATmega128A here I should get some practice programming for. 128k of Flash too, not that I'll ever need it.

>>2519873
the pdf says the c3 and c5 shuuld be tantalum, and c2 ceramic, but i only have the normal film smd caps, i wonder if it will be a big problem if i use those instead, like will it not work at all? or will it introduce just some possible noise?

I want to order from JLCPCB but the last time they forgot to drill my vias and holes, maybe it was my fault for not attaching the holes or vias gerber file, idk, it's been a couple years since I ordered anything from them.
I've now followed this guide on their site:
https://support.jlcpcb.com/article/137-how-to-generate-gerber-and-drill-files-in-autodesk-eagle
I've exported the gerber files and everything seems alright. Can anyone who has used JLCPCB in combination with Eagle confirm this is correct? Do I need to add anything else?

look at this shit
they just put shit into a machine and it does everything
i have to solder shit one by one by hand and then test shit with multimeter like some sort of fucking caveman
https://www.youtube.com/watch?v=-EhDlXx3okU
so unfair

>>2520253
>they forgot t
thats impossible it's all assembly line ran by compuetr, you fucked up, on your account there should be archive of files you submitted so you can check if you included excellon files

>>2519863
welp here we go, i am going to mill a third pcb with for this damn project featuring this new design from TI and if this one also doesn't work then thats it, i'm just throwing the whole project in the trash. I sunk so much fucking time into this already that i'm just done.

>>2520320
What is your supply voltage? You need to have about Vcc/2 on jfet drain. R1 seems to be way too big, at ground bias jfet will pull anywhere from 6 to 9 mA, so at 12 V Vcc you're looking at something like 666-1k.

>>2520331
Pic related is what is going to happen if you don't fix R1 value.

>>2520253
excellon file is all they need for drilling. they shouldn't be looking at gerbers since holes don't necessarily require pads or vice versa. also, i'd recommend looking into other EDAs as eagle is a steaming pile. kicad isn't great but it's marginally better than anything else i've tried plus it's free

>>2520291
it's definitely possible. never underestimate anyone's professional capacity to be lazy or incompetent

>>2520331
9V
pic related, i didn't come up with those values, i just copied them from here
https://www.ti.com/lit/ug/tidu765/tidu765.pdf
This is like some sort of PhD tier level paper on design of that circuit posted above so i'm just straight up copying everything

>>2520320

why arent you showing the connections here?
this is the most important section, and we have to guess how you wired it up.

>>2520365
because that isn't part of the pcb, microphone is connected wia JST connector to the pcb
but it will be like in the pic
the electret plate is connected to the base of the jfet tranny and current from R1 basically flows through the tranny to the ground

>>2520364
Yes, internal JFETs are different to what you have. If you keep values as is it won't work as shown in >>2520339 You can't just copy values, because your mic is completely different.

>>2520367
>because that isn't part of the pcb

it's a crucial part of the circuit so it must be included in the schematic for understanding/completeness/correctness/diagnostics.
see pic for sane method of drawing off-board components.

>>2520376

actually, this is better.

https://www.youtube.com/watch?v=7sGN0-YsiYw
He has run out of ideas.
Buying useless chinkshit off AliExpress is called 'DIY' nowa pparrently.

>>2520364
>>2520373
Here are actual values you need for circuit to work as intended (100dB SPL -> max line input 1.228Vrms). Sticking to same values won't work, you need to change them.

I need a hot air station. Should I get the cheapest one, or just save my money and buy a hakko hot air station?

>>2520399
If you're going to be using it all the time, spend the money and get quality tools. Also, big PCBs with heavy ground planes require bottom heat in the form of a preheater. Take that into consideration when weighing the overall cost of your setup.

>>2520243
If they're tantalum I assume they have a high capacitance like in the microfarads at least.
It's difficult to find film caps with that kind of capacitance and SMD film caps are a rarity so you probably mean MLCC caps you have lying around

>>2520253
OSHpark takes straight eagle files and they generate previews so you can check.
I've never used anything else because I knew your experience would happen to me.

>>2520398
Thanks, i am going to try both of them so i can see the difference it makes

>>2520253
I outputted drill files in gerber format, a drill map gerber file, and a job file. That worked for me with JLCPCB, though it was from KiCAD not Eagle.

>>2520289
A diy reflow plate helps a fair bit, and is a fun project.

>>2519788
I second the Hakko FX-888D as a first soldering iron since it only has 2 buttons so you can't mess up but if you are on a budget check out the PINECIL.

>>2520429
yeah you are correct, i mean those ordinary brown smd ones, probably ceramic i think
is it a huge issue if i use those?

Anyone know how to calibrate these 'high presicion' meters?

>>2520491
you absolutely can mess it up
when you go to change the temp, there's only two buttons and they're not up/down. it's easy to accidentally enter callibration mode and mess up the station.
I have an 888d and I've never changed the temp from the default 750F for this exact reason

>>2520504
For PSU filtering they’re probably fine. But for DC-blocking AC-pass capacitors, or other signal filtration capacitors, you’ll likely want to avoid them. MLCCs like that can suffer from microphonicity and voltage-dependant capacitance, and they’re all somewhat prone to PCB deformation causing distortion. Unless you know they’ve got the right dielectric for your purposes, that is.

Sharts keep boasting about how good their lead solder is but I've only ever used lead-free solder. Is it really that much better and should I order some from aliexpress?

>>2520599
>>2520599
lead-free solder is only good for plumbing

>>2520599
>Is it really that much better
No. In fact there's barely any difference.
20 years ago lead free solder was shit and the boomers still living in 2001 are perpetuating the meme.
Today due to regulations lead free is the industry standard so it has become very good.

>>2520607
zoomer here
lead free solder dont bussin frfr. it's a standard because of enviro regs and it's cheaper.

>>2520603
>lead-free solder is only good for plumbing
you're delusional.

>>2520620
no u, ching chong
you can keep your mercury cadmium cyanide wire

>>2520622
zhongguo superior
seethe & cope gweilo dog

>>2520625
>superior
how does that translate to insectoid?

>>2520599
If you order it from aliexpress it's probably gonna be lower quality than unleaded solder from a western brand. Even if the metal alloy is right, rosin core is likely something low quality and even corrosive.

I would suggest to check small businesses which sell both to hobbyists and for companies. Those usually just mention "only for professional use" or so and thats it

>>2520231
Well at least now I can spell out messages during idle. That kind of versatile display driver makes me want to make it modular from the PSU board, that way when I get a bunch of the boards from JLC I can use them for different projects. Same for the MCU-part (idk I may just use the arduino nano footprint), the half-bridge is already a modular PCB from JLC. If I can get the main PSU board down to an easy single-sided design, then I can make it at home without too much trouble.

>>2520607
tfw you look at what industries have leaded solder exemptions.
>medical
>military
>industrial monitoring and control
wow it's almost like leaded solder is more reliable or something

Need a simple rotating motor for a purpose most would burn out from. Not sure if this is one of those "no electrical solution" applications.

I basically need a motor that will rotate to a mechanical stop, but will continue applying pressure until the stop is lifted, in which case it rotates to the next stop and continues for at least 10 minutes with 1 minute of downtime between uses without degradation.

>>2520698
hydraulic

>>2520657
For hobbyist electronics lead free is more than reliable enough.
If you want to make circuits that comply with military standards for epeen that's your problem.

>>2520708
why would I change my setup when I have pounds of kester 63/37?
switching to the inferior solder would be dumb

>>2520704
Can't. Has to be small, portable, and able to rotate about 20" in one go. Maybe I could use a ratchet mechanism to keep it in place on the stop and create an open circuit when the stop is in place to turn the motor off, but I'd be dealing with less than 1mm of tolerance so finding something that can handle that level of precision is going to be difficult. I won't exactly be able to cut the handle off a hand ratchet and use that.

>>2520698
Standard motors (excluding stepper motors) aren’t designed to run in stall-condition continuously, they draw significantly higher current when not rotating. So either use a stepper motor, or run any other kind of motor significantly below its rated voltage. You can do ohm’s law calculations using the DC resistance to check that the stall current when undervolted would have it dissipate a normal amount of heat.

Undervolting a motor naturally means it will be larger than it would otherwise need to be, a stepper may be more space efficient.

>>2520712
You're free to do as you please. I'm just saying you shouldn't sell leaded to newfags like it's some miracle shit.
The anon asking is likely quite new to all this so for solder to be the weak link in reliability would assume that all his components are quality, makes PCBs instead of using perf boards and follows all sorts of other good practices.

>>2520698
>Not sure if this is one of those "no electrical solution" applications.
A mechanical solution could be to allow some slippage in the connection to its load so that it keeps spinning even while the stuff it's driving is held in place by the stop.

>>2520716
>stepper motor
That's pretty much exactly what I'm looking for as long as it can make a little longer rotations than the typical models I'm seeing. Is ~45° per step a thing?

>>2520729
Not really. That's what the stepper motor driver is for anyhow, just tell it to make like 10 steps or whatever.

I know it's naughty to put BJTs in parallel (especially without ballast resistors), but what about MOSFETs? I have a friend who wants to design a unidirectional motor driver that does 48 volts, 100 amps. Given that that's 4800 watts more or less, I can't see how he can do this without parallel power FETs. What I'm seeing online is that the PTC nature of a MOSFET's RDS_on compensates for current imbalances, intrinsically building negative feedback into paralleled MOSFET circuits.
Is that true? Is it really that simple, getting a bunch of MOSFETs and MOSFET drivers in parallel and just letting them rip?

>>2520698
>I basically need a motor that will rotate to a mechanical stop

you can do this with a motor that includes a clutch.
the motor spins continually, but the load can vary from zero to full speed.
you control it by passing a DC current thru the clutch coil.
the clutch is also a brake. when you have no current, it locks in place.
how you control the clutch is for you to determine:
- maybe using pressure switches, or magnetic switches
- maybe the making of electrical contact between the 2 parts
- maybe on a timer

>>2520786

or a rotary encoder

>>2520768
MOSFETs act in the ohmic region when fully switched on, and have a normal positive tempco, so running them in parallel is fine. Usually you give them seperate gate resistors from what I've seen, but you may want to look into that since I'm not sure it makes sense. If switching losses are an issue you may even want to use seperate gate drivers, as to not split gate drive current up too much.
Saw a welder with like 16 IRF540s in parallel on TPAI's youtube channel, which was kinda funny.

>I can't see how he can do this without parallel power FETs
I'm also designing a simlilarly powerful motor controller, running on 12S lipos, though only to half the current since I'll be driving two wheels. Assuming it's still in stock, I plan on using the TPW1R306PL MOSFET by Toshiba, it's rated at 100A continuous, though in reality that's a function of heat-sinking. It's a real modern part with the ability to strap a heat-sink directly on top of it, that can operate up to 175C. Assuming 25C ambient, that's ∆T=150C, and it has 0.93C/W thermal resistance to a heat-sink atop it. I plan on using 6 of them (BLDC) with a 1C/W heat-sink, so that's 6C/W when split up, 7C/W total when including the FET itself, meaning maximum of 21W. The FET's gate resistance gets up to 2mΩ at 175C, so through P = R*I^2 we get I = √(P/R) = √(21/0.002) = 102A. So yeah it's possible, arguably even cheaper than parallel FETs.
There are better MOSFETs out there, but they don't have the upper thermal pad, so you'd only get that kind of performance if you mount them to an aluminium-backed PCB. Which is arguably worth doing, but note that JLC only does single-sided aluminium-backed boards, so getting your entire driver circuit on there is more or less infeasible. I think PCBway or someone else does multi-sided ones.

Wait shit it's on back-order everywhere.

>>2520698
>>2520716
A gear-motor can provide more torque at lower power wasted than a direct-drive motor. You'd still need to undervolt them to prevent overheating when stalled, but the higher torque may be a requirement. Some stepper-motors are optimised for high stall torque.
It may be worth hooking them up to a smart power supply so the motor is driven at its full rated voltage when rotating, and undervolted when detecting a stall condition. That way it isn't really slow when rotating freely, which may be an issue especially with a gear-motor. On a brushed DC motor this could be done with a simple current-limiting SMPS. With a brushless or stepper motor, some sort of current or position sensor would be fed to the ESC (via a microcontroller?) to tell it to reduce the duty-cycle.

>>2520786
It needs to be able to provide stall torque, not just stop. Dumping that power into the clutch may not be something it's rated for.

>>2520231
>only logic level shifts down
fug

Let's say I needed to drive something a few inches in one direction very quickly. Solenoids do this pretty well, but they only do about 10mm of travel. Is there any reason why I couldn't attach 10 of them front-to-back and create 10cm of travel?

>>2520821
Better just design your own solenoid. Maybe multi-stage solenoids pushing a single actuator. See: coil-gun

How the hell would you even use this IC? If you connect it to a microcontroller, then those inputs are going to feed the Vcc back into whatever is feeding them, with only 8k limiting the current. The damn thing has higher output impedance than input impedance, by a lot, but I guess it's almost forgivable because it's an RS232 driver. I'm also getting the sneaking suspicion that diode arrangement is trying to tell me something.

I just want a damn push-pull level shifter that goes up to like 12-20V so I can use it to drive a VFD. All I can find are high-side transistor arrays like the TBD62783 or the 2980/1/2 (high-side equivalent to good-old ULN2003), which are kinda ok, but I fear not having any pull-down might lead the segments to glow a bit when dark. See:
https://www.noritake-elec.com/technology/general-technical-information/vfd-operation#vfd-operation-06
I can move my average filament voltage a bit above the 0V rail, but without something to pull the segment and/or grid below that filament voltage that bias voltage won't make a difference.

Hence why I thought to look at the MC1488, technically some sort of high voltage level shifter.

>>2520848
oh wait the 8.2k and 3.6k make a divider that prevent the input going too high
so now the input impedance is even lower

Want to use this thing to passively cool a cool. Does this suffice for 20w?

My brain is melting.

Want to use this thing to passively cool a chip.
Does this suffice for 20w?

>>2520855
those are actually pretty good heatsinks. take the fan off if you're not going to use it
my favorite heatsink is from a launch day xbox 360, copper with heat pipe and aluminum fins

>>2520864
You sound like you're expecting to be spoonfed.
A better question is how can you make a dummy load to dissipate a specific wattage for heatsink testing?

>>2520867
>can you make a dummy load

If he has to ask whether an active heat sink designed for a TDP of 60W+ will keep a 20W chip cool, you know he can't.

>>2520864
just try it.. if it gets too hot add a fan. simple.
there are many variables that affect it like enclosure etc so trying it is simplest

>>2520708
what? No. lead free is trash. It flows much worse and requires more heat to solder and releases nastier vapors due to higher soldering temp.

the only reason to use lead free solder is if you are a tree huger

Hello. Which OPamp would best be used with pic related equalizer? https://sound-au.com/articles/eq.htm

>>2520895
Any op-amp designed for audio use, like an OPA2134. TL074s are pretty common too. I recommend EEVblog's video on jellybean op-amps. You may want to check the noise specs to ensure your signal is staying significantly above the noise floor for a given input amplitude. Also specs like common-mode input range, slew-rate, etc. could be handy to check depending on the signal input, but generally a line-level audio signal isn't going to get anywhere near the limits of a common op-amp on those.

Should go without saying but that circuit wants a split-rail power supply. ±5 to ±18 is usually what you're aiming for, but some op-amps (especially JFET-input op-amps) want a wider range rather than a smaller range. Best to use an iron transformer-based power supply with linear regulators and sufficiently large caps. If you absolutely have to use a switching power supply (can be very noisy with audio circuits) then ensure you tie the ground rail to protective earth, and try to use LC filters or capacitive multipliers or zener-pass regulators on the power rails to block that noise. Common linear regulators are bad at blocking high-frequency ripple.

>>2520522
I will use the ceramic ones i have i checked the ali listing and it says they are NP0 so it should be fine. But just in case i ordered a few sets of tantalum caps from ali today so if there are issues i can swap them out (holy crap tantalums are expensive) for what i paid for 20 of them i could have many hundreds of ceramic ones.

https://youtu.be/6uhMOZUWURw?t=390

>>2520342
>excellon file is all they need for drilling
Aight, that was sent correctly. Thankfully they added an option so you can import their CAM processor settings.
>>2520461
I only use oshpark for really small boards. Oshpark wanted to charge me 20$ for 3 boards, while jlcpcb only charges me 7$ for 10 boards, with similar prices for fast shipping.
>>2520485
>I outputted drill files in gerber format, a drill map gerber file, and a job file.
I used to do that as well, last time I probably forgot the drill or via map. Now it seems you can import the CAM settings file from their site, so hopefully all is good.

>>2520864
If you want to calculate it acurately and not just eyeball it, check the datasheets of the IC and heatsink, find the thermal conductivity of the different elements and construct a thermal circuit to check if the temperatures at the IC are between the allowable ranges given a known ambient temperature and the dissipated heat (20W)

>>2520320
fresh from the spinny boi, time to solder this shit up and then moment of truth
i am so tired of redoing this project, this is pcb #3, i it doesn't work again i will crash this thread with no survivors.
Altho i have to say being able to print pcbs in 10 minutes flat is comfy as fuck. If i had to order from jlc and wait two weeks every time i would be swinging from the ceiling already.

>>2520978
btw the pads and traces look a bit rough coz i went a bit hard on the sand paper, but that has no effect on functionality

how hard would it be to make a bidirectional voltage-to-current converter that can do +/- 1 mA with let's say 5-10% accuracy? i don't know enough about electronics to understand if this is a reasonable thought or completely retarded. i basically want to make a discrete version of https://www.analog.com/media/en/technical-documentation/data-sheets/292012fa.pdf if possible.

>>2520982
>voltage-to-current
....what?
you can't convert voltage to current they are two different things if you want constant current source you can do that with something like the legendary lm317

>>2520987
>you can't convert voltage to current

yes you can. https://www.ti.com/lit/an/sboa325/sboa325.pdf?ts=1670906659232

i just don't know how hard it would be to hit my specifications.

>>2520989
you can either have CV or CC both can be done with lm317 very easily.
current is dependent on voltage / resistance,can CHANGE voltage which will change current.
I'm not sure what you want exactly.
circuit where you apply 3 volts and it sets the constant current source to 3amps or something like that?

>>2520994
>I'm not sure what you want exactly.

input a voltage via DAC (0 to 3.3V) and have this device source or sink a current of +/- 1mA into the feedback network of a power supply with an accuracy of 5-10%. pic related is pulled from https://www.analog.com/media/en/technical-documentation/data-sheets/292012fa.pdf and shows the general concept.

>>2521006
>>2520989
>>2520982
found this reference design which claims bidirectional nanoamp output with 1% error, so assuming it (or these types of discrete circuits in general) could be adapted to +/- 1mA.

https://ez.analog.com/amplifiers/operational-amplifiers/f/q-a/545872/ltc6081-ltc6082-precision-bidirectional-current-source

Grandpa passed away last month and left me these decent Bose speakers cuz we were close on music but the left speaker crackles like a motherfucker. I've replugged all cables, done a power cycle reset and switched sources. I took apart the speaker and it was mint on the inside, absolutely perfect. Please advice on what to do, really wanna get these working.

>>2521058

>>2521060
Is your signal coming over bluetooth or a cable?

>>2521064
LM350
same thing more amps

>>2521060
Swap the left and right speakers. Still crackling? Check the 3.5mm signal input jack to the amp and use another 3.5mm patch cable between the amp and source.

>>2521073
>LM350
whoah, slow your roll there elon musk, that thing costs more than twice as much as 317

>>2521060
>>2521072
it's over an optical cable to my DVD player. Bose has and annoying system. I used RCA cables and the issue persisted.

>>2521074
Takes RCA and Optical. Changed cables and sources many times over and no better. Swapped speakers too and same issue.

Wondering if the issue is with electronic interference, I have many power cables and power bricks running behind the wall.

>expertly solder quad package with million legs
>perfect run, single solder tip drag, no solder bridges, looks like if it came out of a chink fab house
>notice the chip is upside down

>>2521078
Check the left RCA output jack on the amp and the input jack on the speaker for bad solder joints. Test the speaker resistance too.

Mechanical connections get flexed over the years and loosen from the PCB due to plugging/unplugging and cable strain.

What would be the simplest/cheapest way to make a bunch of Bluetooth beacons? Doesn't need to be anything fancy, literally just broadcasting a single message over and over. I want to see if I can work out positioning in a room through relative signal strengths.

Is it worth doing any prep before starting an EE course or should i just wing it?

>>2521089
It's a course so it's de-facto made retard friendly. Unless you're somewhere in Korea where your life depends on academic performance there is no need to be too eager about anything. With EE will be learning your whole life, so don't get sucked in having to achieve top results and burn out before you even start to get to the interesting things. Courses are just there to give you the basics, everything else is self learning.

>>2521089
even if I could afford that shit they'd discriminate against me based on gender and race

>>2521116
>even if I could afford that shit
Where i am an advanced diploma in ee (which is what im starting) is free, you just have to pay ~$750 a year in "resource fees" and $20 as a security contribution (fucking stupid i know)
Also the course only considers academic achievements and work history for application
>>2521110
Thanks for the reassurance. Its definitely anxiety about falling behind. I went this route because i dont want to deal with the stress of an ee bachelors just yet.

newfag here
going to make my gramps a crystal radio for christmas and was wondering if you needed a crystal earpeice or if just regular earphones/earbuds would function?

>>2521139
you just want a low impedance *phone. The power of the radio signal alone is all that's going to be available to drive the transducer, so the lower the impedance the better.

>>2521058
just use a LM7809 or, for better efficiency at high currents, use one of many switching regulators

>>2521146
No, you want high impedance phone. Regular earbuds and headphones are tens of ohms, studio earphones hundreds of ohms and crystal earpiece kilo-ohms.

>>2521149
>>2521146
shit my bad, the adderall hasn't kicked in yet
yeah you want HIGH impedance

Talking on crystal radios, have anyone here tried to make FM crystal radio? It seems to be possible and components are much simpler than on AM sets https://web.archive.org/web/20080303203252/http://solomonsmusic.net/FM_CrystalRadio.html

>>2520982
If you're using it closed-loop anyhow, (e.g. can trim for error with the MCU) you could just add a series resistor from the DAC. But that still outputs current as a function of the difference between the voltage on either side of the resistor, so it won't be floating unless you use even more MCU feedback, which might not be fast enough. A single op-amp design would also not be floating.

If you need a floating converter then yeah a circuit like >>2521053 is probably necessary. Would be easy to adjust it for ±1mA, just change the key resistance values, swapping the 10M for 1k should work I think. With such a low resistor you may be able to get away without the buffering bottom op-amp. The ripple-dampening network (1k,100Ω,10nF) is a requirement, though values will depend on your situation.
That schematic is a split-rail design though, so you'll likely want to use a 4th op-amp to buffer a 50/50 voltage divider between 5V and 0V.

>>2520994
Mate he wants a variable current ranging from -1mA to 1mA. Your wanky LM317 circuits can't do that.

>>2521082
ESP32s, probably.

>>2521139
Could likely get standard headphones to work if you used an extremely wound transformer. Like 100:1 or 1000:1. Crystal earpieces are capacitive though, so you'll probably need to get that post-detection capacitance from somewhere. Otherwise the high-frequency shit might mess with the transformer.

>>2521159
not planning on using a transformer, my goal was more or less my pic, if a crystal earpiece is what i need i'll get it
i was kinda hoping i wouldn't have to shell out 7 bucks for a single earphone though but whatever,its for my gramps

>>2520398
>>2520978
>>2520320
of fucking course... it doesn't work... FUUUUUUUUUUUUUCK
fucking why??
i don't understand.
I even tried to use the resistor and cap values recommended but it made no difference.
Talking to the mic outputs absolutely nothing above the noise floor, and only when i blow straight into it i get this, basically it bottoms out the op amp all the way to the rail.. but why? why does it only pick up that but not any normal sound?
The only thing i haven't tried changing yet is the tranny attached to the electret capsule, since i doubt it's that, but i will try it tomorrow to swap it with J111 i have lying around here.. after that there is no more hope left.
I regret ever fucking doing this cursed project.

>>2521172
>7 bucks for a single earphone

if you can find one of these, the earpiece is around 2000 ohms so it'll prob work.
they sell fake ones for use with your cellphone. those almost certainly will not work.

>>2521187
bruh they straight up told you it wasn't gonna work
you have some kind of fundamental misunderstanding about what type of microphone you have

>>2521201
i used the values the anon recommended so it should be working. the tranny is also one recommended for this capsule
in the original circuit i posted before the mic was actually working fine but there was a bunch of 50hz noise, which i assume was due to the shitty chink instrument amp, which is why i'm trying those other preamp circuits

>>2521187
Are you sure someone didn't put a curse on you? Try measuring the voltage at drain of JFET (resistor R1), it should be around 4.5V. Same with opamp, you should see 4.5V on both + and - inputs. Output of opamp should also be at 4.5V. If you measure those voltages then circuit is very likely working as intended and this can only mean that your capsule is fucked.

>used too large of an output resistor on an optocoupler
shit, forgot how phototransistors work

>>2521309
>first the 1k/100k arrangement didn't work
>now the 1k/1k arrangement doesn't work
wtf

>>2520867
not the anon you respond to, but LED that you can buy in various wattages. IDK if that would work, I do have some 10W resistors I could put in series. Im new to this shit so I could be way off.

>>2521333
>alibay ams1117 stopped working and gave 12v to my MCU
it never fucking ends

>>2521219
>Are you sure someone didn't put a curse on you?
not impossible at this point.
I really hope the capsule isn't fucked, it's the most expensive part and the shipping takes forever too.
But then again in the old circuit it worked, and apart from dropping it on the ground once, and not too hard either, there isn't anything else i did that could break it i think.
At the tranny drain i get 9V (my power supply is 9V)

at opamp + and - i get 6.8V and 6.1V... hmm shouldn't those numbers be the same? from my limited understanding the opamp should balance those inputs to be exactly the same. But while measuring it i noticed i possibly fucked up here,
the opamp is connected with + on top and - on bottom, but actually in the original schematic i followed it's supposed to be the other way around (so the pins 5 and 6 should be the other way around) could that be the problem?

>>2521512
a-anon, i think you put the op-amp the wrong way around

>>2521518
yeah, i said that in my post, i am fixing it right now, so i'm curious if it will help or not, but at this point morale is at all time low so i doubt it

>>2521520
OH MY GOD FUUUUUUUCK HOLY SHIT THAT WAS IT IT'S WORKING NOW
i am so fucking tired... this was supposed to be one day project .. IT TOOK ME A FUCKING MONTH

FUCK.

>>2521523
Fuck that one article i read that says the opamp inputs are interchangeable.
I think in this case it would work if the opamp had dual power supply so i the - rail had access to -9V it would work but the sine would be inverted but because the power supply was 9v - 0V the op amp had no way to pull negative voltage from to balance the inputs.

But it still doesn't explain why my original circuit didn't work, because that one was directly copied from the youtube video, had a proper +15/-15v power supply and many people who built it had it working just fine.

>>2521523
still a bunch of noise on the output but right now there is no shielding so thats expected, let's see what happens when is put the capsule in a wire cage and the pcb in grounded tinfoil.
The youtube guy has pretty much zero noise so i really hope it works for me too because i wasted way too much time on this already.

>>2521531
i just wanted to stress how much i fucking hate working with audio.
I have built piles and piles of digital pcbs and they usually work on the first try, no problems, or just some minor stuff. This is my second project with audio and the first one was equalizer that splits audio into channels and it was also a nightmare. Audio engineers are masochistic freaks.

>>2520987
>you can't convert voltage to current
What is a resistor

I just dissassembled an LCD TV with defective backlight. Unfortunately one of those flex connectors partly ripped off.
Can I somehow glue it on there again or is it screwed for good?

i'm learning lt spice and jesus that UI is such a fucking piece of archaic garbage!
all other simulators like falstad or everycircuit aren't as good logically, but their UI are light years ahead

>>2521595
if the traces aren't damanged and you align it correctly then sure why not just don't put the glue under it or it will prevent contact

>>2521603
You're just used to other software. LTSpice's boomer interface is actually very efficient and fast once you learn it.

>>2520398
You have tricked me! ;_;
I was getting massive noise on the output of the mic where i was barely able to hear my voice.
So i tried changing all the resistors and capacitors to match the ones in that original schematic from TI i posted and it now works splendidly. Like 90% of the noise is gone (the remaining 10% is prob just EMI since i haven't added shielding yet) and i can hear my voice very clearly, and crisply.

>>2521668
Not intentionally. I don't have a full picture of what you're doing and I can only guess most likely problems. From >>2521512 values were quite off from theoretical expected, even if it was flipped. That means that there is something with biasing that doesn't quite match the model.

>>2521595
But it would have to press on there to make contact then. What do they originally ues? Do they have some sort of conductive glue?

>>2521619
>>2521718
duh!

>>2521750
even if you had conductive glue, how are you going to put it on the pads without getting any between the pads which would cause a short? you have to secure from above with something like tape

>>2521752
>you have to secure from above with something like tape

tape wont work.
i've used insulation foam to press down on it, and it kinda works.
the foam is squeezed between PCB and rear of the case.

>mounting surface-mount components on the surface
>losing one whole dimension of space
bro do you even 3D circuits?

>>2521782
>free inductor in every branch
>free loop antennas
3DPD circuits are like women. You'll never be able to understand them. 2D is superior in every way.

>>2521782
>cracks your MLCCs
nothing personnel kid

>>2519728
Lmao I'm gonna call them tranny's from now on

>>2521532
Nah man, you're just stupid. And helpless. Everyone can tell you're a thirdie because you ask such stupid questions and expect everyone to help you.

im trying to make this pos work again. Whenever I plug it in, most the bulbs blow. I would like to maybe add a resistor but Im not sure if that is right. what should i do?

>>2522153
The socket is 120V I hope?

>>2522160
kek, yea ( i should have mentioned that)

>>2522153
>>2522185
Those lamps are all in series, right? If they say what voltage they're rated for, you may be able to calculate if they're being overdriven by counting the number of lamps. Generally if one dies they die short-circuit so the whole string keeps running, but that puts a higher voltage across them all. So if they technically shouldn't be being overdriven, try just replacing every single bulb (with ones of the proper voltage) if you haven't already.

One possible solution is to add some amount of dropper capacitance. A motor start cap is likely the way to go, X2 caps are too small and would reduce current too much. Ballast inductance might work, if you're careful and add TVSs or MOVs to kill spikes.

could keeping soldering iron on transistor leg for few seconds kill it?
i tried doing diode test with my multimeter between base and source/drain and i am getting about 0.5V both ways, and i am pretty sure you aren't supposed to have any current flow form say source to base right?

>>2522220
2 seconds probably fine, 10 seconds I wouldn't be so sure. If it's so hot you can't touch the plastic bits after soldering, then it becomes suspect. Though honestly that happens all the time for me, with components that have large thermal pads, such as the D2PAK. Those components not only have much lower thermal resistance from the lead to the die, but also take a lot longer to solder properly. So chances are if getting the silicon up to 200C for a 10 seconds doesn't kill a D2PAK transistor, you're basically never going to kill a TO-92 with soldering heat alone.

If you find the datasheet, hopefully you can see a section about soldering temperature and time.

>source and base
What? If it's a BJT, that's collector and base. Yes there's a PN junction from the base to both the collector and to the emitter, getting a diode check for both of those is expected. That base-collector diode won't turn the thing on like a transistor (a high voltage at the emitter leads to easy breakdown anyhow) but that diode is actually useful in some fringe circumstances. It's apparently really low leakage. Maybe good for sample-and-hold and/or envelope-follower circuits.

>>2522222
>getting a diode check for both of those is expected
thats strange then, i tried testing another identical tranny and i get about 0.6V between base and the two other pins, but no connection at all between the two pins and base (when i put the black probe on the base leg)
Could maybe static kill it or something?
I guess i'll try to change it but this is a damn mystery,

>>2522222
also not sure if it matters but it's a n jfet

>>2522229
oh shit it was that...
i can't believe trannies are this sensitive, it was the one with long legs and i held the iron on the gate leg for 5 seconds and it just totally killed it..
i have to be super careful now since i only had one spare

>>2521139
heydo,
if u are building this from scratch form stuff laying around in the shop then here are some useful tips:
1. u can use Schottky diodes instead of germanium ones; the lower the forward voltage the better

2. for earphones it is best to use a piezoelectric one due to its high impedance (unlike what >>2521146 said) because it will be getting its power from the radio waves themselves, u could create an amplifier with really high input impedance and low input capacitance fairly easily but I really don't see the point since u are going for that old timey vibe

3.for the coil it is better to use a ferrite rod, that usually has better audio and signal properties compared to air-coil designs, but if u are going with it anyways for old timey sake, go with either cardboard or wood as they really don't capacitively load the coil as much as pvc or abs does from my experience

good luck m8, also take good care of ur old man <3

My brothers, i am looking to build a kind of diorama of the national grid, with "generators", "motors", "transformers" and "loads" representative of residential areas and such. I want to make it so that it could be shown in schools to explain how stuff works, so it should be low power and low voltage. Ideally it would be plugged into a single wall socket and could be carried around (like in a car trunk or something).

I have some ideas of how to get it working but help me out here: synchronous motors with separate excitation, 3ph transformers and 3ph induction motors do not exist under, say, 48V-50VA?

>>2522231
>>2522248
Yeah JFETs are sensitive to ESD.

>>2522388
Schottky diodes have higher leakage current. When you've got a high impedance setup like a crystal radio, that leakage current might be a significant issue. On the other hand, the lower capacitance might be a significant boon. There's probably some literature comparing them to germanium and silicon diodes for the purpose of radio.

>>2522424
Firstly, BLDCs are effectively 3-phase motors and widely available to run on low voltages. Issue is, you need to figure out some way of starting them. Giving them a pull-start might be fine, you may also be able to convert them to induction motors by replacing the rotor with a block of iron. Drill some holes in it for copper wires maybe. BLDCs would probably be fine as generators without modification, maybe.
As for transformers, I'd either just use 3 seperate transformers, one for each phase, or rewind a standard E-I core to have a winding on all 3 branches of the E. I'm not entirely sure of the benefits of having the flux of all phases getting mixed together, maybe that diorama will show the benefits.

As for actually generating the 3-phase power in the first place, a mains-powered motor would by my go-to. Be that being a fake turbine that's directly mechanically coupled to a low-voltage generator, or a pump that pumps real water or wind past miniature generator setups.

Adding an oscilloscope to probe around the circuit would be nice.

Can you please recommend me a desoldering station for hobbyist use ? I'm mad as hell after 10 years of manual pump. It see many models review on youtube without being able to get a precise focus at selecting the good one.

I bought a new one but still no spark. What can go wrong with these. Is there electronics in these? What do i measure on the kill connector. Primary coil or what?

>>2522188
> replace them
i did, plugged it back just to have another 5+ burn out instantly, maybe more.
should I pull the cap in the pic and add it in-series?

I have some lights of emergency. Due the problems of importation i cannot buy the battery acid/lead. My question is if i can put any battery 18650 from the notebook directally to the light instaead the lead/acid. The battery original is 4v and the litium is 3.7v

>>2522655
You can probably use them to power the light. If the light tries to charge the batteries (and they are unprotected), that would be bad.

>>2522630
Wait what kind of circuit is that? Series cap only works for simple sinusoidally AC loads. Not DC or PWM, nor anything that changes its load current suddenly.

>>2522655
>if i can put any battery 18650

it'll work but it's not a good idea.
thats coz lipos self-discharge, so cant be left alone for a long time.
especially when it's hot, like in the summer in the trunk of a car.
they get weaker over time, and eventually die without a way to revive them.
so not good emergency use.
you're better off stocking a bunch of these instead, from a reputable vendor.

>>2521443
>killed the alibay max6675 too
shit that’s gonna be a pain to replace, hopefully some dropshippers sell them

>>2522630
>>2522720
needless to say I already wired it is series and it did not work, it slowly lit the lights dim.

>>2522800
Cap too small is my guess. You could try it with a larger one, maybe 10-60µF, but better calculate the current draw (or simulate it) by using the rated wattage/current of the string.

I bought a 6V 4.5Ahr lead acid battery for 3.5$. I'm going to make my own battery management system for it using an atmega8. I plan on having an under voltage cutoff, over voltage cutoff, over current cutoff and temperature sense. I have a few nice 49A IRFZ44N MOSFETs.
I'm going to have so much fun. I've worked on a BQ76XX based electric bicycle BMS before at work but never made by own without an IC

interesting, turns out power rails pick up emi noise too, which then transfers to the circuit, even if i add a juicy cap there to counteract that, so it's worth it running power cables shielded a well (that of course applies only when you have a good power supply, if it's shit it will be way more noisy than emi anyway)

what the fuck is this retardation? half of picture have GND Data- other half GND Data+
So which the fuck is it?????????

>>2523004
one of those pictures is usb 3 and the other isn't

>>2523012

>>2523004
anon...when searching for pinouts you need to be specific and use proper resources and search parameters. not just google image indexing
1. one is usb 3 type a the other is usb 2 type a
2. one us a socket one us a plug

i assume you typed something like "usb pinout" which gives absolutely no indication of what youre really looking for

>>2523033
>one is usb 3 type a the other is usb 2 type a
Not that guy but unless I'm missing something USB A is supposed to be backwards compatible to 1.0. Pretty sure flipping the data lines would fuck with that.

>>2523004
It's not even the same connector.

>>2522930
Cool project. Use an op-amp differential amplifier across each cell to get the cell voltages down to a good level for the ADC pins to read it.

>>2522947
Don't forget grounding, helps remove common-mode noise. Use common-mode suppression chokes too. Common-mode noise is the worst, as you can't get rid of it with filtration caps or linear regulation. The noise may be coming from inter-winding capacitance from the AC-DC PSU (if not the PSU itself), so consider replacing any transformers with ones with shields between windings, or completely separated windings. Marco Reps did a video where he experimented with transformer designs.

>>2523004
Yeah flipping the data lines like that would cause the data to not get transferred properly. Just shows how little you can trust google images.

>>2522502
>Firstly, BLDC
Thanks. Yeah I had something similar in mind, was just wandering if there's some equipment I could buy off the shelf or if I have to rewire or even build stuff myself. Like excitation coils with even brushless excitation i guess are just too difficult to manufacture at small scale, no matter the application....

>>2523247
Wait, would car alternators work?

>>2523254
>would car alternators work?
Yes, a car alternator can be used as a fairly bad BLDC. Do keep in mind that you need to provide 2-4V to the rotor coil because there are no permanent magnets in an alternator.
There's also the issue of portability but I don't think I need to explain that.

how safe/dangerous is it to use a chinese relay?
I've been wanting to setup some automation stuff with a raspberry pi. like, using DC output to toggle the pumps from my rain barrels, or to turn on heaters for chickens.
I don't know that much about electronics but how badly could something go wrong if i use these relays to toggle mains electricity?
Maybe in the worst case it fails closed and I should put a mechanical relay or something upstream as a failsafe?
like, the pi turns on the heater for the chickens, but the heat keeps rising even though the relay should be open, and then the pi flips the mechanical one to cut the power?
is that even necessary?

>>2523264
like pic related

>>2523264
you can get thermal fuse that trips when it overheats and resets itself when it cools down.
chinese stuff might be notoriously shit but you should never design something that can be dangerous from a single failure, always have some fail safe. two relays isn't great either what if the pi crashes and it gets stuck on? fail state systems need to operate completely independently.
rain barrels is one thing, a chicken heater you are talking about a living thing.

>>2523264

For starters, I'd get something like pic-related that claims 60 amps and not use it for more than 10 or so. Secondly, have a fail safe because any SSR can fail closed. In my case, I use space heaters that have the thermostat knob you turn, and I set it just above the desired temperature, and they also have an over heat sensor built in (at least mine do) that protects against any sort of failure.

>>2523264
If I was gonna run a pump I would use a contactor.
But I work in industrial so overkill is the name of the game

>>2523269
>>2523270
i'll see about buying a proper namebrand one, at least for the heater. a huge fire even once would more than pay for a pile of relays.
i wonder how consumer appliances that switch AC work. like, i've never had a microwave oven just stay permanently on, but it must have some sort of controller inside to power the magnetron.
maybe an SSR is not really needed in this situation anyway. even with a heater, it's only being pulsed every so often. it's not like i'm speed controlling a motor.
would a mechanical relay be a better choice? they fail open usually.
i guess i should be more specific with what i plan to do. I am thinking about running something like heating tape under the floor, because what i'm using now is just a space heater and it's just not a great solution.
then i could put a thermocouple in the coop and pulse the heating tape until the temperature is stable.
maybe i should do more research but i don't know what a good failsafe would be in case the heater is running full blast and cannot be shut off. maybe a fuse with less current than the tape is expected to draw at 120v? and then i just never run the heater at full blast?
then if the heater is sucking full wall power for long enough, it'll pop the fuse before the room gets too hot.

>>2523306
>what a good failsafe would be in case the heater is running full blast

when i got heating strips, i just chose them by maximum wattage.
so the 250W one is only gonna use 2A no matter what tragedy happens with the control loop.
it's guaranteed safe coz 250W is not gonna get anything close to its combustion temp.
ditto for the 400W one.

>then i could put a thermocouple in the coop and pulse the heating tape

how about a common household thermostat?
either mechanical or electronic.
used and loved by billions.

>>2523242
>Cool project. Use an op-amp differential amplifier across each cell to get the cell voltages down to a good level for the ADC pins to read it.
Just what i intend to do anon. I'm a bit confused about how to do current sensing in both directions without having a -5V rail.

is this a good setup for cleaning usb power noise?

>>2523552
A differential amplifier's input resistors will pull the voltages towards the ground rail you give it. If you use a +2.5V ground, then I think the small negative voltage you get across the sense resistor will still end up between 0V and +2.5V, without getting out of the common-mode input range of a ground-sensing op-amp (e.g. LM324). Though it depends on your gain resistor selection. For 1:1 gain (i.e. not useful at all) then the ~0V at the sense resistor pins will end up at ~1.25V on the op-amp inputs, both at 0A and at significant currents, so perfectly fine. For 1000:1 gain, you'll get ~+1.25mV on the op-amp inputs at 0A, and if the voltage difference across the current sense resistor is more than -1.25mV that will end up getting outside the common-mode input range. It should be possible to pick a sensible gain that stays within the common-mode input range. If in doubt, simulate it.

>>2523554
Not sure about the choke specs, but probably. 220nH? µH? The inductance you can get away with is a function of how pulsed the current will be. For an analogue supply you can probably go pretty large so long as you add some TVS diodes, maybe clamping diodes across the choke itself. If it's a digital supply then I wouldn't use a choke at all unless you're getting some really nasty noise.
Why not put a 10µ before the choke too?

I need advice on good textbooks about SMPS design and on transformers design and manufacturing. I want to make a compact electrostatic generator, so very high voltage (50 kV to 100 kV roughly) but low power and extremely small currents. I have half an idea for the overall topology, thanks to what I could gather on the internet, but I want to make it as power efficient and as safe as possible, especially in case of sparkover. The diy projects I could find don't inspire confidence in their safety as none are optimized for this kind of application.

>>2523575
that sounds like xrays

>>2519612
Why the red one have a pp resistor as well?
Diy into tranny shit too?

>>2523585
Nope. I want to play with fluid/gasdynamics. Scientific literature on the interaction between the two exists, but it's not great. Besides, I'm fussing over safety, why should I mess with xrays of all things?

>>2523562
>A differential amplifier's input resistors will pull the voltages towards the ground rail you give it. If you use a +2.5V ground, then I think the small negative voltage you get across the sense resistor will still end up between 0V and +2.5V, without getting out of the common-mode input range of a ground-sensing op-amp (e.g. LM324). Though it depends on your gain resistor selection. For 1:1 gain (i.e. not useful at all) then the ~0V at the sense resistor pins will end up at ~1.25V on the op-amp inputs, both at 0A and at significant currents, so perfectly fine. For 1000:1 gain, you'll get ~+1.25mV on the op-amp inputs at 0A, and if the voltage difference across the current sense resistor is more than -1.25mV that will end up getting outside the common-mode input range. It should be possible to pick a sensible gain that stays within the common-mode input range. If in doubt, simulate it.
This seems to pretty well anon! At least in the sims
I'm using an LM324. I'm planning on having three ranges for current measurement, 1mA to 100mA, 100mA to 1A and 1A to 10A
10A ought to be enough current for a 4.5Ah 6V battery. I'm planning on using the last remaining opamp to wake up the MCU on logical change

>>2520895
NE5532 is used almost everywhere for audio and the NE5534 is the single version of the 5532 with slightly better specs and the ability to adjust offset voltage. LM4562 is another option optimized for audio and it's a great part. TI has a good site to research and pick something you'll like.
https://www.ti.com/product/LM4562
Or you can just use a 741 for maximum crap factor.

>>2523602
I hope this is a joke.
Any self respecting audio engineer would never use anything except high grade vacuum tubes for amplification for that sweet sweet matching harmonic frequency noise.

I'm trying to connect 12 pins on one PCB to a 12 connection header o another board. Both the pins and header are single row half pitch (1.27mm). I can't seem to find the half pitch IDC ribbon cable I need so I figured I'd just make one but I can't find the half pitch components I need for the job either. Is there such a thing as half pitch grouped dupont connectors? What's the best way to make this connection? FWIW the boards are only going to be sitting 8 or so inches apart.

>>2523591
I think that's the green one's hand

>>2523604
Ironically I prefer to use pentagrid converters for maximum sonic displeasure. Nothing like an inherent 8kHz lowpass paired with a sufficient amount of odd order harmonics and a helping of IMD, if designed correctly.

>>2519612
I just fixed a $700 Klipsch soundbar with a few replacement capacitors. I bought premium 5,000 hour capacitors instead of the 2,000 hour caps they put in at the factory

What's the significance of almost a straight line parallel to the X-axis in the saturation region, not a slightly sloped line? I read somewhere that it's how you explain high gain or something like that. Or is it just about linearity?

>>2523717
Internal output resistance of transistor. You can interpret slope as dy/dx, which in your case means dI/dV and when it is constant it represent inverse of resistance - conductance. At 0 slope you get parallel line, 0 conductance or infinite resistance - this is perfect current source. At non 0 slope this isn't the case, change of Vds will cause change in current - current source has some output resistance.

>>2520630
Those houses are built to scam the government and not ment to become livable. Even many high quality products are made in the PRC. Your problem is with your own, nonexistent, production.

>>2520630
This is pinnacle of engineering.
>everyone can make a bridge, only an engineer can make a bridge that barely stands

>>2523721
Why do we want a perfect current source? What does it affect, linearity, stability?

ESP32-CAM or any camera board.
So I'm digging into cameras recently, want to make a sketch which checks the color of pixels in certain frame region. All the standard examples work just fine and I'm able to stream video.

But for some reason I can't get the actual content of a frame buffer if I strip an example from all that HTTP server bullshit. What do I do wrong? Here is the explanation from camera datasheet (OV2640), which explains the colors in certain frame resolution. It seems my buffer stays always the same even though the framerate and buffer length changes with every frame. Im stuck for three days with this shit and am confused as fuck. SEND HELP PLS

>>2523728
What does casting an integer to String even do? God, I hate C++.

>>2523730
I have no idea bro... I just do what my heart says because after three days trying to get ONE FUCKING FRAME my brain is dead.
Here is an exemple which does pretty much the same but saves the frame into a file
https://randomnerdtutorials.com/esp32-cam-save-picture-firebase-storage/
And it fucking works. Why mine does not?

>>2523733
It really looks like fb->buf should point to image data of some sort. Try to confirm whether your output is correct. Start by using separate int variables and confirm whether they're sent to the serial output as you expect.

>>2523735
I have a feeling I output it wrong. How to work with pointers properly if *fb->buf[0]* is wrong?

>>2523730
For one thing: additive operation on integers adds the integers like this
>2 + 3 = 5
adding strings the same way usually just concatenates the like this
>'3' + '4' = '34'

>>2523736
fb->buf is probably uint8_t* and fb->buf[0] will be an uint8_t. Aka unsigned char. The real question is whether casting uint8_t to String (that is the "(String)x" where x is uint8_t) converts the number to decimal. From your screencap, I suspect it actually does this, but we can't be sure.

>>2523727
A perfect current source will be able to push current no matter what. If you had ideal current source of 1A and had nothing connected to it and just left wires dangling in the air, it would raise so much voltage that 1A would flow through air. This obviously doesn't work in reality and there is a limit - finite output resistance. Output resistance limits your power output. It does affect linearity, because real transistors are inherently non-linear devices, output resistance is not constant and varies with various parameters - but this is usually neglected. Stability depends on per-circuit basis, but in most cases it won't affect it.

>>2523741
Apparently it does cast it properly, the values definitely look like a color data 0-255. But it stays always the same for every frame, even though I obviously move my camera around and it has to change.

>>2523749
Your buffer length is changing. Maybe the buffer is actually in jpeg format, and the first bytes happen to be the same for every frame (jpeg headers)?

>>2523752
It is jpeg and i thought the same, so tried different random values. Whole buffer is the same all the time. Maybe for some reason it is not updating even tho the length of it does?

>>2523753
Post a screenshot of the http code or whatever it is.

>>2523753
>It is jpeg

a reply to your other post in /mcg/ has a guy converting jpeg to rgb array >>2523743

>>2523754
Got none http, plain camera routine. Here is the code https://ideone.com/OuGlHK

>>2523759
Oh, thanks gotta go check it.

>>2523749
If you are not used to C(++) you might hold an old reference to your object (which makes it not getting released in memory).
Read up on pointers, dereferencing and such again. I'm not very experienced either but remember having had your problem myself.

>>2523749
>the values definitely look like a color data 0-255
First 2 bytes of JPEG are 0xFF (255), 0xD8 (216) which is exactly what you get in >>2523728

If you want to read colors you need to decode whole buffer into raw data.

>>2523761
You said something about HTTP server bullshit.
Yeah, your example makes the camera return jpeg. Jpeg might return the same bytes in some places because of headers or out of coincidence, but if you tried many byte offsets, it's weird that it never changes. A nice trick to check if anything changed without outputting everything is computing the CRC32 of the whole buffer.

>>2523749
>>2523763
fb_buf is object I'm referring to here

>>2523745
So a straight line means that Id remains constant and only changes with a change in Vgs but not in Vds. But why would Vds change on its own?

>>2523763
>>2523768
>objects
Jesus Christ, forget about that. It's a pointer. It most likely does not point to a memory position that has been freed before, because the code is too simple (look at the ideone link), and because you shouldn't use malloc. Though I would piss myself if it turned out that the caller is supposed to release the fb pointer.

Wtf is this cast to String bullshit? What is String even? It is not the same as std::string is it? Is it some arduino specific bullshit? Is it a way to print integers as strings? What happens if you just pass an integer, as if you used printf with %d?

>>2523773
>arduino specific
Probably. I don't know why Arduino chose to confuse inexperienced people new to programming this much.

>>2523770
Should he not let a library do the parsing if he is not familiar with C++ and byte streams? Picking pixels from a stream that begins with the jpg header is retarded...

>>2523775
Here is the Byte #502 printed >>2523753
Its not JPEG header...

>>2523775
He should make the code request a raw format, or decode it manually as another anon suggested. But at this point, I think he's just confused that the output isn't changing.

>>2523769
>So a straight line means that Id remains constant and only changes with a change in Vgs but not in Vds.
If you mean line that is perfectly parallel to x-axis then yes. As soon as you have non 0 slope Vds affects Id.
>But why would Vds change on its own?
I don't know what you mean by this. In >>2523717 Vds is parameter - thing that you yourself change. Only output is Id. Vgs is also input. That plot is made by taking the transistor hooking it up to voltage source on drain and gate, ground the source and sweep voltage from 0 to some value while measuring current that flows into drain. Vds is always fixed because source is at ground 0V and drain is forced to whatever you want.

I am thinking of buying a adjustable power supply for the purpose of charging batteries (and as a bench power supply), and I have a 100w 19v5a laptop charger that I could use with a Riden DPS frontend to charge a 36v battery at 1 amp (could be lifepo4 or li-ion), and I have settled down on 2 options for how I would set it up.
I could use RD DP50V5A which is very compact, but it's only a buck converter which means I need to buy a cheap boost converter to fix that.
Or I could get RD DPH5005 which is a buck-boost converter, but it's a tiny bit more expensive, and I plan on occasionally bringing this onto a bike, and it looks like all the components are probably going to break from vibrations.
But I have a feeling the DPH5005 is probably more valuable because it's bigger and made of better components, but I am not sure.
>DP50V5A
https://www.aliexpress.com/item/32596330856.html
>dumb boost converter (I know, it has no fuse, but the DP50V5A should have one)
https://www.aliexpress.com/item/32658959267.html
>DPH5005
https://www.aliexpress.com/item/32840324731.html

>>2523600
>I'm planning on having three ranges for current measurement
I'd then use three RRIO op-amps, each with different gain settings, powered by the same voltage as the MCU/ADC. That way even when you're seeing 5A, the 100mA range isn't outputting more than the ADC can handle. Choosing an ADC reference below that voltage rail may make for slightly better linearity, idk I'm not too familiar with RRIO op-amps.

Otherwise you'll need to either have some sort of gain-swapping circuit (maybe possible with CD4053) and/or a clipping diode. Also LM324s can't output close to the 0V rail anyhow.

>>2523611
The ribbon cable I use for IDC connectors seems to have 0.05" pitch anyhow, since it works perfectly for double-row 0.1" pitch connectors.

>>2523774
that's why i think arduino is cancer. and especially in conjunction with SoC like ESP8266 etc. the proper way to learn is to roll your own and write in plain c/c++ but unfortunately it is hard to do with ESP8266. it is a weird architecture and it comes with TCP/IP stack and other custom libraries so it is a mess and a big fucking black box. When I look at the instructions like the anon above posted I want to kms. It's hard to learn like that. STM32 is much more user friendly, and you can actually understand what you are doing.

>>2523884
stm32 is incapable of media encoding/decoding.

>>2523779
I meant to ask, in practical terms, when FET is used as an amplifier, why would it matter if the line Ids(Vds) is straight or not. Presumably if it is not straight, then Ids could change with a change in Vds AND Vgs is unchanged (we are on the same curve). But when does this happen?
Typically we set a load line and choose an operating point which slides up and down along the load line as Vgs changes and so does Vds and Ids as we hope from one curve to another depending on Vgs.

>>2519612
I have only one thing in my syllabus to get an university plan complete on engineering.
So I'm going to make an special test for analog devices, and it covers a lot on analog filters, is there any special bibliography on the subject or somekind of recipe to work out the: "design Butterwoth, Eliptical, etc...."; I think this is all come down to the lineal control principles and the second degree general equation, but I wonder if there is something more pragmatic.
Thanks.

>>2523934
well you have to program it. pretty sure the libraries exist you will just have to compile the source code and run it. or roll your own.

>>2523946
Good luck to do it with those stm32 mcu's you think of with 40 mhz or whatever.

>>2523951
there are all kinds of stm32 chips, many are close to 200Mhz and have FPU and parallel dsp instructions and plenty of SRAM. i don't know how they compare to ESP chips

>>2523959
Can you give me an example of a development board with stm32 capable of media encoding with proper pfs, wifi and possibility to hook camera without soldering? Im beginner and I would like to try an alternative to ESP32 CAM.

>>2523727
>Why do we want a perfect current source?
You probably don't. If you get to that level your fet is probably going to blow. When the line goes flat it means you hit the max current it is able to pass before Vds starts to go up, which means it's gonna heat up real fast.
>>2523939
>when FET is used as an amplifier
In that case the "linear" part of an amplifier is usually a resistor. The transistor only needs to be monotonic, the linear resistor + feedback will compensate for nonlinearity in the transistor. If you really care about the linearity of the transistor, you are looking at the wrong graph. You want the Vgs vs Id graph.

>>2523962
>pfs
FPS

>>2523962
Not sure about a development board for that. STM32 didn't support wireless in the past, they have some new series STM32Wx but I think they only support BT and they are pretty slow.
https://www.st.com/en/microcontrollers-microprocessors/stm32-wireless-mcus.html

Sounds like your task is pretty specialized, fast CPU + wifi +DSP+camera, so perhaps ESP32 is your best bet. I roll my own and buy bare STM32 chips and solder and program them but I am not doing anything fancy like that. I needed wifi for one project so I used STM32 with a separate ESP8266 in At mode so I didn't need program it. I programmed the STM32 chip to talk to ESP in serial mode by sending AT commands. So basically using ESP with Wifi as a slave peripheral and STM32 is the main MPU master unit.

Here is one dev board:
https://www.aliexpress.us/item/2255799864726702.html

But it uses two separate chips: STM32F103 (slow!) and a separate NRF24L01 just for wireless.

NRF chips are STM32 based wireless SoC wifi+BT so they could be used on their own and can be directly programmed but afaik they are slow (40-60Mhz).

>>2523978
Raspberry Pi's are also good for my purpose, I would take it to try but ESP32 CAM with all those things costed me a little bit less than $7, so it's pretty hard to beat.

>>2524005
yes but also causes a lot of frustration and hair pulling. and reading those confusing instructions with examples that do not work ugh. but yeah such is life, nothing is perfect.

Anyone have one of these 4 dollar multimeters?
I was wondering if anyone knew if there is a way to solder in a speaker so when you do continuity tests it beeps

ANENG DT830B

>>2524016
It doesn't have a dedicated continuity test. Just resistance and diode. Buy a better meter, even a $10 meter has a beeper, let alone a nice autoranging AN8009 for ~$30.

>>2524020
Alright thank you

>>2524020
Not him but I was wonder if it would make sense to buy the cheapest one you can find just for the sake of the case and controls and the LCD, and then build your own schematics and make it a a decent 3000-6000 count meter.

>>2524030
>keep the brittle plastic housing
>keep the shitty rotary switch
>keep the display that can't display a most-significant digit larger than 1
>ditch the inner electronics
Honestly I'd rather do the opposite, but I say that as a 3D printer owner. I'd like to get a kelvin-measuring meter into a nice portable chassis. Or make a nice case for one of those chinky transistor/LCR testers.

Electronics noob.
I need some pressure sensitive switches/sensors of some kind that detect multiple pressure levels (to be pressed with your fingers). What is the correct terminology for what I'm looking for?

>>2524068
PS2 controller

>>2524068
For something with non-trivial travel distance, you're after an "analogue trigger", or maybe "analogue axis". Like the triggers on a game controller. Physically these are just a potentiometer (rotary or linear) with one or two spring preloads so they return to their original position. Naturally this means fabricating mounts for the static potentiometer, the moving actuator, with spring/elastic mounting points. Doable with a 3D printer / CNC machine. Potentiometers do wear through though, you can use any kind of rotary/linear encoder. Be it magnetic, optical, mechanical, etc.

For something that doesn't move significantly at all, then the traditional way would be sticking a strain gauge to the back and measuring its decrease in resistance. These could be mounted to the back of any panel, so long as it's not too flexible. They're also small and pretty cheap, but do suffer from thermal expansion-based drift if you don't match the tempco to that of the panel it's stuck to.
But if it's a finger pushing on it, you can also measure the deformation of the finger skin using a simple capacitive touch pad. Harder finger press = finger more flattened against the pad = higher effective coupling area = higher capacitance. These would be easy to fabricate into a PCB directly, or by mounting a metal plate on the back-side of a thin insulating sheet. These won't work through gloves, and wet hands would mess with them.
Resistive touch-pads might also work, but I'd tend to avoid them.

>>2524016
You can diy yourself a continuity sensor.
This video shows something that starts off with something really simple, and then moves onto something very overkill.
https://www.youtube.com/watch?v=N2M-p-OGvPg
This was the first video I clicked on, so I am not sure if there are better options, and I personally haven't made this.

>>2524172
Thats cool

I need an opinion.

I have this Razer Blackwidow keyboard from ~2014. One the 's' key failed for no reason (no liquid spill damage, no physical damage, etc.) and I couldn't log into my computer. I originally thought it was just the 's' key, but I discovered that the 's', spacebar and the windows/command keys have all failed.

I disassembled the thing to see if something was going on with the switches physically, but my multimeter and my oscilloscope tells me that the switch for the 's' key is working fine.

I figure I should try to wiring it to one of the unused keys as I *think* that's what someone tried. I attempted the wiring today but via keyboard tester that the 's' and the spacebars are working....except they are somehow triggering the page up and page down respectively and the windows/command key still isn't working at all.

At this point, I am convinced that there is a component that is failing but I have no idea what it is. Has anyone seen this type of shit before? Also, at this point, should I just cut my losses on this? Not like Razer was that good of a brand anyway.

>>2524306
>I should try to wiring it to one of the unused keys

instead of soldering shit, just use software, like KeyTweak.

what is this shit?
i thought i would get into BGA soldering finally since i have been using smd for a long time now and this is fucking gay, there are sizes now?
why not use one size of balls? so you telling me that if i use say 4 bga chips they can all need different balls? is there something like 2.54mm pin header ball size? as in there are various pin header sizes but 99% of all designs use 2.54mm

So I rewound, REWOUND, the shitty little transformer, it still got like 0.1ohms impedance on its windings, almost a short. I'm supposing that's how it's gotta be right? Changed the bad mosfets, it's not working and continues to blow mosfets. The caps are doing shitass they're all alright. Why are the mosfets blowing up? I'm ready to give up, I haven't got anything else to do to this inverter.

>>2524407
Tried to make a sketch for the aliexpress chinks to make me one but I realized I can do it myself. Why it's got no impedance? I'm thinking this is the reason the mosfets are blowing up.

>>2524379
Excellent question. One reason is that I use a windows and a mac, so while the 's' and spacebar is mapped to 'page up' and 'page down' respectively on a windows, I checked and they are mapped to the number '3' and '9' respectively on a mac for some weird reason, so it will most likely work for my windows machine, but not for my Mac (and god hopes I'm not using a windows VM inside a Mac). Second reason is that the windows/command key doesn't work and I use the mult desktop shortcut windows machine like crazy and the cmd is used for a shit ton of shortcuts on a mac. That one is the deal breaker.

>>2524082
Nowhere near enough inputs.
>>2524123
Spring loaded linear pots sound perfect for my application. Thank you.

>>2524394
You can use solder paste and templates usually. Maybe it doesn't work for some small one I dunno

>>2524410
>Why it's got no impedance?

it's got a lil bit of impedance, but zero DC resistance.
coz it's only 23 turns, right, which is like 5 feet of wire.
which works out to 0.102 ohms.

What are these things?

I want to make a dpad for a project. In some dpads they have comb like copper lines and pushing on it with conductive silicone of dpad connects them, which makes sense. But in others they gave pic related

What are these? How do they work?

>>2524068
BOOBA

>>2524544
Those are clearly onions.

This is so fucking annoying.
jlcpcb pays the vat for me when i order pcbs so i don't have to do anything but lcsc does not.
And they are basically the same chink company anyway. If you order pcb + pick and place services they get those parts from lcsc.
So why the fuck can't they do the same ting jlc does?
In order to avoid stupid customs paperwork i would have to basically order a blank pcb and have the pick and place components i want on it and then desolder them at home since that way they would handle the vat fee for me.
Really dumb.

>>2524464
Oh wow.
So why are the mosfets blowing up? One or 2 of them instantly blow as I touch the battery terminal. There is absolutely nothing in short, I even made sure to charge the mosfets corresponding to the circuit polarity so they don't "stall", I think that's what transient response is called. I installed the heatsink on them as well this last time. I don't know wth's going on. I blew 5 mosfets by now. It is to mention I only installed half of the circuit, so 4 mosfets and 1 transformer. Does this have to do something with it? I just want to see it light up and continue fixing the other half. Is the logic part gone to shit and maxing the mosfet gates? Is the voltage requirement too high and the other half required?

Been trying to fix this old AC voltage regulator, tl;dr already replaced some busted components but the thing is VERY sensitive to voltage changes (Supposedly it works 90-130v.)
It's got 3 variable resistors (10k), uppermost I have no idea what it controls, middle controls output (ranging from 110-130v) and lowermost controls the out of range protection mode (had to turn it all the way down, else the OR mode will kick in all the time.) I've messed with these to no avail, still finicky, can't turn any appliances on or else this thing will go into OR mode when the main goes a bit under 120v.

>>2524530
Tactile switch domes. Think there’s a proper name for them that I can’t remember, but I’ve definitely seen them on alibay and LCSC. They’re metal domes that act as both the contact and the spring, and are soldered directly to the PCB. They’re a bit shit, but if you need small or low-profile switches they’re a decent option.

>>2524651
>So why are the mosfets blowing up?

that's for you to determine using your vast experience and collection of instruments.
if you experience is vastly non-existent, you might consider a divide-and-conquer strategy.
i.e. isolating parts of the circuit to see what *stops* the FETs from dying.
you might start with the rectifier diodes, seems there's 8 of them.
disconnect one leg on each of them, then run a test.
if nothing died, try adding an appropriate load after the disconnected diodes, maybe some 12V lamps between an anode and cathode in the same 4-diode set.
if all went well, put back the diodes but cut traces or remove components to isolate other parts.
continue until you find the culprit (or you've destroyed the board, or destroyed your sanity)

one trick you might try is to put a 120V 60W lamp in series with the power input to reduce max current.
this might keep the FETs alive longer.
google for more info.

Does anyone here have a good schematic for a Sound & Light device, also known as Brain Machine/AVE devices??

upgraded my ds1054z to an 1104-xe. the UI is so much more responsive and less aggravating to use. should have bought this from the start.

>>2524665
I always replace old shitty trimpots like that, can't trust them. Preferably swap them for more dust-resistant designs.

>>2524756
Huh, that's really interesting sounding. Can't see a teardown though. Best bet is to look at some US patents, to see what waveforms and such they use. For the hardware, I'd just use an MCU with hardware I2S (e.g. ESP32) and an audio-grade DAC with waveforms stored in an SD card. Plus a transistor for each LED channel you need to drive (or use WS2812s).

It may be good if you get one of those brain-wave measuring kits (like what Micheal Reeves used to control a car), that way you can get live feedback, or just use to tune an open-loop setup.

>>2524891
>I always replace old shitty trimpots like that, can't trust them. Preferably swap them for more dust-resistant designs.
Well now that you mention it, the alloy or whatever the fuck that is is kinda rusty. That's probably messing things up and I definitively don't think scrubbing them a bit with fine steel wool is gonna help at all.
Think I've got some jap-made trimpots (enclosed in hard plastic) from a really busted cassette player, will report back with my findings eventually. Thanks.

This pcb is fucked right?

>>2524943

>>2524943
Yes and no. You should be able to just solder wires onto pads part of the same node, or scratch the solder mask to make new pads. May not be an option if those mounting holes are also mechanical though, looks a bit like a computer mouse's switch footprint, but even in that instance you can epoxy the new switch to the board and run short jumper wires from its pins.

Gotta be real careful with FR2, the paper/phenolic substrate delaminates much more easily than glass/epoxy. Moderate temperature, a bit of extra flux on the board, a bit of liquid solder on the tip, and only keeping the iron on the board for 3s or so before letting it cool down.

Hey guys I'm the lead acid battery BMS anon
I came across this, an application note to measure current in both directions on the low side. Its exactly what I devised on my own
https://www.ti.com/lit/an/sboa223b/sboa223b.pdf
Should I be scared that the opamp connected to the shunt may see some -ve voltages on its input pins (its powered by 5V and 0V). Could that hurt the opamp?
I can implement a simple -ve voltage charge pump using a few diodes using the MCU but I don't want to do that unless I'm absolutely sure

>>2524951
It is my first time soldering. The circuit board (PCB) is from a cheap chinese mouse that I bought from three years ago.

>>2524951
Do you think the circuit board was poorly made, or was it entirely my fault that it is damaged? I'm worried that I will damage my Logitech mouse if I try to change the switch one day."

>>2524943
>This pcb is fucked right?

it's such a silly question.
some people fail to understand that a PCB is just traces.
you could rip off every pad and evry trace on a board, and then duplicate them all using wire (and bondo).

>>2524943
if you did that then yeah, it's time to stop making it worse

>>2524962
>>2524951
> just solder wires onto pads part of the same node
>using wire
I am a beginner, but I will give this a try

>>2524959
Yeah then the mechanical mounting is something to consider. Still, I think any reasonably strong and slightly elastic adhesive should do the trick.

>>2524961
Just cheaply made, it's phenolic not glass-fibre. It's easier to damage, but with proper technique damage shouldn't be an issue. That said, desoldering solid components with three legs isn't trivial, usually the best option is to cut the legs but that isn't always an option if the part is mounted flush to the board. Destroying the part so it can be desoldered in 3 parts is what I'd want to do to prevent delamination, but honestly that's a pain in the ass and the mechanical strain from doing so could also be a problem. Those hollow desoldering needles look useful.

If it's a mouse cheap enough to use FR2 I'd consider just buying a new one.

I found this little number at a flea market for $10. It's an AM/FM radio + Portable Television. You can already bet I'm going to install a custom 18650 battery pack in it. I want to display arbitrary stuff on the screen, but, there's no way to get a video signal in there (without modification) other than broadcasting.

I guess I could modify it to directly inject a video signal into the signal path, but that sounds a bit brutal and almost profane.
How would you guys go about it? It sounds a bit difficult to not only design hardware to broadcast a VHF/UHF television signal, but to also have driving software that encodes a signal as an analog TV signal. But, I don't really see any other way of doing it.

>>2524978
forgot picture de sorry gozaimasu

>>2524976
I already have a mouse, and I am doing this solely for fun at the moment. I might want to watch some videos on YouTube again so that I don't accidentally damage my mouse in the future. Thank you for your help. Cheers!

>>2524978
>It sounds a bit difficult to not only design hardware to broadcast a VHF/UHF television signal,

no, this part is easy as can be.
get one of these, or pull the equivalent box from any C-64 era vintage computer.
a thrift store VCR will also work, if you have lots of room.
note: the quality will be somewhat poor, so it's best to actually inject the signal inside the TV (if you're not afraid of 15K volts)

>>2524989
just to be clear, the only I/O available on the TV is that headphone jack. There is no coax or RCA. Only the RF jack and an antenna. So, one of those boxes will do the job? If I remember right, I believe there was one right next to the TV and I evidently just wasn't aware of what I was looking at.

>>2524978
Chances are you can feed an RF input to the antenna port like you had to with old TVs, for which you just feed a normal composite video signal into an RF modulator (>>2524989). Kinda a hack, but it's the most authentic. If you're modding it anyway, it should be possible to add a switch after the RF demodulation for a normal composite video input. Converting VGA to composite is somewhat easy, doubly so if that's a greyscale TV, but if it is colour there may well be a section past the composite-to-colour decoder circuit to inject RGB signals that could come from VGA with none/some modification.
>digital
On your own for that. I'd just bodge a decoder chip in there, preferably one set to output at the right resolution. Maybe just the internals of a common converter box if that happens to work properly.

By switch I mean a diode switch or some other impedance-respecting RF gate. That kind of signal injection may be kinda hacky as a modification to the existing PCB, so if you're in the game of making PCBs I'd definitely consider reverse engineering the video decoding PCB to make a new one with the switch built-in. If you live somewhere like I do with no analogue TV anyhow, you may consider desoldering some components in series with the composite video signal and replacing them with a custom board on pin-headers, or just wires going directly to a composite video socket.

what trannies would be suitable for the circuit on the right? should be able to supply roughly 2a of current.

>>2524989
>So, one of those boxes will do the job?

yeah, it takes composite video and audio in, and outputs RF on ch3 or ch4 which can be inserted into the TV's RJ-45 jack, or to the antenna terminals using a 75-ohm to 300-ohm balun.

>>2525003

oh crap, it's called an F-connector, not an RJ-45.
i was thinking of the coax cable, which is called RG-59.

>>2524998
What matters more is the power dissipation. What voltage will be across each transistor? P = I*V, so if it's 12V then it's 6V on each transistor, and so 12W dissipated in a transistor worst case. You can probably derate that depending on situation. 2A is a bit much for a BD139/BD140, so I guess I'd use something like a TIP31/32. The minimum hFE of them is 10, so that's 200mA out of the op-amp, which is kinda high. May want to consider using a sziklai or darlington setup (choose whatever lets you common the tabs together on the heat-sink), or finding power BJTs with higher current gain.
Then you'd need the heat-sinking to handle that power. For ∆T=50C, you'd want 50/12 = 4.2C/W, and the transistor clamp thermal resistance is 1.7C/W, which means a somewhat large heat-sink of 2.5C/W. The TO-220 package has a resistance to ambient of 57C/W, for reference. Can maybe run it a bit hotter than that, and derate the power dissipation based off required current output waveform. Each transistor will only heat half of that if the average current is 0, but if both are strapped to the heat-sink it's the same required thermal resistance. Divide that 1.7C/W by 2 if it's even.
Naturally, use your own PSU voltage.

Also watch out for crossover distortion when the current goes from positive to negative, I'd want to use a high-slew-rate op-amp if not turn that into a class-AB topology instead of a class-B. Or just use a power amplifier IC. A big output cap or two would likely help a lot. Actually big enough output caps would buffer the current so it wouldn't have to come from the transistors. Which is why single-ended speakers on single-supply rails usually just sink the other side of the speaker to the negative rail with a DC-blocking cap. Split-rail supplies or differential-output amps are popular to avoid such big caps, and they also work for net-DC signals.

>>2525025
Oh well, thanks for the effort.
I got BD245C and BD246, don't know if the forward current is too high though, I'm afraid.

>>2525025
Also is pic related a viable solution?

Hi erewaan it's me yi lo ma has anyone ad experience with an atx psu having 7v where it should be 12v+ I had this issue after this powersupply would not turn on so I changed a 4 diodes picrel then it supplied 7v so I tried messing around with diode solder and somehow it came back up to 11.94v ish

>>2525086
Also can anyone help me know it that pfc area on the pcb with that shunt across is it important

>>2525047
Yes, if the current draw is AC.
Also you'd better not be using SMPSs for audio.

I've been jumping from one battery to another, but I think I've settled it now.
I think I will go with this battery for my ebike.
https://batteryhookup.com/products/4-pack-a123-3-2v-31ah-lifepo4-cells
The only question I have is, do I need to pad the batteries?
I am probably gonna buy 12 and stack them ontop of each other into a 10x12x30cm box, and I will throw them into a triangle frame bag (it's 8cm wide, but I think there is enough give for it to fit in).
Only worry is that in like 1 month (depending how fast aliexpress to canada is), which is when I will actually buy the battery, the stock might disappear.

>>2525275
if I get a 30a BMS, will the BMS handle the batteries 250a discharge?

>>2525333
I mean, apparently the BMS says it can handle 60a burst, but will it get fried?
https://batteryhookup.com/products/li-ion-lifepo4-3s-20s-30a-bms-with-balance

So, I have a project that requires more (digital) buttons than even the arduino mega has inputs.
What limitations do I have when it comes to multiplexing? I really need to be able to press any combination of the buttons (up to and including all at once).

>>2524904
Alright, following up: Replaced the old rusty trimpots with these equivalents which are good enough (Just noticed it was x2 10k's and x1 100k; didn't have 100k trimpots so I used a 150k one.) Tested with my 120v main under heavy load (induction stovetop, electric oven, blender on max and insinkerator) and it seems to be properly working now; there's only a ~2v drop and OR no longer kicks in at the slightest main voltage variation. Still on testing phase but I'm sure it's gonna be alright, not so bad for something I found in the trash.

>>2525401
PISO shift registers chained together your latency will be negligiable. but why do you have that many buttons?

>>2525444
>why do you have that many buttons?
Because it's a musical instrument and it's ridiculous.
>PISO shift registers chained together your latency will be negligiable.
That's where you cycle between all the different options to see which ones are active, right?
As long as it's faster than we can notice then that's fine.

>>2525451
Parallel in, shift out.

>>2525451
piss in, shit out

I'm trying to build my own controller that's a light gun with a small steering wheel attached on the side of the gun. I think I know where to find the parts for the wheel so I'm not worried about that but does anyone know where I can get a functional lightgun that works on a standard computer screen?

I'm not worried about the technical aspects. I'm pretty quick when it comes to learning new things and am sure I can figure it out if I work hard enough. Just a few pointers on where to get the proper parts would be nice. Attached is concept art for a video game console I am looking to make. Building the whole console is too advanced so I'm settling for just the controller right now. Thanks.

>>2525401
SIPO shift register (e.g. 74HC165) like the other anon mentioned is one way, but I'd recommend going for a conventional key matrix. With a matrix you can split the pins in half and multiply them as rows and columns. For example, if you have 32 pins to spare, you'd split that into 16 rows and 16 columns, allowing for 16x16 buttons = 256 buttons. This is what computer keyboards do. Use a diode on each one if you want N-key rollover.

>>2525438
I'd also check out that diode or whatever next to the trimpots, the one with the darker area of PCB around it. That darker area suggests nontrivial heat production, that component may be the cause of the original fault, or perhaps it's just running close to its thermal limit by design. Either way I'd want to replace it with a better part. There's also the possibility that it overheated as a symptom of damage to other parts (maybe the trimpots, maybe something else), in which case replacing it may not fix all your problems, but it's probably worth doing anyhow.

>not so bad for something I found in the trash
Yeah not bad at all. I've got a UPS someone chucked because the battery was dead, but I'm split on harvesting the power transistors, heat-sinks, and magnetics from it, and using it as an inverter or UPS myself. Having a UPS isn't a terrible idea, especially for my 3D printer, but that 1kVA transformer is looking mighty attractive.

Also my oscilloscope's brightness adjustment pot is giving erratic readings like you described your trimpots doing. Hope I can find a replacement for them too.

>>2525481
>I'd also check out that diode or whatever next to the trimpots, the one with the darker area of PCB around it.
It's a 12v 1w zener, honestly I wanted to replace it but I've got no spares at hand. Tested it unmounted and it seemed fairly normal, though I was worried about about the darkened PCB since it looked the same in other places (can't really see them from this angle) where I replaced other busted components (Swollen cap, burnt resistor, green LED was also burnt out.)
I'll check my scrap depot tomorrow, maybe there's something with that specific zener, though finding something will be a huge pain in the ass.

> I've got a UPS someone chucked because the battery was dead, but I'm split on harvesting the power transistors, heat-sinks, and magnetics from it, and using it as an inverter or UPS myself.
It's amazing the kinda shit people throw away, I found and oldass circa 08 CPU tower just lying on the curb, all it had was a blown fuse in the PSU. Wiped the drive clean (And removed the nicotine gunk+dog hair choking the fans) and replaced a few swollen caps and it's now exclusively used for document edition and internet browsing. I've found CRT monitors too, but hobos tend to pick those clean; they even take the tubes too, sometimes smashing them against the curb, fucking crazy.
People also tend to just throw away old PC PSUs, I've scrapped quite a few of those. Been thinking about turning one into a bench PSU for testing crap.

>Hope I can find a replacement for them too.
GL man, I've yet to learn how to use one of those.

>>2525516
>It's a 12v 1w zener
Ah, then chances are it's running normally. Zeners are often left running rather hot, and it shouldn't be bad for the zener, but judging by the surrounding PCB's discolouration it's quite possible that it could reduce the lifespan of nearby components (like the trimpot).

It's possible that you can replace the zener with a TL431 circuit (with additional PNP shunt transistor), which is arguably easier to find. TL431s are somewhat common in isolated SMPS designs, so chances are you may have some on scrap boards already.

>>2524016
spend $3 more to get a meter that beeps, christ anon

>>2525451
no. its where you chain (N/16) x 16 to 1 parallel in serial out shift registers by using a clock pin to read out the value of each button. N being the total buttons you have, either active high or low. you can do this extremely fast.

>>2525563
>16 to 1
Those exist? I've only seen 8-to-1.

Also just reading off a PISO is basically the same as cycling through each value. Can kinda use SPI hardware to take the load off your MCU, even better if you can use USART hardware's extra pins to output a clock AND toggle the CTS or whatever pin in order to latch the register. DMA or maybe an event system could streamline that data path further.
Personally i'd recommend using entire I/O ports for parallel data inputs (e.g. keyb matrix), so you can do direct port reads for quick data inputs. You could combine these by using SIPO shift registers to drive the rows, and PISO shift registers to read the columns. This kind of circuit would be useful for something with hundreds of pins needed to be mapped, like an LED cube.

>>2525199
What do you mean by AC current?
Look I want to power a chip amp requiring symmetrical power source using a DC jack (well, I want to have flexibility, so I'll put it as an extra next to a big center tapped trafo).

I just saw the picrel >>2525047 solution in a video and it appears almost too good to be true. Just a simple zener and 2 resistors to balance it out?

Otherwise I have BD245C and BD446, so I could try the circuit proposed earlier. >>2524998

Maybe I should get my OP amp basics down before I try something like this...

>>2525598
*AC current draw
*BD246

>>2525598
>What do you mean by AC current?
capacitors alone can buffer an ac current output, since it's kinda a voltage divider using capacitive impedance Z_C = 1/(2πf*C). If f=0, then the impedance is high, and so it can't provide any current. If f is high enough that Z_C is much lower than the speaker impedance, then it acts close to an ideal ground rail.
>I want to power a chip amp requiring symmetrical power source using a DC jack
You can also just use two (at least one of them ungrounded) PSUs in series to give yourself a middle rail. That said, I'd avoid any switching power supply for noise reasons, unless you're adding your own nice discrete transistor pass regulators and/or LC filters. If your barrel-jack PSUs are not switching, then they've got an iron transformer that you can disassemble and wire like your picrel anyhow, they might even have multiple output taps anyhow.
>zener
There's no zener there. Just a Schottky reverse polarity diode and an LED. The two resistors are a requirement, and the caps need to be big enough to have minimal ripple voltage at the desired current. Too small caps and you'll lose bass.

>>2525591
i can see what youre saying. but combining SIPO and PISO for row column access is the wasting the same amount of clock cycles anyway, or possibly more. also, OP said he needs to be able to read everything simultaneously. the problem with modt keyb matrices is that you cant do this. you can only read the buttons in that row. with chaining big PISOs together, you may be reading all of them sequentially, but ALL buttons can be active at once and still be read correctly.

you could use spi hardware or just designate the gpio as 'very fast' in its config and use timer interrupts. additionally you could attach the clock pin to multiple PISO chains attached to say 8pins and on every clock cycle do a full port register read which should give you all the pin states at once. allowing you to read in 8 buttons at a time on every clock cycle. theres many other ways of doing a keyb matrix, but thats my take

>>2525591
I've got 69 digital buttons for the notes alone, let alone knobs for volume and stuff, so each one getting its own pin on the arduino is out of the question.

I was under the impression that a matrix causes problems like if you try to light up two corners of a grid it'll light up a third corner. No bueno for a musical instrument of this sort.

The software will be a lot more complicated, but what I have in mind is to collect the on/off value of each switch as fast as I can, then output them all as a note at the same time for the duration of time it takes to collect/process/output the next one. That gives a minimum note duration which I have to keep as low as humanly possible, so I really need the optimal wiring setup.

>>2525636
kek. 69 is a lot yes, plus all the other stuff. but you could easily fit it all on a 144pin mcu. hell if you still cant fit it, get a 200 pin mcu.

in any case using chained piso shift regs (lets say 8-1) means you only need 9. these things can easily be clocked in at 100khz (or more). which would give you a theoretical read time of less than 1ms, even with overheads you just gotta write the software correctly.

you mentioned leds. do they need to be rgb? if not and you just want a static solid color per button, you can couple the led to the button press and still habe the buttons be fed into parallel inputs.

>>2525668
Yeah, <1ms should be more than fine.
I'm using "light up" colloquially, here, though a lot of the button options I'm looking at are illuminated.
I just mean that if I want to play notes 1, 2 and 3, I can't have it play note 4 or 5 along with them, you know? Hypothetically you could make an instrument where it does that as a feature, kindof like a drone note on a bagpipe, but that's not the purpose of this one.

>>2525671
sounds like a good project anon. good mix between digital and audio/analog. good luck. post progress in future threads.

>>2525676
Probably biting off way too much to chew, but that's alright.

>>2525625
>but combining SIPO and PISO for row column access is the wasting the same amount of clock cycles anyway
Yeah I'd only do it if the number if IO pins weren't enough for standard row/column addressing.
>you could use spi hardware
I think you might be better off using synchronous USART and using one of the extra CTS/RTS/DTR or whatever signals for latching. Assuming your MCU can do that. Or just use an independant counter running off the same clock (or its falling edge). Using interrupts and/or events is doable too, though it would take you another clock cycle or two unless you've got some strange internal asynchronous hardware.
>OP said he needs to be able to read everything simultaneously
I don't think that's an issue. With a 16MHz clock you could read everything within 0.1ms easily regardless of what method you use, having a slight sequential read is less of an issue than getting a delay between keypress and sound output. Gaming keyboards use matrices after all, and gamers can be real anal about latency.
With a matrix and well-thought-out direct port manipulation, I think you could take only 2-3 clock cycles to read 8 keys. 2 cycles for an 8x8 matrix:
>PORTB.OUT = 0b00010000;
>data3 = PORTA.IN;
>PORTB.OUT = 0b00001000;
>data4 = PORTA.IN;
>etc.
I think 8 pullups/pulldowns are required somewhere, probably possible with internal ones. Add another pair of cycles near the end for writing the 9th pin, and you've read 72 buttons in 20 clock cycles, or 1.25µs. Arguably too fast to see a decent signal. This is the speed optimised version, using a loop would take less flash but more clock cycles. Debouncing is left as a task for the reader.

>>2525636
>I was under the impression that a matrix causes problems like if you try to light up two corners of a grid it'll light up a third corner
That's what the diodes are for, look up how mechanical keyboard creators use them for N-key rollover.

I hope you have some plan to implement key velocity. Maybe through key bounce?

I have kinda stupid question. How can I start to learn everything about radio? Frequencies, power, bandwidth, antennas, dsp etc. Not for a student who needs theory, but just as a practical thing. At the end I want to be able to make a radio communication station to communicate with my rc plane using my homemade tx/rx modules and antennas. Maybe you got some really cool youtube channels for me?
I'll get a license if needed, dont worry.

>>2525864
>I hope you have some plan to implement key velocity. Maybe through key bounce?
I think for the prototype I'll stick on a pot that controls that.

>>2525947
There's a ham radio general, ask there. For details on radio-related circuits, some guys like W2AEW have neat content.

>>2525953
Real keyboards just have two switches for each key, and measure the time between each getting pressed, which is pretty doable, hardware permitting. Just one pot means it's the same for all keys and takes time to modulate, so you won't be able to play dynamically at all. Being able to hit a key loudly or softly is what makes a synth interesting.

Theoretically you can measure something akin to key velocity by characterising the bouncing the switch does when it's actuated, though it's a lot more software intensive than the 2-switch method. And I've never seen anyone do it before. If you've got a dedicated MCU or core to handle it I'd say it's still pretty doable, but if you're also producing an audio output directly with that 8-bit MCU you may well get bogged-down in the synthesis step.

Getting into soldering but I live in Europe and found out that I can't get lead solder.
So far the most popular one I can find and actually buy is SAC305, I can't get SN100C which everyone says is way better. Sn96.5/Ag3.5 solder doesn't seem to exist here either.
Ideally something eutectic I guess since I'm starting out but every forum post I'm reading on what lead-free solder to get just says solder-free bad and never any answers

>>2525978
When I bought components from LCSC, I also picked up some solder. It was lead-free stuff from Mechanic (HBD366), and soldered a lot nicer than the local lead-free stuff, but I'm not very familiar with lead-free solder in general so I couldn't tell you if it's better or worse than the ones you're talking about.
Chances are the customs fuckers won't be too invasive and take it away from you if you import leaded solder alongside a bunch of other stuff, but you never know. The EU probably still use leaded solder for military stuff anyhow.

>>2525978
lead free works fine with my 888d, but when I tried it with a cheap iron and it was unusable

>>2525983
Haven't even seen any posts mention Sn/Bi so I'm not so sure about that either but I guess it's another thing to consider. It's not worth buying stuff from US regardless because of import charges, rather just suck it up and deal with lead free solder
>>2525996
I'm not too worried about the iron, I have an ADS200 so that should be fine.
I'm more than likely overthinking this, I might just get the SAC305 solder and call it a day, it's not eutectic which would be nice to learn with but seems like I don't have much options regardless

>>2525969
I would love for the buttons to be analog, honestly, but at that scale I'm pretty sure I'll have to design my own boards. I'm still very much a noob and looking up a lot of what you guys are saying. I seriously have so many tabs open all I can see is the website icons.

>>2525536
Damn this is something I definitively wanna try out but don't know where to start exactly. Probably by mapping the circuit out.

>>2526000
Wait I didn't even know it had bismuth, no wonder it felt so different to this 99.3/0.7 Sn/Cu stuff. It's probably weaker than either kind of solder.
>from US
LCSC is from China. Like $40 shipping fee minimum. Only worth it if you're ordering a bunch of shit from them.

>>2526008
>I'm pretty sure I'll have to design my own boards
Oh, what are you doing at the moment? Just bolting some perfboard with tactile switches under the keys? Or mounting a bunch of switches some other way?
You can always attempt the debouncing method later using the same hardware. Probably an unreliable pain in the ass trying to get consistent values out of it, but you never know.

>>2526157
Pic related, though in your case the resistor on the far left would already exist in your circuit. Just the main square of three resistors, the TL431, and the PNP transistor will replace the zener. Would also give a much more stable output value.

Just repaired this BMS. Did I do well?

>>2526178
The prototype is probably going to be a giant spaghetti monster inside a box, lol.
This is probably a huge noob mistake and I'll end up kicking myself for not just making the boards to begin with.

what is the best way to power a micro from a battery?
When the micro works with 5 volts it's a no brainer i can just hook up to a lipo directly and it lasts forever.
But problem is when the micro is 3v, or there is some other stuff on the board that is 3v, so what then?
I can use LDO, but those use consume current to work, and even if they use say 0.5ma to self power that is a massive waste of energy if rest of the circuit uses say 3ua in deep sleep (not to mention ldos need minimum current to work properly that is like 2ma usually)
Buck is also gay.. often when lipo is down to 3V the buck will output like 1.5V (or nothing), not to mention you have to put a ton of components to make a buck work or a whole separate pcb, and it also wastes too much power.
When the circuit is on 0.01% of the time it consumes 500ma so power wasted by ldo, or buck etc doesn't matter, the 5ma extra it uses is meaningless, but when the circuit sleeps 99.99% of the time and uses only 10ua on average than 2ma becomes a MASSIVE waste.
So wat do? smaller voltage battery? that presents its own issues

Anyone know of a component that is normally open, and then becomes permanently closed after a shock/impact?
I assumed something would exist for package security and such but I can’t find shit :/

>>2526432
>Anyone know of a component that is normally open
yer mum's bung
https://www.murata.com/en-us/products/sensor/shock
these little piezo sensors generate a small voltage on impact, but don't do any switching.

>>2526280
Ok, assuming the pinout is correct, but I'd want to ensure that the conductive underside of the transistor can't short on anything beneath it. Also vibration resistance would be poor. A dab of silicone between it and the board beneath would solve both issues.

>>2526430
Pick an LDO (with a really low quiescent current) or otherwise make one from a low-current RRIO op-amp and low-current voltage reference and a PNP.

>>2526432
There are these cheap vibration sensors that have a little metal filament inside a can, so when you shock it the filament makes contact with the can. Not sure what they're called, just know they're on those cheap shitty 8W USB soldering irons.

For latching, connect it to a 74LVC2G02 or some other SR latch setup.

>>2526462
>but don't do any switching.

so you could combine them with a magnesium flash bulb, which can act as a 3V 1-bit write-once-read-many ''flash'' memory.

>>2526476
>anon posts "repair" so bad it belongs in the cursed pic folder
>"Ok"

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

first time using optocouplers, pic related. i'm getting a constant 5V at the output despite switching the input between 5V and GND (by hand, very slowly.) what am i doing wrong?

>>2526625
Change the output pullup resistor to 5k.

>>2526645
i tried that as well. these are chinkshit 6n137s but i tried like 3 of them. i can't believe they're ALL broken.

>>2526653
Did you try latching the enable pin to Vcc, just for shit and giggles?

shitzen giggles

>>2526655
i just tried putting enable to VCC and enable to GND. literally no difference. what the fuck did these chinks even sell me.

>>2526481
>no corrosion
>shiny solder joints
It's not gore at least. It's a lot better than my janky repairs, picrel.

>>2526484
Gotta multiply instantaneous current by instantaneous voltage, and make some sort of weighted running average. Probably using DSP, since analogue multiplication is hard. Also consider a transformer or isolated ADC+divider for mains voltage, to prevent the majority of your digital circuitry from being mains referenced. A differential amplifier measuring live-neutral and referencing it to ground also works. Naturally you'll need anti-aliasing filters to obey the nyquist sampling theorem on both ADC inputs. Figuring out the power factor and reactive and apparent powers would just be a software problem.

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

>>2526625
Chances are it's some other chink optocoupler with different badges on it. May want to look on LCSC for other optos with a similar package.

>>2526866
>>2526866
>>2526866
THREAD NUVEAU

>>2526860
I posted a reply in the new thread, but just in case you didn't see it, is an anti aliasing filter necessary? If it is, can I use a simple passive RC/CR filter to do so? Wouldn't the filter affect the current wave form and therefore the power measured?

>>2519847
Sounds like a very cool and comfy workshop