/diy/ - Do It Yourself

>>2496398
so single opamp is a no go then? i guess i could shell out 10 bucks on this since i only need one and it's the one intr amp i would and they have in stock
i want to keep the circuit simple

>>2496408
As I said, since input impedance isn't particularly an issue you should be able to get away with a differential amplifier circuit. The one with one op-amp and 4 resistors.

NE5532:
>100dB CMRR
>100dB PSRR
>5nV/Hz input noise
INA163:
>~110dB CMRR
>~110dB PSRR
>1nV/Hz input noise

So the 163 is somewhat better, but not hugely.

Honestly I'd breadboard or perfboard a differential amp with a cheap LM358 or TL072/LF353, just to see what the performance is like. Oh but chink TL072s are also often fakes.

>>2496412
alright i'll just buy 3 5532 and try using just one in diff amp setting and if it doesn't work well i will just use 3 5532 and build a whole ass instrument amp, since they cost only dollar a pop this is not an issue, they will probably be fine since i read they are extremely popular for audio equipment
if i wanted to be able to change gain i could mabe use pot for the Rf resistor?

>>2496413
You could absolutely use a pot for that.
Going to just put it on a seperate board and wire it to the old one I guess?

Also I’m going to try to make a microphone capsule by tensioning some Mylar space blanket. Might be fun.

>>2496418
I'll just make a new pcb since the 5532 is smd so it would be be a pain in the ass deadbugging it plus i want it nice and compant without a bunch of wires shaking around. i'll just use that small blue trimpot;

>>2496402
is lt spice failing to find a solution in ac analysis always an indicator that a circuit is doomed to fail? i'm dicking around trying to simulate an electronic load (op amp feedback on a MOSFET) and am having a bitch of a time getting solutions to actually calculate.

Not quite the right thread, but can anyone recommend an outlet tester? I'm looking for something that can tell me whether the voltage is good, grounding, and diagnose basic problems

I'm looking to buy a cheap multimeter to get into the hobby and while going through the usual recommendations I happened to notice the AN870 is currently sold at less than the AN8009 is.
I don't really see a reason why one should go for the 8009, am I missing something? Do you think the 870 is a good choice for someone just starting out? I see >>2495942 in the last thread got one but I can find no other explicitly first hand endorsements in the archives. Thank you in advance, anons.

>>2496535
Any $20 multi-meter from your local hardware store will test voltage and continuity.
For outlets use a voltage setting higher than what you expect and never use ohms setting on a powered circuit

Probably asked a million times but any recommendation for DIY kit to learn to solder?

I have a dc-dc power supply and an oscilloscope, but I need more ideas.

>>2496420
Then I’d design the board to have a discrete instrumentation amp circuit, since the differential amp part is the same. Just use solder jumpers to bypass the footprint for the first two op-amps.

>>2496525
Nah spice can be funny like that. Try adding insignificant series and parallel resistances, maybe parallel capacitances or series inductances too. Ensure there’s good ground referencing too.
Why are you doing an AC sweep anyhow?

>>2496611
Is there a model with a dual display? Such that you can see AC voltage and frequency at the same time, or current voltage and peak voltage. 20000 count is pretty ridiculous, but whether the noise floor is such that you can use it is questionable. Plus there’s no Kelvin measurement so resistance measurements won’t be that accurate anyhow.
As for the 870, look for a YouTube review first just to see if there are any red flags like a delay before the continuity buzzer goes, or an unintuitive UI. Be nice if the diode check lights LEDs.

is the power output of these things legit or are they scamming?
https://platinumtherapylights.com/products/biomax-rlt?variant=15601444487234
when checking for diodes, the highest output I could find for 850nm leds was 1300mW/sr
I made a model of their product and the maximum amount of leds that have beam overlap is ~175(90 degree beam angle at 6”)
even assuming they carry 100% output for entire beam area, it’s still dispersed on ~730cm2@6” out.
If each led can only deposit 1/730th of 1300mW on a given cm2, so it ends up being something like 0.3W/cm2
Am I doing something retarded in my calcs or is this shit a scam by a factor of like 500?

>>2496663
>Is there a model with a dual display?
Doesn't seem like there is one around this price range.
>any red flags like a delay before the continuity buzzer goes, or an unintuitive UI.
One youtuber complained about AC voltage and DC voltage being on different dial positions, while mV AC and DC are on the same spot. I don't really understand the implications though, I'd be grateful if you could explain this to me.
Every review I read or warched warned against using it for high voltage stuff and complained about the lack of protection, but the same is true for reviews of every other equally affordable multimeters I checked.
>Be nice if the diode check lights LEDs.
Seems like it does.

>>2496749
>One youtuber complained about AC voltage and DC voltage being on different dial positions
Just a convenience thing. Another advantage of a dual display is being able to see ACV and DCV at the same time, which can be pretty handy if you're probing audio circuits without an oscilloscope. Assuming the frequency range covers the full audio range. For that purpose I'd also want to be able to swap from trueRMS to peak or peak-to-peak.

>>2496611
I'm also looking at getting a multimeter. I'm being told I should get this Fluke 101 (hi dad). Someone tell me what's wrong with it. I mostly need a multimeter with a beeping continuity test.

>>2496611
>8009
for me (same anon who bought it) it's the opposite, I paid 2x more for a AN870 than I would've for the AN8009.
the AN8009 is more accurate than the AN870 if you stare at the spec sheet.
But the reasons why i bought it is:
>19999 vs 9999 (very minor, but neat)
>bigger
>uses AA batteries, since I don't have rechargeable AAA ones, and them rechargeable batteries are pretty expensive.
It's very minor reasons but it's what made me wanna buy it.

>>2496748
>LED Power Class: 900W
>150mW/cm^2
>36" x 12" = 91.4cm x 30.5cm = 2790cm^2
>0.15 x 2790 = 418W
That "LED Power Class" rating could be close enough to pass the power to area x intensity test, since it's a class and not an actual value.
>Power Consumption: 380W
237% efficiency sounds like bullshit to me. My guess is either:
A) "LED Power Class" is referring to equivalent incandescent light output, so the actual output is off by a factor of 10-50. But it's kinda odd for a situation where we're talking radiant power rather than luminous power anyhow. Maybe they have their own mock-photometric units for gauging production of red light.
B) The light is pulsed, you can arguably see a bit of rolling shutter in your pic, idk.
C) Actual scam, that red light is kinda targeting the foil-hat market anyhow, and you can sell fucking anything to those guys.

As for mW/sr, you can't really just calculate solid angle from emission angle and multiply, because the intensity-angle curve isn't just a step function. You need to integrate, or better yet just find a total radiant power rating for the LED chip. At 15cm out all those LEDs are gonna approximate a flat light panel anyhow, with minimal light loss to the sides.

>>2496919
Not true RMS? Jacks on the bottom is a bit wanky, but if it comes with non-right-angle probes it's ok I guess. No current measurement is a red flag for a lot of people, but personally I never use that anyhow.

>>2496937
*but note that when i say more accurate, technically it means the AN8009 is more accurate at the lowest ranges, and the AN870 is more accurate at the highest ranges (because of the extra 1 in the display, so like 19 vs 9).
I am pretty sure both meters have the exact same chip, and the chip is also used in like 100 other similarly priced meters (for example the kaiweets km601), just changing the registers of the chip to enable/disable different features (just something I have heard, not an expert).
>>2496919
I think it's fine, but the problem with buying from amazon is that it will arrive in a few days, so technically you could just do whatever you are doing without a multimeter (no meter challenge), and just buy it when you need it. But then if you buy from china, you gotta wait months, so it's more about "buying a meter with all the features that you could ever possibly need", even if it's overkill, but I don't have any imminent projects I am working on, other than far out projects that will take a ton of research before I start (ebike).
like you could buy a $20 or less multi meter if all you want is continuity, and you might also want a thermometer sensor (like maybe you want to check if your soldering iron temps are correct).
Also just watch reviews, and do some research into specs.
Also take with a grain of salt I am not even a hobbist, I haven't even built anything.

>>2496412
>>2496663
is the the correct way of puting in the differentional amplifier circuit?
basically, there is a high pass filter on the same input rail, so i want to make sure simply hooking up the opamp in differentional amp configuration after the filter like this is correct way of doing it

do piezos eventually die? I have this drum synth pad toy from the 80s that the pads are barely responsive. can the piezo surface be cleaned or should I just go buy some new ones?

>>2496611
I have the DM6000AR, which is fine except for two annoying things
- there is button for backlight but will only stay on 5 seconds and then you have to press it again over and over
-the machine auto shuts off even when youre using it. i think it has a 15 min built in timer, and it will beep loudly to warn you its going to shut off, and you have to cycle through modes to reset the shut down timer

roll

>>2497074
>you have to cycle through modes to reset the shut down timer
This is the most annoying "feature" of most MM. It really needs a separate power on/off switch but few MM have it. I don't know if that's something that could be easily DIY

What's up with the stupid "count" spec on a MM instead of resolution/accuracy?
If one thinks in "counts" give me an example. what is the "count" of a 12-bit ADC powered from 3.3v? The resolution is 3.3 / 4096 = 0.0008v. But what about the count, 4096 counts? 33000 counts?

So this is 20k counts.
So in the range from 0 to 2v it can display up to 1.9999v. But who decides on the range? What if the range was 0 to 1v? then it would be 10k counts? Can't we just say that the resolution is 0.0001v and isn't that all you need to know? You'd guess that it is a 16-bit ADC (which is pretty standard I think).

>>2496270
Is this a good example? Looks fairly simple and it is a commercial made device.

So counts is the range divided by the resolution for that range. And the resolution for that range depends on the overall DAC resolution. How? How would you implement a 200mv range? If you have a fixed voltage reference of say 2v? Use a resistor divider after the Vref? And then how would you implement a 20v range? You'd then have to divide the input obviously, something like a 1:10 divider and Vref=2v. What happens with the resolution/counts in this case?

>>2497073
just get a new one
>>2496634
I started with one of those chinese xr2206 signal generator kits and one of those chinese DSO-138 oscilloscopes. Before you gunk those up tho, definitely practice bonding bits of wire and soldering random junk to prototype pcbs first, cuz the technique's a bit tricky

Amazon and ebay sell $15 kits that come with an iron and a bit of solder to get you started

I'm gonna ask in here and /mcg/
I'm reading through the book "Musical Applications of Microprocessors" by Chamberlin. It's effectively a textbook for mixed-signal audio synthesizer design, and also full-digital (almost-)register-level programming of microcontrollers as synthesizers. It's got a shitload of applicable information to what I want to do, but it's also from 1985. Needless to say, it's a bit outdated. I don't know if I can program my stm32 blackpill in BASIC. I like this one because it tells me how to implement synths digitally without relying on libraries. More recent books of a similar caliber are all a bit of a sham. They say "learn how to make digital audio synths!" and then you open the book and it says "download this audio synth library!"
Do any of you guys know of similar textbooks that are more recent?

I'm trying to build a laptop "powerbank". Obviously I can go with 18650 cells, but then I'm reminded of the custom flat-pack lithium cells used in actual smartphone powerbanks that are much more energy dense since it's one or two chunks of lithium instead of smaller units of lithium in a relatively chunky 18650 casing.

Where can I find those flat lithium cells for retail purchase? I can find bulk suppliers, and I can find custom high-current ones for things like RC planes, but never anything like the less-dangerous ones used in powerbanks

>>2497147
Going from basic to c should be trivial, it's just a different syntax, algorithms will still be the same.

>>2497154
oh yeah that's a given. I was just being a smartass. I'm more concerned about whether I should bother trying to re-implement a DSP algorithm from 1985 when there might be/probably are better versions in 2022

>>2497155
I doubt that math changed so much in 40 years. Sure there might be efficiency optimizations, but that's just additional confusion for beginners. People were already starting to dabble with signal processing in 17th century, and digital part already matured in 1980. If you like the book then read it. If you don't, find something else. In the end you will be acquiring your knowledge from multiple sources.

>>2496402
Hello /ohm/ i want to make a current detection led so i can see if a smart lithium battery is still charging itself from the 12V. Current below say 10mA then led should be off, otherwise on. But the reverse is also fine. Current can be as high as 1500mA. Any really simple solutions here?

>>2496996
Without those resistors R8/9, you’ll want the total impedance to ground from those locations to be 3.9k or thereabouts. Since you want high gain, the series resistors need to be smaller than the parallel resistors, so that would put your series resistors in the realm of 39Ω for 40dB gain. I’d check the 5532’s datasheet, but the resulting high output current may result in less than stellar audio quality.

If you do keep R8/9, you would be able to use 1k/100k resistors around the op-amp without issue, which is the most sensible option. I’d at least keep those footprints open.

There’s also the option of changing the capacitors, which may be a good idea anyhow considering they’re probably non-polar electrolytics, and the nontrivial impedance coupling to the JFET circuit.

>>2497074
On my meter you can disable backlight auto-off and power auto-off by holding specific keys when I turn the range switch on.

>>2497093
Usually 2^N, rounded down. The reference voltage can be assumed to be selected such that 1 LSB is the smalllest division available on the screen, makes the internal math way easier. They’ll likely just use programmable gain amplifiers and/or dividers. So with a 4000 count meter in 40V range, the smallest value and hence the LSB will be 0.01V. I have a 6000-count meter though, not sure what bittage that implies, maybe they’re doing over sampling or whatever.

>>2497187
The BJT+resistors circuit diagram from last thread.

>>2497202
Yes, forgot about oversampling.
But I am still confused, 16-bit DACs are nothing special, if it was just about that I'd expect that even the cheapest MM would have 2^16 ~ roughly 60000 count but apparently it is not so simple.

>>2497206
Noise is the main thing. Chances are they already oversample a ton to stamp out jitter, but getting 6 or 7 or more digits means having the low-noise circuitry to actually make those digits meaningful. Which is a lot more expensive than the display or the ADC.

>>2497202
>The BJT+resistors circuit diagram from last thread
Thanks i think that would work. Now i have have to find a BJT that can handle 1.5A b-e current.

>>2497216
So, it needs low jitter clock source? Like what, a special low noise crystal and/or low phase noise oscillator? There are cheap low noise oscillators in a case like picrel. I think all of them are pretty low noise for the job.
Or there is another source of jitter other than the clock jitter?

Has anyone ever looked inside their MM to check the specs of ADC/oscillator/precision opamps, shunts, etc? I've watched some reviews on youtube, and the dude be like oh you know what this is? this. is. a. SHUNT! And then he'd measure the db level of the continuity buzzer. that's the level of those reviews.

>>2497243
Yeah that's the tough part. Also the voltage threshold is probably closer to 0.5V than 0.7V.

>>2497293
I didn't mean clock jitter, rather voltage jitter. Maybe jitter is the wrong term. Mainly noise from digital stuff.
Though I think clock jitter may matter for delta-sigma converters.

As for DMM teardowns, I'm pretty sure Dave from EEVblog has some pretty in-depth ones.

I had an idea for pic related the other day and was wondering, does something like this already exist? What possible applications could it be used for?

>>2497315
Is that a switch? Just use a DPDT.

>>2497316
It's combo-ing a momentary n.o. and momentary n.c. switch in one package; two poles, but only one pole will be active at any time.

I don't need it for anything in particular, I just thought "why has no one thought of this yet?"

>>2497319
>"why has no one thought of this yet?"

Because it's just a less-useful version of a DPDT switch.

>dry film photoresist is underdeveloped on one side, leaving a bunch of residue where i want bare copper
>dry film photoresist is overdeveloped on the other side, causing some exposed areas to flake off
laser ablation and milling are literally the only two methods that aren't a massive fucking hassle

unless you're applied science and make a massive home-shop computer-controlled photolithography setup

>$6-10 each
fucking killing me here. is bulk the only sane way to get these things? is there a cheaper option?

Can any of you guys help me understand this random ass assumption my book does?
That is not one of the saturation conditions as far as I know.

>>2497457

Unless you specifically need the properties of a metallic enclosure, 3D printers have mostly made that kind of project box obsolete.

>>2497457
If you need something shielded, it may be cheaper to order PCBs as walls with slots that you solder together, or maybe fold something together out of flex PCBs. Conductive filament is likely insufficient for shielding, but it is handy for antistatic purposes. If it doesn't need to be shielded, then 3D printing is likely your best bet. If you need something sturdy like for a stomp-box, you're shit-outta-luck.

>>2497466
they arbitrarily decided it would be the operating point in the first section
>to establish a dc voltage of 0.5V

>>2497466
saturation is when V_DS > V_GS - V_th and V_GS > V_th
If V_GS < V_th you will be in the cutoff region, and you won't be pulling the current necessary to get 0.5V at output, instead it'd be at/close to V_DD. So you have to be in either triode or saturation region
You're also looking to set V_D = 0.5V and V_G is tied to ground, so V_DS = V_D + V_GS. This would then mean V_DG + V_GS > V_GS - V_th, or V_D > -V_th . So as long as your threshold voltage is over -0.5V (note the negative sign, and yes it is possible to have negative threshold voltages but I don't think you'll be dealing with those for a while, at my university only EE does in one junior year semiconductor class, CompE does not) you'll be in saturation mode here.

I can't find any datasheet regarding this component, can anyone help me please ?

>>2497614
It's an old IC, maybe a differential amplifier. Where did you find it?

>>2497120
I am still thinking about this. Looking at this screenshot, how do they make sure that the resolution is in exact powers of ten? Because it seems like it is the necessary condition for the display counts to match the ADC counts. Otherwise, the step would be some fractional number and you won't get the continuous counts from 0 to 9999...
In other words, it is necessary that Vref/2^n = 10^-m. For example if n=12 then Vref has to be either 4.096V or 409.6V etc. If it is 4.096V then m=-3 i.e. the resolution is 1mV and that's exactly 4000 continuous counts. But what if Vref=2V. Then the resolution is 2/4096 ~ 488uV. It is still ~4000 ADC counts but the step now is ~488uV. So you will never see 0.5mv or 1mV and 4000 ADC counts doesn't translate to 4000 display counts. How do they get around that to get a continuous count scale?

Also can you make sense of this chart?

>>2497744
never mind, i got it

If both voltage and current specified, does that mean there is no need for constant current regulator and there is a limiting resistor installed already? Otherwise they would just specify the current?

>>2497457
Maybe a 3d printed case with aluminum foil if you need the shielding.

>>2497120
Counts are not the range divided by the resolution, it is the number of different values your device can distinguish. You get the resolution by dividing the range by the number of counts.

The voltage range is selected by scaling the input. The Vref is always the same no matter the range. downscaling is done with a resistance network, and upscaling is done by amplifier stages.

>>2497741
>how do they make sure that the resolution is in exact powers of ten?
There are multiple ways of doing it. The old school way is to scale the input voltage relative to the Vref to get a whole resolution. Or you can use a higher resolution ADC and digitally calculate the actual measured voltage, this is a more modern/expensive solution. For both you need to calibrate the device. In the first case, the resistor networks get adjusted so that the voltages that get fed to the ADC are the right scale compared to the Vref. In the second, the digital values get adjusted so that the displayed voltage is correct. What's nice about the second method is that you can calibrate your instrument without opening it.

>>2497788
>does that mean

no it doesn't.
your datasheet reading skills are sub-par.
the current readings just specify under what conditions the various other readings are specified.
it's like saying water boils at 100-deg-C ''at sea level''.
so you mos' def' need some current-limiting scheme.

>>2497827
But that info seems redundant. Just give me the current. I can apply however much voltage is needed to guarantee the summary drop will be within the limit and the current will not exceed the specs. It seems like it specifies the summary voltage drop across 6 LEDs, not the voltage applied to the driver. Here is a more common example. Most of the time only the current is given. Your LED driver will figure it out.

>it's like saying water boils at 100-deg-C ''at sea level''.
nah If I am not at the sea level I need to know the boiling point at the given altitude. The LED string doesn't care about any other parameters other than current and minimal summary voltage drop. I can apply whatever voltage I want.

>>2497457
cast it yourself

>>2497836
Its saying that when you push 40mA the voltage drop will be between 18 and 20.4, typically 19.2
The reason they tell you the voltage is because there is no such thing as a magic current source that can supply 40mA from zero to infinity volts, you have to know what voltage will be to spec a suitable driver in the first place.

>>2496402
>>2496402
I want to play with a car stereo as indoor soundbox, which are the best to do so
>Sony DSX-A410BT
>Sony DSX-A110U
>JBL Celebrity 100
Or what else?
Also
Recommendations to not blow the unit or speakers (appart of the fuses=)
THX

>>2497788
Forward voltage changes with temperature. At one temperature it may draw 1A at 20V, at another it may draw 50A. It’s also a positive feedback loop, so if running at 1A causes it to heat up a bit, it will start drawing more current, which causes it to heat up even more, and so on until it lets the smoke out. A constant current on the other hand is a negative feedback loop.

>>2497986
Inside a house? Pretty sure the 12V 50A or whatever power supply is going to be prohibitively expensive.

>>2497457
I like 125B better. $4 for 1590 and $7 for 125B ain't bad.

About a month ago I posted in this general about needing a 5v regulator (PCB design) with the same footprint as an adjustable regulator I'd been using. One of you anons designed such a board for me and sent me link on EasyEDA. I replied a day or two later via EasyEDA but anon probably doesn't check those messages since he mentioned he doesn't like EasyEDA.

Anyway anon, if you're still here, please get back to me via that messaging service so I can pay you. Thanks.

>>2498110
>50 A

The 12V are checked but 50 A, where¿?

>>2497614
10 pins is kinda strange. Could be TO-74, 96, 97, or 100, you'd need to measure it to be sure. See the references at the bottom of this article: https://en.wikipedia.org/wiki/TO-5
Could you type out what you can read on it? Do you know where it came from?

>>2498154
Just a guess, car amps draw a lot though. Especially subwoofers.

>>2496996
so since i need 1000x gain, as the signal coming from the mic is very tiny, i don't think one opamp is ideal for this and since the chip has two amps on it i think it would be better to connect them in series so first op amp will do 10x and the second one100x

>>2498195
but for AC i need 3amps.. shit that is 2 chips.. so i suppose i could just copy this as one anon suggested or try if single opamp will work if i use 100ohm and 100k ohm resistors to get 1000x amplification from it. But that seems way too extreme, that will be noisy as fuck i bet

>>2498198
okay so it seems to work on paper, it amplifies just like it should
but can anyone tell me what the reference is for?

the datasheet simply calles it "reference input", do i just leave it connected to ground?

>>2498203
yeah, fuck it, this seems to work perfectly in the simulator , the resistor values i stole from the instru. amp datasheet, but i will use 3 ne5532 to build it, if i assume the condenser mic signal is between 1 - 100mv then i should be easily to get to line level output between 1v - 3v. god i hate working with analog signals, specially AC, but it seems like this is it baby, it will work

>>2498207
man, if it turns out, after going into all this trouble, that the problem wasn't in the shitty chink amplifier, i'm going to kill myself for sure
but it surely must have been since the gain resistor did nothing, and it got really hot when amplifying

>>2498198
>that seems way too extreme, that will be noisy as fuck i bet
60dB isn't that extreme compared to the open-loop gain of the amplifier IC. Check the gain-frequency graph, I wouldn't go more than ~20dB away from the limit since you'll start to lose linearity the closer you go. But you'll likely want to add capacitors in parallel with the negative feedback path(s) on any high-gain stages to cut off the high frequency gain.

>>2498203
You can may be able to get ~30dB gain in both the first stage (currently 5k / 2k) and the second stage (currently 6k / 6k) to get ~60dB total. Not sure what this method does to the common-mode voltage range, but I think it would actually be an improvement.

>>2498207
I'd prefer not to use such small resistors for the real thing. Even 5k can potentially be strenuous on the current supplying capability of some precision amplifiers. Check the amp's datasheet for a rating of something like gain / linearity / noise with respect to output current / load impedance.

>>2498195
>so since i need 1000x gain
You're probably doing something wrong here. You should get about 10mV at 100dB SPL out of your mic, which means to get line level input 1.228V you need gain of ~123. Also getting single ended signal from mic then converting it to differential to amplify it and then back to single ended is pointless.

>>2498223
>I'd prefer not to use such small resistors for the real thing
well i am not nowhere experienced enough to do better than the datasheet so i will just follow their values for now, if it starts getting hot or something, i can always easily change the resistors

>>2498228
Well i actually don't know what i'm getting from the condenser capsule i tried to oscope it but the output is too small to measure since it also has DC offset on it, so i wasn't able to measure it well on my chink scope, so 1 - 100mv is a rough estimate, but i think it probably outputs something around 10mv, from the tranny attached to the condenser capsule

>>2498229
>Well i actually don't know what i'm getting from the condenser capsule
That's why we have datasheets. It says what sensitivity is and from that you calculate how much volts per unit of pressure you will get out of it. Are you sure you're correctly biasing the FET? FET by itself should already act as an amplifier. To minimize noise you should start amplifying a signal as soon as possible.

>>2498238
yeah, rest of the circuit should be fine, i just copied it from the diy microphone video
I already had the mic working, but i had strong noise issues (despite good shielding) and the instru amp was getting very hot (it wasn't a short) and then it stopped,working completely, which is why i am replacing it with this custom made instrum. amp i will cobble together from these cheap audio grade op amps https://www.ti.com/lit/ds/symlink/ne5532.pdf i ordered from local seller
As for the capsule itself, it is a condenser capsule and directly to the capsule i soldered a tranny, and the tranny is being fed voltage so it can amplify the capsule signal. from the tranny i pull the signal via shielded audio cable to the circuit.
The tranny itself, 2N4416, i also bought from chinks so it's possible it is also shitty, but i am not so sure about that, but just in case it is i also oredered J111 tranny from the same reputable source to use as replacement if i run out of options. (Basically hunting down all the parts used in the video was a nightmare, nobody but small chink shops carry them).
And then if after i do all that, + add shielding cage for the whole mic, and if i still get the noise or poor performance, then i will go ahead with the aforementioned suicide, because there is only so much a man can take.

>>2498229
>i will just follow their values for now
Those are values for the internal equivalent resistors of the INA217. If you're using the NE5532, the datasheet says in relation to a differential amplifier circuit:
>Because the transfer function of VOUT–
is heavily reliant on resistors (R1, R2, R3, and R4), use resistors with low tolerances to maximize performance and minimize error.
>This design used resistors with resistance values of 36 kΩ with tolerances measured to be within 2%.
>But, if the noise of the system is a key parameter, the user can select smaller resistance values (6 kΩ or lower) to keep the overall system noise low.
>This ensures that the noise from the resistors is lower than the amplifier noise.
So it would actually be a good idea to use low resistances like that, my mistake. With minimal voltage slew it's probably no big deal in terms of heat.

The NE5532 has a unity-gain-bandwidth of 10MHz, so at 20kHz it will have a gain of 54dB [20*log(10M/20k)], so yeah you couldn't get the full 60dB for the whole audio window even if you tried. I wouldn't trust it to behave well when set to that kind of gain, especially without using a filter to lower its high frequency gain below that. Not that I've really seen issues with running op-amps at the edge of their frequency range. Using a gain of ~31 for both the stages of an instrumentation amplifier would be my intention.
Say, 15k/1k for the first stage, and 33k/1k for the final stage, would give about 30dB each time, pic related. With 2.2nF caps across each 3.3k resistor, I think that will result in a decent frequency response, though I'd simulate it with an AC sweep to be sure though.

>>2498238
Doesn't a JFET only act as a source follower? I can't see how a JFET gives voltage gain, but I've never closely studied JFETs so who knows. In this instance, with more or less equal source and drain resistors, the voltage gain shouldn't be high either way, I think.

>>2498244
That's why I hate those type of videos. They just give you schematics without explaining why and how circuit works. If they did a better job of explaining this you wouldn't have any questions about whether I need to use this or that device.
>>2498248
>Doesn't a JFET only act as a source follower?
Its a transistor so it behaves in same way as BJT or MOSFET on surface level.
>In this instance, with more or less equal source and drain resistors, the voltage gain shouldn't be high either way, I think.
Yes, with same resistors gain will be approximately 1. You can then get your differential signal from source and drain since drain will be inverted. But I don't see why this is necessary or even a good idea for wide variety of reasons.

>>2498164
Don't frighten me anon. The car amps are overvaluated chuncky clunky current sinks.

But Im planning to use it without these.
In fact, I'm trying to mix a car stereo with some older speakers from not working home stereo. Im starting to thinks its all a great ponzi scheme (well not like that but close) its a fraud.
Almost everything between 4-8 Ohms can be used, and with the proper wire, impedance calculated and fuses, It must be a decent sound experience.

>>2497986
>Sony DSX-A410BT
Power: 4 x 55W
>>Sony DSX-A110U
Power: 4 x 55W
>JBL Celebrity 100
Power: 4 x 50W

Make sure your PSU puts out at least 250W continuously if you want to listen at high volume. ATX PSUs can do it, but SMPSs are noisy. Also, the case might need extra ventilation.

>>2498267
Well that resistor arrangement ensures linearity without relying on the small-signal gain of the transistor, and buffers the signal to a low output impedance for further circuitry. And it keeps the signal fully differential.
The only downside is the lack of help biasing, but this is an electret microphone circuit that relies on the electret to bias the JFET, which is already suspect.

>>2498411
>Well that resistor arrangement ensures linearity
Yes, but I doubt it will be noticeable unless you'll be recording someone performing an Italian opera with incredible vocal dynamic range. You can always cope that it makes the sound warmer, like tube people do. It's better to just use proper feedback from the output to get more linearity.
>and buffers the signal to a low output impedance for further circuitry
Problem with that is that source pin will have lower output impedance than drain and they won't be equal. Looking into drain you have a current source with very high output impedance so whatever drain resistor you have that will be output impedance. Looking into source you have a buffer, which means very low output impedance. Source resistor doesn't really matter because output impedance is so low. In >>2498244 this means that there is different impedance on vin+ and vin- lines and if same magnetic field induces current in both lines top one will produce more voltage compared to the bottom one and you'll just amplify the noise which makes the whole circuit kinda pointless from noise rejection perspective. If you need to drive balanced audio lines just use a transformer or at least proper differential line driver.

where the fuck do i get 512kb DRAM ICs for breadboard? only ones i seem to find are for soldering on PCB

Hi, does anyone know of any good communities/forums for electronics enthusiasts?

>>2498441
should is separate the two lines going into the opamp on the pcb with more distance then?

>>2498446
There are various boomer forums, but its difficult to fit in when you have oldfags with 10k posts and whole sidebar of badges and titles. There are probably some zoomer discord servers, but that's also walled garden territory. Other than that you have this place, but electronics just isn't as popular compared to programming. Most of engineers are in their professional communities at work or in academia.

>>2498446
lemonparty dot org

>>2498448
No you should keep lines as close as possible and if you really wanted you could fix the impedance of each line so that they match. If we're talking about >>2498244 if we ignore C2 and C3 looking into the circuit from Vin+ you have R1||R8 which is 1.41k, but if you look from vin- you only have R9 of 3.9k because source of JFET will have low output impedance so it looks like ground for analysis purposes. This means that R9 should be 1.41k. Now that I think of it you're actually attenuating signal from JFET by R1 R8 divider before amplifying it so no wonder that you need a gain of 1000. This whole circuit just rubs me in the wrong way.

>>2496402
I was an electronics technician for eight years and now I’m an electrical engineer. I absolutely suck at circuit design and have been tasked with designing some PCB’s at work. I’ve just come here to say hello and god bless you all

>>2498454
well, i am no smart enough to make any modifications to the circuit like that, so i'm sticking to the original schematic, i heard the mic in the video and it sounds amazing
a tiny amount of noise is fine as i am not making music recordings, i just want high quality mic for voice overs and gaming

>>2498454
one big problem i had with was 50hz noise tho, do you have any idea where it could come from? could it be the shitty chink interument amplifier?
i tried wrapping the mic in grounded tinfoild so it wasn't from the mains

>>2498449
tfw born just in time to have to be a lonely boi

>>2498482
Don't worry, son. It'll all be over–


SOON.

>>2498482
How many hours have you logged with heavy machine guns, Anon?

>>2498473
This is one of those black magic parts of electronics and requires search by elimination strategies and general tweaking to see what works. If shielding doesn't work then it's magnetic in nature and that won't be stopped by some aluminum foil - you would need thick iron for that. Did you try to use your device on another computer? That way you would know if it's a ground problem or if it comes from outside.

>>2498482
Time to start speaking to pigeons. That's what Tesla did and it worked for him.

>>2498110
>constant current
that's what i am saying. voltage is pointless information and is not mention in most datasheets. they just show the LED configuration so if you know the current and there are 6 LEDs you know that you need a constant current LED driver that can drive 6 or more LEDs with the approximate Vcc max about 3v per LED. That's all you need to know.

>>2498449
do EE's even do circuit design? it seems like they just chain obscure chips and let the computer do layout for them

>>2498449
why do you need to fit in? most frequent posters are opinionated narcissistic douchebags with huge egos and they are protected by the mods since they generate traffic. lurk and read and search as much as possible, and when you are stuck, ask a question. then you have two choices: 1. ignore all the snarky remarks, such as "did you check the database", and wrong or condescending responses, until you are satisfied and just go way or 2. call them out, which they don't like, so once they give you more crap, escalate and expose their BS and get banned and create a new account when you need to ask another question.

>>2498519
It depends on where you work. But yes, if you want to design from scratch you need to be in academia or in specialized researcher position that you find in large corporations. Small engineering jobs are mostly about copying over shit from datasheets or application notes while adhering to required standards.

>>2498527
Forums were never that useful anyways. It's exactly like reddit because everyone is obsessed with building reputation. There is enough books and other resources available that you can figure out anything yourself so asking for help is unnecessary nowadays. What >>2498446 was asking for is not help forum, but some place on the internet where people have fun without trying to prove anything, like here, but this place is small and slow.

Can someone tell me if RHW-2 aluminum 250MCM 3 conductor cables exist? I've been asked to find it. Everything I find is 1 conductor or strands (is that different from 3 conductor?) I don't know anything about cables

>>2498544
>where people have fun without trying to prove anything
I don't think it is possible. they always start like that since that's the idea. but as long as you have "castes", i.e. accounts and nicknames and users are identifiable and most importantly you can tell newbies from old timers by the number of posts, account age, various stupid "awards" etc etc, these kinds of forums always attract people who want to prove something to others to feel good about themselves. diyaudio is a good example.

>like here,
because it is an anonymous board so nobody gives a shit and people strictly ask questions and help each other and there is no place for all that bullshit that comes with namefaggotry. but even this board is pretty toxic. /ohm/ is ok, but there are lots of examples on /diy/ of people shitting on other peoples projects just for fun.

>reddit
the question/answer subreddits are ok for the same reason. random people pop in and just answer your question. no room for elitism.

>physicsforums
I liked it initially because it is also a question/answer board for the most part but it is still a circle jerk of the "gurus" who fellate each other and run up their "points" and look down on anyone who disagrees with them.

also keep in mind that places like /ohm/ are unique in a sense that it is based on the willingness of very few knowledgeable posters to answer questions. if those few people leave, this general will become even slower. i honestly don't need a huge message board with 1000's of posters which is just a bunch of noise. i prefer a slow board since if my question is answered that's all i care about even though it may turn out that it is only one or two anons who are able and willing to answer this particular question.

>>2498441
Wait the impedances are different? Christ I don’t like that circuit anymore. Why not just use a JFET-input op-amp instrumentation amplifier on the capsule?

>>2498652
Isn’t EEVblog forum ok? Seems a pretty constructive community, lots of people posting cool shit they’ve made or found.

>>2498670
>eevblog
never posted on it, but yeah once google search takes me there, i don't remember seeing much drama. but it has its share of assholes. picrelated.

they also allow "project" threads i think. you can basically create a thread for your project and post updates and other people will pop in to share advice. i'd invariantly expect some douche to show up and start bragging how he did better.

>>2498677
I don't thing the person in picture is an asshole, I think whoever OP is retarded for not including as much information possible in his post.
The #1 reason I don't help with programming in 4chan is because usually the problem the poster has is probably from another problem that would solve the problem in a much more simpler way.
I guess in DIY electronics posters like these are much more common because many people don't care about their time or money, they want to do things the hard way, they just want to DIY to learn and whatever, but it's not cost effective, or just dangerous, and can just solved by not doing DIY (aka just buying the solution).

I'm building a machine that includes an xy axis moving plate powered by 2 stepper motors. I want to use only one output from my microcontroller capable of pwm.

What do I need? is there a special type of relay in which one output can be 24vdc and determine which of the two steppers receive the pwm? I only need one stepper at a time.

This also leads to another question, why are xy linear guides so damn expensive, are there any cheap options out there? the cheapest i've come across are cnc plates which is overkill (max weight i require is like 1.5 pounds) and like $1200.

>>2498712
are you outputting step+dir signals into a stepper driver board/ic? if so just use a couple of NAND or NOR gates to switch which driver the step signal goes to. if you're using a generic motor driver then you'll need two signals anyhow.

if rigidity doesn't matter as much, use rollers on extrusions like the ender 3. or just buy a broken 3d printer and modify it.

>>2498719
I guess I would have 2 different motor drivers with the dir and power signals individualized. Just the pulses, my microcontroller only has two outputs capable of highspeed pulses so I wanted to use just one for the gantry table.

>just use a couple of NAND or NOR gates to switch which driver the step signal goes to
How would this look like on a PLC? This might be a dumb question, but can I have a single 24vdc output controlling this relay coil and changing the direction of output for the pulses? Would this affect the pulses at all?

For example, stepper 1 would be NC, stepper 2 would be NO. Or are there better alternatives

>>2498725
I guess I should clarify the motor driver PWM inputs 1 and 2 would be NC and NO, respectively

>>2498445
make a pcb yourself that breaks it out onto a breadboard

I figure this is more of an electronics question than a EMT question.

I replaced my control board and for some reason I keep getting a sporadic input signal from my CNC machine when I try to use a probe (input wire to CNC machine tool/frame). I disconnected as many things as possible and pic related is my current fucked up setup. There is supposed to be an optocoupler (separating the 5v and 12v sections) for the input signal. Normally 12VDC is supplied to the board, the "input pin" is positive, and current flows through the optocoupler when bridged with ground.

I have no clue whats going on electrically now. I still get the signal if I remove the 12VDC, but the ground from the 12V supply must be plugged in. It also only seem to happen when the stepper driver/motors are on, which are connected to the frame but are not grounded. All input pins work fine when the frame is not involved and there is a simple switch/bridge.

Somehow there seems to be current flowing from 36VDC -> stepper driver -> stepper motor -> frame -> frame wire -> optocoupler(?) -> wall wart ground. What solution am I looking for here? Other than making sure the stepper leads are properly insulated?

>>2498806
Also I checked for any shorts between stepper wires and the frame and found none. 36v is only connected to stepper drivers, and only the 12v ground is connected to the board. I'm at a loss how current is getting from the 36v psu to the 12v wall wart ground.

>>2498670
Eevblog is alright. I usually browse the Metrology and Dodgy Tech sections.
There are a bit of in fighting occasionally, but nothing super serious.

>>2498851
Oh, and we also have our residential skitzo
https://www.eevblog.com/forum/dodgy-technology/optical-bench-redux-digital-switching-can-have-analog-functions!/

>>2498445
buy breakout boards and/or zif sockets on ebay

>>2498670
>Wait the impedances are different?
It's just horrible, because JFET is struggling to get into active region.

>>2498905
Forgot mic capacitance, but still it's giving out 2mV from 10mV per line this means it's attenuating by 2.5 before amplification just to get some vague differential signal for no reason which gets then immediately converted back to single ended output.

>>2498905
You forgot the electret’s intrinsic negative bias that pulls the gate down below ground.

>>2498909
I don't see how it will pull anything below ground. It's not going to do anything at DC as far as I know.

>>2498907
The differential signal ideally means that you avoid all common-mode noise before the amplifier. Not very important if the preamp is right next to the microphone, but in this case there was a gooseneck between the mic and the PCB with the amplifier.

>>2498914
It's an electret, it has an intrinsic DC electric field. Like how a magnet produces an intrinsic magnetic field. To the low leakage JFET gate, I'm pretty sure it floats down by some voltage, though I couldn't tell you how much.
Electrostatics are strange.

>>2498916
>but in this case there was a gooseneck between the mic and the PCB with the amplifier.
Oh, ok, I missed that. But still there is impedance mismatch between pairs of around 350 Ohm so external fields will produce noise that will get amplified. JFET in this configuration is not a proper single ended to differential converter.
>Like how a magnet produces an intrinsic magnetic field.
Yes, but you can't exactly wrap a wire around magnet to create a battery. You need to move the magnet. Same with electrets - attaching a voltmeter will read 0V until you start moving things and putting in work. It can't produce current from nothing and even with low leakage of JFET it would deplete it eventually.

How do i convert a square wave of unknow polarity to a current able to drive LED? Want to light my model train and in the first iteration i just want to light the wagons if there is power on the track. Though I use DCC

I want to build a diy light fixture with dimming which consumes about 100W. Problem is that the only dimmable 350mA CC supply caps out at 40W. So i would need 3 of them and with each two cables or would it be better to drive them with CV and design a current limiter board using a NSI50350ADT4G and dim them using PWM?

>>2496402
I'm fixing a motherboard and I can't figure out why the 1st pcie slot is not working. Every other one works just fine. I checked the voltages and the small caps coming out of each lane yet it doesn't detect a gpu. Any ideas? It's not like the southbridge is gone, I've got clocks, the others are working.

>>2498955
Reflow all of the solder joints for that slot. Inspect the contacts inside the connector too. Try to find the stress points in the PCB where it would flex from a card being pushed into the slot, and examine nearby components for cold joints.

>>2498955
and make sure there's no funky BIOS setting that's preventing your PCIe card from being initialized.

>>2498955
Look for broken traces under magnification.

>>2498931
Is it PWM? If so you can use a low-pass (LC?) filter with an appropriate corner frequency. What do you mean by polarity?

>>2498950
>the only dimmable 350mA CC supply caps out at 40W
Why not go for a dimmable 1A CC supply instead? And what voltage is this for? 40W at 0.35A implies 115VAC, while the driver you posted can only handle 50V. Best go for a switching driver instead of a linear one either way, much more efficient for 100W. But you can stack CC regulators in parallel.

Anyone using lithium iron phosphate? 3.5V charging voltage vs 4.2V for li-ion, and lower energy density, but immensely more durable... according to Wikipedia 400-1200 cycles for li-ion, 2750-12,000 for li-ip. Tolerant of both overcharge and deep discharge. The way li-ion just dies just sucks

>>2499087
Anyone using lithium iron phosphate? 3.5V charging voltage vs 4.2V for li-ion and lower energy density, but immensely more durable... according to Wikipedia 400-1200 cycles for Li-ion, 2750-12,000 for Li-IP. Tolerant of both overcharge and deep discharge. The way Li-ion just dies just sucks.

I was just reading about Tesla starting to use lithium iron phosphate in mass:
https://electrek.co/2022/04/22/tesla-using-cobalt-free-lfp-batteries-in-half-new-cars-produced/
>Tesla confirmed that nearly half of all its vehicles produced last quarter are already using cobalt-free iron-phosphate (LFP) batteries.

I was just reading about Tesla starting to use lithium iron phosphate in mass:
https://electrek.co/2022/04/22/tesla-using-cobalt-free-lfp-batteries-in-half-new-cars-produced/
>Tesla confirmed that nearly half of all its vehicles produced last quarter are already using cobalt-free iron-phosphate (LFP) batteries.

Anyone using lithium iron phosphate? 3.5V charging voltage vs 4.2V for li-ion and lower energy density, but immensely more durable... according to Wikipedia 400-1200 cycles for Li-ion, 2750-12,000 for Li-IP. Tolerant of both overcharge and deep discharge. The way Li-ion just dies just sucks.

>>2499087
I would be using the NSI driver at 36V with one driver for each cob led. Those run at 33V. I am concerned about overdriving the leds, as they should get max 500mA and run normaly at 350. What do you mean with switching driver? I hoped to use the LCM series or an elg.
also DCC is a rectangular AC wave. So yes PWM.

>>2499147
>Li-IP

These aren't new. LiFe/LFP/LiFePO4 batteries have generally been the preferred chemistry for "good" EV battery packs for a while. Their major downside is the lower energy density, but it's not by a huge amount. The radically increased cycle life is usually worth the tradeoff.

>according to Wikipedia 400-1200 cycles for Li-ion, 2750-12,000

Note that cycle life, in general, is extremely heavily influenced by both depth of discharge and charge profile. A cell that is discharged to near zero and then recharged rapidly to full will not last very long. Last I checked, even LFP won't last more than 500-1000 cycles (to 80% initial capacity or less) in those conditions.. A cell that is rarely or never below 20% and deliberately "undercharged" to a voltage lower than normal can last an order of magnitude longer than one in the aforementioned scenario. There's also the separate issue of "at what point do you consider a cell 'end of life'?" In any event, yeah, the cycle life of LFP is much better than Li-Ion. Too bad lithium titanate cells and their 20,000-30,000 cycle count lives are so expensive...

>Tolerant of both overcharge and deep discharge.

I assume this is in reference to the fact that they won't necessarily erupt into a fireball (unlike lithium polymer cells), because overcharge and overdischarge will absolutely make a big dent in the service life of the cell. Ask me how I know a shitty Chinese charger that tries to float the battery will kill an expensive pack in a couple hundred cycles or less.


Anyway, to answer the initial question, yes, I am using lithium iron phosphate.

>>2499220
So, tell me, how you charge them, is CC/CV to 3.5V at <0.2C good? I generally charge li-ion to 4V, so do you think I should aim for 3.3V? At what voltage would you consider them discharged, typical and conservative values?

I buy these cheap little protection circuits for li-ion off aliexpress, I see there are some for li-ip too, which is good, I think there'll be li-ip in a future project of mine soon. What are the smallest size cells you've come across? I have hundreds of 13mm x 40mm li-po batteries from vape pens that people drop like cigarette butts, but they all contain a 1340 550mAh li-po battery

>>2499227
>CC/CV to 3.5V at <0.2C good

Yes and no. It will "work", but you really need a 3-stage charge profile. They charge with a normal CC/CV to start with, but they don't want to sit indefinitely at the and of the CV portion. The charger needs to either reduce to a lower hold/maintenance voltage, or just stop supplying power completely. Ideally, the charger would be integrated with the balancer and immediately enter 3rd stage after balancing was complete. This isn't usually possible, so sitting at end-of-charge for an extra 5-10 minutes for the balancer to work is typical (assuming the balancer only does its job during charging, anyway).

Voltage for fully charged/discharged will vary slightly depending on the exact chemistry the manufacturer uses. Ideally, you can reference a detailed datasheet, complete with test information, but not all manufacturers make this public, if they even bother at all. Typically, though, this will be in the range of 2.4-2.6V (2.5V being the most commonly cited) and 3.5-3.7V (3.6-ish being most typical). Staying within that range should be safe for most any LFP cell, even if it's not necessarily optimal. Charge voltage may also vary on how fast you're trying to charge the thing.

The most recent pack I made for my ebike, using Headway cells, cuts out at (I think, I almost never run it empty) 2.6V. I modified the garbo charger to top out at a little over 3.4V/cell instead of the ridiculous 3.85V/cell it was going for to start with. Charge rate is about 0.4C. Lower is obviously better, but returns diminish greatly after a certain point, and what point that is exactly will depend very heavily on the quality of the cells.

This gives me only about 85-90% of the nameplate capacity, but should radically increase cycle life; at least 4x what it would be at 3.6V. If they also aren't regularly discharged fully, that also significantly increases cycle life. I should get a minimum of 2k cycles to 80% initial capacity using these parameters.

>>2499237
I have read about that finishing voltage thing with li-ion, actually this pdf is linked to at the top of this post:
http://www.elteconline.com/download/pdf/SAFT-RIC-LI-ION-Safety-Recommendations.pdf

Hmm can't copy-paste so I'll make a screenshot, pic related. So I figure leaving a li-ion connected 24/7 at 4V is ok (actually battery university frequently mention 3.92V as a li-ion charge voltage, something like "the military" use that voltage iirc). Similarly I suppose li-fepo at 3.3V or something should be o k

I suppose li-ion at 3.92V would give you many of the advantages of li-fepo. like-faecespoo. No. Iron Phosphate. Good.

>>2499237
>>2499237
I have read about the finishing voltage thing with li-ion, actually this pdf is linked to at the top of this post:
http://www.elteconline.com/download/pdf/SAFT-RIC-LI-ION-Safety-Recommendations.pdf

Hmm can't copy-paste so I've make a screenshot, pic related. I figure leaving a li-ion connected 24/7 at 4V is ok (actually battery university frequently mention 3.92V as a li-ion charge voltage, something like the military use that voltage for maximum power recovered from the batteries taking into account their state of charge and the number of cycles they'll last iirc). Similarly I suppose li-fepo permanently connected at 3.3V or something should be o k

Because it seemed to me all that you'd do by leaving li-ion connected at 4.2V for days would be similar to charging it to 4.3V for hours, just you let more current trickle charge in than a normal 4.2V charge with the days of trickle charging. So I imagine an advanced charging algorithm for li-ion would skip the CV stage of charging altogether, eg what if you CC to 4.2V then stop the charger. Is that equivalent to a CC-CV charge to 4.05V or something depending on the battery internal resistance?

I suppose li-ion at 3.92V would give you many of the advantages of and maybe a similar capacity to li-fepo.

Has anyone built something like this? Are those real instruments? or some gibberish like "primary power overdrive"
https://www.youtube.com/watch?v=axXrs55X9kE

Alright, I had a debate with my brother over whether resistance heating had more to do with resistance or current. After that I looked into it more and now I think I understand it, but please correct me if I'm wrong. So in picrel there would be 1 amp moving through the circuit. Voltage and total resistance both decrease proportionally as one moves further up the circuit. In this case, the voltage after the first wire drops to 11, then 1 after the load, and to zero after the second wire. Since the load has 10 volts applied to it, that means it is consuming 10x1 amp to make 10 watts. Each wire is consuming 1 volt x 1 amp to make 1 watt each. Therefore the load is experiencing 10 watts of heating and each wire one.
TLDR: Heat distribution is determined by voltage drop and current, and voltage drop is determined by resistance. Am I right or just retarded?

how do you calculate the gain of this

>>2499217
Well the IC you posted is a linear current regulator, while most dedicated CC LED supplied are switching. Switching means complicated but also more efficient.

Please articulate what LEDs you have. What their typical voltage is, how much current they want, etc.

>>2499271
Ideal op-amp rules and a lot of algebra. Maybe millman’s theorem too, judging by that middle node.

>>2499262
you got it

>>2499271
KCL and assume V(-) = V(+)

>>2499262
Do you not know what a watt is?

>>2499262
it's about power, which can be related to voltage, resistance, and current. Power is what matters.
>P = IV
>P = VV/R
>P = IIR
In practical systems, current is usually what you have more control over, but ultimately it's about power.

Im having an annoying issue with an Op-Amp. Im trying amplify a steady voltage but when the voltage going into the Vcc of the op-amp changes by just a volt it screws up my amplification by too much. Whats a way I can keep the voltage going into Vcc to stay steady enough?

>>2499384
And to add, the steady voltage Im amplifying is coming from a zener diode so its not coming from directly from the source voltage into the op-amp.

>>2499217
would be 10 Cobs, each running at 33.7V with 360mA and a max current rating of 720mA. Well the lamp would be running potentially all day at max power. So efficiency is kinda important.

>>2499413
If they're all in parallel you'll likely want to use a seperate CC driver for each LED, they can get out of balance if you try to run them all off one due to slight manufacturing differences. If you don't have an AC-to-DC power supply already then running them in series is an option. 5 in series is probably too close to rectified mains (in 120V countries) to work for a non-isolated regulator, but 2 in series would work fine. In that case you'd only need 5 total >~70VDCmax 360mA regulators.

On the other hand, you could run them all off a >~40VDCmax 3.5A supply with a 10Ω resistor in series with each LED COB. You'd lose a few watts to heat in the resistors though.

Here's an example of a switching LED regulator:
https://www.jaycar.co.nz/16w-12v-1-25a-led-power-supply/p/MP3372
Only 12V though.

>>2499415
I dont have any PSU. Bit concerned about cost as the 4-5 CC drivers would cost around 80-100 with more messy wiring, whereas a CC driver for each LED would total 60 and only require 3 wires. But if you reccomend using a resistive dropper, would it then not be better to directly use a linear driver to be more accurate?

>>2499434
No I'm saying just to use a cheap resistor for each LED (dropping maybe 1-5V) and a single 3.5A CC driver for all the LEDs. The resistors will keep the COBs somewhat balanced. Pic related, where each LED is a COB.

>>2499271
Well that Rg is kinda annoying so you can get rid of it by setting it to infinity.

Output of first amp would then just simply be vin- * (1+R1/R2). Due to superposition you can then separately calculate what second amp does for vin- and vin+. For vin- set vin+ to 0 which just gives inverting amp with gain of -R2/R1. Now set vin- to 0 which pulls out of the first one to 0 and vin+ gets amplified with factor of (1+R2/R1). So with Rg at infinity

vout = (vin- * (1+R1/R2)) * -R2/R1 + vin+ * (1+R2/R1) = (-R2/R1 - 1) * vin- + vin+ (1+R2/R1)
vout = (vin+ - vin-) * (1+R2/R1)

Now we need to add back Rg. This you can do with EET method and figuring out driving point impedance on Rg terminals. When inputs are 0 then Rg is connected to virtual grounds and feels 0 Ohm. Next we need to figure out Rg terminal impedance with output nulled. If you calculate this by writing 2 node current equations for - termnals of amps you get Z(nulled) = 2R1/(1+R1/R2).

With that you can get whole transfer function
vout = (vin+ - vin-) * (1+R2/R1) * (1+2R1/RG/(1+R1/R2))

>>2499271
that is why i always connect everything in series
much easier to calculate
anything parallel should be forbidden in circuits

>>2499480
Maybe you should start being responsible and wear an electrostatic grounding wrist strap when you're with people you don't know very well, then you won't, you know

>>2499529
but it feels better without a strap on

Can someone help me understand the point of the double diodes here?
I'm making the schematic from part of a PCB that controls a current sensor.
There's an ASIC that senses the current from the sense winding of a current transformer and then drives power through a compensation winding that's measured with an op amp.

Anyway, this is the push pull circuit from one end of the compensation winding. The ASIC is 5V logic.
I'm assuming if the top ASIC signal went to 5V and the bottom one to ground, Q1 will turn on.

This will saturate Q3 I assume but what will happen at the base of Q3?
The base will remain at a lower voltage than the emitter so it stays on but not lower by a factor of the forward drop of the 2 diodes?

Are the diodes just there to provide some sort of snubbering with the two capacitors?

>>2499562
>Can someone help

after you trace out the circuit CORRECTLY, it'll make a lot more sense.

>>2499577
It is correct.
I have a bare board and can see all the traces

I'm making a battery level indicator for a flashlight. Used an LM3914 and I got it to work more or less.
Now I'd like to make it turn on via a button and stay on for about 3s before it turns off. What are my options?
>inb4 arduino

>>2499562
Looks like some strange negative feedback. Since there’s no protection against crossover distortion in the form of a Vbe-multiplier, maybe the pushing and pulling of the first bank of transistors pulls enough current through the capacitors to influence their emitter voltage and cause feedback that way.

I’d simulate it.

>>2499654
MOSFET + RC circuit. Gate power to the entire circuit with a small-signal FET like the BSS138, tie the gate of the transistor to ground via a resistor and parallel capacitor, and connect them to VCC using a button. So when you press the button the cap charges up immediately and the FET turns on, but discharges through the resistor until the FET turns off. Adjust R and C to get 3s or so. Importantly the idle current is basically zero, so it won’t drain your battery. If you use an N-FET it needs to be between the load and ground.

What is the most efficient buck converter that you can get on AliExpress that:

>input voltage: 11-16V (let's say 10-18V to be safe)
>output voltage: 5-10V (HAS to be adjustable with a potentiometer or something similar)
>current: less than 1A

The buck converter will be driving a small brushed DC motor. I am currently using a LM2596 based converter which has a horrific 59% measured efficiency, mainly because of the low current and input voltage.

I have order a MP1854 and MP2307 converter to test them out to see which will be more efficient, as well as one that I have no idea on which IC is based on:

https://www.aliexpress.com/item/32770347590.html

Is there any other, more efficient alternative that I am missing? The webench power designer suggests a bunch of ICs that I can't find on Ali nor are sold as ready to use units.

>>2499663
Buy a TL494 and use it to control a half-bridge driver. That way you get synchronous buck with high-spec external transistors. You can ditch the driver if you use a P-channel switching transistor and maybe a totem-pole for driving, plus a diode for asynchronous rectification. I got chinked on IR2184 half-bridge drivers from ali, so watch yourself. All ali switching ICs besides the most ubiquitous TL494s and garbage MC34063s I’d be suspicious of, even the 2596s.

An internal-MOSFET switching regulator that uses synchronous rectification going to be more efficient than anything that uses BJTs as switching elements, at 1A it may be more efficient than external transistors.

>>2499562
>Are the diodes just there to provide some sort of snubbering with the two capacitors?
Yes, that's what seems to be the case. Circuit works just fine without diodes and capacitors. But when you start increasing L big spikes start to appear on collectors. Now let's say there was +15 on inductor and we switch to negative -15. Problems happen during the switch if there is a time when both are off. This means that inductor is basically floating and because there is magnetic energy stored in it voltage must change. If you remove positive voltage, you'll get negative spikes until bottom transistor turns on. Before it turns on its collector base junction acts as a diode and happily conducts negative voltage to base. If you leave that shit there it will destroy transistors. Now when negative spike happens it will go through CB junction of bottom transistor and through diodes to ground. Now capacitors are there to speed up transition so that they help pull input pins into same direction. At least that works if you swap direction of diodes and put capacitors to bases of input transistors. Then ringing basically disappears.

>>2499691
No snubber.

>>2499692
Just diodes.

>>2499691
In the pic >>2499562 the caps are connected to the emitters of the BJTs, not the bases.

>>2499723
Yes, I assumed pic is wrong, because caps are pulling in opposite direction and make transients worse instead of helping reduce the spike.

>>2499723
>>2499726
I really doubt this is the intended mode of operation.

>>2499726
>>2499727
Well I say it depends on whether it's being used as a digital totem-pole (idk man just use a half-bridge driver and mosfets instead) or as an analogue amplifier. Since it's for a transformer, maybe it feeds a sine wave or a sawtooth.

Actually how is that switching going all the way to and from ±15V with just 5V logic? Feeding voltage into the transistor emitters feels pretty janky, are you sure an absurd amount of current isn't going in the wrong direction back into the ASIC when it's outputting 5V? At the very least I'd suggest the 4.7k makes more sense connected to the emitters than the bases.

>>2499733
>Actually how is that switching going all the way to and from ±15V with just 5V logic?
That is the clever bit here - for a transistor to turn on you need to have BE junction open. Nothing forbids you from forcing voltage on base or emitter because when top pin is on and bottom one is off top transistor is open, but bottom one is closed, so what happens is that current flows from top 1k resistor, through transistor and to emitter back to pin that is set to 0. Base of the output PNP is pulled low which means junction is open and Vcc gets dumped to the load. Same thing happens in reverse when you swap values of the input.
>are you sure an absurd amount of current isn't going in the wrong direction back into the ASIC when it's outputting 5V?
No it's operating within the parameters, 1k resistor limits how much current IO needs to sink. And 4.7k limits base current and it doesn't make sense to put it on the emitter because that would rise required effective Vbe to turn on.

>>2499737
>1k resistor limits how much current IO needs to sink
But it will flow through the BE junction of Q2 and Q4 instead, won't it? All you'll get is a

>>2499738
all you'll get is a 0.7V drop, it's basicaly just a diode in parallel with those 1k resistors. The only current limit into V2 will be R2 multiplied by Q1/3's hFE.

I have a variable DC power supply that can output 2400 watts, 600 volts max. I need to convert this to 1200 volts, 2 amps.
What would you guys do?
I've considered an H-bridge DC-DC converter, a boost converter, and some other strange designs involving several tapped full wave rectifiers and a smoothing cap. Each one has drawbacks, but I don't know enough about power electronics to know which one would be best given the drawbacks.

>>2499738
>>2499740
Yes, on that one you are correct, it will have to sink current through 1k + base current of output transistors. Emitter/R2 resistor will be necessary to limit this to safe levels. But then again they could be doing something in the ASIC to address this issue.

>>2499747
2400W is pretty extreme for diy switchers. I think most would opt for a multi-phase synchronous or even quasi-resonant converter. See if you can find some appnotes, I have seen some real neat looking multiphase switching ICs by TI with integrated high-current gate drivers.

Though you might be better off modifying or replacing the transformer inside the variable supply. Or rewinding a big transformer from an old welder.

gonna make a dev-board based around the eg2131 15c chinese gate driver with decent specs but a shitty pinout
hope it works, gonna order them from jlc and make 3 of the 5 for testing bldc driving algorithms
note: schematic not designed to be pleasing to look at

>>2499793
just added optional diodes in parallel with the gate resistors. a lot of these components are optional, hence the jumpers and parallel capacitors. not sure what having such a massive bootstrap capacitor will do, but it will be an interesting test

>>2499765
it's gotta be a bit extreme. I can't say too much without doxxing myself, but I'm part of a team developing an ion thruster. The DC power supply is on loan from a corporation, so we're not allowed to modify it.

The NASA design used this circuit, which is fine for a final product. We just want something that's simple and "just works" so that we can have a working prototype. The refinements will come later.

>>2499816
They don't list any IC names, do they? Could also just copy the design of your existing PSU. Can't really say if a DC-to-DC converter would be much simpler than going for a whole new AC-to-DC power supply, considering you're dealing with high voltages either way.

>four phases in series instead of in parallel
Huh, haven't seen that before.

>>2496919
if all you need is continuity testing capabilities they sell ones for that purpose, but 1 probe lights up if cost is an issue. Most come with alligator clips but you could replace that if you have a soldering iron

>>2499834
>They don't list any IC names, do they?
not anywhere that my friends or I could find.
One of the main limiting factors is that me and my entire team consist of undergrads, and of those undergrads, only me and one other person are EEs. We're both pretty advanced for undergrad as far as EE goes, but unfortunately neither of us have done any power electronics.

Based on this datasheet can someone teach me how to determine what the max output current is for this op-amp?

https://pdf1.alldatasheet.com/datasheet-pdf/view/228176/TI/TL074C.html

>>2499916
It's fine with shorting the output indefinitely, as seen on page 4. And it has an output impedance of at least 128Ω + 64Ω, as seen on page 3. So I'd guess no more than that would be at the rated output voltage, perhaps 80% to be safe. Since none of the ratings are quoted at high output currents, it would be a good bet that you won't be getting the full linearity and output range at high currents.
Use a BJT or two to buffer the output current if you're worried.

Can anybody recommend a Chinese/Mexican/Indian company that does thru-hole soldering? I've got to make a few thousand boards that need a human assembly line. I've had companies like JLCPCB do pick and place for me before, but I've never seen one that makes organising thru-hole simple and easy.

>>2499953
god gave you 10 fingers and 10 toes

How far away can I practically detect power lines using capacitive coupling? Is there a better way?

This is going to sound absolutely retarded, but could i use a standard 110v breaker on a low voltage DC circuit (let's say 24v). From what I've read some people say absolutely not, some people say yes but the current values change and it still has to be within a certain voltage range. How exactly are these tripped? Is it purely current based, or rather is it power based? Also i hear DC can cause arcing issues at higher voltages.

>>2499965
>but could i use a standard 110v breaker on a low voltage DC circuit (let's say 24v).
Yes you can do it, they are tripped based on the current flowing through and it doesn't matter if the voltage is lower. They still trip with DC.

You should use a breaker which is dual rated by the manufacturer for AC/DC use. I recommend the "QO" series breakers from SquareD/Schneider for this purpose. (they're readily available and not expensive)

HOWEVER, the current interrupt rating will be a lot different than it would have at 60hz AC. This probably doesn't matter to you unless you're hooking it to a very high current source like a large lithium bank. This has to do with the fact that the AC current reverses flow 60 times a second and that tends to aid in attempts to extinguish an arc inside the breaker. Only some product lines from some manufacturers are rated in DC as well as AC.

https://download.schneider-electric.com/files?p_enDocType=Catalog&p_File_Name=0100CT1901_SEC-07_Aug2022.pdf

>>2499964
With a high-Q narrowband band-pass filter, probably multiple metres without too much difficulty. Might have a better time with magnetic fields though.

would I save money in the long run if I invested in a battery spot welder for making my own batteries for cordless tools and ebikes? I just had a Stihl battery fail just out of warranty , and am kinda pissed about being vendor locked.

>>2499562
>>2499691
>>2499723

So there are actually another pair of double diodes before and after the inductor that are used as flyback diodes so the lower ones must be for something else

>>2499991
What if I use a larger "antenna", say 1m in length? I've seen videos of people powering lightbulbs using capacitive coupling.

>>2500000
idk but i'm just stopping by to check these quints

>>2499659
Thanks for the answer. I guess that would work but my battery indicator includes a boost converter and those tend to leak power so I'd like to use high side switching if possible.

>>2500008
How does the output waveform look like? If they're not for protection then they might be used to help shape output in certain way. You will really need to trace full circuit for that and really confirm that those are really BAV99 diodes - even if there is A7 printed on them there are dozens of other devices with same marking.

>>2500016
When you're significantly below the wavelength of the EM you're trying to pick up, your received signal amplitude will be directly proportional to the size of the antenna. In the case of a short capacitive antenna, output voltage is proportional to length. Increasing the area I think might increase the total power you can receive, but don't quote me on that.

>>2496402
what interesting stuff can be made with phone drop wire terminal unattended?

There is a terminal post more or less easy to reach (since I'm a monkey), guessing must be sometime like picrel.jpg
>3M 10 pair boxes N501186A0000 / N501187A0000 type
half opened but, operative.
I mean, it could be easy cake to jump plug a couple of wires and all that, but then what?
AFAIK, the to be expecting something usable, these hooked wires must be attached to a secondary cross conection boxes, isn't it?
Several years ago, from these terminal there was phone booth conected, (with the prepaid chip cards), but later nobody used it and got vandalized and finally retired.
It's woth?
I understand these carry dsl signals, but I have no intention to spy on neighbouring builds,,,(I mean why if you can tell at first glance, you are just wasting time...)
Can I hack/eavesdropping something interesting?
Can you make phone calls?
It's damage into private propoerty?
It seems the technicians dont give a fuck about the state of the devices...So I starting to wonder
Also
Whats the nominal voltage leves on these terminals?It's funny when there is a blackout the old landlines are still up.
Thanks

>>2500026

Just made my first jlc order
Quickly whipped up 4 boards to order at once so the shipping isn’t too significant
Can’t wait for inevitable errors

>>2500000
Maybe, if you can source quality brand cells in bulk. You should test the cells in the pack you have and swap any bad cells out just to get it going again. Also, Harbor Freight spot welders suck ass. FYI

>>2500033
>Can’t wait for inevitable errors
That's the fun/suicidal part.

>>2500022
I'll do up a full schematic of the circuit.
I'm pretty sure they're BAV99s.
I pulled one off the board and get a forward drop of about 0.6V from pin 1 to 3 and from pin 3 to 2 and double that from 1 to 2.
Nothing in the reverse direction.

>>2500030
vielleicht, that's why I'm asking, but it will be funny to have something to do apart from freezing while going to work or at work, if it it worth.

>>2500052
The short answer:
You might be able to use the POTS landline with an old school telephone. DSL & VoIP are out of the question. Using the line as a power source is possible but only for low power stuff. If you end up knocking out service to customers, you're probably going to get fined and maybe even a short stay in jail.

>>2500022
Right, done up a full schematic of one of the current sensor circuits.
There's four of these on the board and pin 14 of the 4 ASICs are connected so I assume it's an enable pin.
Other two going to the MCU have a separate pin for each ASIC.

There's a cap in parallel with the resistor and double diode that I didn't have in the original schematic.

>>2500016
>What if I use a larger "antenna", say 1m in length?

still, too short.
you need 750 miles for a quarter-wave antenna.

>I've seen videos of people powering lightbulbs using capacitive coupling.

they use the power of the megahurtzs.

anybody knows whats the IC used here? burnt mine and its sanded lol.

>>2500187
MT3608 maybe

I'm making a symbol in Kicad. The IC itself is 14 pins (SN74ACT74 Dual D-Flip Flop) but only 8 of them are used.
Do I still need to add the other unused pins to the footprint? Should I?

>>2500216
>only 8 are used
that is, only 8 are used in the circuit that I'm designing.

>>2500219
Ground any unused inputs for sure.

>>2500216
If you're planning to make a pcb from it, you'll want to at least map it to the correct footprint.
Probably safer to just add all the pins to the symbol and then put no connect flags on the unused pins in your schematic

look mom elon is dropping /ohm/ memes

What do you guys specifcally understand by the phrase "digital development for electrification"? I'm applying for a grad program
at Caterpillar and want to clock in some relevant skills before I submit but this seems pretty vague. It's a mutlidisciplinary program soi it probably doesn't matter much but I still would like to have something that seems particularly useful.

>>2500216
Yeah, make then NC. If they have no internal connections then you can route traces through them if you're low on space.

I'm planning to install a 4 channel amp and an active spare tire sub in my car. I measured cable lengths and came up with pic related setup for powering the system. Anything I calculated wrong regarding cable gauges?

The part of the cable that will be in the engine bay will be protected by convoluted tubing, and the part travelling in the interior will be wrapped in cloth tape and zip-tied regularly to prevent movement. Anything I'm missing?

>>2500403
Pic related

>>2498152

Bumping this post in particular. I received the 5v regulator PCBs today. Haven't stress tested them yet but they've passed the most basic test (5v at output when 12v is at input). Once again - if the anon who designed the PCB for me reads this, please send me a message on EasyEDA so I can pay you.

I need to replace my keyboard cherry mx LEDs

I want the new ones to be as dim as possible. I've heard that I need "3mm flangeless" but I'm not finding good ones at AliExpress (low volume price or color options). Or 2x3x4mm. LED stands for "light emitting diodes", but some site say there's a difference between LED and diode. So how do I know which one is safe to buy?

what are these cute orange thingies? is that a rare find worth salvaging? some kind of ultra inductors/beads?

I'm playing with high power RF transmitters.
I have this application note reference design. The RF amplifier (the DRF1200) is driven by a 13.56 MHz clock signal with a specific pulse width, generated by the crystal oscillator and D flip flop.

Given that it's just a PWM signal, what's stopping me from replacing that subcircuit with a microcontroller? A microcontroller would allow me to implement a feedback loop for adaptive PWM to always have the optimal pulse width, and all kinds of other things to boot. Something tells me though, that if it WAS the correct/better way to do it, the circuit designer would have went with a microcontroller to begin with.
so in your guys' wisdom, would this be a viable alternative?

So I've got some arlo security cameras I'm not wanting to charge anymore. I bought a small 5v 170ma deer feeder solar panel hoping it would be able to charge them, but all the cameras end up throwing an error and not charging after a day or so. I think the variable voltage is triggering the error. Would putting a 5v zenor diode in the circuit fix that?

My thoughts are that the zenor diode would allow current to flow to the camera only if the voltage were above 5v.

If anyone has any resources to help understand electrical engineering, I'd appreciate it.

>>2500529
Non light emitting LEDs are used with keyboard matrices for N-key rollover, completely differently to the LEDs used to produce light. Though I guess you could use LEDs for N-key rollover, might look kinda neat.

>>2500536
Do they have reference designators, because I can't see any. Are they all in parallel like it looks from here? May be PTC resettable fuses.

>>2500547
13.56MHz is pretty high for a MCU to output as PWM at any decent resolution. Say, if the clock is 100MHz, then your smallest time step is 10ns, so the closest you'd get to that 13.56MHz would be 1/(10ns * 7) = 14.29MHz. You'd need to match the MCU clock to some easy multiple of 13.56MHz.
That said, you could manage it with some fast clock dividers and a PLL IC like the CD4046 if you can express the clock ratio as a relatively simple rational fraction. Or just use an Si5351 like everyone else.

>>2500565
Sounds like you want a series regulator, not a shunt regulator. Consider a simple 7805, or the AMS1117-5.0. Dropout may be an issue. A buck-boost, SEPIC, or inverting converter may be more convenient.

>>2500604
>Or just use an Si5351 like everyone else.
how have I not heard of this
the only thing is I can't find any info on if it has variable pulse width.

>13.56MHz is pretty high for a MCU to output as PWM at any decent resolution.
I've heard that the actual attainable GPIO pulse output is typically much lower than the highest nominal clock frequency capable of the MCU. However, I'm looking at the ESP32 which has a 240 MHz clock, AND a 64 bit timer. Unless I'm misunderstanding (I probably am), this seems to solve the resolution problem handily

>>2500606
>I've heard that the actual attainable GPIO pulse output is typically much lower than the highest nominal clock frequency capable of the MCU
That may well be the case for higher spec MCUs than the AVRs I'm used to, since they probably use internal PLLs for clock multiplication anyhow.

>>2500536
Could be mica capacitors.
Looks like they're all in parallel

It's 11/11 so the chinese has a lot of deals.
I haven't gotten around to buying a soldering iron station yet, which are the good chinese ones right now?

>>2500662
>BUY $100 GET $1 DISCOUNT!!!
yeah no thanks
the bugmen deals are trash

>>2500662
I want to know this as well

>>2500662
>>2500707
Hakko FX-888D
order direct (fuck scamazon)

>>2500054
>POTS landline with an old school telephone.
Yo, thats it, I have it clear.

> Using the line as a power source
That's what I was asking, how much can you get from it 5vXmA=mW, perhaps as a back up charging thing.
>end up knocking out service to customers
It is not worth the fuzz, I was expecting to get some gossip out there, nothing important,but I don't want to blew it out (first because my service comes from these box).
Perhaps its more effective to make a fake 3g/4g hotspot.

>>2500662
T12 station is pretty good, especially the ones with the OLED screen. The built-in PSUs are kinda sus though so go for one with a plug for an external power brick and get a more reputable 24V ~3A brick, or even just the barebones control board (get the nice thin blue handle though) that you can mount in a custom case.
The TS101 is also really good looking if you want a more portable iron, though even with 11.11 it might be cheaper to buy a Pinecil if you're in the USA.
For higher spec irons there are a few that use the JBC T245 tips like the Aixun T3A, plus the more modular control board kits, rather similar to the T12 but can pack 200W of power with the right PSU.

For really cheap irons those ones with the 936 tips and digital temperature control on the handle are probably decent. The ones with a dial on the side are often just TRIAC dimming with no actual temperature feedback. But cartridge tips are superior either way so I'd prefer the T12 or T245 tips.

>>2500801
>how much can you get from it
48VDC, usually 23mA. Not much.
>I was expecting to get some gossip out there
Are you the next Jeffrey Epstein? lel

>>2499249
Actually now that I think about it, I use "Battery Charge Limit" android app to limit charge to 80%, but then the battery life seemed a little short so I extended it to 85% and it got a lot better. So suppose you charge a fresh li-ion battery to 3.92V, and as its capacity drops you increase the charging voltage gradually up to 4.2V (or higher), keeping the approx same battery capacity during the battery's whole usable life. And once you're charging it to 4.2 and its capacity is too low then it's spent. Like a crusty 50yo hooker.

>>2500843
>48VDC, usually 23mA. Not much.

right, but not 48V and 23mA simultaneously. voltage drops with load.
think of it like a 48V battery in series with a 2400-ohm relay coil.
watch out for the ringing signal which is like 90Vrms. it provides painful shocks (happy face) and kills electronics (sad face).
but it wont ring if there's a load below some value of xxx ohms.

Not sure where to ask this. I'm interested in making super nintendo repros. Programming something like a 27c801 and replacing the maskrom on an snes cart with it. Pinouts are always different when working with snes carts though.

Some people have made programs that byte swap the rom so you don't have to physically rewire all the pins. Pic related, the binary data for address pins A19 and A17 are flipped. I want to modify the code to support different eeproms than the 27c801.

>>2501325
This is where I'm struggling because the source code isn't documented. The pictured code looks like it's simply moving around 65536 byte segments. Where can I find out which part of the rom corresponds to which data/address pin?

>>2501326
Gonna have to look into the memcpy function I think, and figure out what dst and src are being fed as. I'd guess that the memcpy function is pushing data onto a stack sequentially, or at least that it's summing a bunch of clipped/rounded values.

Maybe post on /mcg/.

I'm looking to make a solar powered device and decided to use 18650s. Am I correct that I need both a charge circuit (with overcharge protection) and a discharge circuit (with overdischarge protection)? or do the 18650s usually come with the protections built in?

I would use something like image related to charge it, but I'm not sure if I can just connect my device to b+ and b- directly or if I need discharge protection. Either way I'm thinking about using a zener to ensure a minimum battery output voltage is used, as I will be powering a microcontroller.

Do low power (<500W) synchronous motors with wound rotor exist? Or is it permanent magnets all the way to industrial scale?

>>2501347
TP4056
https://www.best-microcontroller-projects.com/tp4056.html

>>2501326
>The pictured code looks like it's simply moving around 65536 byte segments
Seems like it.
>Where can I find out which part of the rom corresponds to which data/address pin?
Depends on the architecture. It could be contiguous (same pins), seperate (separate pins) or virtual (same pins, but hardware does translation). and with same pins I mean same pins as ram.

Solar question:
I have access to a lot of used solar panels, and I've been planning on building an off-grid system for my house for some time, then, eventually, cutting my grid connection.
Trying to optimise for cost per panel connected to the system, well, more cost-per-watt, with free panels, fixed mounting costs per panel, what's the cheapest way to get energy out and elsewhere.
Lots of my equipment happily runs off 10-14v, but I'm flexible and would consider 48V

I use from 2 to 5 kWh a day, less than 1kWh overnight, and would want around 10kW of panels installed to give myself a reasonable safety margin during winter, able to top up batteries with the little light available.
Thus, a lot of fucking panels, and keep in mind the return per panel is maybe 200w if I'm lucky.

The problems I face is that while grid-tied or AC solar controllers seem more than capable of sucking in this power, off-grid ones usually only want to service a few hundred watts of panels max.
Secondly, my panels are a bit of a mixed bunch, even ones from the same original site vary quite wildly in their output..
I have tried hooking a few up to a Victron MPPT unit, and while it works at a small scale, many of these would be prohibitively expensive.

There is an insanely tempting option, going "full retard" as I call it, and connecting a cheap piece-of-shit PWM charge controller to every single panel, then feeding their outputs to a battery bank like a firehose, and in a small, dedicated shed I probably wouldn't worry, but on my roof, this kinda feels like a fire risk.
While there are losses from using PWM rather tham MPPT, because the cost to slap a panel on the roof is so low, $20 to get a hundred or so watts doesn't actually seem that bad.
Anyway, please point my ignorant ass in the right direction.
(I saw a youtuber do exactly what I'm doing a while back, but didn't see how they decided to finish it, and forgot their channel name...)

>>2501347
You can buy 18650s with discharge protection, but most do not. Look for "protected" cells. They cost more and are slightly longer because there is a PCB attached to the end of the cell.

>>2501405
nothing wrong with this approach really, especially if you have free/cheap panels. Just be sure to fuse each charger individually in case the minimum bid solder job shorts out. MPPT and high voltage is mainly a benefit when your panels are all well matched and get the same light, if you have old mismatched panels it will basically be limited to whatever the worst panel is.

You will need a lot more wire which is going to be expensive. That's the main reason 600v is used in professional installations.

>>2501418
>Just be sure to fuse each charger individually in case the minimum bid solder job shorts out
I was thinking of something silly like diodes, thanks anon, fuses make far more sense here.
It would be nice to be able to remotely isolate a panel from the controller just in case one goes up in smoke, even if isolated from the main output I wouldn't want a panel dumping 200 watts into a smoking charge controller for too long, but I can't think of any practical way of doing this, perhaps just run leads from each panel some distance away, then into some fireproof boxes that hold the charge controllers?
>You will need a lot more wire which is going to be expensive.
To make things worse I'm installing on a double story building, and the batteries are probably no less than 10 meters away, so I'm thinking lots of thick, high quality leads in parallel for minimum losses.
Can't see any way of cheating the physics on this one.

Anyway, thanks for your help and advice anon, appreciate it!

>>2501351
>>2501410
Cool this helps me a lot, thanks.

>>2496402
>/DIY/ this confuses and engages the roastie

why aren't cable connectors rigid as fuck? like solid unbendable material that doesn't stress the internal soldering and shit?

what's the best material to put between the cable and the connector?

I'm just fucking tired of headset cables breaking on me. I move them a lot and there's only so much care you can take

>>2496402
Just who designed these generic isolator boards and the IC. Ain't it stupid to have a speed switch? why would you want to slow your usb device down?
It's based on ADUM3160
Also, is there a reason to add an 817c optoisolator?

Is enameled solder wire a scam or I just suck?

I tried a hair thin one. It kinda works but it is a strugle to make it stick. It makes the solder very dirty and thus does not stick

I got a thicker one and not sure if my solder melts its coating, although havent tried soldering it yet. Just pressed my solder against it and it didnt work

>>2501326
>Where can I find out which part of the rom corresponds to which data/address pin?

gonna require some big think, but i dont think it's too hard.
- first, do everything in hex, so you're dealing with nice round numbers like 0xC0000 instead of disposable-human numbers like 786432
- make a table (16 columns by 2 rows) of addresses
- first column you enter the start of each 64K block in the 1 meg ROM
- second column you enter the same thing but where you binary flip A17 for A19, A18 for A16, A19 for A17, A16 for whatever you connected it to.
- that should give you all those same offsets as in the program

>>2501470
Because even when you're really careful they break. They're putting a lot of stress on the board.

If you push them too hard and they're too strong, then tge internal connector will break or the solder will rip off the board.

It's a design feature if it keeps breaking yet the jack is still usable.

>>2501509
Scrape it off with a razor blade or hot wire stripper first. Dump a bunch of Flux after your solder melts what it needs to

>>2501509
>Is enameled solder wire a scam or I just suck?

both.
there's 2 types of enameled wire: normal enamel wont burn off with an iron, and chink enamel which will -- sometimes.
anyway, the non-retard way to do this is to use wire-wrap wire instead.

>>2501519
Scraping it kinda defeats the purpose imo, at that point i can just cut a jumper wire

>>2501523
Yea the thicker one i have is the wrong kind i think. But even if it was the right one, not sure if i could make it stick because how dirty the solder become when enamel melts and is mixed to solder.

I will get another thick wire that says it solderable but not super hopeful

>>2501518
I mean the wiggle between the cable and the connector, not the connector and the board
should I epoxi that shit? kek

>>2501566
Pump it full of hot glue. lmao
or silicone sealant.

>>2501566
Epoxy is a decent method to prevent strain, but even then some tensile strain may make it through from the wires. A proper clamp or crimp over the insulation is the only real way to avoid tensile strain. Either way you still have sideways strain relief to think about, for which I’d cast the cable end in a cone of silicone, such that the thick part by the connector is rigid but the thin tail gets more easily bent. That way you prevent any one point being the focus of strain.

Can't believe I didn't fuck this up during soldering (yet). I am making a keyboard.

There are 6 buttons and I divided them to 6 groups of 10. All button outputs are connected together and goes to analog input of my MC (attiny44). And I power groups one by one and can figure out which button is pressed by reading the voltage value.

Right now I only soldered 2 groups. I am out of 220 ohms resistors

I also want to be able to hit 2 keys at once for some combinations. For example you will be able to shift + A etc. So the combinations I want should belong different groups (in this case shift and A should be in different)

>>2501679
there are 60 buttons*

also arduino is for programming & serial output. the keyboard will be controlled by attiny44.

my overall goal is making a small "computer". So I will make a screen and a computer case (where my main MC will be and communicate with other peripherals) and maybe a mouse but that sounds tricky

>>2501682
>maybe a mouse but that sounds tricky
what about a trackpoint module?

>>2501685
yeah, that is probably more realistic

Should I learn to use FPGAs? I'm an embedded dev at a small company. Before i die, i wish to design an ASICs.
I can design circuit boards and write firmware for micros but i know nothing of sillicon fabrication

jesus christ i have so much free time why can't i finish one fucking project?
i have been kicading a pcb and it was supposed to be a one day project to design and build and i am on day 10 and still haven't even started putting down traces, simply because i work on it for 2 minutes then click off to watch a youtube video or write this post and nothing fucking gets done.
I have like 5 other projects i want to do after this one and at this pace i will day of old age before getting to number 3.
Fucking procrastination. Internet was a mistake. I would get so much shit done if my brain didn't have access to infinity amount of easy dopamine at the click of a button.

>>2502005
Just gonna have to train yourself out of it. Working in the morning seems to be better for motivation I find, but there’s no perfect solution. A background podcast may help decrease the desire to watch something.
Do you have a job? Following a schedule for going to work definitely lowers my procrastination.

Also watch out for feature-creep.

>>2501962
I would recommend doing some FPGA dev because it is fun, good chance you never actually get to use it at the place you work. I recommend one of these: https://www.latticesemi.com/icestick.. And the open source icestorm toolchain there for these things. Kinda doubt asics will teach you much about silicon production though, asics are often just an fpga with the connections burnt in after all. For true silicon dev, I think you are gonna need a bit more study to go to a company that actually specializes in that, or a lot of money to do it yourself.

I want to provide some extra structural support to an USB port. Anyone here got any advice for any epoxies / other adhesives that work well for electronics applications like that? Everything I have seen so far is either really expensive (and only has a shelf live of a few months) or is conductive which makes it kinda inconvenient, or flexible, which kinda defeats the point. Would just normal superglue work?

>>2502022
Just put a short usb extender cable there and leave it there and use extender to connect your devices. Replace (cheap) extender if needed

>>2502033
Simply not an option in the space constrained design I am making.

>>2502034
How? a cable is a cable. Instead of connecting your cable directly now you are connecting your cable with an extender on it

>>2502038
You mean outside of the case of my product? that'd defeat the point, I the point is the mechanical stress when bumping into the connector, also, during transport, I want to have the entire thing disconnnected.

>>2496402
how do constant power and constant resistance modes of a commercial / high end electronic load work? obviously it's derived from ohm's law but i'm having trouble understanding how it's implemented with circuitry. in effect this is just P = V*I and R = V/I. but how is this actually done? i doubt it's with op amp multipliers or op amp dividers since that seems really arcane and complicated. i know the software implementation of this is trivial but i doubt high end electronic loads with fast transient response do this in software.

>>2502076
On modern electronic loads, it probably is software doing it. Software can be pretty fast for pretty cheap these days, and it is not like analog electronics is infinitely fast.

>>2502119
Though, there might be some application specific analog front end IC's out there that do it with some (probably opamp) logic.

What's the advantage of Hall sensor based current sensors vs a simple shunt resistor? They seem super expensive and not very accurate. Is it just safety?

>>2502145
I think the main thing is the output is isolated from the thing you are trying to measure. You could be measuring mains AC or whatever and reading it with a 5v microcontroller and not have to worry about it. Even with a shunt monitor there are usually pretty strict limits on what you can do (no negative voltage, no negative current, <XXv to ground, etc.)

>>2502076
Constant R is simple: instead of a fixed Vref, add a voltage divider so Vref will change with Vin. Your R is then Rshut times your voltage divider ratio.

>>2502165
will have to study this, my low IQ brain can't make immediate sense of it. thanks for the schematic.

>>2502120
Might be worth it for high-speed applications. Constant power would require multiplying the measured voltage by the measured current, for which you’d need a double-balanced DC-compensated mixer. Could probably pull it off with exp amps and log amps, but things like tempco and the like will mess with you.

>>2502165
>>2502191
the more i think about this, the more i'm inclined to believe that >>2502119 is right. the pain in the ass that is physically swapping out components as the design becomes more and more complex and the probability of control loop instabilities increases is off-putting.

software control loop with a fast ADC and bare basics analog hardware (comparator for overvoltage/overcurrent shutdown, good op amp for differential amplifier) seems to be the easiest route.

>>2502191
it is a slight modification of this standard building block with an error amp driving a mosfet or bjt and a sense resistor and feedback loop. but instead of a fixed voltage reference on V+ it varies based on Vin.

>>2502269
>software control loop
I don't think you need a software control loop necessarily. You can still have the same standard error op amp based feedback loop, you just need some logic in your software to change Vref according to your mode: if constant current, do nothing, if constant resistance or constant power, sense Vin and set Vref accordingly based on the Ohm's law.
dave covers all that in great deail:
https://www.youtube.com/watch?v=8xX2SVcItOA

>>2502269
Don’t forget over-temperature protection.
And yeah, doing PID in software is pretty easy.

>>2502308
I’ll be sure to watch that, doing constant power without some sort of analog multiplier sounds like magic.

>>2502398
>And yeah, doing PID in software is pretty easy.

i'm reading more about it and it's harder than it seems. i've read that one can't just blindly adhere to the nyquist theorem and assume ADC sample rate divided by 2 is the maximum system bandwidth, because there's a lot more overhead in the system. i found one source guessing 1/10th of ADC sample rate might be a better rule of thumb. you have to think about ADC sample and hold delay, ADC oversampling/filtering, PID loop calculation time, DAC settling time, etc.

GD32E503 has a pretty fast onboard ADC (12 bit, 2.5 MSPS, 0.4 us total conversion time @ 12 bits) and DAC (12 bit, 4 MSPS, 0.5 us settling time) and i can get them for like $5 each on LCSC. my goal is to build a relatively high performance electronic load that will work over a broad range of input voltages and commanded currents without requiring a shit ton of circuitry (or soldering and unsoldering) to ensure stability over this wide input range. i want to test switch mode power supplies, among other things, so having good bandwidth is important. it sounds like i might be able to get 100 khz out of a GD32E503 which should be fine.

>>2502022
Can you cram a panel-mount socket in instead? Then run wires from that to the PCB. That way all the stress is carried into the case, and no PCB contacts are stressed. THT USB A and B ports are usually strong enough in my experience, but if the thing plugged in is getting banged about those sharp shocks could damage the board.

just noticed ohm hero Ben Eater (now Jen Eater) is back on youtube after a year-long break to transition into a female.
she has an 8-day old video on a new and exciting communications protocol called AR-ES 232.
sounds fascinating. cant wait to try it out.

>>2496402
hi there
I need an ESP32 board or similar, something with bluetooth connectivity, working at 150ºC ambient temperature (actually it will be submerged in hot oil)
google isn't helping so far
any idea?

What is the best way to desolder SMD component without hot air station?

>>2502661
Oh god, look at that hot thing
I will blow your fuse very well, baby girl, if you catch my drift

>>2502688
hot plate method. if you can't do this, soldering iron with braid and a sucker.

how long does it take to charge 3 batteries with a BMS 3S 25A?

what's the difference between a current transformer and a voltage transformer? I always knew a transformer transforms both...

>>2502661
>Jen
Still better than Men Eater

Does anyone recognise this manufacturer?
Can't find the marking in the libraries I know.

>>2502822
current transformer is a horrible misnomer.
"transformation" means a totally different thing in this case. it is more like a current sensor. or you could call it a transducer i guess.

guys what is this component in a dimmer switch? It shows no continuity or resistance

>>2502967
Diode.

>>2502967
>what is this component

best guess is it's a diac.
it's like two back-to-back zeners, so it wont pass any current in either direction until you hit 20V or whatever.

>>2502981
>>2502985
thanks a lot!

>>2502736
1 hour per battery

>>2502736
How many amp hours are the batteries? If they’re 18650s then 25A charge current sounds way too high.

>>2502822
Current transformers are put in series with a load, and have an insignificant voltage drop due to their very low primary inductance. So the magnetic field within is largely proportional to current. In a conventional transformer, you put a significant voltage across the transformer, so while the magnetic field is still proportional to current, that current is now mainly a function of the voltage, inductive reactance, and the load on the other side of the transformer.

>>2502985
Doesn’t a DIAC actually act more like a neon, with a negative resistance section of its V/I curve?

>>2496402
Just got a Sony DSX-A410BT because of >muh AM, short wave, etc. but the sytem comes with the same AM/FM, roaming over there I just found in the russian site an firmware update enabling the short wave and all that.
It is safe enough to try to update to the short wave version?
I understand the boards and components must be basically the same, and if the unit does not work I guess it can be downgrade/upgrade to a working version.
Is this assumption safe?
Thanks for your kind comments.

How do all the electronics youtubers make all those nice solder bridges on perf board?
It seems so effortless, it doesn't seem like they're doing anything special. But when I try that shit it's a real fight to get 2 neighboring pads to connect. Do they use some special solder or something?

>>2503380
>Do they use

soldering is 100x easier when you cover all your bases.
clean copper traces + quality leaded solder + good quality clean tip + eye-hand coordination + a bit experience

>>2503394
Your intention is good but that incredibly general advice doesn't help me here.
Soldering individual pads is easy. I've done enough soldering that I can get good joints with no effort at all. But when I want to connect 2 pads with a solder bridge it just doesn't work. The solder just wants to stay over the pads, not bridge the gap across them.

>>2503380
>>2503399
It's harder to bridge 2.54mm pitch pads without a jumper. Use 0-Ohm SMD resistors.

>>2503415
Idk m8 this guy makes it look effortless.
For reference I have the same style of perf board.

>>2503426
>makes it look effortless
That comes with a lot of practice. Still better to use jumpers or even a single copper strand.

>>2503380
Use tinned uninsulated copper wire for the traces instead

>>2503399
>The solder just wants to stay over the pads, not bridge the gap across them
That's a sign that the iron is too hot. If it's too hot the solder will be really fluid and will flow back around the pads.
If it's cooler then the solder will be thicker and won't flow well so it will stay bridged

>>2503431
>That comes with a lot of practice
This is non information. He first needs to learn the proper technique, then he can practice.
Webm at >>2503426 shows the technique but it's easy to miss if you don't know what to look for.
Basically you don't add solder to the pads or the iron, you place the solder wire in the gap and bring the iron over it to melt it.

>>2503445
>He first needs to learn the proper technique
Solder blobs all over isn't proper technique. Waste of time compared to just using wire, unless there are only a couple to do. This guy jerks off with flux for the home audience.

>>2503449
I just told that guy how to do what he wanted to do. The choice of using solder bridges vs wires is up to him, not me.

>>2503456
fair enough

Is it safe to use brand new resistor as toothpick? I just realized I might have been poisoning myself with lead from solder. Are the unused one dipped in solder too?

>>2503476
I knew a guy who ate lead paint chips as a kid. That guy went on to become the 46th president of murica.
t. cornpop

>>2503476
Chinky resistors may well have leaded plating.

>>2503476
Probably but stop wanting to do stupid or silly things.

>>2503896
>stop wanting to do stupid or silly things.
guess i should just kms

Is this the place to ask about home wiring and similar topics? Asking for a friend.

>>2503931
Did you read the OP? It says:
>household/premises wiring?
More rules-driven than engineering, try /qtddtot/ or sparky general first

>>2503933
Thank you anon.

Working on a Halloween prop. It runs on batteries and I want to convert it to run off a wall wart. Unit uses 4 AA batteries.
Battery pack voltage is actually split: the unit takes voltage from 3 AA batteries (4.5V) to run electronics, and all 4 (6V) to run the motors.
My thought is to get a 6V ~1A wall wart, step down voltage to 5V with an LM7805, run the electronics off the regulated supply and the motors off the full 6V.
Anything I'm missing here? Assume any electronics that can take 4.5V battery can take 5V... Or perhaps should just run it off a 5V pack and be done with it.

>>2504054
Use a 5V supply and you'll be fine.

>>2504056
It's funny b/c as I typed it out, I'm thinking why wouldn't I just use a 5V supply. The 4 batteries is probably just there to ensure there's enough current to maintain voltage.

>>2504058
Yep. You could test it with an adjustable PSU so you can find the cutoff voltage and current draw. Then you'd know the minimum operating parameters.

>>2504062
Yeah, I think I’ve taken this as far as it probably needs to go. If the 5v supply doesn’t work I’ll go plan B.
Whole thing is a plastic and hot glue nightmare. Main work is Adapting over metal gear drive motors to replace this abomination in two spots on model.

>>2504068
Are you replacing the gear/motor assembly entirely?

>>2504069
Yep. The plastic motor on left is this weird belt and gear driven thing. The plastic gears inside have stripped off the shafts... irreparable.
Motor on left is basically same voltage and RPM, but with metal internals. Output shaft is close enough to direct replace but if it wasn't I'd just 3D print a new one.
I'm a MechE by training and just know enough electronics to do simple stuff like this. 3D print has really enabled a lot more sophisticated repair work on these sorts of things. I'd be playing around with washer and screws and would never get the alignment quite right without ability to create mid-accuracy (0.01") parts.

>>2504076
> motor on right is metal
What I don't know is if torque output is same, but the old one was so crappy I figure anything is better. Ofc these Chinese motors are absent any spec's. But they are inexpensive.

>>2504076
Having a 3D printer removes a lot of hurdles. Nothing worse than (unknowingly) searching the internets for shit that doesn't exist anymore or non-standard dimensions.
>>2504077
Just try to keep the physical size of the motor similar to what you have and torque shouldn't be an issue. Plus, you'll be driving it at 5V so RPM will be lower.

I need an asynchronous motor for an application. I keep coming across the YJ58 but the RPMs are too high. Is there a slower alternative that anyone knows off the bat?

>>2504231
>YJ58
What RPMs do you need? That type of motor is somewhat limited just bc of its design.

>>2503399
>But when I want to connect 2 pads with a solder bridge it just doesn't work.
you have to let it cool a bit, surface tension fucks with the bridge and you need a bit of a heat gradient

>not using strip/vero board and jumpers
ngmi

>>2504256
thanks for the tip, im new to motors. I would not be opposed to dc a motor, I could run a PWM.
I just need a motor that withstand running 24/7, I was hoping for a small fan motor type @ 1000rpm. Figured motors like the YJ58 fit the bill, if I could lower the RPM I would be set

Desperately looking for an extremely long push/pull solenoid or something that works similarly. It needs:
>4-8" of travel
>relatively thin (can't use a wheel to actuate)
>VERY fast, like 6 ft/s
>at least 4 pounds of force
>easily tuned to stop at a specific point
Does such a thing exist, or am I going to need a drastic redesign?

>>2504416

doubt you'll find an electrical solution.
sounds more like a job for pneumatics.
https://www.youtube.com/watch?v=iXNNjtK5hp8

>>2504484
>>2504484
>>2504484
NEW THREAD

are there any safe open source defibrillators?

Is it safe to directly connect two audio sources? My speakers only accept minijack input.
I keep switching them between headphones port of my monitor to get sound extracted from HDMI of PS4 and headphones output of my laptop.
I own one of the three-way minijack connector. Is it safe to use it to combine computer and ps4 sound lines?

>>2505355
see new thread: >>2505399