Thread was put backwards and exploded: >>2920495

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

>OP source:
https://github.com/74HC14/ohmOP
bake at page 10, post in old thread

>Comprehensive list of electronics resources:
https://github.com/kitspace/awesome-electronics

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

>Books:
https://libgen.is/

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

>Recommended software tools:
KiCAD 6+
Circuitmaker
Logisim Evolution

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

>Most relevant YouTube channels:
EEVblog
W2AEW
Moritz Klein

>microcontroller specific problems?
>>>/diy/mcg
>I have junk, what do?
Shitcan it
>consumer product support or PC building?
>>>/g/
>household/premises wiring?
More rules-driven than engineering, try /qtddtot/ or sparky general first
>antigravity and/or overunity?
Go away

>>2927795 (OP)
>storm in parent's town
>solar installer did a poor job, 2 panels cracked broke because the grout was like 10mm deep and one of the corners came loose because of the wind
>they replaced the two panels with new ones
>now we have 2 panels (48V 400W) which have cracked glass but are otherwise working perfectly
I want to make a 98VDC to 300VAC inverter. It shouldn't be too hard to build, I just need to get a nice EE core and wind a 1:3 transformer and build a forward converter with it and then have an H-bridge stage to make AC out of it.
I will start working on this. Gods willing, I will finish it before this thread dies

Just playing with the idea of driving the base with just a suitable amount of current to keep the V(ce) relatively low. But not sure how to measure the V(ce) because the reference would be at +24 V instead of 0 V. Maybe a differential amplifier could be used here. And some kind of comparator circuit(s) controlling the base drive switches.

>>2927798
> a 98VDC to 300VAC
How are you going to make a sine wave out of 98 VDC? Resonant oscillator? DC motor turning dynamo? Is it going to run at 50/60 Hz?
Why 300 VAC instead of 100/120/240 VAC? (Japan, NA, rest of world)
Have you considered 3 phase output?

t. conceptualizing similar project

How does
>picrel
current sensor work?
There is 6 pins and at least two coils (besides the core itself). Looks like one bobbin had ome coil on it and the other bobbin has 4 connections.
There seems to be no IC, no sensor or anything in there.
If it is a lame ass current transducer only, why the 6 pins? Some sort of calibration feature? How would it be connected and used? Am I right that this will only react to AC?
The manufaturer produces a current sensor under the same model name, but it is of the 3 pin variety and the manufacturer makes it clear it's hall effect. So they must have updated the product line and just kept the name while changing the technology entirely?

>>2927939
> the base drive switches
Relays?
I’ll be honest, I love the satisfying star-trek-like clicks my my oscilloscope makes when you change ranges.
Interested to see what the current limiters look like, but at 100 mA relays should last 100s of years.

>>2927946
> why the 6 pins
Two for each coil and two for the current-to-be-tested path

I'm >>2927946
I think I found my answer. The working principle is probably called 'flux gate' and the external circuitry required kind of wants me not to whip something up real quick.
Can someone recommend a sensor? I have 1 conductor. Theres a signal in the range of 50-100A, it's got a very steep rising edge of less than 5 nS and the falking edge varies but is generally slower and seems to bounce a bit. I expect no DC component but there can be a fault where a DC component will appear and it would be nice to catch that.

>>2927939
Are you saying you want to avoid saturation? You can just clamp that with a diode, you know?

>>2927950
Nope since the current to be tested obviously runs through a conductor that is external to the device and fed through the laminated loop. But I could imagine that one bobbin has two coils for the flux gate method and the remaining longer coil is for calibration purposes - so yes a test current if you will - but I suspect the long coil could be for the AC component and the two other coils for the fkux gate stuff.
I should mention its rated for 50 A nominal.

>>2927798
With or without a battery? Having no battery in between means the minute-long fluctuations in sunlight are going to influence the maximum power output significantly. Unless you’re powering some sort of load that you can tell to draw less power, or a load that draws way less than 800W, it will be a difficult task. I’d suggest trying to seal up the cracks, permanently mount them to your roof, and try to use them for a specific purpose that doesn’t need an inverter, like hot water heating. You could even build a water jacket behind the panels to keep the panels cooler and longer lasting, and to preheat the water before you electrically heat it. Powering an air-conditioning compressor might be an interesting task that doesn’t need a battery also. Swap the motor for a BLDC with an appropriate Kv rating.

Either way you’ll also want maximum-power-point-tracking capability. The smart way to do that is to vary the PV current to map out the V/I curve every now and then, but a simpler way is to just set a maximum-power-voltage and use feedback to keep it at that, which gets you like 90% of the way there and can be done via analogue feedback loop. If you’re just powering a heater with it, you can use a bunch of series and parallel diodes as your heating element and their V/I curve WC an be made to get maximum power out of the panels.

>>2927939
Beware, transistor current gain varies with temperature.

>>2927952
A regular current-transformer should be fine for the AC stuff, so long as the magnetic core is properly rated for all of the high frequency components (use ferrite). Chuck a DC current clamp around it as well if you want to catch a fault threshold.

Assuming you can’t just shove a 100A shunt in there.

Can I just pull 5v out of usb-c connector like this without any resistors?

>>2927958
I've discovered that my shunts behave unacceptably inductive. I am aware that shunts exist that do not but they are rare, expensive and in the end I feel an indirect measurement might be better all around.
I've looked at some chip like hall sensors, datasheets said the bandwith is 120kHz plus minus.
Why do you suggest I use a core at all? I was thinking if I would have the AC only I might just go with a coil around my conductor, slap an OP on there and see what I get. Cant I get the same a core would give me with just N turns in a coil. It should give a 1:N transformer. Without any saturation issues. But for the time it takes to whip it up and get it to read accurately I might aswell buy some integrated shit. Shit like picrel.

>>2927973
Yes. I think the resistors are superfluous for 500mA or less power draw. Though most phone chargers will provide more than that current with no resistors, as far as I understand.

>>2927981
How the hell is a shunt inductive? It’s just a piece of straight metal with a known resistance, pic related. Heck, you could even just use the wire itself as a current sense resistor, though if you care about absolute accuracy you’ll need to compensate for the temperature of the wire.

A core is used for a current transformer because of the way magnetic flux flows. With the right-hand-grip rule, magnetic flux wraps around the conductor, it does not move coaxially to the conductor. If you wrapped thin wire around the conductor, the magnetic flux would be wrapping around alongside the thin wire, while the thin wire would only be able to pick up magnetic flux running down the conductor. Those are orthogonal, you wouldn’t sense anything. If you put a toroidal core around your conductor, the magnetic flux from the conductor will wrap around inside that toroid. Thin wire wrapped around this toroid would perfectly pick up that magnetic field.

Ok I have a square wave oscillator at about 1kHz, made from a 74HC14 inverter, with on-time and off-time pretty equal to each other. The wave at the inverter's input is about 1V peak-to-peak, oscillating between 2V and 3V as per the schmitt hysteresis. The output is 0V-5V, as a CMOS part operating on a 5V rail.

But I want to have a signal line that stretches the off-time to be significantly larger without influencing the on-time at all. So far I can think of using a FET or analogue switch (or part of a CD4007) to switch a resistor in parallel with the main timing resistor, with a bypass diode in antiparallel with the switch if there isn't a body-diode there already. But needing a new IC for this sounds like a pain (though i may be able to use the inverters of the 4007) and the 5V rail doesn't give me much headroom for turning on a normal 3-terminal MOSFET or a JFET. Pic related. I might get better switching if I put the NCH and PCH fets in this 4007 stage in parallel, i.e. connect pins 1-4, 2-5.

Trying to modify the capacitance to ground would be easier to do with a transistor due to the ground rail, but this seems impossible to use since any extra capacitance or resistance I add will be equally effect both the rising and falling time constants.

>>2928020
The way I usually do this:
Non inverting schmitt trigger followed by integrator feeds back into the schmitt trigger and also feeds into output comparator.
You set the frequency by adjusting the resistor into your integrator (use pot if you like) and the inverting input of your output comparator comes off a resistive divider to set the duty cycle.
In case my desciption sucks or the principle isnt clear:
The schmitt trigger is either 1 or 0. As such the integrator will creep up or down. Since its output feeds back into your schmitt trigger it will eventually flip it. This will then cause the integrators output to creep into the opposite direction and this repeats ad infinitum. This is how you create an adjustable triangular wave using only one dual comparator for example. If you now feed your triangular wave into a comparator and set its inverting input using a resistive divider you get to set the duty cycle.
Use comparators, 1kHz is easy peasy, two LM311 is all you need and you'll have plenty of room to GO FAST. You'll also have 1 comparator left for other purposes. Do not use OPs, the slew rate is always a pain in the ass. Stay away from 555 shit.
Picrel shows one such thing I whipped up real fast. So all you really need is comparators, a few diodes, resistors and a small cap. Two pots if you want to adjust frequency AND duty. Ignore the added on push pull output stage.

>>2927985
Are you really asking how the shunt youre showing is inductive. You gave the answer. I tried. Again: We're going from 0 to 50+A in a few nanoseconds or faster. I need to go to uni soon to get better time resolution than I get at home. I don't have a wire either only transmission line (impedance controlled trace, BNC, coax, BNC business end aka spark gap) and I was hoping I could get the shield go on the outside and the core on the inside of a split core sensor. That would still result in some complex impedance but I am just hoping it wont be that bad.
Regarding the magnetic flux: Yeah obviously but I'm manly concerned with the AC component, as I've mentioned. So I believe it should be possible to construct a transformer like that. I'd still simply prefer a known good solution with no guesswork or calibration.

>>2928034
Like so. One of the first things an image search will show.

>>2928035
Sorry. for spamming ITT but this discussion made me think and yes it turns out they do make coaxial shunts for exactly my application and for exactly the reason that >>2927985 this shunt is inductive af and >>2927981 those things have a BW of 120 kHz.
Guess I'm breaking the bank for a piece of copper again.

>>2928033
How is that duty-cycle changed by a signal line? Regardless of whether the Schmitt trigger is switching based off a nice integrated ramp triangle-wave, or off a curvy RC wave, your duty-cycle is controlled by a resistance. I want a 0V/5V signal to change the on-time only, and the only way I can see to do this is to switch a resistor in parallel with tan existing resistor. Attenuating the output voltage to not go all the way to one of the rails would work too, but I can’t see an easy way of doing that.

Also integrators rely on negative feedback, they should use op-amps not comparators. Schmitt triggers rely on positive feedback, so comparators are a good idea, though CMOS logic probably gives an output that’s closer to the rails than a cheap bipolar comparator.

>>2928038
Just bite the bullet and buy a precision mini metal lathe like clickspring.

>>2928057
I do have a clapped out large lathe, a second operation lathe and a toolroom lathe.
But manufacturing capability is not the determinibg factors when you make components. By the time you'll have bought half of the test equipment needed to verify your products you'll be begging your local loanshark to spare your life.

>>2928034
Oh I didn’t read the part where you said it was a near GHz level signal. You actually want to measure that high-speed pulse, and allow it to travel through these wires instead of just buffering it at the load with a bunch of expensive capacitance? Microwave engineering applies, here be dragons, etc.

>>2928035
I have no idea what that’s meant to do, but I don’t think it would do anything. Source?

>>2928038
Looks like your high-frequency impulse will just capacitively couple between those walls instead of flowing coaxially. Check the spec sheet of one of those for a frequency response or corner frequency or something. If you had a layer of shielding between the coaxial conductors there it would probably work at high frequencies, but you’d want to match the impedances properly.

Is there anything stopping you from just measuring the voltage between two points on your 50A conductor? You’ll see can calibrate out the tempco of the copper later. Even if it’s some sort of high-power coax cable for minimal self-inductance, it should be possible to measure the voltage drop across the shield alone. I think the tactic is to run the sense wires alongside the main conductor to the middle of the sensed length, then lead them out while twisted together. I think you’ll still get a delay between the voltage going up on one side of the shunt compared to the other just due to the speed of propagation through the copper and the surrounding electric field. If that’s a problem, run a wire to a third equally-spaced point along the conductor and do some math on the two waveforms.

>>2928060
Yes I want or 'need' to quantify current that discharges in a spark gap. Ionization happens fast and once it did the gap is expected to behave like 1R parallel to some C.
My last job was in RF, so thats good.
What it's meant to do? Its a 1 wire transformer. Think of it as a wire running parallel to another because that is what it is. Only that the field has a different leverage because of angle introduced between the wires.
I will avoid buying components without reading the datasheet. I wont rely on it but it's relatively evident that the whole idea is that the shunt behaves like a transmission line, as in R+jX where X is very close to ± zero. The ones I found are good for dI/dT of 10^12 A/s so it has the bandwidth.
>If you had a layer of shielding between...
I have no idea what you mean here. Also it sounds like youre implying you expect something to not work? What is it you expect to not work?
And naturally I'm not running my shit mismatched. But that wouldn't affect my measurements. A mismatch will have zero influence on the voltage drop across the shunt. The reason why I match things is because I dont want to loose expensive dB that I had so much trouble getting in the first place. Don't care about signal integrity tho.
Nothing is stopping me from measuring the voltage drop across the transmission line or two arbitary points on it. The problem might rather be that certifying an arbitary length of coax for its resistance might be a bit overboard and might not cary over to another piece of coax too well. The datasheet might say something completely useless like 'max. 40R / km'. Too much uncertainty there, even if you went through all the trouble of actually rating it. A piece of coax is just not a resistive element. You'd also introduce problems when you find the length of conductor is significantly longer than your edge. I rather throw out the money to get the proper device and save all the time and effort and later discover it causes more problems.

>>2928020
I thought about your issue a bit more and maybe my last input is a bit lacking if you want to remain purely digital, 0 or 5, so no triangles, or want ro stick with the 74 family.
Have you considered driving a ripple counter (74 series has at least a decade ripple counter if not more options) at a higher frequency and diode or-ing outputs together to give you the needed pulsewidth? You wont be able to co tinously vary pulsewidth but thats the plight of the digital realm.

>>2927943
There are 2 parts to this.
First is the forward converter part of the circuit which will turn the 98VDC to 300VDC.
Second, for making the 300VDC or so into 240VAC, I will use an H-bridge stage with an appropriate LC filter to make it into a nice sine like wave. I will use those 600V 1uF "induction stovetop" caps. If possible, make the H-bridge SPWM to make it even more sine like.
300V to 340V is just the peak voltage in 210VAC to 240VAC (these are RMS voltages). No plan for 3 phase output, doesn't make sense for 800W of power honestly.

>>2927958
I don't have a plan for a battery, just a large capacitor bank, a few mF to maintain the voltage rail in case of heavy transients. If I did use a battery, I would have to put the two 48V panels in parallel to match with a 48V li-ion battery pack (12S or 13S). That will make the entire system be limited to 48V. Transformer will need to be 48V to 300V, a worse ratio, primary winding will need to be thicker to carry more current, switching FET will need to be even lower Rds to prevent higher losses (or even worse, I may have to put multiple FETs in parallel). Everything just falls apart.
I do agree, I should probably look into filling the cracks in the panels though, no harm in making the "free" panels last longer. Probably potting resin or something similar? I have no idea.
I am planning on making this project MCU based so MPPT will be a software feature. Anyway, I'll be using good old TL431 + 817 optocoupler to get feedback from the secondary side.
I would use DC but its not adventurous enough for me.

I am really excited. Its been quite some time since I did some serious power electronics. I feel manly already. I finally get to wind a big boy transformer

>>2928063
>Think of it as a wire running parallel to another because that is what it is
It's definitely not parallel, but you're saying there's a non-trivial parallel component because of the angle of the winding? I've never heard of that kind of current-transformer before, is there a name for it I can look up? Intuition tells me that a single layer of helical windings going down length L of the main conductor is no better than a single secondary "winding" wire that's fully parallel to the main conductor for length L, but I wouldn't know for sure.

>I have no idea what you mean here
Pic related. Turning a coaxial resistor into a triaxial one. No clue if it would improve anything though, it's just my intuition.
>What is it you expect to not work?
The narrow clearance between the coaxial faces in the image of the coaxial current shunt you posted will have significant capacitance at the impulse frequencies you talk about. Say they have an area of 5cmx10cm and a spacing of 1mm. That's 44pF, which at 0.5GHz is 7.2 ohms. I don't know how many significant figures you're trying to measure, but you may well find this RC constant to be smoothing out the sharp edges of the triggering spark-gap. Doubly so since the round-trip distance through the resistor is almost a nanosecond longer than the short capacitively-coupled distance.

>The problem might rather be that certifying an arbitary length of coax for its resistance might be a bit overboard
Did you know that you can just put a known current through the wire and measure the voltage? Could even build that into your machine.
>You'd also introduce problems when you find the length of conductor is significantly longer than your edge
Yeah ideally you'd have whatever current sensing resistor be as short as possible, but the smaller it is the more heat will build up in it, or the higher gain you'll have to put in your sense amplifier. Below 1uV absolute accuracy and GHz speeds means really expensive ICs.

Bergoz makes fast CTs

>>2927554
anon is it by any chance an online course? i may be interested

>>2928065
Counters get big and ugly pretty quickly, and the time difference I want to add/subtract from the on/off-time isn't an even multiple of anything nearby. I don't see why integrators and triangles would help, but I am trying to solve this with analogue methods. My oscillator is basically just a simpler version of the integrator oscillator you already mentioned, and any methods for stretching on/off-times apply equally. I need 560us normally, but 1640us immediately when a digital output goes high (that's a ratio of 41:14, though maybe 3:1 is close enough...). That digital output does go high synchronously with this counter. Both the frequency and the duty-cycle are changing, so a purely digital solution wouldn't be easy.

>>2928073
>forward converter
Why not a push-pull or even an H-bridge transformer driver?

>I would have to put the two 48V panels in parallel to match with a 48V li-ion battery pack
You could piece together a 96V battery. Or even use a rotary phase converter to turn solar DC into AC, since that gives you some flywheel energy storage. Parallel FETs isn't the end of the world anyhow, they're in the ohmic region when in saturation so no need to worry about thermal imbalance. The primary windings for a lower primary voltage would be thicker but fewer, ultimately taking up the same volume.

>I am planning on making this project MCU based
I would recommend having an analogue feedback loop (e.g. TL494) and use the MCU to control the voltage/current reference DACs, as opposed to using the MCU to directly tell the transistor(s) when to turn on and off. You don't have to keep it isolated, so you could ditch the optocoupler for direct linear feedback.

>I am really excited. Its been quite some time since I did some serious power electronics
Go for GaN FETs if you want to try something new.

>>2928063
Just put a rogowski around the spark gap, it's easy to make yourself and way cheaper than any high frequency CVR

>>2928090
Not him but I've never heard of those. Now that I look at what they look like, I realise we had handheld ones at our shop for years sold as 3000A AC clamp meters. Explains why the loop was so flexible, it wasn't made of some sort of ferrite-filled elastomer after all. Only rated between 50 and 400Hz though.

what does Input Common Mode Voltage Range even mean? the AI mentioned this as a potential caveat. the incoming audio could potentially be driven to +24dBu or around 12V. the min (V-) -0.1, (V+) -3.5 V spec reads like gibberish to me like of course it would accept signals with a +3.5V peak and the input doesn't have to be -0.1V?

>NJM072
>min +/- 10 V

>NJM5532
>min +/-12, typ +/-13 V

>OPA1612
>min (V-) +2, max (V+) -2 V

>OPA1642
>min (V-) -0.1, (V+) -3.5 V

>>2928092
>>2928093
>>eatmyass
If you power an op-amp from +/-15V, like a lot of datasheets specify at the top/bottom of the DC electrical characteristics table, then the op-amp will only function properly with inputs somewhat within those rails. If it's a rail-to-rail input op-amp, then it can basically go all the way, and there are some special op-amps that can go beyond, but generally there's a volt or two dropout. A value of +/-13V means it can go within 2V of the lower rail and 2V of the upper rail when powered from +/-15V, and it's pretty safe to assume that 2V of dropout will remain roughly constant at other voltage rail sizes. Some datasheets even have a graph of this. Like this misspelt one from my Philips NE5534 datasheet. These graphs are kinda useless because of the zoom and coarse gradations, it would be better if they graphed the difference between the rail and input voltage (e.g. 2V) instead of the total voltage.

Some op-amps have an asymmetry on this feature, the OPA1642 looks to be capable of measuring 0.1V below it's negative voltage rail, while has to be 3.5V below its positive rail. At the specified voltage.

>>2928094
ok i got it, so at +15V minus 3.5V, the acceptable input is up to +11.5V. that's a whopping 0.37dB less than if it were +12V (the minimum spec of NJM5532).

now how do i modify the stereo link of drawmer DL241

the original implementation is a bit cheap because it just uses the INBUFF audio input from channel 1 to control the gain reduction of both channels. it's convenient in a way so that you only need to turn the knobs of channel 1 without the channel 2 causing a mismatch between the stereo channels. but it's not detecting the peaks that are exclusively in channel 2. the INBUFF of each channel can't simply be summed because you want just the peak value to be detected, so the correlated signals would be detected by a disproportionate amount compared to a peak in just one of the channels. i would want something like an analog equivalent of an OR gate, either to the INBUFF audio signals (but then i'd want the negative peaks as well) or allow each channel to run separately but then the control voltages would be combined with some diode setup.

lol this is a bit creepy, i had already found the diode "analog OR" but the AI is pretty much on point

>>2928083
Yes its what I said, it acts like a wire thats parallel. You said it sourself now.
I dont really see why I would need my shunt to be triaxial. I consider this a solved issue since I looked found out there was non-inductive coaxial shunt. It was your comment that got me thinking about shunts again and wondering if such thing exists. Thank you.
Naturally there must be a capacitance it's the only way how they arrive as a shunt with neglectible imaginary component. The manufacturer designed it to be so, did their math, verified their result and rated their devices. I simply rely on their datasheet and should I find a device is not to soec it goes right back to where it came from. I won't second guess their work as long as I dont have to. I'm pretty sure your exemplary dimensions are far of, do you mean 50 mm for the part that is the circumference? So your hypothetical central conductor was about 1 cm in diameter? Like I said I wont do the work someone else already put into it. I need to allocate my time to other questions that have no known answers yet.
>Run current measure voltage.
It's not like the ohmic law was news to me. Can you please not always imply I was about to just bux components without reading datasheets, or the manufacturer of components got thwir shit all wrong, or people didnt know about Ohms law and that? Measurement is only that simple in theory. Please read about measurement and a chain of proof and all that yourself if interested. I'm not lecturing.
>crazy fast
an OP like that would still be less of an issue than getting a tevhnology where the bandwidth typically seemy to be 120 kHz to work in my favour.
>>2928088
Sorry its no use. But the triangular wave, aka integral, works in yiur favour through a principle you're already employing. I'm just suggesting a method using 3 comparators to allow independent and continous control over frequency and period. I do admit it violates your constraint of great simplicity, just like the counter.

>>2928090
Thanks I never considered checking right at the gap. I'll look into it but the spark hap doesnt exist in 'a vacuum' there's other things going on there, which may be why I never considered putting anything there but maybe I can conceptualize something.

yo!!! is COMP_OUT (bottom right) already fitted with the appropriate diode for the "analog OR gate" (top)? i'm not sure how it works because the D18 diode before COMP_OUT is also connected as feedback to the op-amp. does the altered wiring through the switch (bottom left) look plausible?

>>2928155
wait i see that the resistor ground in the altered wiring isn't right but other than that??

>>2928156
or is it right?

>>2928157
COMP_CV already has the resistor to -15V so i guess no resistor needs to be added, just connect the CV from each channel together?

ok holy shit, COMP_OUT has a precision rectifier aka super diode, so i guess it will work to just link them together

>>2928155
yes, that will work for analog OR. the only requirement is that it can source current but not sink it (or vice versa, as long as the other input is the same).
keep in mind that the diode, whenever it’s reverse biased, will prevent the op amp from equalizing its inputs like op amps are normally supposed to, which will impact the “speed” of the op amp because it’ll slam against the rails and saturate, and potentially damage it if its not designed to have its inputs vary that much. in the datasheet, look for a graph called “overdrive recovery” and a figure called “maximum differential-mode input”.

>>2928182
hmm maybe that's another reason why the stereo linking isn't done like that from the factory, besides the convenience of the controls being synced off of just one set of knobs. maybe i could convert it into a "super duper" diode to avoid saturating the op-amp.

https://www.swtest.org/swtw_library/2013proc/PDF/SWTW13-22.pdf
Why is this not used for prototyping in the era of accessible and accurate 2D/3D machines? What are the downsides? I'm sick of waiting for custom PCBs

wtf i love AI now
it insists that the control voltage should be buffered with a voltage follower

it also insists that the diodes should be in addition to the super diodes. but now it mentioned comparators or op-amps configured as maximum selectors which i might be more interested in than the diodes.

>Which is best for your DL241 modification?

>For combining control voltages from precision rectifiers for a compressor's stereo link:

>The Precision Diode OR Gate (method 2) is generally the most suitable. It's a classic and highly effective analog circuit for selecting the maximum of two continuous analog voltages, precisely, and without the contention issues of simply tying outputs together. It effectively creates a "precision maximum" detector.

>The comparator-based approach (method 3) is usually overkill and adds unnecessary complexity and potential artifacts for this specific application, where you need the instantaneous higher value of two continuously varying control voltages.

>So, when I mentioned "active OR gate" earlier, the precision diode OR gate is often the intended meaning in an analog circuit context where you want to overcome diode voltage drops. You are indeed still using diodes, but the op-amps work with them to create a precise "maximum" function.

>>2928073
an H-bridge stage with an appropriate LC filter to make it into a nice sine like wave
> just a large capacitor bank, a few mF to maintain the voltage rail
This is practically a class d amplifier

Components:

OpAmp1, OpAmp2: These are standard op-amps (e.g., TL072, NE5532, OPAx344 for low input bias current) used to create the precision rectification and selection.

D1, D2: Standard small signal diodes (e.g., 1N4148). These are crucial for the OR logic.

R1, R2, R3, R4: (R3, R4 missing in diagram but implied for input biasing/feedback if using different op-amp config) These are feedback resistors. In this common configuration, OpAmp1 and OpAmp2 are often configured with unity gain feedback around the diode.

R5, R6 (optional but common): Current limiting or mixing resistors before the final buffer.

OpAmp3 (Buffer): A third op-amp configured as a voltage follower (unity-gain buffer). This is essential to prevent loading the diode OR gate.

How it Works (Step-by-Step Explanation):

Let's assume we want to output the maximum of Input A and Input B, and both inputs are positive control voltages (which is typical after a precision rectifier).

Individual Op-Amp Branches (OpAmp1 & D1, OpAmp2 & D2):

Consider OpAmp1, Input A, and D1. This part of the circuit acts like a "precision voltage follower with a diode."

The op-amp will try to keep its inverting input (-) at the same voltage as its non-inverting input (+). So, it tries to make V(-) = Input A.

For current to flow through D1 to the Output_Combined_Pre_Buffer node, D1 must be forward-biased. This means the voltage at OpAmp1's output must be (Output_Combined_Pre_Buffer + V_forward_diode).

Crucial Point: The feedback loop for OpAmp1 is closed through diode D1 and back to its inverting input. Because of this, the op-amp will output exactly the voltage needed to make Output_Combined_Pre_Buffer equal to Input A, effectively canceling out the diode's forward voltage drop.

The same principle applies to OpAmp2, Input B, and D2.

The "OR" Logic at the Output_Combined_Pre_Buffer Node:

Now, imagine Input A is 5V and Input B is 3V.

OpAmp1's Role: OpAmp1 will try to drive Output_Combined_Pre_Buffer to 5V. To do this, its output will rise to (5V + V_forward_diode). D1 will be forward-biased.

OpAmp2's Role: OpAmp2 will try to drive Output_Combined_Pre_Buffer to 3V. However, if Output_Combined_Pre_Buffer is already at 5V (because OpAmp1 is driving it there), the voltage at the output of OpAmp2 would need to be (5V + V_forward_diode) for D2 to conduct. But OpAmp2's output can only go to (3V + V_forward_diode) (or slightly above if it tries to push). Since (3V + V_forward_diode) is less than 5V, D2 will be reverse-biased.

Result: Only the diode connected to the highest input voltage will be forward-biased and conduct. The other diode(s) will be reverse-biased, effectively disconnecting their respective op-amp outputs from the common node.

Therefore, Output_Combined_Pre_Buffer will accurately reflect MAX(Input A, Input B).

The Final Buffer (OpAmp3):

The Output_Combined_Pre_Buffer node, while having the correct voltage, still has a slightly elevated impedance due to being driven through a single diode.

OpAmp3, configured as a voltage follower, has extremely high input impedance and extremely low output impedance.

It samples the voltage at Output_Combined_Pre_Buffer and outputs an identical voltage (Final_Combined_CV) but with the ability to drive the subsequent stages (like the VCA control input) without loading down the precision OR gate. This maintains the accuracy and provides a robust, low-impedance control voltage.

>>2928185
Seems like a nightmare for crosstalk and for making impedance-controlled traces, if any wires are overlapping. Also aren’t you limited to the wires only contacting vias from the one copper side? Or do they drill it and plate vias through the board afterwards?

>>2928204
>This is practically a class d amplifier
All high frequency inverter stages are basically class-D amps.

>>2928199
now it's backtracked after i got it to make a simplified schematic and it agreed that the existing op-amps can be repurposed so that only one extra op-amp is needed (the voltage follower) and i clarified that the existing op-amps are configured with diodes in the feedback path
>If the DL241's existing precision rectifiers are properly designed "super diodes" that operate as ideal half-wave rectifiers with their output being the compensated cathode, then you can indeed directly connect their CV_CH1_Rectified and CV_CH2_Rectified outputs together, and feed that common junction into your single voltage follower (OpAmp3).

>>2928213
building the circuit on a breadboard > analytically quantifying the circuit with known constraints > brute-force spice simulating it > circuitjs simulating it > asking an ai model > asking tarot cards
Climb the ladder

>>2928230
nibba i just want to link two control voltages together, they're already super diodes and the AI reliably gives reasonable explanations for why to use a voltage follower. just gonna be a bit of a pain to build a precision full wave rectifier to avoid saturating the op-amps but it's not a necessity, just chasing marginal gains
>For ultimate performance (less distortion, faster response): If you're an advanced modder and want to push the performance envelope, then replacing the existing saturating precision rectifiers with non-saturating versions in each channel before the ORing junction would be the next step. This is a significantly more complex undertaking, requiring careful component selection (fast op-amps, matched resistors) and potentially PCB layout changes.

>>2928233
My point is you can easily simulate anything the AI suggests, instead of second-guessing whenever it contradicts itself.
I'm unsure how the feedback resistors in a normal half-wave op-amp rectifier react to each other when they're ORing a signal, but considering this full-wave rectifier design does it it's probably fine. The other full-wave rectifier design has a push-pull output and so isn't suitable for ORing together.

>>2928020
I found a better solution. If I pull down to ground with a logic-level FET, the diode preventing backflow ensures that the timing cap isn't influenced. Then I can just duplicate the diode resistor circuit above but backwards to make the duty-cycle 50% when the FET is turned off.

>>2928235
where you even finding this stuff rofl i'm talking about the "improved" precision rectifier on wikipedia which is then connected to a summing amplifier so it's positive instead of negative, no push-pull

>>2928209
> All high frequency inverter stages are basically class-D amps
Checked. You can get 1kW class-D monoblocks for the same price as an inverter.

PUT THOSE MOSFETS DOWN….

I found this insane single-chip 600W asskicker from TI, it’s $2.00 in quants of 1k… two dollars?
What the fuck have I even been doing with my life…
AC power supply, backup inverter, high power sig-gen, motor driver, piezo driver, lighting PWM…etc, etc, ad nauseam.
The government should be giving everybody an allotment of 12 of these per month that come with my food stamps.

>>2928246
The summing amplifier's output is a push-pull stage though; that LM358 at the bottom-right can both source and sink current to the output, unlike the LM358 on the upper circuit. So you couldn't just connect the outputs of the two summing amplifiers together.
But in >>2928235 you could connect the left and right channel full-wave peak-detectors, because the sinking of current is blocked by the output diodes D2 and D4.
Again, simulate it.

>>2928248
Impressive, but I wonder how those handle loads that aren't purely linear. Both rectified loads and switched loads. Also you'll need a massive 50/60Hz 600W iron transformer to convert from the 50V-peak wave into a 170V/340V peak wave.
>Applications
>High End Soundbar
What the hell kind of soundbar shits out hundreds of watts?

>>2927795 (OP)
I just installed a new head unit into a car(>picrel.jpg), I have seen the specs and has a couple or rca connectors (for left/rigt/bass(2channels?)), now supposedly this connectors have a signal of 2V, this makes me wonder
>usually how ample is the audio signal that comes from the power harness?(the color cables)
>the commercial amps (alpineand all that), how much raises the signal(say the rcas output the 2V and the how much raises the 1000W rated external amp?
>I have seen small factor amp boards with supposedly good reviews on outube, how much raises the power,nsr, thd and all that values?
at naked eye,imho, these small ones can't handle the 1k power (they would be fried before half the output).
And lastly
>these speakers with too much "excursion(?)", how much volts are these receiving?(isn't degrading the signal and the speaker?
Thanks anons.

>>2928256
forgot to add, the head unit is rated to 10A,12V,
>120W PMPO?/doubt it is rms
thanks

>>2928250
i'm not gonna use a literal LM358 it's just an example

>>2928256
2 V p-p is the input.
> Car
The output is going to be 12 V.
But at 1000 W, that’s 80 Amps.
You’ll need thick wire friend.
> iHeart radio
Poison. Just listen to regular FM or CDs or yout iPod through Aux, or bluetooth.

>>2928260
Are you familiar with what a push-pull output for an IC actually is? It means it can sink and source current. All op-amps do this. Stop blindly trusting your shitty AI, it's conflating a push-pull stage with a discrete BJT totem-pole circuit. Read the datasheet, does it source and sink current? Like the OPA2134 being able to output +/-40mA short-circuit current. That means it's got a push-pull output stage, and so the outputs of a left-channel full-wave rectifier would conflict with the outputs of a right-channel full-wave rectifier.
>just an example
Yeah I got that, I was specifying such that you could more easily see which op-amp in which image I was talking about.

You need two rectifier circuits, one for left and one for right, correct? Try putting two of these >>2928247 together, and comparing the result to putting two of these >>2928235 together. See what happens.

>>2928262
>Poison
I say the same, this is for a friend and were his bucks.
>80A
then the units like >picrel.jpg are a bunch of boost converters centered in current/mirrors?
Then all that amperage isn't going to blow the fuck of the speakers?
How it's the power or current to db ratio?

>>2928262
and thanks for the wire tip, I was to use 14/16AWG

>>2928020
>I want to have a signal line that stretches the off-time to be significantly larger without influencing the on-time at all

this was solved around 100 years ago, according to google
use trimmers if you need to trim it exactly

>>2928302
I don't understand where the signal input wire goes in your diagram, I don't want something where the duty-cycle can only be manually adjusted. I figured it out BTW, it's fed into a transistor that pulls down one of the resistor paths: >>2928237

>>2928302
>where the signal input wire goes in your diagram

dont see that mentioned anywhere in your original post
anyway, another way to do it is using a vactrol, either commercial or hand-made
advantage of that is that you have linear control, not just on/off

>>2928312
>I want to have a signal line that stretches the off-time to be significantly larger without influencing the on-time at all
It's the very thing you quoted. The "signal line". Maybe this alone is a bit ambiguous, but all the talk about trying to use a FET or analogue switch should imply something along those lines. So would the fact I was already using parallel resistors and a diode in my original schematic.

Vactrols are not fast enough for me, this thing needs to switch on and off for single milliseconds at a time. A FET-based optocoupler would work in place of the analog switch IC or 4007, but I don't have any of those, and it's solved now anyway. Time to breadboard it.

It's for a discrete logic NEC IR protocol sender.

>>2928248
>>2928250
wait i can do an inverter from an old transformer and a cheap d class amplifier? lmao
i would have never thought it

>>2928293
> all that amperage isn't going to blow the fuck
It’s the total power (V times A) rating for the speakers (within reason). Don’t go above 300 volts because insulation breakdown.

Home stereo amplifiers can run at 60 volts output, so for 1000 W, that’s 16 amps, so 14 AWG wire is going to be fine.

Cars have very low voltage, so you need thicker wires… it’s the amperage that heats up the wires, so for (potentially) 12 V, at 80 A you need 5 times the wire thickness. In theory.

It may have a voltage booster, so you might want to check that output. Many don’t, which is why a lot of car speakers were always 4 ohms or lower, when the home stereo standard was 8 ohms.

>>2928327
Yeah, especially the big transformers you find in UPSs and non-inverter welders.

>>2928279

Part A: Precision Full-Wave Rectifier (PFWR) - Two Channels Identical
(e.g., for Channel 1. Channel 2 would be an identical circuit with "CH2" labels)

Explanation of PFWR (Part A):

OpAmp_P1_CH1 and D_P1: This is an inverting precision half-wave rectifier.

When Audio_In_CH1 is positive, OpAmp_P1_CH1 output goes negative, D_P1 is reverse-biased. No feedback through D_P1.

When Audio_In_CH1 is negative, OpAmp_P1_CH1 output goes positive, D_P1 is forward-biased, closing the feedback loop. V_half_wave_CH1 becomes positive and proportional to the negative half of Audio_In_CH1.

(Note: There are many PFWR topologies; this is one that demonstrates the two-stage idea.)

OpAmp_P2_CH1 (Summing Amplifier): This op-amp combines V_half_wave_CH1 with a scaled version of the original Audio_In_CH1.

By carefully choosing resistors (R4, R5, R6), this summing amplifier inverts the half-wave and adds it to the original signal such that the output V_full_wave_CH1_rectified is proportional to |Audio_In_CH1| (the absolute value).

The output of this summing amplifier (V_full_wave_CH1_rectified) is an active, low-impedance output that can source and sink current.

Part B: Active OR Gate (Precision Maximum Selector)
(Takes inputs from the outputs of the two PFWRs)

Explanation of Active OR Gate (Part B):

1. Inputs (V_full_wave_CH1_rectified, V_full_wave_CH2_rectified): These are the already full-wave rectified and active outputs from the two channels.

2. OpAmp_OR1 / D_OR1 and OpAmp_OR2 / D_OR2: These two branches form the core of the precision OR gate.

Each op-amp (OpAmp_OR1, OpAmp_OR2) is configured to act as a "super diode" for its respective input and associated diode (D_OR1, D_OR2).

The op-amp will drive its output such that the voltage at the cathode of its diode (the Common Junction) precisely matches its input voltage (V_full_wave_CH1_rectified or V_full_wave_CH2_rectified), effectively compensating for the forward voltage drop of D_OR1 or D_OR2.

3. OR Logic: As explained previously, because the op-amps actively compensate for the diode drop, the Common Junction will always rise to the voltage of the highest input (MAX(V_full_wave_CH1_rectified, V_full_wave_CH2_rectified)). The diode connected to the lower voltage will be reverse-biased, isolating its respective op-amp.

4. OpAmp_Buffer: This final op-amp is a simple voltage follower. It takes the highly accurate, but still somewhat high-impedance, Common Junction voltage as its input. It then provides an extremely low-impedance output (Final_Combined_CV) to drive the VCA control inputs of both channels reliably, without loading down the precision OR gate.

Revised Stage 2: Non-Saturating Active Maximum Selector (for positive voltages)

Explanation (This is a simplified representation of a known non-saturating Max circuit. Values and specific internal connections might vary slightly depending on the exact variant):

This class of circuit leverages summing and differencing. The key is that the op-amps remain in their linear region.

OpAmp_MAX_A & OpAmp_MAX_B: These act as individual processing stages for each input.

OpAmp_FINAL: This combines the outputs of the previous stages to produce the maximum.

The exact resistor values and topology for R_A, R_B, R_C, R_D are critical to ensure that the sum of the inputs and outputs correctly resolves to the maximum value while keeping all op-amps within their linear range.

lmao

should i look at those crazy high slew rate op-amps with like 600-3000V/us slew rate
>I used Linear Tech LT1358's for the input, National LM6172's around the VCA and another on the output. A that 2180A vca was used with the 47 ohm - volts rail bleed resistor removed as well as the 100k resistor off the THD trimpot. Sockets were added as well as local power supply decoupling and larger psu mains caps.

>>2928346
There it is. Again, consider this circuit: >>2928235, as it bypasses the need for an active OR gate.

I'm not going to attempt to read those ASCII diagrams, also I'm going to sleep soon. Maybe you can ask the machine to make netlists.

I think AI models will be really useful for turning schematic images from datasheets and appnotes and articles into netlists, though I wouldn't trust that kind of specialised task to an LLM just yet. The same goes for converting the graphs and tables into machine-readable data to go into spice models, though that one is even more the realm of a conventionally written computer-vision-based program. LLMs would still be useful for extracting information out of paragraphs of technical writing. Maybe you'd spice sim the results, or maybe you'd train a PINN on the spice sims to speed things up at the expense of some absolute accuracy.
Imagine an enormous dataset of circuits and sub-circuits you could search through by their simulated characteristics. Like maybe you want to make a +/-5V off-line switched-mode power supply for a synth, you could specify the allowable ripple as a function of frequency, maybe some radiated noise specs, cost, size, efficiency, etc. and you'd get a set of separately tested sub-circuits you could trivially assemble into a circuit that would match your specs. It would be like browsing octopart, but for schematics. And because they're all publicly published for free reading by component manufacturers, there would be less concerns about IP laws.

>>2928235
>I'm unsure how the feedback resistors in a normal half-wave op-amp rectifier react to each other when they're ORing a signal
i think the op-amp with the lower value saturates, which might be acceptable if it recovers quickly enough, but i was hoping that saturation could be avoided. oh well. now i might just use the existing rectifiers but perhaps with extra fast op-amps.

>wondering why low-pass filter on 5V square wave only gets up to around 2.5V
>swap the 1M resistor to a 10k resistor and the problem goes away
>start to think the input of the CMOS chip somehow has an input impedance in the low megaohms
>start to think I got dud chips
>remember my scope has a 1M input impedance
>swap probes to 10x
>problem goes away
Feeling rusty. And a bit stupid. And now that the 100n/1M low pass filter is giving me a time constant of 70ms I'm really doubting myself again.

ok i'm satisfied now with just ORing the existing precision rectifiers. the math doesn't add up for why i would need a gorillion slew rate to recover from saturation in the "OR gate". the AI calculated 0.02V/us for the control voltage's maximum rate of change assuming 10V voltage range and 0.5ms attack/release so even mid-tier op-amps are orders of magnitude faster in this instance.

also, compression is typically very subtle and you hear it as an overall sound rather than picking out minute details or defects in the gain reduction
>Psychoacoustics: The human ear is generally much less sensitive to very short, infrequent artifacts in the control voltage path than it is to distortion in the direct audio path.

>>2927795 (OP)
probably a dumb question, but do the intermediate stages of this cascaded OR arrangement need pull-up/pull-down resistors? e.g., pins 3/9, 6/10, 8/12. (i know that i need a pull-up/pull-down on pin 13.)

>>2928341
fuck, the best i can get for 60v is only 687w
shit i wanted to make an inverter for my edirtbike battery

>>2928379
the output stages of the CD4071 are push-pull, so no you dont need pull-ups/downs.
>(i know that i need a pull-up/pull-down on pin 13.)
if you say so, it depends on whats driving it.

>>2928379
>>2928396
also the pins in this diagram are different than the ones in your schematic; you may want to double check your footprint.

>>2928396
thanks
>>2928399
nice catch, thanks again.

>>2928370
Yeah you don’t need full-wave rectification’s speed unless you’re doing very fast attack compression. But it can still be useful if your audio signals have different positive-going and negative-going amplitudes, which is always possible. Take a 1% duty-cycle square wave as an extreme example, when AC-coupled that little spike will be much more positive than the base of the wave is negative. I believe some compressors use full-wave rectifiers for sampling signal amplitude for this reason. Look for some existing schematics to see how they handle it, maybe they just amplify the signal to like 10Vpk and use regular diodes.

>>2928386
Wouldn’t be very portable if you use an AC transformer for it. 600VA is a fair bit, enough to plug in your power tool chargers, laptops, test equipment, soldering equipment, etc. What would you want to use it for, boiling a kettle?

>>2928379
You know you can just combine the outputs of the open-collector comparators together to get an AND, no? Just flip the inputs to each comparator, common their outputs with a single pull-up resistor, then add a NOT gate (or transistor) on the output of that to invert. Because OR(A, B) = ~AND(~A, ~B). Use a diode pointing away from the pull-up resistor for that extra fifth input.

>>2928495
its 68w... i am writing like shit today, not enough sleep
680w would be pretty decent for general use, but honestly i would love to have around 2kw to be able to use the power washer and shit like that.
i was guessing putting some of those in parallel could be somewhat viable, but not in the 60w range, there would be to many of them, so not saving any complexity at all

>>2928504
Old dead welders from 2nd hand vendors are probably the best place to look for power transformers. Maybe find a big transformer inside a high-power audio amplifier.
It would be kinda funny to take a home hi-fi audio amplifier, then use its input power transformer as an output transformer. But anything other than a class-D would make a lot of heat.

I wonder what the best sine generator is?

>>2928328
Thanks anonbro, I really appreciate it,
>80A
Whoa!! thats pretty impressive, this could be dangerous.
I was trying to connect the rca-s to the amplificator but using the cables did not get any sound from the amp/speakers/sub directly (hence the 2vpp idea) so try soldering directly to >picrel.jpg, from the schematic these are the solder points for the rcas, but still no signal at the amp and no sound using directly a pair of the speakers.
I could get a multimeter due the space reduced and the volts at the speaker terminals are zero.
So how can enable the rcas output, I guess this is no fault, simply these outputs arent enabled.
Looked for the service manual, and there is no data about to enable these.
So either it could be a fault or misconfiguration?
by the way I told my friend about the poison and since he got from a pawnshop and there was chance to select other, I suggested a sony dsx-410bt(once I got it for a while, but never used the rca, so Idk how to enable it).
Thanks a lot again!!

>>2928534
Is there a fader knob or setting? The RCA outputs are line level signals (pass-through meant to be run to separate amplifiers). Connect your speakers to the speaker output terminals, not the RCA jacks. Adjust the fader controls to bring up the rear-audio channels.

>>2928534
Section 4.1
https://manuals.jvckenwood.com/download/files/IM381_Ref_K_en_00.pdf

>>2928537
>>2928539
The faders are set balanced, midle and front/rear.
But I get the idea all this serve me.
Im rewiring it.
Thanks
>>2928539
>picrel.jpg

>>2928495
https://www.analog.com/media/en/technical-documentation/obsolete-data-sheets/105738070ssm2120.pdf
oh yeah the drawmer uses a SSM2120 as in >>2928116 for level detection which has full-wave rectification. their product page says "Proprietary log level detector (not RMS)". it takes the log average and control out from the SSM2120 and does the attack/release processing. then the final output goes through the D18 diode. the AI thinks that the D17 in parallel with the 47k resistor might be part of the processing whether it's for a soft knee curve or something else so the stereo channelss shouldn't be linked at that point. i do want to replace the D18 diode with a 1N5711 schottky diode, which although the benefit of the increased speed over a regular silicon diode is probably marginal, it also won't be harmful in a significant way when driven by a good modern op-amp such as OPA1642, the 1N5711 has negligible reverse leakage current which the AI calculated might affect the gain reduction by some one percent of one decibel (considering that the dbx 2151 VCA has a sensitivity of 6mV/dB), it also has less capacitance than a typical silicon diode such as 1N4148.

hmm hmm? modders like to use nichicon fine gold or something like that for the power supply but the AI is saying that solid polymer caps might be better

maderchod, now it's saying it would 1000uF is the ultimate, after i asked it to clarify the issues with it. jim williams (of audio upgrades, not the guy on wikipedia) put 4700uF input caps and 1000uF output caps in the power supply of the drawmer so it seems to work fine but you don't see this being done in typical commercial products, perhaps due to diminishing returns. LT3045/LT3094 don't take issue with this capacitor size, it's already using 10uF ceramic for stability and then the output cap is just to stiffen the supply.

I'm trying to repair the power supply board from an old monitor, just because I can essentially. From what I can tell, the filter cap, 120uf 450v went open, blew the fuse, and that's how it went out.

I have a modest amount of shit I can try to scavenge another capacitor from. I already found a suitable fuse. The first two items I opened though, one was a junk VCR and only had a 200v cap. a 350W PC PSU had 2x200v caps.

Is anything more likely to have a capacitor closer to 450V? I have a bunch of laptop PSUs I can take a crack at.

>>2928754
you can connect them in series for higher voltage rating but lower capacitance

>>2928754
If you’re on 120VAC then a 200V rating would generally be fine, but it’s possible they’re using a pair of single-diode rectifiers to double the peak voltage up to 340Vpk so they don’t have to design it any differently to run on 240VAC. Follow the diodes to see.

>>2928526
>Old dead welders from 2nd hand vendors are probably the best place to look for power transformers.
i though the thing that died in those were the transformer

>>2928768
>CRT (Cathode Ray Tube) internal voltages are significantly high, with the anode voltage often reaching 20,000 volts (20kV) or more. These high voltages are necessary to accelerate electrons and focus them on the screen to create the image. CRT monitors also have lower voltage components, such as the filament voltage (typically around 6.3V AC) and the G2 voltage (around 460V), used for the electron gun and other functions.

>>2928776
The monitor isn't a CRT, it's a CCFL backlit LED.

>>2928788
why in the frick did they use a 450V cap then

>>2928790
Same PCB used in 220V countries. 200-250V caps on an LCD PSU would be fine on 120VAC.

>>2928790
So you can plug your monitor in when you go overseas without frying it or needing an isolation transformer that costs as much as the monitor itself.

>>2928773
Oh, maybe? I figured it was the power semiconductors that died in those, if they had any that is. I guess you don't want to go TOO old.

>>2928793
Switching power supplies create roughly 340 V DC usually. Same in the UK.
…and switch that. It’s the one time north america doesn’t loose to to eurobros

>>2928840
I’ve seen switching PSUs that just rectify the input voltage regardless of whether it’s 120 or 240, and switching PSUs with a voltage switch that presumably swap their diodes around to double the input. Maybe there are some switching supplies with relays and switches that detect 120V in order to actively change the diodes to rectify. But I don’t think this practice is normal for 120V only supplies.

i love perfboard design

>>2928850
Yeah, I had this big ass switching power supply on my desk for a while from IBM that died in a power spike. It was 120 VAC and had 450 V caps.

>>2928889
it's getting worse

>>2928889
>perfboard design
huh, are you in kicad? following any tutorial or any tips?
i should do that instead of just randomnly do it, my designs are getting too complex for just doing it

>>2928889
>>2928921
Cool, what are you designing?

>>2928923
Yeah just KiCAD. There’s software called VeroRoute that apparently can import KiCAD schematics, but instead I had to import a netlist, and then it wouldn’t recognise a 1N4148, and the text and UI scales got messed up by Wayland or my broken graphics drivers so I gave up. It has the nice feature where you can select a resistor and change how long it is at the click of the button, be nice to have that in KiCAD.

I’m not following a tutorial, just laying things out in a 0.1” grid and drawing vertical traces. It’s best if your jumper wires are unimpeded and only run horizontally, but so whatever you have to. Often you will need to replace a footprint with a longer or shorter one, for this I click on the component in the layout, hit E to bring up the edit window, hit Change Footprint, and open up the dialogue for just the selected component. There I may swap a 7.62-long DO-35 for a 10.16-long DO-35. Then I’ll hit my keybind for “update schematic from board layout” that I set to F9 and save both the schematic and the board. That way if I change the schematic (e.g. swap two identical logic gates) and propagate that to the board, it doesn’t revert my footprints.

>>2929013
Programmable NEC protocol infrared remote. Without a microcontroller or ASIC. Just using four 74HC165s, two 74HC14s, and a hundred passives. The Schmitt inverters are used with resistors, capacitors, and diodes to make the 9ms, 4.5ms, and 54ms timing intervals, and to generate the edge triggers needed to set the time delays off properly. Might have been more compact to use dedicated timer chips (maybe NE558s) instead of the inverters, idk. Maybe more immune to different supply voltages too, but I trust 7805 regulators.

This stupid project is feeding my autism, it’s the next step after my logic-IC-based stepper driver.

I want to make a long-lifetime power-bank out of LiFePO4 cells, does anyone know of any chips to use for this? Even just a manufacturer to start looking at. Not afraid of chinese datasheets. I’m not sure if I want to make a small unit with just one USB C port and nothing else on it, or a unit with multiple ports on it, but if it can use a single USB C port for PD charging and PD discharging that would be cool. I don’t need QC, but USB A port and capabilities would be nice if I have the room. A modern take on the single 18650 is my current thought, but a dual-18650 or other form-factor is an option. I don’t really want to use a seperate charge and discharge chip since that means I can’t use the same USB C port for both.

I’ve had two modern power banks full of pouch cells that expand and I’m looking for something more rugged and longer-lasting, and figure I might as well make one. Adding an amorphous solar cell or other energy harvesting method to prevent self-discharge from taking the voltage below a critical level is a good idea too.

what do you think about this? seems pretty accurate. 470nF is a practical inexpensive size for my purposes. searching for this info elsewhere was like trying to get blood out of a stone.

>>2929140
Have you done the calculations to see what noise level you’ll have at the power rails, and hence how much will leak into the audio signal via the op-amps? 470nF feels more like bulk capacitance at the power supply than a bypass cap placed close to the op-amps.

I’d also consider a 10-100nF ceramic paired with a 1-10uF multilayer ceramic. The voltage-dependant capacitance of ceramic capacitors doesn’t matter for power supply rejection, the only behaviour to worry about is microphonicity. I wouldn’t know how much of an issue that is, but it does vary with differing dielectrics.

>>2929143
i don't know how much noise there will be but it's strongly suggested to use at least a 100nF ceramic per supply pin as close as possible to the op-amp because of noise that gets picked up via the PCB traces. the AI did suggest values of up to 10uF (polypropylene film caps) paralleled with a smaller ceramic but 10uF isn't a practical size for polypropylene film caps. 470nF is slightly bigger than its initial suggestion of 330nF (when it assumed 1MHz PSRR dip) or 100-150nF (with 2.5MHz PSRR dip) but it seems to approve of the 470nF polypropylene + 10nF ceramic solution, even when trying different assumptions like 20nH ESL with 1.6MHz resonance frequencies for the 470nF polypropylene cap, paralleling with 6.8nf or 22nF ceramic etc.

>>2927795 (OP)
>tasesr for my sisters
>inverter for home
>emergency battery powered lights
So many projects, so little time

>>2929305
>tasers, inverters, emergency lites

so you live in a violent place with terrible public services
Somalia or USA are my best guesses

now it likes 470nF PP with 100nF C0G, even compared to the manufacturer's suggestion of just 100nF ceramic, after i got it to calculate the impedance at different frequencies and uncovered the anti-resonance peak

>>2928526
>I wonder what the best sine generator is?
i have been slowly researching this since we had this convo lol
i found this document from ti on diferent designs https://www.ti.com/lit/an/snoa665c/snoa665c.pdf

>>2929358
THD probably doesn’t matter much, what matters most is voltage stability. So a triangle wave-shaper looks sensible. If frequency stability is the name of the game, then maybe a divided-down 32.768kHz crystal should be integrated and shaped instead.
It would be nice if there was a soft-start option, whereby you could slowly ramp up the voltage when it’s turned on. An option to set the output voltage to 110V or 220V would be nice too.

I hope this is right thread. Some time ago, someone has showed me a kind of gadget/toy that is an essentially a fast changing led strip that you can wave around to create patterns and do light painting. It was ridiculously expensive though, so I have decided with a friend to make on ourselves since it seemed pretty easy to do.
Today I finished second prototype. It uses ESP32 to drive a led strip that allows for an external clock so it can get much faster than typical addressable strip. For now I'm using bitbanging for testing so it's only 150 fps, but with SPI this should allow for much more detailed patterns, especially if you swing it around on a rope.(I think it should support up to 3000 fps IIRC)
The circuit is quite simple, just the esp board, BMS, stabilizer, capacitors and buttons and leds.

>>2929416
Here are some of the better photos I managed to make.

Just bought some ESP32s for a project I am working on and I have like four more left. Give me some idea what I should use them on.

>>2929418
Very cool. Are you sensing rotation speed or angular position at all? Are they WS2812-like LEDs?

>>2929421
Depends on what you’re interested in. ESP32s have a lot of RAM, DMA, and I2S, so audio DSP is definitely an option.

CNLohr has some neat but esoteric project videos with ESP32s that I’d recommend.

>>2929421
Make 4 penis enhancement devices and advertise them on MeTV.

>>2929421
https://www.bottango.com/
Make a Hitler puppet and a masturbation machine.

>>2929421
Make a Hitler puppet that tells jew jokes written by AI. Green screen and put him in random scenes.

>>2929447
>Call the show Max Hitleroom
>Upload it to dik dok
>?????
>get debanked

>>2929445
>>2929447
I wonder if the Germans have equivalents of the “AI Presidents Play Video Games” videos with the moustache man?

>>2929391
point is to keep it simple honestly.
>It would be nice if there was a soft-start option, whereby you could slowly ramp up the voltage when it’s turned on.
that would be more complex
>An option to set the output voltage to 110V or 220V would be nice too.
you select that by using an adequate transformer... is not like you could get it directly to the mains voltage

>>2929441
>Are you sensing rotation speed or angular position at all?
Not at all. I didn't really meant it as a rotatory display or something, it doesn't move fast enough for your eyes to see persistent image anyway. It just leaves some trails behind.

>>2929441
>WS2812
They are SK9822. WS2812 are significantly slower.

>>2929542
>simple
Yes, but also safe.
> that would be more complex
Literally just a self-heating thermistor in the feedback path of the output op-amp.
>you select that by using an adequate transformer
Yeah ideally you have two taps that you can wire in series or parallel, but there’s nothing stopping you from just switching a different resistor in the output op-amp to lower the amplitude. It will be less efficient and less capable of driving high VA, but it’s easy,

I've got this SFF PC I want to use for emulation, after some testing its main bottleneck is being limited to 135W power draw, the charger I've got is rated for 200W so its got plenty to spare but the board is capped to 135.

The whole thing is cooled by one 20W fan, is there any problem with me tacking some wires from the DC port to a buck converter so it's not powered by the computer rails but the same charger to squeeze out a little bit more juice?

>>2929585
That will work but the fan will run at full speed unless you use PWM to lower it.

>>2929590
and it will run when the PC is off.

I used to really like electronics/circuit design when I was in university (Computer Engineering), but after I got my degree I just became a web developer and never touched a circuit again. I want to get back into it, but I really absolutely suck at soldering. I've read all I could and I've watched a thousand videos, but I need practice as I always fuck up the things I'm trying to solder.
I vaguely recall someone telling me of "training kits" for soldering where you're just soldering/unsoldering things for practice. Can you guys recommend me a cheap aliexpress one? I looked into OP's resources but none are recommended there.

>>2929595
Learn and get better by doing.
“Training kits” are the most retarded thing ever.
Go get some trashed electronics, desolder, then re-solder everything on there (after ensuring capacitors are drained).

Or, just make a kit hat does something useful.
Soldering ain’t hard unless you’re re-balling a 1024-pin gate array with heatsink pads.

>>2929590
>>2929592
Be pretty easy to add a transistor to the original fan’s output wires.

>>2929595
Surface mount or through-hole? Best practice for through-hole is probably those kits with lots of LEDs, like chasers or dot matrices. You could also just buy some cheap ICs and LEDs and resistors and solder them to some dot-board, The green double-sided dot-board is noticeably different to solder than the brown single-sided dot-board. But there’s also kits for more useful things, like an OLED clock, Bluetooth speaker, FM radio, calculator, even shitty little oscilloscopes, for ten dollars or less on aliexpress.

Desoldering is significantly harder then soldering, but it’s still useful to be able to do. Usually it’s post important to keep the PCB intact not the component, but sometimes it’s the other way around,

>>2929590
> pwm, full speed
You can just cut the voltage. That’s pro grade.

Chipquik makes pic related but sadly the datasheet says only the center pad is silicone, base is made from PVC.

However, the base look suspiciously similar to eBay "silicone" soldering mat which begs the question: are all those eBay "silicone" mats really made from PVC or has Chipquik gone full cheapskate here with PVC base when silicone would have been a better choice material-wise (heat resistance etc)?

>inb4 just get a wooden workbench
I would, but they come with a house that's financially unavailable for me at the moment.

>>2929595
an old school mechanical keyboard with direct soldered cherry mx style switches could be an option. you'll have numerous switches to desolder and solder and you can enjoy the end product with your preferred choice of new switches. a lot of the newer keyboards have hotswap sockets but maybe you can find a used older keyboard.

>>2929659
Or you can just make your own custom macro pad. If >>2929595 is all in on his computer engineering having a custom keyboard to help with frequently used commands could speed up his workflow, as well as giving him a chance to practice his soldering. It's a super easy project.

https://www.instructables.com/DIY-Arduino-Macro-Keyboard-Increase-Your-Productiv/

>>2929655
The blue chinky ones online are will not dissipate static, but they are fully silicone, and they shouldn't generate static like rubbing a balloon on your hair does. At the very least, if it doesn't have a button to clip a probe onto it definitely won't be sufficiently conductive to dissipate static. That chipquik one likely is static dissipative. See:
https://elimstat.com/how-to-use-anti-static-mats/

>>2929590
>>2929592
Thanks good to know I'm not too retarded, was planning on passing the PWM and FAN_TACH from the original 4pin cpu fan header and tying the original 12v to a thyristor with a current limiting resistor to toggle the fan on and off+control by the computer.

Was more concerned whether this might lead to any adverse effects to the power brick or computer since I'm essentially just ramming another 12V rail in the thing albeit only for 1.6A

I used my arduino nano to unbrick my attiny85 after following a guide. I feel very smart.

>>2929710
Facts don't care about your feelings.

>>2929710
Buy/make a USBasp and use that for all your AVR programming needs instead of the USB serial bullshit. Bootloaders and printf debugging were a mistake.

ok turns out i made this remote control perfboard design 2.3 times longer than my perfboard can support
but now i have the option of folding the thing up to make it more compact but thicker, like 90x80mm and 50mm thick, instead of being 80x200mm and maybe 15mm thick.
ergonomics, hmmm

In the software world, there's hands-on books that teach you by giving you a block of code, explaining what it does, and then telling you to tweak it to your liking. I'm thinking of books like "Learn Python The Hard Way" and "WebAssembly from the Ground Up".
Is there anything similar for electronics? I'd like a hands-on guide on building different opamp circuits, or building DCDC converters, or similar.

>>2929822
>anything similar for electronics?

no
programming is creating logical sequences
electronics is witchcraft and deviltry

>>2929837
Physics, chemistry and applied materials science held together with math & engineering. The only limit is your budget. What's the quote about advanced technology being indistinguishable from magic?

>>2929822
Breadboards.

There are guides on places like Sparkfun or Make that might go through simple breadboard circuits one by one, but seldom do these entry-level electronics tutorials touch on DC-DC converters. Maybe because the currents and frequencies can be too high for a breadboard. Read appnotes I guess.
Moritz Klein does deep-dives on designing audio effects/synth circuits, showing why he makes each decision with clear diagrams and feedback at each step. Same for Ben Eater to a lesser extent, because digital is harder to get intuitive feedback on.

>>2929885
It's painful as I have to watch repair videos to understand how circuits work. Spent several days tracing our a smps still don't quite understand how it starts.

what about some cheapo 1MHz oscilloscope like FNIRSI DSO153, what are the fucking odds that the circuit will oscillate at 10MHz with all the proper decoupling and whatnot, when it's already rolling off at 33.86kHz

the 10MHz DSO510 is pretty cheap too

i probably wouldn't mind some USB scope without a built-in display if that exists as a good price/performance albeit somewhat inconvenient to use solution

or 18MHz DSO154Pro

>>2929999
>>2930000
nice quints
consider buying an old analog scope second hand, those can be found really cheap, and come with real probes and capabilities
of course they have their own downsides, the first one the size and weight, but against super cheapo scopes the ergonomics of having actual buttons and knobs win

>>2929999
I just got the SCO 2 10M scope. It’s the cheapest 2-channel scope I could find, and it’s a battery-powered unit with a decent enclosure and even a current monitoring sensor. The UI is pretty shitty, with just 8 push-buttons, but it has all the features I’d want in a portable scope if you can navigate to them. Don’t think it has any math or FFT or protocol-decoding functions, maybe for audio you’d want the ability to see THD readings, but at least it has the suite of waveform stat readouts like peak and rms voltage.

If you’re serious about hifi audio, I’d want to buy a scope with a high-resolution ADC over high-frequency capabilities. Some scopes boast 16-bit resolution, which you may want to look out for, I doubt they go up to 24 bits. Even a (sufficiently isolated) USB sound-card with a protection filtration and range-setting front-end might be a cost-effective and useful option.

>>2930006
I’d agree, if you don’t need single-shot measuring capability and can actually find one near you, you have to spend at least $250 before you start approaching the ease of use of an old CRT scope.

I got the best of both worlds by getting a hold of a CRT scope with a digital mode where it passes the waveform into an ADC, into RAM, then back into a DAC and onto the screen, giving me the ability to pause and measure a transient signal. But it’s getting badly out of calibration now so I have to do something about that sooner or later. Each time I turn it on I have to change the zero offset of the channels by a lot, and the zero-offset changes from analogue to digital mode.
Well, the scope was free so I’m hardly complaining.

Which one do I need for hobbyist desoldering?

https://www.amazon.co.uk/YIHUA-Multi-Purpose-Electronics-Soldering-Temperature/dp/B0D6FZXTPK?ref_=ast_sto_dp
https://www.amazon.co.uk/YIHUA-858D-desoldering-Adjustable-Temperature/dp/B0BYJPDNX4


I'm going to buy one, pump/wick isn't enough for me. Price difference isn't huge but not sure if having the station is necessary or not. Any help welcome thanks

>>2930008
i have an hybrid scope too, but mine is broken, the digital parts apparently work, but not the analog ones, it doesnt trace a signal at all.
got it for 5€ lol, should be relatively easy to fix, but i dont have a workshop until house renovations finish

>>2930009
what are you struggling with exactly?
have you tried the pumps with the integrated irons?

my electric drill battery wasn't charging anymore so i oppened it and it's 5 18650s in series with some elctronics shit on top and turns out they somehow became unbalances as FUCK
two were 3V, two were like 2.1V and one was 1.5V, so i'm balancing them manually by charging each one of them to 4.2V separately. i hope that fixes that shit since i don't want to have to buy a new aku battery

>>2930063
18650s discharged past 2.5V are likely dead.

>>2930016
Removing analogue sticks on controllers. It's really hard to do with just pump/wick because there are multiple joints that need heating up
Trust me, I tried everything. I haven't got one of those pumps but I did try heating the solder and putting the pump directly over with the iron still there and trying that way. It just doesn't wiggle free because of all the connection points. I can't tell you how long I spent on it. I don't want to deal with anything like that ever again ergo the hot air gun kek
I just don't know which ones are good+cheap

>>2930121
for that usecase, and i'm understanding that sticks are through hole, i really recommend the heated sucker. it makes way more difference than you think
the air guns are for smd work
but, as you mention
>I did try heating the solder and putting the pump directly over with the iron still there and trying that way.
maybe you need more power, but you could put some leaded solder, or even lower temp solder to unsolder the joints, and flux shit gets way easier that way, adding low temp solder lowers the melting point and can made a ton of difference when you are dealing with modern non leaded solder that has a high melting point, not meant for the /diy/ enthusiast
also maybe a board preheater?

>>2930011
Do you at least have access to a service manual? Hope the tube is ok.

>>2930016
These Yihua 929D-V desoldering irons are good. Saw a teardown of a cheaper one and it was pretty sketchy.

>>2930063
>>2930074
If that 2.1V cell is still holding any voltage at all, it may be chargeable up to 4.2V without significant self-discharge, but chances are its capacity has dropped, so I wouldn't expect it to last well in a balanced pack. Maybe it's only 10% lower and you'll just end up with 10% lower capacity, but I'd probably want to shove a cell balancer in there, or replace its BMS with a balancing BMS. That kind of bms mod would be cheaper than replacing all the cells, but I'd probably want to do both. Apparently even the pricy milwaukee batteries don't do cell balancing. They have the chip do balance the batteries, but it's never enabled, which is odd.

>>2930141
Gentle hot-air plus a cheap desoldering iron is probably the easiest way forwards, besides buying a pricier desoldering iron. I would kinda like to have a higher-power temperature controlled desoldering iron, and after some browsing I found pic related by searching for "T12a desoldering". $3 for the tip, plus another $10-$15 for a pump and foot pedal, it's pretty cheap if you're already using T12 tips like I am. Even cheaper if you just put the other end of the tube in your mouth. Though I'd need to make a 3D printed adapter from the hollow tube to a silicone hose.
I can't find anything similar that will fit into T245, TS100, or 900M tips, but with the M4 threaded tips you can buy for heat-set insert dies, you can probably make an adapter block that fits the 929D-V tips. Get it 3D printed in metal by JLCPCB for $8 or so.

>>2930154
Ah shit aliexpress's new user bonus was tricking me, they're actually more like $15 for the desoldering tip. Or $66 for a new station with the pump built-in. Honestly a good recommendation for new users, since they can just swap out the tip for a normal soldering tip. I'm tempted, but I already have a T12 station, a T12 USB iron, and the bits lying about to make a third one too. So I'll just shell out the 15USD for the cartridge tip alone.

anyone know where's a good place to source piezo stacks, never had a project that needed them before
would prefer not to spend $50+/unit but can't seem to find a trustworthy alternative that's cheaper

>>2930141
>>2930154
>the air guns are for smd work
doesn't that apply for the controller boards?
As for the other suggestions, I tried all of them.
I scoured Youtube and didn't find a single person taking analogue sticks off boards without using a hot air gun. The only method that worked in the end was breaking the stick module into pieces so I could take out portions at a time. I wasted so much time anons and it's time for another stick replacement.

I'm not sure what a desoldering iron does, but afor a heated solder sucker, can I buy any old cheap shit? I'm sure I've seen them on AliExpress for like 3 dollars. I just think a heat gun could come in handy for any future work as well

>>2930192
>for a heated solder sucker, can I buy any old cheap shit?
As I said in >>2930154, I saw a teardown of a cheaper one and it was pretty sketchy. See
https://www.youtube.com/watch?v=xR8k89t4diE

A desoldering iron is a heated iron with a hollow tip, and typically also with a vacuum pump that sucks solder through that tip when a trigger/button is held down. Though I'd probably consider those heated solder suckers to also be a type of desoldering iron.

What iron are you using at the moment? Looks like some people are making custom tips to reflow all the pads at once, pic related. Otherwise I've had some success on difficult desoldering jobs by running thick (e.g. 1-2mm) solid-core wire around all the pins so you can melt them all at once.

>>2930227
>solid-core wire around all the pins so you can melt them all at once.
I tried this too at the time but I don't think my wire was thick enough
The iron I use is some basic one. I have some random chinese one. I have a Ryobi wireless one too but I've never opened it.
I guess I won't get the cheap desoldering pumps.

How much are decent heated desoldering pumps and desoldering irons? Not too decent but enough to not be sketchy?
A heat gun would still do the trick, no?

>>2930001
I bought this opti-tek scope a long time ago.it was cheap useful but I wish for a better one. It can be inconvenience with board work and for close zoom the scope has to.get way to close to the work. Other than traces. no real complaints useful for helping me follow those tiny pain in the ass traces.

>>2930233
As I said, the Yihua 929D-V is good, and it's fairly cheap. Otherwise, if you're in the market for a better iron too, that $66 desoldering station I posted here >>2930156 can be used with normal T12 tips just like a normal T12 station (what I've got, it's good). That make for a decent 70W temperature controlled machine, AND presumably a good desoldering machine too. Get another normal pencil if you swap between tips often.

Never used hot air myself so not sure on that front.

>>2930154
>Do you at least have access to a service manual? Hope the tube is ok.
no, company had a flood some years back, which is apparently true
tube is fine, it shows the digital parts, like the scale i am on, and the pointers trying to measure a specific value, it just dont measure anything, so it should be the analog part of the signal acquisition.

>>2930249
I have a Hitachi VC-6545 and found the manual to a very similar scope, the VC-6145. Or maybe it’s for the 6045/6025. For my case it was on Elektrotanya and archive.org, though I think I originally got it from t some sort of radio museum’s wiki. Is your unit’s manual not anywhere online?

What is this machine for, and how much could I get for it? I have no idea if it works properly, but it turns on.

>>2930266
That's the computron for a gas chamber, you plug it into the wooden door and insert zyklon b pellets.

>>2929790
Should have my discrete logic programmable NEC remote control fully wired up tomorrow, then I’ll test it. Printing out the front and back layers made it nice and easy to figure out where to score out the copper, and where to place the jumper wires and components. Shame it will look awful once I try cleaning off the flux with IPA, which will soak the flux into the paper.

Some stretching of button pins and diagonal capacitors did occur. The passives sticking high out of the first board are the ones setting the critical time-constants. If I don’t see a proper waveform resulting on an IR receiver, it will be easier to adjust them this way. Maybe I should have used some multi-turn trimpots…

I kinda want to 3D print a 2:1 scale TV remote control to house this stupid thing.

>>2930260
its a promox od 442, this https://youtu.be/rPeOR3mljs8 but at 20mhz instead . I just send an email to the video dude if he has the manual.
there is service manuals of the analog models, so i hope those i can use as a reference
like the 417 https://www.radiomuseum.org/r/promax_osciloscopio_od_416.html
and 402 https://elektrotanya.com/promax_od-402_oscilloscope_schematic.pdf/download.html

>>2930266
>Mains-powered ultrasonic device (transceiver)
>Used in industry for locating defects in metal parts

>Krautkrämer’s Origins Rooted in Post-War Innovation: Founded in 1946 by brothers Dr. Josef and Herbert Krautkrämer in a garage in Cologne, Germany, the company began as a small electrophysics lab focused on repairing and developing measuring instruments
>Pioneering Ultrasonic Testing for Industrial Applications: The brothers responded to a challenge from Krupp-WIDIA to detect subsurface defects in metal carbide dies, leading to the development of one of the first ultrasonic flaw detectors
>Breakthrough in Non-Destructive Testing (NDT): In 1949, Krautkrämer introduced Germany’s first ultrasonic testing device capable of visualizing internal flaws in steel, marking a turning point in industrial inspection technology
>Legacy of Innovation in NDT Technology: Over 75 years, Krautkrämer has become synonymous with precision and reliability in ultrasonic testing, influencing global standards in quality assurance and safety
>Foundation of Waygate Technologies’ NDT Leadership: Krautkrämer’s pioneering spirit and technological breakthroughs laid the groundwork for Waygate Technologies’ continued leadership in advanced non-destructive testing solutions

maybe you could loot just the knobs and dials from some of those vintage measurement things, they sell for $5-10 to use as decoration pieces. pic rel costs $10k and it's totally overrated for its sound, people say it has a nice front panel though.

>>2930268
>>2930276
>>2930277
>>2930279
Thank you anons.
Harvesting parts (especially the case) was my main goal when fishing it out of the trash at my uni. However, machines from the same brand go for several thousand bucks on eBay. They all seemed to be smaller, I did not find a listing (running or ended) of this monstrosity.

>>2930276
that would actually be useful
>>2930266
>>2930284
do you have the probe? its probably not worth much anyway
like the kind of shit a dude buys for his garage workshop for some bucks
maybe try to sell it to a welder

>>2930284
If you’re sufficiently curious, I’d attempt to get it working before deciding to sell. It’s that or try to sell it to someone else sufficiently curious, like with a $250 reserve to weed out the scrappers.

If you do try to get it working, I’d first disassemble the unit and clean it out, checking for any expired or burnt components. I’d guess it’s a vacuum tube unit from the size, even if it’s solid state there’s still a Braun tube, so it may well have broken tubes that aren’t even worth repairing. If it is intact or easily repaired, see if you can get yourself a schematic, be it by manual or by reverse engineering. Then you’ll know the pinout of the probe and its mode of operation, so even if you can’t get an OEM probe you’ll be able to cobble something together.
Maybe contact the university to see if they have any info or anecdotes about the unit to share.

>>2930248
So the 929D is the same price as one of their hot air guns
Isn't the latter going to be more effective?

>>2930293
From what Voultar said, hot air should apparently only be used for surface mount desoldering, while desoldering irons/stations a should be used with through-hole parts. But with plated through-holes, it can be pretty tough to remove all the solder connecting a pin to its hole, so I’d personally be leaning towards a technique that can reflow all the pins at once.
So hot-air, or a funky tip-shape, or something like chipquik alloy.

You can just buy bismuth online and dissolve it with regular leaded solder, you’ll get an alloy that melts at about 100C. Indium and gallium are expensive.

>>2930391
>So hot-air, or a funky tip-shape, or something like chipquik alloy.
or use a preheater and a more regular technique
like aliexpress.com/item/32833389729.html
a hot plate is also useful for more shit and works with this aliexpress.com/item/1005007102432729.html
for small shit you have these cute ones aliexpress.com/item/1005009295505186.html
or if you want to goo super poorfag route (please at least ground this crap, i have one lol) aliexpress.com/item/1005008246249465.html

>>2930393
It’s a THT component, so conventional hotplates aren’t sufficient. Radiative preheating might help, but because there’s like ten pins to reflow you’d want to actually melt them all at once, not just preheat. If those radiative things will actually melt solder, then with some shielding foil tape you could probably do a pretty good job. Even cheaper if you just buy one of the ceramic heating elements and tie it to a TRIAC dimmer. And mount it somehow.

As for those shitty PTC reflow plates, are they safe to use without any temperature control without damaging components or PCBs, while actually melting the solder? I was building a temperature controller for one, but recently I realised they have PTC elements within so maybe I don’t need to. I’m sure the changing tempco would mess up the PID loop, not that I couldn’t correct for that.
Instead maybe I should just build a little box for a power switch and ground wire termination.

I’m surprised nobody makes a large but cheap reflow plate that can set a proper reflow curve.

>>2930397
anon please could you share an actual photo of the pcb?
the idea is just to heat the board before like at 100º degrees, or whatever temp is safe, way bellow the temp to melt the solder, so you only need to add 200ºc with the heated sucker instead of 300ºc, that makes a big difference
forgot to say you could use also a hardware store heat gun, or maybe even a hair drier, but i dont recomend the last one
or you can go really old style with an old pan filled with sand as a preheater
any hot plate or ptc becames adequate to tht if you use sand, it adapts to the irregular height of the components, and the more sand mass you the longer it takes to heat or to cool. just get a themistor probe for like 2€, and you can adjust the temp manually, or use a pid or whatever

the ptc heaters are made for lead soldering, they are not really a good idea to use directly, but could do shitty reflowing if needed. I plan to use mine as an awful, yet small hotplate with some sort of overheating protection

I was trying to mount an IP5306 on a board today and I just couldn't. My hands have been shaking lately and today it was just impossible. I have stopped drinking coffee but it isn't enough. Is there anything other than drugs that can stop those minute tremors in your fingertips?

>>2930409
anon please go to a doctor not this mongolian basketweaving forum

>>2930401
No I'm not him. I'm this guy: >>2930248.

>the idea is just to heat the board before like at 100º degrees, or whatever temp is safe, way bellow the temp to melt the solder
I think his problem is that, even with wick and suckers, there's fragments of solder around each hole. Even without a preheater, a desoldering iron will do the same job, it will just take a bit longer, and maybe be more likely to overheat the board/components. For ground planes it would help I guess.

>the ptc heaters are made for lead soldering
>lead
Do you mean Pb, or do you mean component legs? I intend on using it for my own boards with my own leaded solder-paste. I bought a 100x200mm one because I have a 100x100mm board.

>>2930409
Man, my coworker has really shaky hands (like, moving more than an inch), and it's gotten worse in the last few years. He's in his 30s.

>>2930429
>Man, my coworker has really shaky hands (like, moving more than an inch), and it's gotten worse in the last few years. He's in his 30s.
How does your friend cope with the shakes?

>>2930432
No clue, he's not my friend.

>>2930409
>>2930432
nm, I figured out the solution. I got frustrated and drank two glasses of wine. Went back to the bench and my hands were magically still. I rarely drink more than 3 drinks a week so it wasn't DTs making my hands shake, but the solution was still the same as if I was.

>>2930391
>>2930393
Sorry I'm silly
Am I to understand you'd advise against using hot air, or are these just alternative methods? A lot of them seem to cost about the same as a hot air gun/station, and I'll be honest, I only considered one because that's what I see in Youtube repair videos.
The desoldering gun is an alternative but I would cry tears if I get it and still am unable to remove an analogue stick which is my worry. I did use a hair dryer last time as well and heated that board up. I tried literally everything but unless I could heat up all points at once, that stick did not come out or even budge.

I found some radio's on the curb. Both were intermittent power, only fully power on if like upside down or one the side, so I imagine some power connections have gotten loose. Though I don't really care about that as I am more curious as to this. When I turned the volume knob on in this case, a RX 5040 it will screech and you have to fiddle with it side to side to get it to stop.
I have a Tandy Patrolman of a similar vintage, and it does the exact same thing, though maybe not as bad. What causes this? Is it just the switches going bad or something else?

>>2930477
>Am I to understand you'd advise against using hot air, or are these just alternative methods
Again, I've never used hot air so I don't think my opinion is worth much on that front. I've used the 929D-V to decent effect, and it is much better than a solder-sucker being used with a normal iron, but it can take a while.

>unless I could heat up all points at once, that stick did not come out or even budge.
Yeah it sounds like you're getting solder remnants stuck in the plated holes. You may be able to get rid of these with hollow desoldering needles (they're cheap as shit, at least give them a try), and maybe something like the 929D-V will be sufficiently better at getting rid of those remnants, but a way of melting all pads at once is likely what you want to go for. If I were you, I'd probably buy the hot air gun, and some foil or kapton tape to mask off and protect areas from the hot air (this is important), but also pick up some desoldering needles and maybe $5-10 of bismuth. Hopefully you have some dead boards to practice on.

>>2930479
Dodgy volume knob probably just means the potentiometer is full of gunk and needs to be cleaned out. Easy job, the older pots can be pried open and put back together pretty easily, if not it's still a relatively cheap replacement. The same applies to switches, especially rotary or slide switches. Toggles and rockers are less prone to this. For the orientation-dependent power, I'd assume something heavy has cracked a solder joint. Especially an iron transformer or heat-sinked transistors. Any components connected to the chassis of the unit and the PCB at the same time are also suspect, like connectors and controls.

shit i keep getting intermittent faults with my big perfboard circuit
i press a button 10 times, and get different results for 3 of them
i kinda want to order a pcb, but then i'd want to redesign it to use better timer chips than 74hc14s, which would mean more breadboarding, which sounds like hell right now

Im into analog computers now, any tips?

>>2930477
i mean you can use hot air station, i have one myself, but see what i said here >>2930401
please post the pcb, there is something that i am missing here
>>2930429
>there's fragments of solder around each hole.
i dont understand
>Do you mean Pb, or do you mean component legs? I intend on using it for my own boards with my own leaded solder-paste. I bought a 100x200mm one because I have a 100x100mm board.
i mean smd reflow soldering with leaded solder paste, yes, ptc stabilizing temp should be somewhat around the temp needed for lead soldering. Of course the reflow profile will be shit, but for that price what do you want

>>2930511
Get rich enough to afford an FPAA, or read declassified WWII documents on how ballistics computers worked.

>>2930522
>i dont understand
You know when you use a solder-sucker to suck the solder out of a plated-through-hole, but the component lead still won't come out? And then you look closer, and there's a partial layer of solder inside the hole, connecting to the component lead? And it doesn't easily get sucked out because the component lead basically touches the walls of the hole and the surface tension keeps the molten solder there? That. Sometimes you can break it out, but sometimes you delaminate the pad. Desoldering is a pain.

>leaded solder paste
Gotcha.
>Of course the reflow profile will be shit, but for that price what do you want
To turn the switch on, wait for the paste to go shiny, and turn it off again. There's enough aluminium and copper for it to cool slowly enough, probably idk.

>>2930531
>You know when you use a solder-sucker to suck the solder out of a plated-through-hole, but the component lead still won't come out? And then you look closer, and there's a partial layer of solder inside the hole, connecting to the component lead? And it doesn't easily get sucked out because the component lead basically touches the walls of the hole and the surface tension keeps the molten solder there? That. Sometimes you can break it out, but sometimes you delaminate the pad. Desoldering is a pain.
oh, for that you can use desoldering needles, i think, you get the lead inside the needle and the solder doesnt stick to the outside of the needle
aliexpress.com/item/32785718126.html
i have one kit but never actually needed to use them

Bought a ESP32 and decided to run some tests with a 4-digit 7-segment display. I wired it and uploaded the sketch from this
https://microcontrollerslab.com/esp32-74hc595-4-digit-7-segment-display/

It works fine for any number up to 5555, but entering any number higher than that into the serial monitor causes errors. Random numbers, symbols that aren't even numbers, and blank screens. I looked through the code and I didn't see anything wrong with it, and I swapped out the 74HC595, resistors, and wires incase there was any fault there.

Is my ESP32 just borked in a completely retarded way?

>>2930563
anon wrong general, while most people here will be familiar, there should be more help
>>>/diy/mcg/

2uF C504 is charged slowly because of the 10M R529 between V+ and GND? but then the discharge rate through D509? 15V/220Ohm=0.068A? so a 15mA 1N5711 would be an unsuitable replacement for D509? or the IC506A takes up most of the current or what?

10kOhm R514 limits the current through D509?

*2.2uF

i don't understand the signal flow, apparently it can go through a backwards diode? idk. anyway, the AI estimated a fraction of a millamp max current from C504.

i tried to get the AI to explain the diode signal flow and it stubbornly gets the anode vs cathode wrong when reading schematics, overriding my written instructions. garbage.

now it accepted it but it thinks the schematic is wrong
>If the schematic is drawn literally and correctly, D506 is connected in a way that would generally prevent the positive rectified output from IC505A from reaching IC506B. It would only allow current to flow backwards from IC506B's side towards IC505A's output if IC506B's input were at a higher potential than IC505A's output. This is highly counter-intuitive for a signal path in a compressor's detection circuit.

>Therefore, it is most probable that D506 is drawn with its polarity reversed in the schematic, and its anode should be at IC505A output (pin 1) and its cathode at R520/IC506B input (pin 5) to function as a normal output diode passing the positive rectified signal. If it were oriented correctly, then the previous analysis of current flow would hold true.

now it got it i think, it passes the signal or 0V depending on the signal vs the threshold

AAAAAAAAAAAAAAAA SON OF A BITCH

>>2930599
Define a signal.

Nobody wants to read a stream of consciousness blog about your struggles and triumphs with ChatGPT. Shut up.

What book do you recommend for an absolute beginner who can't even get the difference between the negative and ground

>>2930663
Getting Started in Electronics - Forrest Mims
How to Build an Obstacle Course for your Hamster - Rube Goldberg
The Great International Paper Airplane Book - Jerry Mander

>>2930585
Why would you use a schottky? Leakage current of this one is in the realm of 1uA or less, but at 15mA the forward voltage is all the way up at 1V. 1N4148 is unironically the best diode for this kind of circuit.

At high frequencies, C504 will act like a short-circuit, so the instantaneous current through D509 could get as high as V/220, where V is whatever voltage is produced by IC505B. 15V sounds kinda high, but that might be a fault condition you want to plan for.

>>2930626
RIP

Okay, so I was fucking the the radio a bit and I noticed the case was cracked. Overall kinda a piece of shit but I might give it a little more time to fuck with it before tossing. Jesus Christ I had no idea how complex the internals were, then again it's a cassette deck and has 50 outputs so I don't know what I expected.

Is there anyway to hotwire a battery powered device to run off a power supply by just hooking up some wires to the terminals or something? If I could see if it ran on batteries better it would help narrow it down,but I don't have a dozen c or d batteries around?

>>2930672
the AI favors schottky for lower forward voltage. it's true the forward voltage goes up to 1V at 15mA but the actual current will be more like 1mA. not sure if C504 can go full circuit, that's why i was trying to figure out if D509 needed a certain current rating. R514 won't limit the current draw? this is for the rectified control voltage so the frequency won't be too high as it has been averaged with transistor array magic in a previous step. the attack is in the millisecond range and the release is in the hundreds of milliseconds range.

*full short circuit

>>2930683
Yeah, the batteries are generally in series, and are 1.5V each, so if it uses 6 C/D cells that’s 9V. You’ll want to clip the negative wire to the spring terminal at one end of the series chain, and the positive wire to the nub terminal at the other end of the series chain. I once clipped a PSU to the spring and nub that were internally connected together at the wrong end of a double-row battery bay and melted the wire, good learning experience.

>>2930689
Shottky diodes can have bad leakage compared to silicon diodes, and the extra speed isn’t necessary for audio frequencies. You could just choose a higher-current diode, maybe one with a lower forward voltage too since 0.4V is still not brilliant. Or you could replace the 220Ω/2.2μF filter with a higher-impedance filter, though it would start to influence the stuff after the analogue switch chip.

>>2930531
thanks, and what's an "fpaa"?

>>2930715
reverse leakage is a drop in the ocean compared to forward voltage, the AI calculated less than a hundredth of a decibel in the context of a VCA with 6mV/db control sensitivity
>You could just choose a higher-current diode, maybe one with a lower forward voltage too since 0.4V is still not brilliant.
name some examples, preferably with axial leads. and why are you only concerned with the immediate 220Ω-diode connection for estimating the current and not the 10kΩ and op-amp that comes after it?

anons getting outclassed by an AI lmfao

1N4148 is used for soft knee so i would leave them in the section labeled "soft knee". but in the other parts you'd probably want less forward voltage drop for better precision, like the AI says. of course you'd listen and compare the results, perhaps do the modification in one channel and compare it to the unmodified channel. though there's a huge amount of subjectivity and people find uses for even some of the bottom of the barrel compressors. most respectable professionals probably use at least an SSL channel compressor for workhorse duties and fancier outboard comps though.

>>2930716
FPAA is a field programmable analogue array. Op amps and shit that can be internally rewired on the fly.

>>2930761
>reverse leakage is a drop in the ocean compared to forward voltage
Worst case you're switched to IN1 for over 1M of resistance from D509's anode to ground, so 1uA would pull the anode 1V higher than it should. I'm not sure to what extent negative feedback via IC505B will amend this, but it seems like bad design to use a schottky here.
>name some examples, preferably with axial leads
1N5819, or maybe BAT46.
>and why are you only concerned with the immediate 220Ω-diode connection for estimating the current and not the 10kΩ and op-amp that comes after it?
Worst case scenario, switched to IN0, C504 is charged up to some Vmax (V+SC maybe), while IC505B tries to pull down to some Vmin very quickly. You also get a current spike from C507 but the capacitance is much lower. The 1N5711 has no SOA diagram, so it's hard to say what kind of non-continuous current it can handle, though intuition suggests it's fine.
Steady-state, the most current D509 would see would be through R519, R524, and P503. So worst case 20220 ohms, plus the 5 ohms of the 4053, and the ESR of the op-amp IC505B, so less than a milliampere.

>>2930762
What is this in relation to?

>>2929585
>>2929701
Hey it's this guy here, I've mocked up a quick diagram for this idea.

It uses a separate power jack instead of a buck converter for testing, the resistors are to step the 12v voltage down to 2v from the CPU header for the MOSFET gate but I feel like I've gotten that part wrong.

Does this look right to you more knowledgeable anons?

>>2930773
thanks

>>2930788
I'd want to use a PFET and switch the positive side instead, that way you can safely connect the 0V wires of both PSUs together. The TACH and PWM signals are referenced to GND, so when the low-side transistor is turned off they'll float up to 12V, potentially damaging the chip on the computer that interfaces with those signals.

>>2930788
>>2930799
Something like this. I don't think you can avoid the second logic-level transistor, though if you use a small NFET like a 2N7000 you could avoid the base resistor and get yourself a bit more isolation between the two voltage sources.

>>2930801
>>2930799
Thank you so much, so am I correct in understanding with your design GND is unified across both power sources and since logic is handled with a PFET the reference to GND isn't impeded so TACH and PWM won't float up?

This is my first time actually designing a circuit so a little out of my depth, it looks quite similar to some of the solid state relay designs I've seen which I guess is essentially what I'm trying to do?

>>2930805
>so am I correct in understanding with your design GND is unified across both power sources and since logic is handled with a PFET the reference to GND isn't impeded so TACH and PWM won't float up?
Yes. Simulate in Falstad or LTspice if you're curious.
Solid state relays generally use an optocoupler of some sort for extra isolation. You could use a solid-state relay here as it would simplify things a bit, or you could use an optocoupler, but either way to get those TACH and PWM signals you'll need to connect the GNDs together.
One thing to consider is what will happen if the computer is turned off but the external PSU is turned on, or vice-versa. It's possible that the PWM signal could damage the fan if the fan isn't getting 12V, so it might be wise to add another MOSFET to that wire, picrel. The fan shouldn't get 12V while the PC is off because of the MOSFET I guess.

>>2930806
oh, with a pull-down resistor from 12Vext to GND too.

>>2930806
>>2930807
Awesome, I'll definitely simulate it and look at adding that extra MOSFET.
Again thanks so much for your help.

>>2930773
ok fine i got the AI to favor low reverse leakage current with better prompting

lmao
this is for >>2928116

>>2930819
Where did 0.3uA come from? The 4148’s datasheet says it’s 25nA. Maybe 50-60nA worst case at 40C.

>>2930820
25 nA is worst case at 20V at 25C ambient temperature, it gets worse at higher temps

>>2930821
Read the chart. Half way between the 25C and 75C lines (i.e. 50C), in-line with the 20V mark, and you get maybe 50nA. No way are you going above 50C.
If you want something even lower leakage, use a reverse biased BJT:
https://sound-au.com/appnotes/an018.htm

>>2930925
lmao i misread the 1N5711 STMicroelectronics graph, it's not graduated in tenths like the others

LMAO can someone please debunk this or what a beautiful coincidence, even though the graph is in fifths the math works out

i'm trying to read fig 4 at 10V, 75C
https://www.st.com/resource/en/datasheet/1n5711.pdf

and if you figure that the voltage is only up to 0.1V in this specific application (a generous 16.67dB gain reduction considering 6mV/dB VCA control sensitivity) the reverse leakage is very very low for 1N5711 at 25-50C

the farnell sheet has the 1N5711 on par with 1N4148 in reverse leakage current at 10V 25C anyway

I need help identifying a shorted resistor. I'm pretty sure the 3rd ring is transparent and what we see is the winding.
If it's any help it's part of a Weishaupt WRD 0.2 heating controller, but I can't find any schematics for it.

>>2930964
I'm guessing 4.7 Ohms assuming that stripe is gold.

>>2930964
Btw it looks fine to me. Give it an anal probe with your multidongle.

>>2930969
>>2930970
Thanks, it's probing close to 0ohm but maybe there's something else interfering. I'll take it out and check again tomorrow.

>>2930964
Looks fine. Today I had a customer with a dead Kenwood cake mixer and it had a dead carbon composition resistor. Properly cooked and cracked, thankfully the green-blue 56 ohm colouration was legible. If it were all browns and oranges I wouldn't be so sure.

>>2930943
I have no idea how you'd read a 5-step logarithmic scale. Normal 10-step scales are obvious:
>1 2 3 4 5 6 7 8 9 10 20 30 40 etc.
But 5-step? Would it be:
>1 2 4 6 8 10
The step from 1 to 2 is the same size as the 10-step while the rest are different, so it feels really shoddy.
Odd numbers make more sense.

>>2930961
8nA is pretty good. That seems to be unusual for a schottky diode. The common 1N5817 has 1mA leakage current, and even the BAT46 is 800nA. I should get me some of those.

my neighbor gave me this box of vacuum tubes as he heard that I'm interested in electronics. I do'nt know much about that stuff. anything cool I can do with them? can I trade them for LM4562?

>>2931028
that one looks pretty funky. seems to be some kind of nixie

>>2931028
Cool stuff. Almost all vacuum tubes are basically obsoleted by something these days, but that doesn't mean they can't make for good experience, or a fun project. Those big tubes might be suitable for large high-voltage amplifier stages, especially for radio transmission, in case you're a ham. Indicator tubes are still awesome, so if you have any I'd recommend making them into a cool (but at least somewhat practical) project.

For the ones you have no use for, look up their part numbers individually and see what they sell for. Chances are most are just a dollar or two each, but you might have some real gems in there.

I want to make an dimmable LED light source for a microscope I have. If I get this LED assembly kit https://www.ledsupply.com/led-kits/10-watt-led-light-module-kit, or maybe the 5W one instead, what else might I need? Clearly I need a dimmable power supply and a dimming switch. Is this a suitable PSU https://www.ledsupply.com/power-supplies/mean-well-pwm-series, I guess I'd take the 40W 12V option?

>>2931028
you can look up the value by the general type like 12AX7 or EL34
some people might want to buy a bunch of 12AX7 from different brands to try them for subtle differences in their audio setups

>>2931037
>Chances are most are just a dollar or two each, but you might have some real gems in there.
>>2931056
yeah I just checked ebay with the part numbers. those idicator tubes go for like $60 each. the four big boxes in the right corner have EL34 and GZ34 tubes. they go for like $200 each. the ECC83 and E80CC sell for $100 each. rest of the tubes is more like $4 to $20. that's a really nice outcome!
thank you for your help anons

where is the component in LTspice? i've added the spice directive and refreshed the list in the components selection

.MODEL CA3046 NPN(IS=10f BF=145.76 VAF=100 IKF=46.747m ISE=114.23f NE=1.483 BR=100.1m VAR=100 IKR=10.01m ISC=10f RC=10 CJE=1.026p MJE=333.33m CJC=991.79f MJC=333.33m TF=277.09p XTF=309.38 VTF=16.364 ITF=1.7597 TR=10n) VCEO=20 ICRATING=50m MFG=RCA)

>>2931067
it wont show up in the components list. instead, you add a regular NPN and set its value to “CA3046”.

>>2931068
it's not showing up though

then there's this other format? the ltwiki is nonsensical

*SRC=1N5711;DI_1N5711;Diodes;Si; 70.0V 15.0mA 1.00ns Diodes Inc. -
.MODEL DI_1N5711 D ( IS=315n RS=2.80 BV=70.0 IBV=10.0u
+ CJO=2.00p M=0.333 N=2.03 TT=1.44n )

oh the "other format" still has the .MODEL declaration

>>2931069
double-click or right-click (i forget) the actual value field and manually type in the name.

if the reverse leakage current isn't too bad, looks like the knee curve is similar enough, just the threshold is shifted which can be compensated in the circuit design or perhaps even with just the threshold knob because you don't particularly care about the exact value of the threshold, just the resulting gain reduction. and different knees can work fine, for example SPL IRON has 6 different diode/capacitor options to choose from for slightly different sound characteristics out of the same device.

>>2931073
i changed the name but wouldn't it update these values like the manufacturer

>>2927795 (OP)
i'm trying to DIY a multiphase converter. i have a master clock that runs at e.g. 100 khz, 50% duty cycle. i want to tee this off into 16 clocks each with successive phase delays of 22.5 degrees. how do i do this IRL? conceptually i think i need to

>100 khz x 16 multiplier = 1.6 mhz clock for timing
>clock1 = triggered at 1 rising edge from 1.6 mhz timing clock, HIGH until next falling edge from timing clock
>clock2 = triggered at 2 rising edges from 1.6 mhz timing clock, HIGH until next falling edge from timing clock
>etc

ive read at PLL ICs should be able to do clock multiplication so i could probably figure this out. im struggling to understand how to do the next part, where "something" needs to wait arbitrary numbers of rising edges before outputting a signal.

>>2931029
You lucky bastard, wish I got gifted stuff that cool
That's an E1T counting tube (unrelated to nixies). While it doubles as a display, its main purpose is counting. They're pretty rare.
https://www.youtube.com/watch?v=PxZV_0mYZAU
https://www.youtube.com/watch?v=WAyJMB7fD80
>can I trade them for LM4562
They're worth a lot more than modern components. I personally would hang on to them, but that's just me.

>>2931107
that was interesting. sadly, I only got two of them, so no clock :(

>>2931079
works on my machine

>>2931123
i have version 24.1.9, for example the move command is M instead of F7 and it's not a hand symbol. also the ground isn't rotating (is that a ground on your V1?)

i can rotate the ground after it's been placed but not while placing it

>>2931053
Any old 12V power brick would be fine. To dim the light, you’d just need to take a parallel branch off the 12V supply that goes into a potentiometer, in order to provide the 0-10V dimming input to the Luxdrive BuckBlock. Add another resistor if you want to limit it to 10V, e.g. if you have a 10k potentiometer use a 2k resistor in series with it on the 12V side.

>>2931092
Some switching controllers can generate clock offsets internally, though the most I’ve seen is for 3 or 4 phases. 16 is a lot, man, are you doing tens of kW?
This kind of phase offset would best be done with a ring-counter, recently I tied two CD4018s together (with an inverter in between) to divide by 12, dividing by 16 would be doable too and you’d have all the phase offsets from the 8 output pins and their inverses. Then just feed it a 1.6MHz clock. 100kHz is kinda low for a switching supply too. If you use 74HC14s for your inverters, they also work for oscillators.

>>2931126
Fuck I hate LTSpice.

>>2930970
>>2931013
Yeah the resistor was fine and had nothing to do with my problem, it was just a corroded switch.

>>2931132
>16 is a lot, man, are you doing tens of kW?

it's actually the other way. i'm trying to DIY a 12-16 cell NiMh/18650 charger that i want to power from a laptop power supply. each cell will have its own buck IC. if i don't spread out the clocks (phase offset) the input current draw will be massive and trip the laptop power supply's current limit.

>>2931134
Ah. You could just use soft-start plus a bunch of input filtration (and a higher frequency so you can use smaller filter components), but I guess multiphase is an interesting way to go.

how do i even run this and probe the output in the bottom right? it's unreasonably difficult to find instructions on how to use this. they just say start the simulation and probe but they don't mention the "configure analysis" horseshit and the errors

and i got the AI to interpret the connections correctly but the calculated output seems unreasonable whether is V3 is +1 or -1V and no matter what the R5/R6 variable resistor value

the AI thinks the op-amp will either saturate to -15V (-14.3V after diode drop) or otherwise the output is 0V. that isn't how the circuit should work but i don't know how to debunk it.

i'm getting something with .tran 9999999999999999999999999999999999999 but the output is nonsensical

double negative on the -10V voltage sources

still nonsensical output like it's the inverse or something

it might be right now after i defined the supply rails to the op-amp but i haven't fiddled with the diodes yet. it's meant to be a variable threshold.

>>2931180
Watch a basic tutorial. You can use SI prefixes before your letters (e.g. 30k instead of 30000), and you need a spice statement along the lines of “.tran 0 0.1 startup”. Not sure what the R is for after your resistances, shouldn’t be necessary. You need a power supply for your op-amp too. As for the signal source, change it to a sine wave.

>>2931191
You put the negative side of a voltage source on the positive power input for the op-amp. You can use symbolic labels instead of wires for your power rails, like you’ve already done for GND. Load one of the included example files to see how they do it.

BROS, I need help. I solved this circuit using KCL and superposition, but I'd like to know if there's a general expression for it. I understand that it's configured as a differential op-amp, but R2 and R3 seem to prevent the direct application of the standard differential op-amp formula.

>>2931201
awesome
this part is only for the threshold processing of the rectified and smoothed control voltage so i'm satisfied with testing it with DC for now

>>2931225
Generally you want to mirror the resistor networks. So you should set R6=R5, R7=R4, and add a copy of the R2/R3 divider between +3.3V and R4, if you even need that divider in the first place. Differential gain is a function of the R6/R7 ratio. Only then will you minimise error due to nonzero input bias current.

What’s it for? The 1k load is kinda low resistance.

>read about DC bias effect on capacitance with MLCCs
>never really internalize it
>finally look into it
pic related is samsung CL31A106KBHNNNE, 10uF 50V 1206 X5R. fucking absurd.

>>2930481
thank you for all the held
i did end up with the hot air gun because i saw a deal on prime day

but got a question, is it advisable to use heat proof gloves when soldering/using the hot air gun? More the latter than the former, is there a decent risk? I've never used one so I dunno

>>2931246
I ended up with the general expression with more KCL and superposition, lol, but thank you very much.

It's part of an Analog Front End for a scope, developed by a japanese engineer.

With the analysis I found out that it justs offsets an (already attenuated) input signal into the 0-3.3v range, the design allows the user to select the range of input voltage by selecting different R2 values with a switch

I'm still wondering if this topology doesn't ring a bell for somebody, it's just a differential opamp with a voltage divider in the input of the inverting terminal

I want to know if this configuration already has a name (if that's the case, please tell me about it!) or if it's product of the talent of the japanese designer

>>2931276
it's over...

>>2931385
sofa king over

>>2931387

>>2931389
Ask it about MLCC microphonics.

>>2931391

some studio gear, both highly sought-after vintage/discontinued stuff and poorfag stuff, can easily be decades old

>>2931276
That’s way worse than I thought. Even at 10% of the rated voltage it’s very significant, though fine for 5V decoupling caps. I wonder if my 555 RC time constants are getting messed up by that.

>>2931409
>I wonder if my 555 RC time constants are getting messed up by that.
my brother in christ, most capacitors only have a 20% tolerance.

>>2930626
LMAO
to be fair I fucked up custom footprints where the documentation was just confusing before too
even at work when some spastic before me made an 'adaptor' and put that in a lib and used it in all designs instead of uhm... updating the footprint

>>2931415
whats even fucking stupider is that this board was supposed to replace a previous board that i also fucked up by connecting the EP of a different HV op amp to ground instead of V- like it was supposed to be. i ended up swapping that op amp with the 4522 for some reason AND THEN I FUCKED UP THE RAILS AGAIN, I DID IT AGAIN.

>>2931414
They’re changing by more than that just within 5V.

Honestly the curve is so extreme I want to use them as varactors.

heres another fuck up
>use SOT-23 BJTs for everything
>essentially every SOT-23 BJT is BEC
>select the Q_NPN_BEC (or Q_PNP_BEC) symbol in KiCAD and set the footprint to SOT-23-3
>everything works completely fine because why wouldnt it
>fast forward a couple months
>eventually need something with a little more oomf
>SOT-89 looks like a popular footprint thats still fairly small and cheap
>set footprint to SOT-89-3
>look at datasheet
>wait a minute
>these are ECB, not BEC
>use the Q_xxx_ECB symbol instead
>phew, that was a close one
>board shows up
>SOT-89 BJTs arent working
>look at datasheet
>everything seems fine, keep troubleshooting
>cant find issue
>look at datasheet again
>i definitely set the symbol to ECB, and the datasheet clearly says these are ECB, so why the fuck arent these working
>keep troubleshooting
>im at my wits end
>look at datasheet
>it suddenly hits me
how the fuck was i supposed to know to check for that :(

>>2931430
I don't get it anon
Care to elaborate?

>>2931485

>>2931486
LOL

What's the function of your design, I'm interested

>>2931487
voltage regulator, nothing interesting

>>2931067
>>2931123
>>2931132
it threw syntax error when trying to start the simulation because of the last part because it has an extra closing parenthesis. maybe version 24.1.9 is more strict about syntax. this model is from LTwiki

>>2931489
damn, im on 17.0.37
maybe its time to update

>>2931486
You sure that's the top view and not bottom view?

>>2931541
KiCAD footprints are all top-view, and the datasheet indicates it’s the top via the dotted lines.

>>2931544
I see, I suppose the dotted lines are a giveaway, but I still prefer the datasheet say it explicitly. Usually there's recommended footprints near the bottom.

When did Fluke stop including Hebrew in the documentation they ship with their multimeters?

>>2931634
This is antisemitisms caused by Tucker Carlsons and Dave Smith, fellow nooticer.

>>2931488
Voltage regulator is nothing interesting? Nikka, it's the foundation of pretty much any electronic circuit you build.

Can someone ELI5 how to replace this chip in pic? It's my DAC and this chip died. My PC doesn't see it when I plug in my USB cable. It's a well known issue and the fix is to replace this chip.

I have access to my friend's hot air station and soldering equipment. I'd ask him to replace it but he's a noob and doesn't know what he's doing either.

Can someone give me a step-by-step procedure on what to do? What temps to use, how to remove, how t place a new chip on etc.

Thanks in advance!

>>2931660
https://www.youtube.com/watch?v=YUryJOAiPa4

>>2931661
Thanks. So no messing around with soldering paste? I did a bunch of googling and vid watching and some people say you must use soldering paste.

Also, I'm really nervous about the removal...

>>2931660
Protect the surrounding components with Kapton tape. Pre-heat the board in your oven for 2 hours at 250F. Secure the board with a stick vise or helping hands. Blast the chip with high heat and airflow. Remove the chip with a suction tool. Put flux on the solder joints and use desoldering wick with your iron to remove the rest of the solder from the pads. Remove the Kapton tape and clean the board. Place the new chip in the correct orientation and tack-solder two adjacent corners to hold the chip in place, then drag solder the pins. Clean it again, inspect and test.
>>2931662
You don't need solder paste.

>>2931653
jej

Seriously though, the last Fluke I bought was back in the 00s and I distinctly remember the documentation included Hebrew translations. I was browsing the Safety Information that shipped with my new 17B+ and there wasn't a Hebrew section. So it got me a little curious if anyone knew when they took that out so I could make a conjecture on the reasoning behind it.

>>2931677
250C is hot enough to kill aluminium electrolytic capacitors, so if there’s any of those go cooler. Not sure if any other components have a low critical temperature like them.

>>2931683
Maybe they print the instructions in certain languages depending on where they were manufactured, or the files got destroyed/lost, or there was no need because those filthy heebs peaky Engly, or the guy who does the translations is half-Palestinian and half-Iranian. Maybe Fluke wanted Israel to pay for Hebrew instructions and it wasn't in the US budget.

>>2931662
>no messing around with soldering paste?

that's a diff technique which requires less skill but more equipment
that is, you need a specific metal stencil to apply the paste to the PCB and a tiny flexible spatula
and, even if it's lower skill, it's still way way out of your league
there's a 100% chance you'll send that machine to silicon heaven

still, you can watch someone who *can* do it here: https://www.youtube.com/watch?v=ls2_IagvLfM
he occasionally mentions what temp and airflow settings he uses
also watch https://www.youtube.com/@northwestrepair/videos
and also https://www.youtube.com/@dellpartspeople/videos

and while you're there, check out my own channel, https://www.youtube.com/@ThatVeganTeacherYouTube

>>2931808
>that is, you need a specific metal stencil to apply the paste to the PCB and a tiny flexible spatula

idk, i've never used a stencil in my life. surface tension and the presence of a solder mask helps a lot. if you do solder bridge adjacent pins it's not hard to fix with some solder wick. it's only a pain in the ass with things like flat no-lead packages.

that said, a lqfp package or whatever is in his picture is easier to do with a soldering iron than paste.

I used the circuit found from pg. 12 of the ST lm317 datasheet [1] for setting the output voltage. It would be nice to be able to set the output current as well. Ideal would be a continuosly adjustable current limit, but I think a digitally selectable limiter with a couple of settings (e.g. 20/100/250/.. mA) would be okay too. Any idea how to add that feature?

[1] https://www.st.com/resource/en/datasheet/lm217.pdf

>>2931677
>>2931808
>>2931830
Frens, chip arrived today. When I looked at it, I couldn't believe how small the pitch is.
NGL, my confidence level of pulling this off is pretty low right now.

>>2931964
believe or not, with enough flux solder tends to go by itself to the corrects parts
but maybe you should get a practice board first, if only to get more confidence

>>2931964
its not hard, anon.
start by taking the other chip off. dont bother practicing unless you have a second identical board lying around, since every board is different and the board itself is the thing most likely to break. you can rip pads/traces off the board if youre too quick/forceful.
turn the hot air up to max speed. temp is whatever, i think mine is set to 600 F. heat up the chip by waving the wand in little circles around the pads. keep the nozzle close to the chip or its going to take forever. occasionally tap the chip with some tweezers to check if its loose. be patient, itll probably take at least 90 seconds from room temp. once youre convinced EVERY pad is loose, you can remove the chip. make sure you have a plan for how youre taking the chip off, namely pick a direction that minimizes the chance of you slipping and knocking a bunch of other shit off the board. use the new chip to practice handling it with the tweezers, some chips can be deceivingly difficult to pick up with tweezers.
you can protect stuff with Kapton/copper/aluminum tape if you want, but theres nothing in your original pic that i would be super concerned about (those two metal things are crystal oscillators, btw; exercise caution if youre planning to ultersonic the board afterwards). post a pic of the other side of the board if you can. also it doesnt have to be tape, you can use anything that can withstand the heat and will block airflow.

>>2931978
>(those two metal things are crystal oscillators, btw; exercise caution if youre planning to ultersonic the board afterwards)
what? never heard this one and i have an ultrasonic cleaner
i barely use it for electronics, that is true... yet i will like to know what precautions i need

>>2931965
>but maybe you should get a practice board first, if only to get more confidence
No other boards... I'll have to 'do it live'.
>>2931978
>start by taking the other chip off. dont bother practicing unless you have a second identical board lying around, since every board is different and the board itself is the thing most likely to break. you can rip pads/traces off the board if youre too quick/forceful.
yeah, thats what I thought.
>turn the hot air up to max speed. temp is whatever, i think mine is set to 600 F. heat up the chip by waving the wand in little circles around the pads. keep the nozzle close to the
use the new chip to practice handling it with the tweezers, some chips can be deceivingly difficult to pick up with tweezers.
Great advice! I'll def take it slow and I'll plan ahead.
>you can protect stuff with Kapton/copper/aluminum tape if you want, but theres nothing in your original pic that i would be super concerned about (those two metal things are crystal oscillators, btw; exercise caution if youre planning to ultersonic the board afterwards). post a pic of the other side of the board if you can. also it doesnt have to be tape, you can use anything that can withstand the heat and will block airflow.
My DAC is slightly different than the original pic which was taken off the web.
Luckily, the chip is the same.
However, it's way more complicated. Lots of caps around too.
Not sure how to protect all this shit.
Anyway, here's top and bottom of the board.

I don't have an ultrasonic. I have a brush and a bottle of IPA tho.

I'll be doing it tomorrow. I feel tired af and my self-confidence level is at near zero.

I appreciate any advice between now and tomorrow evening!

>>2931989
>Not sure how to protect all this shit.
Aluminum foil and Kapton tape.

>>2931986
the crystal oscillator has a resonance frequency, and if you force it to vibrate at that frequency it can cause problems. any oscillator over 32 kHz should be fine, but lower frequency ones could be at risk.
>>2931989
>Not sure how to protect all this shit.
aluminum foil. you dont even really need to tape it down, just shove it in there enough to stay put. really the ONLY purpose is to block the airflow.
>I don't have an ultrasonic. I have a brush and a bottle of IPA tho.
good nuff
>I'll be doing it tomorrow. I feel tired af and my self-confidence level is at near zero.
sleep tight anon, keep us posted!

I know jack shit about electronics and I don't know if this really belongs here. /o/ sucks major dick.

Can I use this:
https://www.aliexpress.us/item/3256808381048938.html
Cut off the gyro, hook up the old gyro leads to an adruino or something, hook it up to can on the ecu, and get a cheap adaptive headlight system?

Alternatively the gyro might actually work ok.

>>2931999
If the gyro leads are a digital protocol that the microcontroller can speak (likely), and the CAN bus can give you info on steering direction (idk bro) then yes. If you want to properly send and receive packets on CAN, you’ll probably want a microcontroller that can speak CAN directly, which the Arduino cannot. But an ESP32 can if I recall, you just need to wire its TX+RX CAN wires to a CAN transceiver chip/module, or maybe you can even find a dev-board with all that shit already on it. Arduinos and such can also be used with a different type of CAN transceiver, one that speaks normal serial or whatever back to the AVR microcontroller, but these are less flexible. I’d search for a sufficiently popular CAN MCU dev-board and first test that you can receive steering data and show it on some little LEDs before buying the chinky headlights. Is it even legal to use aftermarket LED headlights where you live?

>>2932018
Commiefornia. LED's in halogen housings are not but LED's in projectors are in a grey area I think.

There is a steering wheel angle sensor my vehicle so all the controller would be doing is reading the current steering angle and wheel speed and control the projector based on that.

Thanks a bunch on giving me a good head start.

>>2932029
>There is a steering wheel angle sensor
You might be able to read that with any old arduino directly, instead of having to read it from the CAN bus.

>>2931958
> Any idea how to add that feature?

the datasheet shows you exactly how
it involves putting small resistors between 2 pins
however you cant do both voltage and current on the same chip
so, you'd use one reg to limit current, then, in series, one to reg voltage

>>2932072
Don’t R4 and R5 imply the quiescent adj-pin current is constant and reliable? I thought you had to vary the low-value current-sense resistor.

Also use an op-amp circuit with a beefy pass-transistor instead.

>>2932074
>quiescent adj-pin current is constant and reliable?

the formula they use suggests this current is so small, you can just ignore it in your calculations

>>2932079
So then putting resistors like R4 and R5 in this image >>2932072 wouldn’t do anything. It would stay at the fixed current defined by R3.

>>2932090
>wouldn’t do anything

it would definitely do something coz the resistors would be in series
in one position it would put 1.1 as the current limiting resistor, giving a limit of 1.13A
in the other position 125.1, giving a limit of 9.99mA

>>2932094
The current limiting resistor needs to be between the load and the output of the LM317. R4 and R5 are not. The load current does not go through them, the pyre in series with the ADJ pin so only the ADJ pin’s current goes through them.

>>2932095

i asked my 3-year old and she agrees with you

>>2932098

now she's claiming the circuit could be slightly rewired to produce desired result but wont tell me more until cookie ransom is paid

>>2932100
uuoohh

>>2932060
Originally I was thinking of CAN because it might have been easier to wire from the headlights instead of directly to the sensor. I guess I don't need to factor speed like cutting off if slower than 5mph, might still be useful at any speed.

What connector is this?

>>2932382
https://www.ebay.com/itm/253689509617

>>2927795 (OP)
how to derive an expression for vO in terms of vS from >picrelated.jpg?
It seems more like a voltage divider but the gain makes me lost the idea.
And because these are two separated circuits I don't see how to apply Kirchoff's laws.
It's a representation of a linear voltage amplifier as a VCVS.
Thanks.

>>2932480
They’re just a pair of voltage dividers connected to ideal voltage sources. Calculate them just as you would any voltage divider.
Vout = Vin * Rbottom / (Rtop + Rbottom)
Specifically:
Vi = Vs*Ri / (Vs+Ri)
Vo = (Aoc*Vi)*RL / (Ro+RL)
Then plug equation 1 into equation 2.
If you really want to explicitly use Kirchhoff’s laws, then you can only do that for each isolated circuit separately. Vs + I*Rs + I*Ri = 0, for example.

>>2932488
Thanks, it's easier going with the voltage dividers.
>Kirchhoff’s,isolated circuit separately.
As I learned these laws, the catch was all the voltages and currents were in the same circuit not separated ones.
So I don't get how to make a common variable to relate one side with other. As you noted
>Vs + I*Rs + I*Ri = 0
AocvI+I2Ro+I2RL=0
where is the bridge for both sides?

>>2932492
>Vs + I*Rs + I*Ri = 0
The other one is:
>Aoc*Vi+I2*Ro+I2*RL=0
And Vi=I*Ri, which is already in the first line.

I got a discounted solar cell with a battery from my local DIY shop. Initially, it's made for fountains, it will output anything between 6.4 and 9V. I'm happy that there's some extra space for my electronics, but the amount of epoxy and hot glue lets me think that it's not exactly waterproof. How can I waterproof what I want to put in there?

>>2932527
>Vi=I*Ri
This are the kind of details that always fail to pick.
Thanks.

>>2932409
Thanks big dog. I searched for appliance molex but nothing came up, what'd you look for?

>>2932633
“molex 2 pin female”

>>2932615
Conformal coating on static components, and dielectric grease on connectors and potentiometers and such.

>i have done extensive thermal tests of the LM3046, PMP4201, THAT300, and MAT14 in exponential converter applications, and found that the high emitter resistance and low beta of the LM3046 made it the worst of the bunch, regardless of its monolithic process. the PMP4201 is so small, that another layer of insulation around the whole IC makes it pretty much the same temperature inside the plastic packaging. the PMP4201 and THAT300 performed identically, with the MAT14 being miles ahead of the rest.

>It's lower beta and (seemingly) higher emitter resistance than the duals only introduces a static linear error which can be perfectly compensated for.

i'm not sure i want to learn all this transistor stuff right now when i'm really only tweaking existing circuits. is LM3046/CA3046 sloppy enough to warrant an upgrade or even just replacing it with a newer manufactured equivalent that might have been manufactured with tighter tolerances than what they had 40-50 years ago? for example pic related is probably some averaging and logarithmic conversion of the full wave rectified audio signal to use for sidechain detection, i have no idea if replacing the CA3046 would improve performance.

fun fact: PMI MAT04 is from like 1986 and MAT14 might just be the RoHS-compliant version but maybe they improved the manufacturing and maybe the specs have changed slightly, i haven't looked into it. but it's basically a 40 year old design that still slaps the alternatives.

fun fact: valley people TA101 is pretty obscure but basically it's an array of thermally coupled transistors to within 0.1 degrees celsius, i don't know how good it is compared to the others because the others don't show specs for the thermocoupling. this guy has been trying to remake them for like 10+ years
>Making the TA-101 is not as easy as you all think. We own Valley. We have all the original milling machines to mill the PNP and NPN's to the specified thickness for thermocoupling. We have all the original potting trays as well. You have to trim the leads to a specific spec. There is more to it as well for matching everything. We are making more TA-101's, but it takes a lot of time to get done. We are also going to release some of the old products n both 500 Series and rack mount in due time. Specifically for now, The Comander, Dyna-mite, Gain Brain and MaxiQ. We tried to do them a different way and it does not work. You have to follow the process done by Paul Buff and Doc Morgan....

>Just so all of you are aware... Allison Research, which become Valley People, and Valley International was purchased by PMI Audio in 2006. All of these documents and schematics are our intellectual property. I am not concerned about the documents being posted, however, please be aware should any of this become a product for commercial use, I will defend our intellectual property rights. We are beginning to design and produce Valley again by the end of this year. We are currently producing the original TA-101's using the original Valley jigs, milling machines, and test gear, which we will use the TA-101 on the new Dyna-Mite, and Commander. Other models will follow..... We continue to develop these with the original assistant to Paul Buff...Doc Morgan. Doc has been a wealth of knowledge on how to incorporate these designs and VCA builds in accordance to the original designer...Paul Buff...

he said pic rel in 2008 lol but looks like it's happening this year

>>2932705
Haha... they actually called their IC as "complimentary" transistor array.

>>2927795 (OP)
I'm interested in Ham radio and wanted to design some circuits, the software I used didn't seem to have an option for a loudspeaker, what's the best circuit designing software I can use for designing simple radio circuits?

>>2932812
holy shit it's really 0.01 degrees celsius thermocoupling
https://www.proaudioparts.nl/wp-content/uploads/2021/04/TA-101-NPN-PNP-array-datasheet-small.pdf

focusrite red 3 allegedly uses MAT04 both for its VCA and sidechain processing although schematics aren't available

some others use dbx/THAT VCA ICs in their sidechain (e.g. gyraf GSSL)

HUM LAAL has thermally compensated detection

the AI said that one option would be to take discrete transistors and manually match them and thermocouple them basically like what valley people did and this would potentially be even better than SSM2212 or the MAT series. but then you have to thermocouple them which might be difficult especially if you want to file down the plastic for closer coupling. thermal paste like for computers tends to dry out and weaken in performance eventually.

Thermal paste at home
Blend the following in a bowl:
>1/2 cup synthetic motor oil
>2 cups zinc oxide (cosmetic grade)
>1/8-1/4 cup mineral spirits
In a separate container blend the following:
>1/3 cup Vaseline
>1/4 cup boron nitride (cosmetic grade)
Then mix everything together and blend thoroughly.

DIY contact cleaner
1 quart (1000 ml) naphtha (white gas)
2 oz. oleic acid
1 tsp. mineral oil (substitute with 5 tsp. 91% isopropyl alcohol)
Mix into container and shake thoroughly. PETE or HDPDE plastic only.

>>2932885
*HDPE

>>2932879
Are you looking for a validation tool, with simulation or other built-in analysis techniques to estimate the efficacy of your circuit? I’ve seen people use spice for this, and I believe you can use wav files in signal sources within some spice simulators, but it’s a brute-force numerical simulation that might take a while to resolve and has its own quirks. The common spice simulators are LTspice, Pspice, and NGspice if I recall. There’s NGspice built into KiCAD.

If it’s circuit design and layout you’re looking for, then Altium, Eagle, and KiCAD, plus a few other more obscure ones. KiCAD is FOSS and it’s pretty good, not only does it have plenty of features for a conventional EDA workflow, but has plugins that make it easy to export to JLCPCB or autoroute or curve your traces or panelist your boards.
All of these EDA softwares ware intended for conventional PCB design in their layout section, while I’ve used it to a reasonable degree of success for veroboard routing (see >>2929790), it isn’t obvious to me how you’d use it for manhattan design. If nothing else, you could place the components about on a ground-plane with no traces in between, just to get an idea of where to put things and what their pinouts are.
If you’re doing more serious RF design you’ll be using PCBs anyhow, with microstrip calculators to optimise your printed impedance lines and also to implement inductors and capacitors as patterns in the printed copper if your frequency is high enough. Probably best off sticking to 70cm and longer for a new ham though, I say this as someone who got his license two weeks ago.

>>2932885
You want contact cleaner to evaporate out, or at least leave behind residue that switches and pots don’t care about. Mineral oil seems like it wouldn’t work, but I’m unsure as I have seen electrical cleaners that lubricate. What’s the oleic acid like, does it evaporate at room temp?

>>2932892
for potentiometers you might want lubrication?

Why do you keep spamming the thread with off topic unsolicited LLM bullshit? Nobody needs you to make five posts in a row about your latest AI conversation. Nobody wants recipes for thermal paste and contact cleaner that ChatGPT just hallucinated and you haven't even tried. Half the thread is just you copy pasting and babbling. Have you even made anything?

>>2932915
sneed, that wasn't even me. i'm getting much better responses from the AI than from anons as to why you'd want better transistor matching and how to get it etc. this is immensely valuable meanwhile most anons don't have the slightest clue as to why you would want analog studio equipment for music production/mixing/mastering in terms of the sound and not just the tactile workflow of turning the knobs and reading the dials.

>>2932915
I posted the recipes. They're based on MSDSs of manufactured products. I've tried both and they work fine. The AI guy is annoying so we agree on that at least.

>>2932920
Good for you, keep at it. Just stop posting random snippets of your chat logs here. I can use ChatGPT too. If I want to know its opinion on something, I'll ask it myself.

>>2932929
My bad. I assumed it was sloppa because most sane people wouldn't offer a recipe for nearly a quart of thermal paste out of nowhere. That's pretty neat if they're tested and working.

>>2932902
Only if it’s thin enough, like dielectric grease. Either way it has the risk of attracting abrasive dust.

>>2932915
Hey wait a second, those unsolicited recipes are kinda cool, assuming they actually work. I wouldn’t want 50 of those posts in a thread, but once or twice a week is fine by me. I guess the sad thing is that you can’t tell without looking it up.

>>2932920
Oh yeah it’s useful. But that doesn’t mean you should spam posts. If you want to post your AI insights here, do it in condensed format, with specific reference to how it is useful in this case. Use a local writing application for storing the raw output.

>>2932935
i asked anons first but you've got nothing
you can barely learn this stuff if you spend years in a traditional school setting, this is quite niche stuff. most musicfags can't even use a soldering iron and what i'm doing is basically rocket science tier that rivals the top studios (you haven't seen all of what i'm doing)

>>2932942
>i asked anons first but you've got nothing
Have you considered that that might be because no one here cares enough about the subject to know? You admit it's niche.

made a basic composite sync generator on epm240 as a practice on how to use a fpga.
now to figure how to make a simple dac in the ~100 LEs left.

>>2932951
transistors are some of the basic electronic components next to resistors, capacitors and diodes though. matched pairs are more specific to audio probably but i found some god-tier low-noise individual transistors anyway. most people would service audio gear probably just by replacing electrolytics with the same value and maybe upgrading the op-amps so replacing individual transistors is pretty hardcore compared to that.

>>2932978
Most people don’t waste their time upgrading dubious audio gear when the alternative either is buying a better one in the first place, or building a better one in the first place. Your time would be better spent making quick experimental designs to learn what they do and don’t do well, building up a base of analogue circuit intuition.

As for the Village People’s transistor array, it’s just a dollar worth of off-the-shelf BJTs, hand-matched with a simple matching jig, and stuck together so they’re thermally coupled. Diy electronics hobbyists have been doing that for decades.

>>2932998
https://www.youtube.com/watch?v=-nmVx091Hjo
they don't make them like they used to, some things are fucked to buy new even if you have unlimited money to spend. for example chris lord-alge uses an old SSL console and discontinued focusrite red 3 among other things. some vintage stuff is completely unique and overlooked with no modern equivalent.

>>2933002
I didn’t say buying a new thing in the first place, but buying a better thing in the first place.
But if you’re after the very best performance, modern digital signal processing techniques make everything analogue obsolete so long as you get the ADC and DAC nailed. I think that’s the part to really work on. I suspect the anti-aliasing and anti-imaging filters of a delta-sigma DAC/ADC pair will cancel each other out, even any suboptimal phase response, but I’m not sure. Less fun the digital way, of course.

How do I wind
>picrel
into a coil.
You know, along it's tall side.
I usually do my coil winding under power on the lathe. But I never did rectangular. I'm worried it would 'buckle' and lay down flat if not really tightly constrained.
Do you guy's think I really have to make something like a guide that fences it in on at least 3 sides and a very strong mechanism to clamp the start to the arbor or is there a trick to it?
I alsp dont know how hard or soft it is. Should I take it to red heat first to be sure?

>>2933061
like using that as wire?
how powerful is your late? better be very
give us some sizes and shit, not only of that thin g but what are you actually trying to do

>>2933065
Wind a coil. I got several lathes and the one I'm looking at is decent sized, I have zero worries about the power. I abuse it for such things because it's already clapped out and literally bent.
I'm trying to make a coil.
>picrel
kinda shows what I expect

>>2933078
you need a tool to extrude it in the toolpost, basically a die, hope my super cad is understood. I think you need the 4 sizes for this
but i am unsure it will enough just to pull through it, you may need to come some way to actually push to extrude it, look for copper extrusion

>>2933081
i dont know why i draw the feed material as a roll instead of a bar... this is just confusing
the starting material is on the right, sorry

>>2933081
I thought like
>picrel
to be honest
weld a washer to some bar stock
braze the end to the washer or maybe just out a woodscrew through
clamp something in the toolholder so theres a gap that just fits the copper strip and doesnt let it buckle
run the toolpost as close as possible
high feed per rev
lube the copper up, let her rip and hope for the best if no one has a better idea

like so
the thing can be pulled through but cant lay down on its side
im just not sure about the unsuported length
(no complaining about the beater lathe nor its tools this isnt the toolroom)

>>2933089
i mean you can just try that, worth a try. put something bellow it so it can get as straight as possible to the hub
but i would expect that you will end with it deforming when it bends, that is why i proposed to define all the movement, it shouldnt be dfiicult to do in a mill.
hell it should be easy to mill it in the lathe using the toolpost as a tool support

>>2933092
>tool support
i mean support for the piece which is the tool that you are making to do the bend
i need a walk, i am really not focused today

>>2933057
most gear can be disregarded, for example alesis 3630 or yamaha GC2020, but the handful of pieces that i've chosen all have some interesting sonic character, feature or design architecture and you just have to bring up the circuitry to modern specs with a solid power supply and lower noise op-amps etc, then with clever tricks with how you route the signal flow or more extensive modifications you can make it sound very reasonable, some people said that analog audio technology peaked in the 80s and digital started to take over, now companies like wesaudio mainly differentiate themselves with digital controls and digital recall so engineers can recall the settings when needing to make revisions but still maintaining an analog audio signal path. digital signal processing for music is a joke, you're relying on self-taught programmers to tackle modeling of complex non-linear behavior of analog gear, it just doesn't sound right, most songs with billions of streams have a significant amount of analog processing, for example the vocals are tracked with 1176 and/or LA-2A or similar, and 'despacito' and 'shape of you' were mixed in analog. serban ghenea being the most prolific digital-only mixing engineer but he has a far better ear for emulating the analog sound than most people, and the motivation for mixing digitally is for workflow reasons as they can process tracks faster than real time and they can instantly recall settings for revisions, it's not because digital can perfectly replicate the pleasing qualities of the analog sound.

Well I guess you learn something new every day :D
I did keep the thing red hot while welding the fixture and that definately helped in terms of softness, the difference was crazy. Maybe even too soft. I guess in la factoria they either use a former of some kind or they really do have a guide that wraps around most of the way. With a few cm of unsupported material going to the mandrel it just aint gonna work.

>>2933100
holy moly

>>2933103
funny enough shes back to hard as she ever was after just one operation
kinda telling: got to anneal copper anew for every step
anyways wont try again today

Sometimes I just like to type NIGGER in a public forum.

>>2933145
NIGGER

>>2933153
N

>>2933100
yeah that is why i told you to do some sort of die that can go on the toolpost

How do I fix my GPU?

It works fine, except for the fans and lights, and it seems like it's only the 12V supply that is dead. I bought it like this, and I don't know the history.
I tested the fans with a separate supply and they work fine. Plugged a fan into the gpu but with the 12V wire pulled out of the plug and fed from an external supply and I was able to control the fan speed from the OS.
I tested the lights with a tarduino, and they work fine too. There's an MP1475 step-down converter on a daughterboard going to the LEDs, so I assume that is supplied by the same 12V supply as the fans.

I tried doing some research, but everything I find is just MUH POWER PHASES.

The fans connect at the top, and the lights on the right. More pics here:

https://www.ixbt.com/3dv/palit-geforce-rtx-3080-gamerock-oc-review.html

>>2933161
It looks like the board in your pic is missing a capacitor, switching IC, and inductor on the bottom left.

>>2933167
And two memory chips as well.

Fug :DDD

>>2933172
Two niggers just stole your chips.

>>2933161
See if you can find a board view schematic of your specific card (maybe if you're lucky your card doesn't deviate too much from the reference design) and you could easily trace the circuit from the connector using that. Maybe you could try voltage injection after tracing out the circuit to help diagnose, but I can't really give a lot of advice because it's been a while since I've dabbled in graphics card repair.
>>2933167
I'm pretty sure that's just a picture from a review site, missing components aren't really an indication of anything on a GPU unless they've been torn off or something. It's pretty common to see unpopulated pads on a lot of third party cards, Nvidia/AMD release a reference design and a lot of their board partners decide what to add or cut corners on from there. You can see this a lot with budget vendors who often use less VRM phases than the reference design.

>>2933175
Thanks, anon.
I actually looked for a board view when I first got it, but I guess I was looking for one for this specific model/manufacturer, which I don't think is available.
Plenty of others, so I'll take a look at where the fans are supplied from on those.

>>2933161
>fans
remove the old fan and shroud, ziptie a case fan to the heatsink, and power it from the motherboard or a SATA/Molex adapter
>lights
why are you gay?

>be me last month
>replacing switches on keyboard
>somehow ripped off 4 diodes during the process
>probably while moving the board around on my desk as I was pulling off the defective switches
>they're dual diode array with common cathode so I got 8 dead keys
>by some miracle 2 of them have intact pads
>one has a pad ripped off but trace running in a good spot to solder directly to it
>last one is worst, one pad is ok but last two ripped and traces running too far to solder directly to them
>no choice but to create pads using wire
>thankfully all on the numpad so can still use it while I order the parts
>arrived today
>first time I have to do this on an actual repair though I did practice a bit on dead boards
>do the other 3 first and finish with the hard one
>takes me a good 15 minutes fiddling around, poking, pulling the wire with tweezers
>finally get them in a good position
>secure it in place with uv mask
>poke at it gently
>the patch of dried mask immediately releases from board
>wtf.png
>realize I forgot to wipe off flux
>try to gently wipe everything off
>wire breaks off
>fuck
>oh well time to clean it better this time I guess
>spend another 15 minutes placing wires, again
>apply mask, again
>this time it holds
>solder on diode
>poke at it gently
>holy fuck it's actually on there pretty good
>apply a dab of hot glue on all of them to secure them to board
>poke at them gently
>as solid as they're ever gonna be
>plug keyboard in fully expecting at least one of them to not be working
>everything works

I am fully erect.

>>2933182
I literally just have my case lying down on the side with two 140mm fans against the heatsink for now. My case doesn't have a window. I will sell the gpu again at some point in the future thoughever, and some people like gay things.
If the fix is too much trouble I will just tap 12V from one of the PCIe power connectors and call it a day.
I appreciate your concern.
>>2933184
(You) did it.
>no pics
shame.

>>2933187
Just some light post-bumplimit banter.

I bought my GPU with a bodged fan. It's kind of nice because it stays quiet. The GPU market is so fucked that you won't lose much resale value as long as you guarantee that you've stress tested it and the temps stay safe under maximum load. I think mine was only about 5%-10% cheaper than an intact one.

>>2933184
Nice, anon. Congrats!

>>2933161
>>2933175
So it seems like none of the models that have boardviews/schematics are that similar to mine, but I assume the basics are the same.
Looks like the 12V for the fans is coming straight from the PCIe power connectors, only passing through a shunt and a choke. So I guess I should start poking around in the highlighted area in pic related?
>>2933191
Luckily I got this one real cheap. It was a bit of a gamble but it works perfectly fine. 60 degrees full tilt with the improvised fans.

>>2933389
Isn’t there a transistor and some sort of driver circuit to modulate the speed of the fan? I’d have thought that would be the most likely part to die. Well, any solder joints there could have cracked, on the choke and connectors too.

>>2933406
According to the link in my original post, the holtek mcu generates the pwm signal for the fans (and the data signal for the LEDs, I assume).
My initial thought was that it was dead, but no. The fan pwm signal is there. It's just the 12V supply missing.

I'll take another look at it today.

>2933537
>>2933537
>>>2933537
>>2933537
>2933537
NEW THREAD
MIGRATE NOW