/diy/ - Do It Yourself

After ages of tinkering with arduinos, cameras and various electronics I'm looking to build something more ambitious involving lasers and/or transducers, maybe even radar.
The first issue I came across was how to switch a transducer between the powerful transmit circuit and the sensitive recieve circuit. Before starting soldering I figured it was a good idea to simulate stuff so I installed LTspice which is very much on the clunky side. My question is: is it worth sticking at it and learning to use it?
I see that here, Circuitmaker is recommended, but I tried it out and got frustrated quite quickly. My purposes are really quite simple: simulate a ultrasonic transducer circuit that can emit and listen for return pulses. I think this is quite similar to a radio or radar circuit that emits powerful pulses and then switches to another circuit that amplifies, filters and digitises a return signal.. I'm a novice at using circuit simulation and pcb design software so I'd appreciate some pointers so that I dont waste time learning inferior software.

>>2658058
Get used to it. Most electronics software is old ancient boomer computing artifact from era that doesn't exist anymore. Zoomers aren't really interested in this so nobody is using modern tech to update anything. EEs never really cared about software and especially about making it user friendly. Only programming language that matters is solder, computers lie to you, so you throw them off buildings
https://www.youtube.com/watch?v=vrZ_fMqd8k8
Simulations aren't really what you're looking for. Software will just bog you down and make you hate the hobby. You need to do this in real world, to touch, to smell (solder fumes), to measure... Use simulations when your thing is already working to tweak various values. You should already know the approximate answer before you run it. What you really need is to learn fundamentals behind whatever you're trying to do. Then it doesn't matter what shitty software you use. You can do your simulations in excel, matlab, ltspice, on paper with old HP RPN calculator or with a slide rule, that's how EEs do their work. If you want to be good at electronics you need to be more like Bob
https://www.youtube.com/watch?v=AEJtajaRj_s

>>2658021
I want to record analog audio on a CD-R, for fun. I don't mean sampled digital audio, I mean true analog.

The ordinary frequency of pit/lands on a CD is about 2 mHz.

The idea: could I just frequency modulate some audio to ~1.5mHz, put that signal through a comparator, and feed it to the laser on the CD burner?

On playback I would tap the signal from the laser pickup and demodulate it with an FM radio circuit, making my audio listenable again. Would this work? I can clarify any part of my plan

>>2658110
It's a dumb idea, but audiophiles would love it. I don't think it's really practically possible as there are so many unknowns to consider.

>>2658110
my guess is that it's going to sound shitty. the point of going digital is that you have a lot of leeway on how you encode your data. physical size and depth of individual pits, speed of rotation of the CD and all that kind of shit can largely be ignored because the data itself will still read accordingly. Same thing as a TTL signal being "high" from 2 to 5 volts, that allows "real" analog voltages to be waaay imprecise while still retaining perfect data. hell, you'd loose the error correction too (it might actually be needed for normal functions ? idk) .

I have absolutely no clue about optical storage systems but methinks engineers working on it wouldn't ignore the huge advantage of digital encoding, so there might be issues on playback because each drive don't spin exactly at the same speed for instance.

Also, i don't even think you can read/write analog values on a drive. You'd have to find out if the laser burning the data does so linearly for depths (as in, at 1V it burns 1µm deep then a 2V it burns 2µm deep) AND if the read is linear too. there might be some optical trickery since when reading 1s and 0s you just care for the timing and transients, not absolute values.

Keep in mind that you don't actually need to modulate in the first place though : you're not actually "shrinking" the length of the music by modulation, so if the CD has an usable length of 20 minutes you'd have 20 minutes of music even if you modulated to 2MHz. (you meant mega not milli right ?) so why not just skip the modulation ? also since the frequency is for digital data, it's completely irrelevant to your analog endeavors.

if you can figure all of that then yes, your project would most likely work.

visiting /o/tist here
are there any good and affordable instruments for measuring current and voltage over longer time periods
perfect would be a stand alone device with multiple input channels and clamps for current measurement so i don't have to cut into the wiring
smart phone connectivity would be great for a quick overview and .csv exports for detailed analysis
basically pic related but in one box
don't need extreme precision and willing to spend a few hundred but not thousands on lab grade stuff

>>2658110
>I want to record analog audio on a CD-R
You could use a lightscribe drive to burn the groove pattern/waveform into the label side.

>>2658110
>put that signal through a comparator
so you want to record digital music in the most complex way possible?

>>2658189
>/o/
dc? in a car? no its all shit. clamps dont work well on dc
ac in a house? search current transformer intstead of clamp meter

>>2658189
Would you be so kind as to describe what in the ever-loving fuck is going on in your pic? Pi GPIO pins put out 3.3V at up to 16mA. They aren't a power supply.

>>2658196
>dc? in a car?
yes, the ut210e i chose for my mock up works great for car stuff and is cheap as fuck
i don't need scope tier speed for current 1-10Hz is just fine
same for precision 10% is plenty low currents also aren't that relevant 100mA+ is fine
but i guess the option to set up in line measurements when needed would be nice

I have no problems with the measurements themselves what i am looking for is an easy solution to set up and record them
more software than hardware

>>2658110
That just sounds like a worse version of Delta-Sigma encoding, as used by DSD/Super Audio CD.

>>2658199
no power supply
gpio is just digital input to read and record the signal coming from the meter
18650s to power the pi
it's just a mock up but in theory that would be enough hardware

>>2658203
Try "data logging clamp meter".

>>2658204
can you recommend something?
they all seem to be stuck in the 90s regarding their logging, setup and interface capabilities
also a single channel costs like 300 bucks

I am basically looking for a company like Saleae but for clamp measurements not logic analyzers
someone who lives in the twenty-first century, offers easy to use software knows that digital power and memory cost nothing
and therefore sells a good product at a reasonable price and still makes money

>>2658208
>can you recommend something?
No, but here's a link instead:
https://www.ti.com/lit/eb/slyy154a/slyy154a.pdf

>>2658212
thanks that's a good read but honestly I am looking for a tool not a project

>>2658213
Sir, this is a /diy/ Wendy's.

>>2658058
I absolutely love LTSpice. I love it so much that's pretty much all I do. Yes the interface is not perfect but it is very easy to get used to. It is extremely powerful and can give you extremely realistic results if you care to learn it. The only side effect is I rarely even build anything since once I run a bunch of simulations and make sure everything works I lose interest. I generate Bode plots, check power dissipation, simulate noise, plot FFT, etc. what else do you need ?? It has a bunch of special models, like transmission lines, arbitrary signal sources, gauss noise, voltage/controlled voltage/current sources, parametric R/L/C values, and a lot more. You can simulate antennas, motors, 2D girlfriends, anything.

>>2658058
>simulate a ultrasonic transducer circuit that can emit and listen for return pulses

done it before.
i send 200V 100us pulse to transducer.
have high-gain amp and comparator across transducer.
a resistor and diode limits input to amp to 0.7V.
the amp detects the echo, and comparator converts to digital signal.
you need a monostable to create a dead zone that ignores comparator output for a lil bit after transmit pulse.
that lil bit determines the min distance to object, which was like 6 inches for me.
so effective range was aprox 6 inches to 10 ft.
to get lower than 6 inches you'd need separate send/receive piezos.

>>2658215
you still need tools for diy
and my hobby is diying cars not building measurement equipment
for example i'm in the process of replacing my mechanical power steering pump with an electric one
and it would be nice to measure and log how much current that consumes in different driving situations
so i can appropriately size the wires, fuses, relays and know how much extra load it puts on my system

>>2658220
Maybe a CAN bus logger? I have no idea, Anon. Somebody else will have a good answer eventually.

>>2658229
that will work only on newer cars and not for diy additions
is it really that weird of a use case?
I'd assume most electrical tinkerers and engineers need a easy to use low precision multi channel current and voltage logger to evaluate the performance of their projects
like if you were to build a 3d printer wouldn't you want to measure the power drawn by the various heating elements and motors over the course of a 20hour print

>>2658236
multi-channel DAQ

>>2658240
can you recommend a good one?
budget 200-300€ preferred but up to 600€ for a great product
at least eight channels four with current clamps
should be easy to use, phone app would be a great plus for quick'n'dirty field work but that is optional
but it must run stand alone without a pc connected and should have plenty of storage (ideally just a usb stick or sd card slot)

>>2658110
Very silly idea, could be fun. Biggest issue is going to be hacking the optical pickup head to feed asynchronous data that’s not Reed-Solomon encoded, while keeping all the handy bits like beam tracking and speed control. Read some data sheets, I think if you go for an older CD player you can probably find the encoding IC to be seperate from the beam control IC. Stereo is possible if you do the same trick as stereo FM radio.

If it’s too difficult, try doing it with a floppy disc instead. You can just make your own pickup for those, and by having two offset pickups you could probably turn it into a guitar delay.

Use a CD4046.

>>2658121
>>2658194
He’s doing frequency modulation of a square wave, not direct analog writing.

>>2658201
I gotta look into that. Delta sigma modulators are cool. Wonder if there are delta sigma class-D amplifiers?

god i hope it's real

xbox controller (model 1914), have joystick snapback overshoot problem, found a guide which said to add a capacitor, in the guide OP said he tried 0.1 uF capacitor, which did not work and then he tried 1uF which did work. Since I have only one try, should I go for 1uF or 0.47uF?

>>2658411
Just use logitech's controllers. All the pros use them.

>>2658411
I'd go for the higher one.
>Since I have only one try
What's stopping you from chopping one cap out and replacing it with another?

>>2658422
im on the left side of the bell curve so i have no idea about this stuff, but I have a buddy who works in electronics, he has all the tools and all the capacitors and the skills to do those mods correctly. But apparently his wife is due today or tomorrow, he will gtfo for a month when she gives birth
1 uF seems like safe bet, but I dont want to overdo it tho, then I couldnt play vidya and will be forced to do cringe stuff like socializing etc

>>2658439
>have a buddy who works in electronics

ask him to put two 0.47 caps in parallel with long leads.
if 0.94 is too much, just chop one off with nail clippers.

What is the turn on and turn off resistance in mosfets?
The turn on resistance is 2.3 ohms and the turn off resistance is 3.4 ohms of the mosfet I have
It's all the info I can provide

>>2658412
Too soon, anon. Too soon.

>>2658408
Yeah, I just auto-piloted to the store in my flying tesla using graphmeme batteries and paid for everything with NFTs. KY-69 is totally relevant to my interests.

>>2658412
This wouldn't have happened with madcatz controllers.

>>2658408
So far all of the academic attempts have resulted in failure, but one russian catgirl amateur has produced a levitating clump about the size of a grain of sand.

>>2658588
if all the reputable labs have failed then the russian catgirl, by ocams razor, is likely lying

what is your favourite bullet type?

>>2658408
My guess is it being real, but only at really low current densities / magnetic field strengths.

>>2658520
What? There's more or less an "on resistance" and maybe an "off resistance", but I've never heard of resistance being different during turning-on compared to during turning-off. Kinda hard to specify it since the whole point of turning it on or off is that the resistance is changing significantly. Or is it the effective resistance into the gate? If so, 2.3Ω sounds really high for both small-signal and power FETs.

Anyone else interested in LVDTs?

>>2658854
they fit on a rather skinny gun

>>2658859
It's a mosfet for a 300kW 3 phase inverter
I don't understand between which 2 points this turn on and turn off resistance acts and why they are different values

>>2658887
On resistance is milli-ohms, not ohms. More than likely. Yes?

>>2658890
I think I figured it out
It's about the resistance added to the gate to control how fast they turn on and off
The gate capacitance on these is almost 2000nC

>>2658897
Total gate charge *

>>2658868
Only just found out they existed. They look pretty cool though, what kind of use case do they have compared to linear magnetic/optical encoders?

>>2658897
Ah, so it wasn't a spec of the MOSFET but rather a spec of the required driving circuit. Good to know. Some drivers have one line for sinking and sourcing, while others have seperate lines. Even if they're the same line you can use a diode to have a different current in one arm compared to the other. In this case you could have a 3.4Ω resistor directly from the driver output to the gate, with a diode plus 7.1Ω resistor in parallel with that. Or use two diodes.

Do you think this is a cool kit that I bought? You can replace 'sensor' with whatever you want to trigger it on

>>2658929
Looks pretty neat, the price isn't even terrible by Jaycar standards. I'd just wonder about the strength of those outlets on the ~2mm plastic case.

>You can replace 'sensor' with whatever you want to trigger it on
I'm guessing it feeds a 12V line out to the sensor? It uses a 6 pin connector to the sensor, I wonder what the other pins do?

Could you post the schematic? Be interesting to see if the timer begins at the start of the input pulse or its end, to see if continuous retriggering is possible. I work at Jaycar but we don't have the schematics for that one available.
FYI there are some interesting items at Jaycar that have been discontinued and can be gotten for cheap, like the diy ultrasonic antifouling kit. I got one and will try to make an ultrasonic cleaner out of it. Also there's an IMU sensor module, a battery saver kit, and a "PCduino" SBC with accessories that include a SATA cable and camera.

wonder what a keyboard looks like if you use leds for the n-key-rollover diodes? seizure inducing as it scans, or would keys or entire rows/columns light up as you press keys?

>>2658938
>It uses a 6 pin connector to the sensor, I wonder what the other pins do?

i think it's so they could make it backwards compatible with an RJ12 cable.

IC1A in pic is a schmitt trigger

>>2658929
What does it do?

>>2658189
i actually designed actually what you are asking for as part of a uni project.

https://files.catbox.moe/y2eclq.pdf

>>2658200
https://thearduinoshed.wordpress.com/2020/04/13/uni-t-ut210e-current-clamp-hacking/

>>2658944
transformer and relay in a box. i think the schmitt trigger is so it can behave like a latching relay

>>2658946
you actually did that?

i don't really understand all this conductor shit, what would it mean if it turns out to be real?
cope is very cheap to make and has very low resistance per meter at room temp so not sure the improvement would be so big

>>2658948
>the schmitt trigger is so it can behave like a latching relay

the 2 cross-coupled NAND gates are a set/reset flip flop (or latch).
those dont need to be schmitts.
but the other two do need to be schmitts to correctly handle logic levels that have very long rise/fall times.
like the 10K slowly charging the 100nF cap on pin2.
or the differential spike caused by the cap/resistor on pin8.

>>2658942
IC1A is a schmitt NAND gate like the rest of IC1, they've just ignored pin 1. Chances are it's electrically connected to pin 2 or to +12V though, leaving a CMOS input floating is a recipe for unpredictability. By tying inputs together or tying one input to Vcc, a NAND gate turns into a NOT gate, as you can see with IC1c.

>>2658946
Not really sure what R8 and R18 do. Are they just for offset null? Better op-amp choice would fix that, though with such low voltage rails you're somewhat limited. Maybe there's a nice single RRIO chopper amp out there somewhere looking for a guy just like you. A dedicated instrumentation amplifier package might behave better if your resistor precision is becoming an issue. Is the dual gain stage due to limited GBWP per amp?

>>2658948
No the schmitt trigger is just for noise immunity. The latching feature is he SR-NAND latch built by IC1b/d. When you pulse pin 13 low it sets (turns on), when you pulse pin 6 low it resets (turns off). IC2 is a clock dividing oscillator that provides a delay between the latch turning on, and pulsing the latch's reset pin.

>>2658957
Did you mean to write "cope" or "copper"? Anyhow, it's a ceramic so it won't be being used for low-cost wiring anytime soon. Maybe inside transformers and other magnetics.

>>2658957
Let me list just one example above all this would allow.

Pocket sized portable MRIs.

Also all trains would be maglevs now.

>>2658965
>Pocket sized portable MRIs
it sounds like you're not familiar with the inherent current density limitations of superconductors. like, why do you think modern superconducting mri machines need to be whole metres across in the first place?
>maglevs
maybe, but if the demonstration video is anything to go by, the maximum magnetic field strength it can resist is pretty low

>>2658964
yes it's for the offset. They're potentiometers so the board can be mass produced for cheaper than it would cost to buy better op-amps.

the output of the 2nd op-amp is read when the device being measured is asleep and has a lower current draw.

The output of the first op-amp is read when the device is awake and has a higher current draw.

>>2658967
>why do you think modern superconducting mri machines need to be whole metres across in the first place?
because they need to be cooled to like -250degs so their supercdonductiors actually superconduct and that takes a ton of space

I have a midland 400 mxt that uses 40 watts of power. Was wondering how to power it in my 2009 ford f150. Tried using a piggyback and used the ignition switch feed fuse slot which is 15a. I have it grounded and everything plugged in but the radio didn't power on. That you guys might have a better idea if the piggyback is good or should I just hard wire it to the battery.

Anybody plays with e-ink displays? Can I cut it into smaller pieces and make each individual piece work?

>>2659013
>Can I cut it into smaller pieces and make each individual piece work?

It should work. I have used a skil saw to make small LCD monitors from one large monitor, and the concept is the same.

>>2659013
Are you making a e-ink flip book?

>>2658969
>supercdonductiors
ubersupraleitermenschen

>>2659021
Hardware time tracker like https://timeflip.io/ but I want to use some kind of no power display instead of lame stickers.
There's this project https://learn.adafruit.com/time-tracking-cube but it's too flashy and the only upgrade from the stickers is etched plastic parts.

>>2658642
well, it seems to be rather plausible. At least the very specific arrangement of the lattice.
https://arxiv.org/pdf/2307.16892.pdf
Can't wait for the debunked videos to begin posting focusing almost entirely on the published videos by the korean lab rather than the released(unpublished) papers.

>>2659005
If you cannot power it using the cigarette lighter then perhaps direct wiring is the best solution, make sure to FUSE IT if you power it directly off the battery as close as possible to the battery terminals since if you only fuse it on the radio side, a short in the wire leading up the line will cause a fire. Maybe also have some form of relay that opens/closes depending on when the car is on or off to prevent draining the battery when the car is off. You can do that by tapping into the voltage of one line you know is only "on" when the ignition is on, build the circuit with 1 relay, 1 transistor (or some buffer, used to not perturb the other line) to control the relay, 1 diode for relay back emf, 1 fuse and a few resistors for biasing transistor or buffer + wires from the battery and a wire that senses the voltage on a power line that is only present when ignition is on. This way, it prevents you from messing with your car's electrical system and accidentally damage it.never piggy back power from a circuit without knowing if it is safe to do since it might at best cause the fuse to blow constantly, worst case damage or shut down whatever is down the line. If that system is critical (for example, some temperature sensor to prevent the engine from melting, or airbag sensor) failure of such systems can be expensive or deadly.

One last thing, make sure the battery voltage is good for the radio. Remember, different car can have different battery voltages, and usually when the car is on, the voltage goes up due to the alternator, so if you design your circuit for say 12V, it might be 14V when the engine is running, so if that's an issue, you could also have some voltage regulator to drop it back to 12V.

>>2658969
> don’t need to be cooled.
Exactly. And all the power we need can be generated by the graphene batteries we’ve had for years by a Single “AA” cell 480Ah @ 10 kV.

I’m going to replace all my house wiring with superconductors to avoid those Cu losses… it can amount to ¼¢ per week!

>>2659005
Just splice off of the stock radio wiring harness.

>>2659041
You know that there are other things where properties like this are useful, right? Besides wiring. Also, it's either a ceramic or a crystal, so you really wouldn't want it in wiring.
think electric motors and tranformers/inductors rather than romex.

>>2659041
This whole "superconductor breakthrough" thing makes me think THE TRUTH IS OUT THERE until I remove the foil helmet. Call me a sayer of nay.

>>2659046
i bet back in time you would be one of the sheep standing in the crowd cheering as they were burning that one nigger who claimed the earth is round

>>2659054
That's a bit of an emotional response, Anon. What you do not smell is called hype powder. It is odorless, tasteless, dissolves instantly in sunlight and it is among the more profitable poisons known to le SCIENCE.

>>2659058
excuse me for being excited about the biggest invention since the atom bomb

>>2659062
Were you excited to get your Warp Speed vaxx too? Did you pitch a tent in your local CVS parking lot waiting for the midnight release? Will you test the hypotheses yourself?

I have an keyboard from the 90's the keeps blowing fuses. 3rd time recently. I replaced it with the certified kind, what else can be going wrong here?

>>2659093
Are there LEDs for caps lock and num lock?

>>2659094
I mean, a musical keyboard.

>>2659096
Are you able to use it for a time before the fuse blows, or does it happen immediately?

>>2659098
For a time. First time I got it, I was using it occasionally within the span of a year, until the fuse blew. Replaced it, it lasted a whole month. Now just today, it blew after a few weeks.

>>2659099
You have an intermittent short probably caused by a failing component or conductive path left by a liquid spill. Is there a battery compartment, and if so, are the battery terminals corroded?

>>2659101
Yes there is a battery compartment on the underside for AA batteries, there was some corrosion inside. How bad is that?

>>2659106
It grows like crystals, so you should open her up and check for damage spreading from that area of the PCB. Neutralize the corrosion with vinegar and a soft brush, then test for continutiy and resistance of any traces that look iffy. Finally, replace any components in corroded areas of the board and power on to test.

>>2659107
After the vinegar, clean it up with isopropyl alcohol.

>>2658216
thanks! I followed a couple of tutorials and now it seems realistic for me to use it to model large cicuits.

>>2658219
I dont want a single pulse. Wouldnt it then only work in a quiet environment? Can you show me the circuit you used?

Currently use pneumatic or water tubes split along its length and curled into a hole to protect cabling. Im not willing to make 15 mm holes for cable glands that accept 5 mm cables, also they cost a lot. Whats a better way to do this?

>>2659119
grommets

>>2659120
>grommets
i regret looking at the image search results to this

thanks

>>2658081
First of all I hate boomers. I learned almost everything I know from the internet after unlearning everything taught to me by boomers.
My goal isnt to be good at EE, my goal is to built devices for use in various robotics applications because I have a lot of ideas. I tend to learn the very basics of a given field and then build something, solving problems as I go. I simply dont have time to become some kind of an EE guru.

>>2659124
lmao
Sorry, should have said "wire grommets" or "rubber grommets".

>>2659107
Got it open. The fuse box is in the top left (bigger one). The motherboard doesn't go under the battery holder. The keybed is underneath and from what I saw, peeking without fully removing there isn't any extended corrosion.

>>2659127
Under the battery holder

>>2659127
>>2659128
Yeah, it looks undamaged. The next part to focus on is the power supply caps, diodes, and regulators. Start with the diodes. Another thing is to touch sus components to feel if they're burning hot.

>>2659131
Alright, stuff like that is out of my knowledge, I'd probably have to take it to a service center at this point, thanks for the help anyways.

>>2659134
Before you do, is there a hard disk connected?

>>2659136
There was actually, a floppy.

>>2659137
Disconnect it. It very well could be the problem. Something is drawing too much current.

>>2659107
> grows like crystals
Could be LK-99
I’m going to fire up Word, trot out my comic sand, and write a paper about it.

>>2659138
also

>VR5, C186 (F-1) provide +5Vana and VR2, CL88 (G-1) provide +5Vref; both circuits are sourced by VR7, C207 (+12Vana 0 TP23, D-1 & E-1).

>VR3, C193 (D-1) provide +12VDC for the optional hard disk at TP19 (H-1) and pin 1 of J18 (I-l).
R200 is required to lightly load VR3 when the hard disk is not installed; this prevents
oscillation.

>VR6, C206 (D-2) provide -12VDC for the audio circuitry at TP24 (E-2) and becomes -12Vana.

>VR5 and VR4 provide quiet +5 and -% volt power to the analog switches in the audio path. VR2 provides an isolated and quiet +5 volt reference to the A/D and D/A converters on the scanner microprocessor and several performance controls.

>VR1, C182 (D-2) outputs -ISVbias for the LCD Vee generator and muting circuit. This regulator
may cease regulating at low line voltage, however the only effects are a possible reduction in LCD contrast and premature clipping of large negative signal peaks.

>>2659138
Will try again with a new fuse, i wasn't even using that anyways.
I know the service manual says not to but would it be smart to use a higher spec fuse than stated? Recommended fuses are 250w 1 amp. As a matter of fact the first fuse that blew was a bit stronger, I think 1.5 or 2 amp.

>>2659143
Use the proper type and rating always, esepecially now. You can wire a 40W light bulb in series with the AC input so you have a visual indication of a fault giving you time to power off before the fuse blows.

>>2659142
5 . 1 . 2 . 5 DIGITAL REGULATORS

>The circuit of D25 (dual Shottky, B-2) and C202 (B-3) provide high current availability of +8VDC (unregulated) 8 TP17 (C-2) . This current source is used by the digital regulator circuit(s).

>Due to the lower forward voltage drop of a Shottky rectifier, less heat is dissipated by the
rectifier and the transformer voltage can be lower; however because of the heavy current drain on the +5 volt supply (up to 3.5 amps), it is necessary to heatsink D25 to the rear panel. C200 and C201 suppress rectifier switching noise.

>The circuit of U62 (C-3) and Q18 (E— 3, a VMOS power FET) makes an ideal pass transistor for a linear regulator and provides +5VDC for all digital circuitry at the output end of L5 (1-2).

>This circuit is wired as a voltage follower, thus the final output voltage at the source of Q18
is forced (by the gain of the op-amp, U62) to equal the reference voltage at its non-inverting
input. This voltage can be monitored at TP16 (H-2) . The output is adjusted at R180 for exactly 5.00 volts (C-4) to optimize performance of the custom VLSI chips on the Engine Board.

>It is extremely important for the Engine board to know when there is sufficient voltage for
proper operation so as to reset itself. The mute circuit (s) must also be asserted when there is
insufficient voltage. Simply put, when the output of U62 is within 1.2 volts of its supply
(close to its saturation point), Q16 and Q17 cut off saturating Q15 and discharging C176 quickly.

>The reset circuit (described later) immediately asserts RESET to all microprocessors and mute circuit(s) .

>R192 and C174 stabilize the feedback loop to prevent oscillation of the regulator. C178 and C180 improve regulation during sudden load changes. R193 provides a minimum load if the power supply is ever operated with no other components on the board.

>>2659145

5. 1.2.7 LCD BACKLIGHT POWER SUPPLY

The backlight for the LCD requires approximately 100 volts rms, at approximately 360Hz operation. The circuit of Q25 and T1 (G— 4 to H— 5) fulfill this need. The backlight cable must be plugged into J21 (1-5) for Q25 to oscillate. A reasonably good sine wave appears at J21-4. C208 and R218 provide the positive feedback route.

Q24 and surrounding circuitry provide a controlled higher voltage from the unregulated +8 volts so as to improve LCD brightness. The output voltage can be adjusted to balance between brightness and the acoustic noise level from the panel by changing the value of R216 (F-5).

>>2659146
5. 1.2.1 POWER SUPPLY

The rear panel and bottom shield are used as a heatsink for the five TO-220 packaged power
semiconductors. For testing in unrestricted free air, just the rear panel is acceptable as a
"short-term" heatsink although it will get hot to the touch. NOTE: TESTING WITHOUT A HEATSINK IS NOT ALLOWED AND WILL RESULT IN SHUTDOWN OR DESTRUCTION OF SOME OF THE POWER SEMICONDUCTORS IN A
MATTER OF SECONDS. The surfaces between the shield plate and the rear panel should be coated with white thermal heatsink compound for improved heat dissipation.

The rear panel is at analog/chassis ground potential; thus all power semiconductors must use an insulated mounting. The sketch below (figure 5. 1.2.1) details the mounting hardware necessary to accomplish this. When installing this hardware it is critical that the semiconductor mounting hole be aligned with the panel mounting hole. If they are not and the screw is forced in, the screw threads will cut through the wall of the shoulder washer and short-circuit against the semiconductor’s mounting tab.

>>2659125
>I simply dont have time to become some kind of an EE guru.
Then you definitely don't need to mess around with simulations. If you want to get useful/accurate answers from SPICE you already have to be EE guru. Just stick to circuitjs, it's plenty good enough for understanding the basics, unnecessary details will just bog you down. You should understand how circuit works before you put it in a computer. Pen and paper is still the best way to go about solving electronics problems.

>>2659037
I found a video, apparently when you use a piggyback you use the original fuse and one that is lower. Tried that out and it works. Yeah the power cable has three fuses, I tapped into the starter fuse which is 15A and added a 7.5A with the piggyback. Seems to work good.

>>2658216
One question I do have is regarding specific components, like, for instance, a IRF510 mosfet. If I do not find that specific model in the library, I have no idea how to proceed.
I figured out how to make a "spice directive" for that mosfet from the datashet. I also know that I can choose various mosfets from the library that have similar spice directives next to their name. How do I create a new mosfet with my spice directive next to it and insert it into the library? This should be straightforward, yet after 20 minutes of trying I am no closer to having my desired mosfet in the schematic. I dont know what people are using ltspice for if not for modelling real circuits that one can build by ordering components and wiring them together..

>>2659215

>>2658216
Where did you learn so much about ltspice?
I want to get a better job position so I also want to learn more about analog electronics
All of my projects are kind of based around MCUs so I wouldn't even know what to simulate...

>>2658903
>what kind of use case do they have compared to linear magnetic/optical encoders?
Well I'm told they're a popular sensor for aviation use since they're very reliable yet relatively precise devices that can be tailored to a variety of sizes/scales from microns to inches. The site I borrowed the pic from is attempting to use them as throttle and joystick sensors for flight sims. What I'm wondering is if a guy could make a simple shop micrometer with one. The commonly quoted linearity is 1 part in 400 (0.25%). So theoretically an LVDT with a 0.2 inch range could give an error tolerance of +-0.0005 inch, which is about the same as a typical shop dial indicator. I'm just curious whether anybody here has already tried something like this and can tell me if it's worth pursuing or not. If I can only get, say, +-0.01 inch tolerance, then there's no point when off-the-shelf tools already do better than that. But if I could get something precise enough to compete with commercial dial indicators, then it suddenly becomes a very useful tool for my purposes (small scale machining). I could, for example, build a Talyrond-style cylinder profilometer to electronically measure the true deviation from roundness of bearings, shafts, rods, and pinions.

>>2659306
I suspect attaining linearity along that range is going to be the tough part.

>>2658236
Why don't you just use pic related?

It's a known resistance piece of metal.
You have to put that in series with the conductor powering your steering wheel.
And then use a microcontroller's ADC to log the voltage drop across that resistance.
Write your software to convert voltage drop to current using V=IR

>>2659117
>Wouldnt it then only work in a quiet environment?

it only listens to 40Khz sounds, so it's always quiet unless someone is using a very specific dog whistle.

>Can you show me the circuit you used?

find some 20-yo schematics? would require a minor miracle.

>>2659124
>>grommets
>i regret looking at the image search results to this

try a different search term.
$5 for 100 sounds like a steal.

>>2659197
First of all, there are two types of models in LTSpice: intrinsic models (.MODEL) and subcircuits (.SUBCKT).
Intrinsic models are preinstalled and that what's you select from the list when you add a generic MOSFET device (nmos or pmos as opposed to a specific part).

They are typically found in lib\cmp. For example, all "standard" mosfets are an lib\cmp\standard.mos. I do have IRF510 in there:
.model IRF510 ...

If you don't have it, you can download the entire lib file from here:
https://ltwiki.org/index.php?title=Standard.mos

But you most likely have it too.
When you add a component, instead of searching for a specific part (which would be a "subcircuit"), just add a generic mosfet: either nmos or pmos.
Then right click on it, and select "Pick new MOSFET". Then select from the list of the intrinsic models (you may want to sort it). See pic.

>>2659197
Now, when you are adding a specific part by name instead of a generic "nmos", you are actually adding a "subcircuit". That's a third part library.
LTSpice ships with a bunch of 3rd party subscircuits, and you can also download individual models for specific devices from the web.
It could be one part per file, or a bunch of parts together.
It could have an arbitrary extension, typically .lib or .txt. But it doesn't really matter. The first line should look like this:
.SUBCKT RFP3055 2 1 3 ; rev 10/26/93
So you know it is a subcircuit, not an instrinsic model.
You can place it in sub or lib or any subfolder inside lib, for example: LTspiceXVII\lib\sym\EXTRA\Fairchild\NMOSFET\RFP3055.lib
Next, find the existing intrinsic symbol such as nmos.asy (it is in lib\sym) and copy it to the same folder where you placed the lib file and rename nmos.asy to your component name, ex RFP3055.asy

Open it and make the following changes:

1.change SYMATTR prefix to X
2.Edit value and description
3.Add SYMATTR SpiceModel <path>

For example:

SYMATTR Prefix X
SYMATTR SpiceModel ..\sym\EXTRA\Fairchild\NMOSFET\RFP3055.lib
SYMATTR Value RFP3055
SYMATTR Value2 RFP3055
SYMATTR Description RFP3055

SpiceModel path: relative to the "sub" folder
if the model is in the "sub" folder just use the short file name

Restart LTSpice and you can now add that part by name: Edit -> Component (or F2) and just start typing and it will show up on the list.

>>2659218
>Where did you learn so much about ltspice?
Autism. I spend too much time tinkering with it. It is addictive. I just added a frequency dependent resistor that simulates LC resonance but actually dissipates power, thanks to the laplace transform command. That kind of a "frequency dependent resistor" is (supposedly) used to model ferrite beads.
If you need something to simulate, simulate LRC circuits, and do an AC sweep to learn about filters and resonance and Bode plots, or simulate an amplifier.

so i was testing one of those fuckers and it suddenly died. now it only has resistance between legs 2 and 3 and leg 1 died and reads nothing. are they normally this fragile?

>>2659419
What currents were flowing during your "test"?

>>2659421
id say max 1A

>>2659433
>id say max 1A

that's crazy.
they're typically rated 1/10-th of a watt over the entire surface area.
and much smaller over a shorter distance.
making them adequate for milliamps, not amps.

>>2659433
They are not fragile, you simply destroyed it. 1mA is OK, 1A is ridiculous.

>>2659435
>>2659437
Oh no...

>>2659391
Can you explain why phase plots are so important?

>>2659524
Stability.

>>2659526
>>2659524
i've never used bode plots to check for stability. I use them to check what frequencies are suppressed or magnified

>>2658946
>>2658947
wtf anon that's perfect thank you so much
you don't happen to plan on starting a small business and sell a semiprofessional kit?

>>2659344
>You have to put that in series
i don't like hacking up perfectly good wiring just for a measurement
it takes time and no matter how much effort you put into getting it back together there's always a risk of introducing new failure points

my headphones were draining battery even when off so o opened them and on the pcb there is this weird white powder like substance, what is that?
i just poured some isopropil on it and gave it a good brushie brushie with my old toothbrush and it's gone so i hope i didn't break shit

>>2659540
Probably flux residue.

>>2659547
>image
I fucking love these.

>>2659536
you will have to resort to using a hall effect sensor then

>>2659551
There's a whole series of them somewhere.

>>2659556
Yeah, but it doesn't seem like there's a readily compiled album.

>>2659560
We all should collect them like Garbage Pail Kids and compile an album. Then upload somewhere and put it in the sticky.

I would also like to compile a "people intentionally burning themselves" album. Please submit all entries to /ohm/ as you discover them.

>>2659572
She is going to have some horrible burns on her thighs.
There's a reason most welders don't wear skirts.

>old thread gets archived
>switch to new thread
>3 days old and over 100 posts

>>2659614
144 to be precise. What's your point?

>>2659574

>>2659560

If I have a transfer function 1/(s+1/s) the usual way is to convert it to s/(s^2+1) to get rid of 1/s and infinity, but matlab can plot both functions just fine and the result is identical. What gives? If I plug in s=0 in s/(s^2+1) I get 0, not infinity.

>>2659547
I hadn't seen this one before; they're literally me.

>>2659616
>What's your point?
I snoozed and so I loozed.

>>2659626
Because it's the same function just rearranged
If you divide 0 by anything you get 0
How many apples does each kid in a group of 5 get when there's 0 apples?

>>2659645
>Because it's the same function just rearranged
Forget this part
Can you show your Matlab code?

>>2659616
Overlapping threads are a waste of energy. Am I supposed to jump to the new thread and miss old replies, or browse two threads for the same general at the same time?

>>2659574

>>2659626
>but matlab can plot both functions just fine and the result is identical
>plots identical functions
>surprised when result is identical
If you plug in s=0 into 1/(s+1/s) = 1/(0 + infinity) = 1/infinity = 0. There is nothing wrong if you have infinity in your expression. Only problem is 0/0 or infinity/infinity.

>>2659645
>>2659673
perhaps i am not making myself clear.
here is the expression:
g=1/(s + 1/s);
bode(g);
yes it is supposed to be infinity at s=0, and it is graphed correctly as shown in the picture.

my question is when i rearrange it as
g = s / (s^2 + 1);
it looks like i've just removed the infinity
since if i plug in 0 i get 0/1.
but bode(g)
plots exact same graph as the original function as if the infinity is still there. which is the expected result for this type of transfer function but I am not sure what's the point of removing the infinity if it is still there.

>>2659682
this is the graph of both functions

>>2659682
oh wait that's because log(0) is not defined?

>>2659684
Are you interested in the direct output of the model or its frequency response? The bode plot isn't showing you g(0)

>>2659682
What do you mean by "infinity still there"?
Also that's not how you rearrange 1/(s + 1/s)

>>2659694
I am not sure I understand your question.
I am just curious why
1/(s + 1/s)
s / (s^2 + 1);
produce exact same bode plot with an infinity point at at 10^0.

>The bode plot isn't showing you g(0)
I didn't say anything about g(0)
Do you mean s(0)?


>>2659695
>that's not how you rearrange 1/(s + 1/s)
how would you rearrange it?

>>2659705

They said many of you are scared of the other threads...

I'm trying to discuss "HV Switched Capacitor Converter" as potentially superior alternative to normal buck converters for where (fixed integer nominal ratio) non-isolating DC/DC can be used.

I.e., the equivalent of an autotransformer with a single turn secondary (or just generally an integer voltage ratio) but for DC.

Note, we're talking kilovolt here: the weakest incarnation would spit out 300V DC for a server rack (most universal power supplies don't require zero transitions for anything; when the voltage is high enough that the uneven heating in the bridge rectifier is covered by it also being sized to tolerate North American 110~120 V AC supplies) from a medium voltage DC distribution grid.

Think at least going for 3000V technology (switchgear (circuit breakers!), grid-tie rectifier/converter stations, etc.) on the HV side, if not something matching common MVAC distribution grids like 10~30 kV; that's for the 300V computer/server PSU target, not the SiC JFET stack with kV output and double digit kV input.

>>2659718
Initially it sounded like you were trying to siphon or t-off the rail lines to heat your house for free before winter comes in germany.

>>2659718
Are those relays? Will they get welded together by arcs? How long will they last?

>>2659718
>other threads
you are kidding. i never check the catalog because it is close to 100% bait threads.
/ohm/ is an island of wisdom and sophistication in the sea of retardation and low effort baits that is /diy/. it is /pol/-lite, too.

>>2659682
>>2659684
>>2659690
You aren't plotting 1/(s + 1/s), you are plotting 20*log10(abs(1/(s + 1/s))), where s=j*omega (imaginary number). If you take limit of said expression you get +infinity because of abs function, which is really a magnitude of a complex number technically speaking.

>>2659724
> relays?

No, that’s a conceptual diagram.
In practice we use SCRs.

>>2659724
They aren't (mechanical) relays!

Why do so many here assume that, instead of treating them as abstract unidirectionally-blocking&reverse-conducting switches?
Mentally replace them with MOSFETs, and you're only loosing out in the HV-stack-module internal construction of them when they need to block kilovolts. For the 300V output scenario they'd probably be 650V GaN "MOS"FETs from e.g. "GaN Systems" at least where they need to block double low side voltage (and the same or if cheaper less-voltage-rated ones for where they only need to block (single, not double) low side voltage).

>>2659775
>Why do so many here assume that
Maybe because you drew the symbol for relays with flyback diodes all over the "conceptual diagram".

>>2659734
Actually, here, for this to be superior to transformers, you need unipolar devices due to frequency.

Also, the topology requires turn-off-capable devices, so it's typically gonna be either lateral pHEMT eGaN (EPC, GaN Systems), vertical SiC MOSFETs (various manufacturers, only ~600V plus), vertical SiC JFETs (UnitedSiC; the competition died, but rumors say Cree/Wolfspeed is working on some JFETs), or a supercascode of such JFETs with a LV transistor at the bottom (such stacks have been demonstrated up to at least 30 kV, but they aren't suited to more than about a single MHz switching frequency due to issues of wide-band matching gate impedance for pushing unity current gain (at full amplitude, not just some weird bias point) beyond 10~20 MHz).

>>2659780
No, a relay would need the coil/actuator drawn. As-is, this is just a generic switch with an antiparallel diode (assuming conventional drawing orientation w.r.t. voltage potential; it's more about making clear that they are to act like having this diode): https://de.wikipedia.org/wiki/Datei%3ASymbol_closing-switch.svg

>You aren't plotting 1/(s + 1/s), you are plotting 20*log10(abs(1/(s + 1/s)))
Yes of course
>If you take limit of said expression you get +infinity
Yes but I am not questioning that. I understand why I get infinity in 1/(s+1/s).
My point is when I plot s/(s^2+1), I still get the same infinity even though there is no zero in the denominator anymore. But like I said it may be because if s=0 s/(s^2+1) = 0 and log10(0) is not defined so matlab converts it back to 1/(s+1/s) or something like that.

>which is really a magnitude of a complex number technically speaking.
well if you have a+jb then the magnitude is sqrt(a^2 + b^2) but if a=0 like in our case then it is just b.
So we are really plotting 20*log10(1/(w + 1/w)) or alternatively 20*log10(w/(w^2+1)
But since the X-axis is log10 too we are really plotting

20*log10(1/(10^w + 1/10^w))
and
20*log10(10^w/((10^w)^2+ 10^w))

w is purely real, there are no complex numbers in these functions.

>You aren't plotting 1/(s + 1/s), you are plotting 20*log10(abs(1/(s + 1/s)))
Yes of course
>If you take limit of said expression you get +infinity
Yes but I am not questioning that. I understand why I get infinity in 1/(s+1/s).
My point is when I plot s/(s^2+1), I still get the same infinity even though there is no zero in the denominator anymore. But like I said it may be because if s=0 s/(s^2+1) = 0 and log10(0) is not defined so matlab converts it back to 1/(s+1/s) or something like that.

So let me try one more time
abs(1/(s + 1/s)) = 1/sqrt((s + 1/s)^2) = 1 / (w+1/w).
So we are plotting
20*log10(1/(w + 1/w))

This is a purely real expression, no complex numbers involved.
If we take the limit of w->0 we get infinity.

Now if we convert 1/(s + 1/s) to s/(s^2+1) then
abs(s/(s^2+1)) = w/sqrt(w^4+1)
So we are plotting
20*log10(w/sqrt(w^4+1))

But technically since the X-axis is also log10 we are really plotting
20*log10(10^w/sqrt((10^w)^4+1))

correct?

>>2659798
>may be because if s=0 s/(s^2+1) = 0 and log10(0) is not defined so matlab converts it

But if 20*log10(10^w/sqrt((10^w)^4+1))
then it can't be zero because of 10^.
when w->0 this expression tends to 1/sqrt(2).

>>2659798
>I still get the same infinity even though there is no zero in the denominator anymore
It doesn't matter because numerator is 0. Limit when s->0 is still 0 (before log). Log is what causes infinity not the expression.
>and log10(0) is not defined so matlab converts it back to 1/(s+1/s) or something like that.
It's not that smart. Type log10(0) into matlab and see what happens. It is defined.
>>2659801
>So we are plotting
>20*log10(w/sqrt(w^4+1))
And you still get infinity at s=0 (because of log). So plot is exactly the same.
>>2659812
It's not an exponent axis. It's log-log axis. You need to plot with log10() instead of 10^w. Then you get your infinity/sqrt(infinity^4+1)) which is undefined in this case. It's best to stick to 1/(s+1/s) form in this case because 1/(log(s)+1/log(s)) = 1/(infinity + 1/infinity) = 0 and then if you log it again for y axis you again get infinity.

>>2659798
Just to make sure that you're reading your plot >>2659684 correctly. That peak at 10^0 is not infinity. 10^0 = 1, s=0 is not even present on that plot. s=0 is infinitely far away to the left. At s=j*1 you have j/(j^2 + 1) = j/(-1 + 1) = j/0 = infinity. Then everything checks out because 20*log10(infinity) = infinity. +Infinity happens at s = j*1, -infinity happens at s=0.

>>2659526
care to elaborate on that? i have a high gain amplifier i've been dragging my feet on for a while now, i'm afraid to start cuz i know it'll be a nightmare to debug

>>2659826
>i'm afraid
Imagine Oprah Winfrey over one shoulder and Mike Obama over the other, and they're both telling you to "be better".
hope and change
class and grace
cock and balls
gayniggers from outer space

babe new electronicos fantasticos instrument just dropped:
https://youtu.be/GMD2DcWGKh8

>>2659815
>It's not an exponent axis. It's log-log axis. You need to plot with log10() instead of 10^w
No, first of all it is not log-log, it is linear-log.
20log is what makes it log-log, but the actual Y-axis is linear. But the X-axis is log10 which is why you need 10^x (the reverse of log) to make it look logarithmic on a *linear* axis: for example if you use desmos which doesn't support logarithmic axis. But of course you don't need any of that in matlab since you are plotting in a log space. Anyway I know it is confusing but it's the way it is.

>>2659826
Only important for feedback.
If you get negative (over) unity feedback with a 180° phase shift along the feedback loop (amp and feedback network), your negative feedback for converted into positive feedback, sustaining oscillation.

Just do a higher-order low-pass for your feedback and you're typically in easy mode (for the amplifier stability; the feedback network itself suddenly left "trivial", though :( ).

>>2659842
That's Buckethead's nemesis Fortunecookiehead.

>>2659857
>Anyway I know it is confusing but it's the way it is.
It's not confusing, you're just making it confusing. If you want to plot data on logarithmic axis x axis that is displayed on a linear x screen you need to take a log10(x). Just take actual numbers as an example
you want to plot x
>1 2 3 4 5 6 7 8 9 10
on linear axis, but you want to compress it so you take a logarithm of x
>0 0.3010 0.4771 0.6021 0.6990 0.7782 0.8451 0.9031 0.9542 1.0000
Notice proper spacing as expected from log axis. Values get more dense as you approach 1. If you take your 10^x you get
>1.0000e+01 1.0000e+02 1.0000e+03 1.0000e+04 1.0000e+05 1.0000e+06 1.0000e+07 1.0000e+08 1.0000e+09 1.0000e+10
which is completely wrong. You're expanding in x axis to 1^10 instead of compressing as you should. Imagine if you're plotting frequency and you hit 100 MHz. With your method you're going to plot that point at x = 10^10^8 units away from 0,0 vs. just 8 as it should be.

My house has some wall mounted speakers, i'm looking at this combo to enable bluetooth and change to some up to date speakers.

Amp: https://pt.aliexpress.com/item/1005004605000268.html?algo_exp_id=a0485c7d-eb34-4e35-966d-939e4fc9d85c-2

BT receiver: https://pt.aliexpress.com/item/1005001771487719.html?gps-id=pcDetailTopMoreOtherSeller&_t=gps-id:pcDetailTopMoreOtherSeller,scm-url:1007.40050.281175.0,pvid:747e41c6-5cc2-4ef3-a02c-516891822687,tpp_buckets:668%232846%238116%232002

Should i go for it?

>>2659869
That's not what I mean.
Here, take a look. This is a simulation of the body plot of a low pass filter on Desmos which doesn't support logarithmic axes.
Notice how the X-axis is compressed since every value is a power of 10. That's why you take 10^x.

>>2659934
Compare with the actual Bode plot done with matlab. Same powers of 10.

>>2659574

>>2659815
>It doesn't matter because numerator is 0. Limit when s->0 is still 0 (before log). Log is what causes infinity not the expression.
Oh, so you are saying log10(0) is actually infinity? I tried log10(0) and the answer is -Inf. Wow, that's unexpected. But why is a negative infinity? Also if log10(0) then why is 20log(0) results in some large but finite positive number on the Bode plot?

>And you still get infinity at s=0 (because of log). So plot is exactly the same.
Again, it looks like +Inf on the Bode plot, but log10(0) is -Inf. Weird. Do you get -Inf or +Inf if you try log10(0)?

>>2659815
Wait no, it is all wrong, there is never log10(0), because 0 is never passed to log10, because the X-axis is strictly positive: it starts with some tiny positive number which is a negative power of 10.

So infinity happens BEFORE the value is passed to log10. It doesn't have to be s+1, it could be any positive number which is the resonance frequency w of a parallel LC circuit. For example, 1/(s+w^2/s) will hit infinity at the resonance frequency w, BEFORE log10 is evaluated.
Here is an example of 1/(s+10000/s). g = infinity at w=100. And g=s/(s^2+10000) produces exact same result. So somehow s/(s^2+10000) equals infinity when w=100.
This doesn't make sense unless s/(s^2+10000) is evaluated as 1/(s+10000/s);

>>2659574
i have such a subcategory in my "/ohm/ op images" folder

>>2659574
another

>>2659900
Instead of wasting money on bluetooth, just take a hammer and in-wall speakerd. Same effect.

>>2659218
Not the quoted poster, but also quite comfortable with ltspice.

Look up the hotkeys, and you'll be a lot quicker when working with it. Practice modeling a bunch of textbook circuits and experiment with the various plots. After this you'll be able to work quickly in it, and thus much more productive when working on large circuits.

>>2659934
Ah, I see, I was thinking of doing conversion the other way around where you start with x,y vectors that hold values of function and then scale x with log so you actually plot log(x) y. You are starting with log(x), y and then doing conversion in function.
>>2659965
Yes, this is exactly what I meant by >>2659823
>>2659968
>So somehow s/(s^2+10000) equals infinity when w=100.
This is where complex number comes into play. We are evaluating s=j*100, imaginary part is important because j^2 = -1. Then you get j*100/(j^2*10000 + 10000) which is j*100/0 or infinity.

HOLY SHIT IT'S REAL LK99 JUST GOT REPLICATED
https://twitter.com/andrewmccalip/status/1687405505604734978

>>2660124
WE HAVE LIGMA-99
AAAAAAAAAAAAAAAAAAAAAAAAA
(gentle fart)

What is the circuit diagram of this flimsy keypad? I see 9 traces in the flex cable. There are 6 rows + 4 columns therefore it should be 10?

>>2660125
shut the fuck up you dumb fuck you are so dumb you have no idea how huge this is we are literally about to have fucking flying cars and hoverboards

>>2660124
That's an inchworm.

>>2660131
Maybe I am dumb, but the material needs to be tested. Let us wait.

>>2660124
So how do they know it's not just some diamagnetic material like pic related? Just because it floats on strong magnets it doesn't mean that it's also superconductive.

>>2660137
because it also works when you turn the magnet around

Fucking hell. What could be causing that power drain? When the headphone is off the battery drains slowly over the course of the day, no doubt there was some corrosion caused by sweat and rain but i cleaned out all the rust on the pcb and it still drains.
So I tried again today, i drowned it in isopropil and scrubbed the shit out of it, but the pcb was already clean so not sure if that will do anything. Also i managed to rip one of the tiny wires off so i soldered it back on, but if it still drains then i have no idea what could it be and i'm probably fucked.

>>2660139
Orientation of a magnet doesn't really matter. Iron isn't repelled if you turn your magnet around, it still sticks to a magnet regardless if you point N or S pole at it. Same with diamagnetic materials, they'll always be repelled no matter how magnet is orientated.

>>2660142
You are missing the key observation here my dear Watson. The inch worm in the video is fully erect, one side attracted, other repelled. If you try that with a piece of iron it will simply lay flat on top of the magnet.

>>2660137
There's more to this than just meissner effect. If anything, the videos do it a disservice (just look at Dave from EEVblog's imbecilic take) as the isolated samples are very likely not pure enough and sputtering it onto copper then making the copper wiggle about may not have been the best idea without also showing the copper by itself to have something to compare to.
There are two ways to approach it, one is in simulation as it was done in this paper >>2659034 and the other is to take actual measurements for magnetic fields and resistance.
>>2660139 diamagnets are polarity independent. If anything that's their defining feature.

>>2660146
>Dave from EEVblog's imbecilic take
>I can wiggle a piece of copper clad FR4 with a magnet, therefore the entire experiment is busted! Click on my video!

>>2660151
daves argument was legit
they tried to showcase sc on top of a copper plate when the magnet will wiggle the copper plate itself making the whole demonstration meaningless

>>2660153
He's starting to backtrack on his meme video. Yes, the videos may have been a disservice, the copper deposited one more than any of the other two, but there's still a fucking paper with measurements attached to the entire thing.

>>2660156
>here are people expecting me to debunk this superconductor video when i know nothing about superconductors
>so heres this video i made about debunking this superconductor
thats it, i'm done with that aussie kangaroofucker, not only does his voice sound like a rusty hinge but now hes making clickbait videos like some common free energy device pajeet channel

>>2660163
All he’s saying is that it’s obviously due to the eddy currents and the people behind the video claiming it to be superconducting are stupid, liars, or both.

I’m sorry to those that lost all their money on monkey NFTs, but you probably won’t make it back on this lk-99 scam.

>>2660163
You know, the problem with getting your hopes up is that they'll most likely be crushed. It's typically a youthful trait. You will become more cynical as you age unless you're an absolute retard. People are full of shit more often than not. You'll see it eventually (hopefully).

>>2660163
Don’t worry, elon musk is going to take credit for inventing it, rename it ‘X-69’, offer to buy it for 40 billion, and with that, promise to go to mars and make floating teslas by next year, then cash out more pumped stock. Then go nowhere and do nothing except move on to the next scam.

>>2660172
Enthusiasm really has nothing to do with this. There's hardly anything emotional beyond just pure curiosity. If you can't dissociate your emotions at the very least don't project that onto others.
I understand where dave is coming from with regards to how embarrassing it is to have the copper substrate video front and center, but the way he's handled some minor and quite valid criticism is asinine.

>>2660175
>Dave noticed something
>Dave made a video about it
barkbarkbarkbark
>If you can't dissociate your emotions at the very least don't project that onto others.
lmao

>>2660174
> floating teslas
Does it cost more to re-make all the roads and interstate highways out of neodymium or LK-99?
Just asking the question. Probably will cost at least $100, no?

>>2660005
the speaker is already installed

>>2660065
I am not sure if we are talking about the same thing, but basically I see what is going on. I am retarded. I made a few mistakes when converting from g(s) to abs(H(w)).
We have g=1/(s+1/s). Since it is SS s=jw.
1/(jw+1/jw). And this is where I missed j in the denominator which means we change the sign: 1/(jw-j/w) = 1/j(w-1/w). And finally abs(H(w)) = 1/(w-1/w). So we get infinity at w=1.
And likewise
s/(s^2+1) = jw/((jw)^2 + 1) = jw/(1-w^2). Same infinity. And same mistake of not changing the sign.
And it also checks out from the point of view of a parallel resonance. wC and 1/wL canceling each other out so it is open circuit = infinity at 1/sqrt(LC).

>>2660065
I see, you probably said the same thing here >>2660214
but you are evaluating in terms of g(s) and passing imaginary values but I just thought it would be easier to first properly convert g(s) to H(w) and thendeal with the real values. But either way it all checks out. Thanks anon.

Any recommendations for a book/textbook/youtube series that would get me enough electrical knowledge to become an electrician? I work controls and I think I know a decent amount but I want to fill the gaps that are definitely there. I don't plan on passing a cert or switching to electrician work as a career, I just hate not knowing what TTHN or TFFN wire is, and there's definitely a remedy that doesn't involve apprenticing myself again. Electroboom is nice to watch while I eat dinner, but he's not going to teach me everything.

>>2660235
Just find a copy of your local code/standard. Here in Europe we use IEC 60364 so we can use something like https://www.electrical-installation.org/enwiki/Main_Page for reference.

>>2660235
This board usually has way too many threads about electricians and electrician wannabees and electrician haters, and you managed to miss all of them. We don't pull cable here.

>>2658854
ones that dont corrode (nonexistant)

>>2660001
russians get a pass

>>2658021
>https://github.com/Rocheez/4chan-electronics-challenges/blob/master/list-of-challenges.png
roll

>>2660129
>There are 6 rows + 4 columns therefore it should be 10?

no.
physical rows and columns dont matter, only electrical ones.
there's 17 buttons, so min number of lines is 9.
if there was just one less button, smallest matrix would be 4x4 (8 lines).
if you have the patience, you can trace it all out and figure out the exact matrix.
post result here, and win a cookie.

I decided to repair my old MP3 player, specifically the battery didn't work, so I wanted to buy a new one. However, upon inspection I noticed that this battery has 3 leads (black, blue and red), while available batteries all have 2 leads. Have any of you had to solve such an issue? Will the device work correctly if it's of identical voltage, without connecting the blue, temp sensing lead? How do I solve this?

>>2660140
It could be that one of the ICs has gone bad and in that case you're fucked unless you wanna bother with BGA soldering

>>2660362
>Will the device work correctly

no one can predict such a thing until they've tried it on that particular device.
chances are 93% it'll work fine.
if not, what you can do is measure resistance between the temp wire and ground (or temp wire and positive) and replicate that with a resistor.

>>2660140
>What could be causing that power drain?

an old/cheap battery will self-discharge.

>>2660374
Thanks, I'll get on it.

>>2658219
I thought I'd contribute somewhat to this; given that this is between audio and switchmode converters, and that diode trick offers log10(200/0.7)*20=49.1dB maximum attenuation, I'd expect a hybrid with a simple small ferrite (or powder) toroidal core to work, especially if you throw a decent matching network on the piezo side of the hybrid.

No idea how sensitive the matching would be to changes in acoustic impedance on the phonon side of the transducer.... any guesses, by chance? Pretty much just out of curiosity (not that a full-duplex ultrasonic phased array wouldn't be cool)...

>>2660214
Damn and technically it is still wrong, 1/(w-1/w) should be 1/abs(w-1/w). Same with jw/abs(1-w^2). I assumed that sqrt(x^2) = x but it is wrong, still needs abs(x)

>>2658189
Can you find multimeters with USB or w/e for which you also find an open source program to run on computer for data acquisition (not just an SDK, unless you're fine writing the logger script yourself)?

If so, take Pi-or-so (something low idle power with USB that can run your software), 12V-to-USB power adapter that tests as not causing low voltage moment whenever you're cranking the engine (hook it on some line from the fuse box that isn't disconnected by the ignition, so almost certainly not the cigarette lighter socket; dashcam-that-records-while-parked installation guides should help), USB-hub/USB-to-serial, and your army of multimeters.
Make it auto-start on boot and ideally arrange for an off-button that triggers an orderly shutdown (literally normal behavior for ACPI power button on PC, might need mapping to an available button); or if you have WiFi, make it just connect to your network (or phone hotspot, if you need away from home) and remote login (remote desktop or ssh console as appropriate and easy) to turn it off (and possibly even turn logging on if you don't want to configure it to do that as autostart).

>>2658247
Honestly this concept sounds amazing :D

You're AFAIK limited by minimum pit length due to recorder spot size along the track, but unless clock recovery (and thus length quantization) is integrated so tightly you can't really use the read/write head tracking without clock recovery, only the AGC (if even present! i suspect it only affects the threshold the read binary quantizer uses) significantly limits your run length.

As for ΔΣ Class-D, that's AFAIK the normal way (perhaps with minimum run length tactics in case of power stages that suffer too much switching losses to just ignore this aspect).

>learning how to use multimeter
>measure AA battery as a test
>set meter to 200m DC amp setting, which I assume is 0A-200mA
>reads 50.0 (0.05A? Apparently AA batteries typically have 0.02A-0.03A, but it's not an outrageous difference)
>set meter to "10" DC amp setting which I assume is 0A-10A
>reads 5.00, which would be 5000mA
I figure I must be either doing something wrong or understanding something wrong.
I tried the fused 500mA port instead and it gave me no reading at all on the 10 setting and an out-of-range error on the 200m setting. Ordinarily I'd just assume that trying to read very small amperages on the 10A setting just results in a screwy reading and leave it at that, but I get the exact same discrepancy when trying to test a DC power adapter that's supposed to deliver 4A, and I don't know whether the adapter is faulty or not because I don't know which reading to trust.

>>2660452
RTFM

>>2660452
You never mentioned a resistor in series and its value which would make it clearer what kind of readings you are supposed to get.
You are not measuring current across the battery terminals, are you? You would have blown the 200mA fuse.

>>2660472
I did. It says the resolution of the 10A setting is 10mA, but would that mean a measurement of 5.0 would mean 50mA? If so that still doesn't match up with the 200mA setting which, with a resolution of 100uA, would be giving 50.0 = 5000uA = 5mA
>>2660477
>You are not measuring current across the battery terminals, are you?
I am, I just put the probes on either end of the battery.
>You would have blown the 200mA fuse.
I don't think it has a 200mA fuse, only a 500mA one. In any case, the fuse is still intact.

>>2660408
Maybe there’s some sort of self-mixing you can use? Or a conventional mixer, where you change the IF between transmission and reception.

>>2660436
It probably uses clock recovery to ensure the motor is spinning the correct speed for X bit-rate. But then you could measure motor speed as your audio output. The only issue would be writing a variable frequency in the first place.

Also some brief reading suggests that most (especially high-power) class-Ds use conventional triangle-wave PWM, apparently they’re better for noise and switching-loss reasons. Single-chip low-power class-Ds though may use delta sigma.

>>2660525
A: the first thing they teach you in multimeter school is to never make current measurements in parallel with something. That’s how you blow a fuse. If you absolutely have to measure the short-circuit current of something (it usually isn’t very good for that something), consider the maximum possible current you think it will draw, and set your range switch above that. If it’s more than 10A use a clamp meter, and a welding mask. Otherwise, you’re measuring the current drawn by some load, in which case you put your meter in between the load and the source, in series not parallel.
B: the measurements on the screen follow the SI prefixes on the range switch. The decimal point is put there for a reason. The minimum resolution is just saying what the smallest unit it can differentiate is.

Watch a video on using a multimeter. Stick to measuring volts and ohms for now, they’re more useful in general.
The second thing they teach you is not to stab the probes inside your body, FYI. Number three is arc_flash_fatality.wmv.

>>2660525
You are short circuiting your battery and drawing significant current which could easily blow a 500mA fuse unless the battery is drained and therefore has high resistance. I guess you could measure a fresh battery directly for a couple of seconds if you need to check the max current or something. An AA battery would probably draw a few amps briefly so your 5A reading on the 10A setting sounds about right. What are you trying to measure?

>>2660548
>never make current measurements in parallel with something
If I'm just connecting the multimeter to either end of the battery, wouldn't that be in series rather than parallel?
>the measurements on the screen follow the SI prefixes on the range switch
That was my original assumption, but then doesn't >>2660452 still apply as far as the readings of different resolutions not matching each other? Or are all of the current readings unreliable anyway because there's no load on the circuit I'm measuring?
My assumption was just that the multimeter itself would pull the maximum amount of current that the power source was capable of providing, up to what the meter could handle.
>you’re measuring the current drawn by some load, in which case you put your meter in between the load and the source, in series not parallel
How do you use a meter to test the output of a power source/adapter that's designed to be plugged into something else, though? I can shove the positive probe inside the barrel connector and then touch the negative probe to the metal bit in the center of the power port on the device, but what do I use to complete the other half of that circuit?
Or am I out of my depth and I should already have things like a spare barrel connector with stripped wires and one of those cables with small clamps on either end and use those to run the circuit back to the connector of the adapter I'm actually testing?

Even if my meter's fine, what are the odds that I've already damaged the DC adapter by shorting it multiple times in trying to test it? I did notice that I could hear the brick quietly buzzing every time I connected the probes.

How do you mount flange mount connectors like this one? Does the flange go on the inside and then bolted to the chassis from the outside? Can't find any examples.
And what's the purpose of these connectors as opposed to simple through hole threaded connectors?

>>2660551
>which could easily blow a 500mA fuse
The fuse is fine (at least on a visual inspection, the wire inside it is unbroken) but I'm mostly using the unfused 10A port anyway since I wasn't getting any reading on the 500mA port. The battery is brand new.
>What are you trying to measure?
Well the battery was just because I knew very little about using a multimeter and figured it'd be a good low-power power source to mess around with, but the actual reason I pulled it out is because I've got a 12V 4A DC adapter that appears to not work (when plugged into a 12V 4A DC monitor, can't power it on for more than a split second before powering off again, whereas a different DC adapter I have works fine with the monitor.) However, the barrel connectors of the two adapters are slightly different and the one that works gives 4.5A, so I was wondering if maybe the "faulty" adapter was actually fine and simply didn't fit well enough in the socket or didn't provide quite enough current or otherwise didn't work for that device specifically for whatever reason.
Thus I thought I could use the multimeter to measure whether it was actually delivering a constant 12V and 4A like it says it should. Like I said in >>2660555, due to not knowing shit about using multimeters, I have apparently already shorted the DC adapter multiple times in attempting to test it, via shoving the positive probe into the end of the barrel connector and then touching the negative probe to the outside of it. Like the AA battery, reading on the unfused 10A port, the meter's 10A setting read something like 5.50 and the 200m setting read in the 50.0-55.0 range. I still don't know how those two can match up with each other considering 1A is 1000mA, unless both readings are just wrong because I'm doing things wrong.

>>2660557
They're for very small items that you don't want to use up space inside with a nut. see pic related.

I'm going to be making my own antenna soon so I want to hear you guys opinions on element material. The internet is full of pissing matches about how/what to use as the element

Copper is generally agreed upon as the best most affordable material, but it will corrode fast. steel sounds nice because it's cheap, but it rusts and has a bad skin effect. Aluminum is "Okay" but because it's very corrosion resistant people recommend it.

Any reason I can't just use copper and spray paint it with protective enamel? At the mention of painting antennas most forums turn to shit flinging instantly.

>>2660557
The screw hole chamfer is on the front side, so the while flange goes on the outside.
Mfg put it inside because it’s mounted to the pcb for mfg so you don’t see the through holes. Sometimes the through holes are inside, mounted on an inner metal chassis wit a false plastic front that varies depending on model.
Rf connectors need good shielding.

>>2660555
You should never attach your current meter directly to a voltage source. You are short circuiting it and drawing huge current because the ammeter has very low resistance. And no your meter doesn't have any current limiting capabilities except for the fuse but the 10A input is unfused.

>How do you use a meter to test the output
What do you mean the "output"? If you are measuring voltage, you measure it directly. But if you are measuring current, you attach the actual load whichever is supposed to be plugged into that adapter, for example your phone, and then attach the meter in series, between the voltage source and the load. Or you could build a simple circuit that consists of a battery, a resistor, and an LED. And then attach your current meter in series to measure current drawn by the LED.

>>2660561
Again, to test your DC adapter you need to attach a real load to it, something that draws power, not your meter. And then you attach your meter in series with the load.

>>2660567
Everybody is using LK-99.
I spent all my savings and took out a loan and reverse mortgage to pre-prder 1 lb of it on release day from mt. Gox.

>>2660565
>>2660569
oh, cool thanks. it looks so aesthetically pleasing like most RF stuff.

>>2660570
>What do you mean the "output"?
I mean the voltage and current that it's supposed to be capable of delivering.
>you attach the actual load whichever is supposed to be plugged into that adapter
I don't know how I'm meant to do that when all I have is the meter, the adapter, and the device. At least not without tearing apart the adapter's wiring so I can stick the multimeter between the two halves of one of the wires. I don't have anything necessary for building test circuits aside from some spare solid-core wire.
Also even if I did cut up the adapter's cable to test it, is it possible that I damaged the adapter by shorting it multiple times earlier and now risk damaging the device by plugging it into the adapter to put a test load on it?
I likewise don't have anything simple that's meant to draw a load for test purposes like a DC motor or light, I'd have to just plug it into the monitor.

>>2660567
Shid I thought this was /ham/

>>2660567
Maybe an oxide layer or electroplating for corrosion resistance.

>>2660582
You could also just tin the copper element.

>>2660572
Would a couple of AR-15's work as a dipole? I wonder if chrome lining in the bore could be an issue.

>>2660555
>wouldn't that be in series rather than parallel?
Well it’s both, since there’s only two components they’re the same. But it’s definitely parallel.
>Or are all of the current readings unreliable anyway because there's no load on the circuit I'm measuring?
I don’t know why it’s giving those values in the manual, could just be a misprint. But there is a load on your circuit, it’s the multimeter. It’s draining the maximum amount of power that thing can provide.
>My assumption was just that the multimeter itself would pull the maximum amount of current that the power source was capable of providing, up to what the meter could handle
It does not, the multimeter is a dead short in current mode. There’s basically a wire connecting the black and red probes together, plus a fuse.

>How do you use a meter to test the output of a power source/adapter that's designed to be plugged into something else, though
It can be tough, worst case you need to get an extra male and female plug to put the meter between. You connect one side together (e.g. negative load to negative power supply) and bridge the other side using the multimeter (e.g. positive load to negative DMM, positive power supply to positive DMM). You may be able to do this with just an alligator clip lead or three, some plugs may require a paperclip jammed into a small hole. Watch that you don’t short anything though.
Some people might sell easy current measuring jigs for an assortment of plug types.

>what are the odds that I've already damaged the DC adapter
Modern DC adapters are generally pretty tolerant of that, usually they’ll just detect that and shut off. But I wouldn’t do it too often. Measure it’s volts, it should be in spec.

>>2660602
I just realized the correct term for the power brick I'm messing with is "AC adapter."
>there is a load on your circuit, it’s the multimeter. It’s draining the maximum amount of power that thing can provide
The maximum amount the AC adapter can provide, you mean? Wouldn't this mean testing current via shorting would work, even if it's not great for the device? The meter can handle up to 10A on the unfused line, and all I'd want to find out is if the adapter is capable of putting out up to 4A like it says it can.
If it's giving me a correct reading though, then I still don't know what the reason is for the mismatch between the 10A and 200mA range outputs is. If the circuit were giving me 4-5 amps in the 10A range, shouldn't it show an out of range error on the 200mA range instead of reading 50mA?
>Measure it’s volts, it should be in spec.
I did measure its volts and they are in spec, even now.
As far as it failing to power the monitor, do you think it's more likely that it's faulty and not providing the 4A it's supposed to (which I don't have the setup to verify,) or that it's fine and the monitor despite being listed for 4A can't make do with just exactly that much, hence the 4.5A adapter working fine? Or the barrel connector doesn't fit quite well enough or something like that?

>>2660608
>The maximum amount the AC adapter can provide, you mean?
>Wouldn't this mean testing current via shorting would work, even if it's not great for the device?
Yes, though more so for dumb things like batteries or solar panels rather than power supplies. Power supplies can provide significantly more than their rated maximum for a short period of time, so for them it’s not really useful to test their short-circuit current. You’re better off testing them with a variable load, which you can set to draw a certain amount of current. Poor man’s solution is to get a power resistor of the right size (R = V/I), or a length of nichrome wire that you change the resistance of by moving alligator clips. I do this with the wire wrapped around a wooden jig and immersed in water for long-term testing.

>shouldn't it show an out of range error on the 200mA range instead of reading 50mA?
No clue, you should measure a current with a current limiting resistor to see if it gives correct results at all any more. You may well have damaged it.

As for the monitor, open it up and measure the voltage on its power socket while it’s plugged in.

>>2660611
>open it up and measure the voltage on its power socket while it’s plugged in.
The shell is snap-fit rather than screwed on and looks like it would be a pain to dismantle, so I think I might just put a note on the AC adapter and deal with it in the future, since I've already got another one that works for the monitor.
For future reference though, would it be possible for the voltage on the adapter itself to measure fine when under no load, but the voltage across the terminals of the device, when plugged in and under load, to fluctuate or otherwise not be up to spec?

>>2660627
>would it be possible for the voltage on the adapter itself to measure fine when under no load, but the voltage across the terminals of the device, when plugged in and under load, to fluctuate or otherwise not be up to spec?
Absolutely. It’s not a common failure mode, but it certainly can happen. Especially with older PSUs with ageing passive components, or ones that have been run hot for too long.

>>2660608
>not providing the 4A it's supposed to

you're missing something important.
a thing that's rated 4A at 19V means it'll provide up to 4A while maintaining 19V, and not get too hot over time.
you CANNOT test for that by shorting it with a meter, coz then it's putting out 0V.
what you do instead is hook it up to a 4.5ohm load, then measure voltage to see if it stays at 19V, and check temperature after 10 minutes.
(ideally you'd also look at the 19V using a scope to check if the output voltage isnt super dirty.)
and where you get a a 4.5ohm power resistor? two toasters in parallel.
or, instead of toasters, you can use the original load.
cut up an extension cord to have access to the wires, measuring voltage across the 2 wires, and current in series with 1 of the wires.
power supplies, especially china ones, do go weak-in-the-knees: i've had 3 diff ones go weak in last 6 months.

>>2660629
>Especially with older PSUs
Just to confirm we're talking about the same thing, by PSU you're referring to pic related, right?
For me that term makes me think of the internal component in desktop computers that hooks up to the motherboard and everything else.

>>2660658
yes

>>2660580
>is it possible that I damaged the adapter by shorting it multiple times monitor.
Possible, but a robustly-designed adapter would have short circuit protection.
>I mean the voltage and current that it's supposed to be capable of delivering.
Those are two things are measured differently: Voltages exist between two points- 'across' components (parallel). Current is a measurement of charge through a single point; It has to be measured 'in-line with' (series) a load.
>I likewise don't have anything simple that's meant to draw a load for test purposes like a DC motor or light
Then you're gonna have to cut up the power cord to wire that multimeter in series with a load. The multimeter has to 'interrupt' the circtuit to measure current.

>>2660567
Depends on frequency.
Aluminium is fine if longer than your forearm.

Dielectric corrosion proofing on copper is allowed if thin. If it stays on, "waterglass" should work well.

>>2660557
Purpose: shielding. Also you now have a flail/mace to swing at 5G hoaxers.
Teach them how 5G towers can hurt them.

I'm powering an led strip with a metal housing through a 20V laptop power supply
Why can I feel electricity when I touch the housing of the led strip?

>>2660790
Capacitive byass for EMI reasons. Try flip plug in main socket if not polarized.

Why are there no SMA jacks on signal generators and similar instruments? All of them seem to be BNC. Not N, not SMA, not TNC, always BNC. Not very diverse.

>>2660807
It depends on frequency/wavelength and transmission cable characteristics.

>>2660794
Capacitive bypass from where to where?

>>2660807
>All of them seem to be BNC.

a BNC jack will outlive both the instrument, and you.
SMAs are made to be smol not tuff.

>>2660817
>from where to where?

google Y capacitors.

>>2659734
Can I use pic related to make a coil gun that is "totally legal" in EU? With like fuckton of electrolytic capacitors for those joules.

>>2660819
Isn't the N-jack the master race that will outlive anyone else?

>>2660828
>N
What did they mean by this?

>>2660831
It used to be called a slave connector. People didn't like the connotation so it was changed.

and that's your /ohm/ trivia for today kids.

>>2660827
PIC related is kinda rather meh for that task, they do not allow nice fast turn-on unless you trigger via extremely sudden overvoltage spike so fast the thing (Thyristor) turns on before current actually flows, because current crowding (more current less voltage drop!) otherwise blows a crater in whatever part turns on first.

Also they are literally like 2-3 inches diameter.
Go to where you can discuss the macro particle accelerator stuff.

>>2660827
every euro cunt has different gun laws and coil guns are guns in name only, they sure as hell aren't firearms. I doubt any country has laws about what's currently a toy or a science experiment.

>>2660883
You'd be surprised about places that call barrel-possessing macro particle accelerators guns unless the device has a good excuse for why it's not gun despite having gun functionality.

>>2660851
Good point.
Is this also why master is called initiator in the SCSI parlance? As in "the master initiates sex with a slave"?

>>2660903
Something like that, yes. Also, in the book "Roots", Kunta Kinte gets whipped with a 68-pin Centronics cable as punishment for putting a peanut butter and jelly sandwich in the tape backup drive.

>>2660790
you sure it's connected to ground? laptop chargers have some weird pinouts

test

anyone here privy on how keyboard circuits work? i found a mechanical keyboard at a goodwill, and it seems pretty good except the p key types this weird shit on every strike. holding it down repeats the p key, but the first strike always does this weird shit. its slightly different from time to time, and it happens on both windows and linux so im convinced its electrical. i took out the circuit and cleaned it but no avail. i can replace components, i can drill the pcb, i just need to know what the issue is.

Are all electromechanical signal relays low capacitance, < 1pf? I checked my scope and it has a bunch of NEC UD2-4.5NU relays. Are those special low capacitance relays? Contact capacitance is not even mentioned in the datasheet.

>>2660959
They use a row x column matrix, pic related. The rows are pulled low one at a time, and when they are, if a switch on it is pressed, the corresponding column will go from Vcc to ~0V. This "qwertiopuy[]\+" could be all the keys on a certain "row" or "column" of the matrix. I use quote marks since they don't necessarily line up with the physical rows and columns of keys. The only thing that comes to mind is the pullup resistor for P's column being open-circuit, as that would cause it to think all keys on that column there are being pressed at all times. That doesn't make complete sense, but it's not outlandish to think that pressing that one key there is what pushes it past the threshold of floating vs "low" for a CMOS input. Why just P? Maybe its diode is dead, or maybe something is shorted to between P's switch and diode.

I think it has to be sequentially seeing each of those keys as pressed (e.g. they're sequential rows in the same column as per pic related), since that's more that 6 keys, and the rollover limit of a conventional USB keyboard is 6. Though there are ways around this, which you could likely verify via a (command line) USB device inspection to see if it's actually a USB hub or whatever the trick is they use.

Also check what happens when you exceed the rollover limit with keys other than P.

Aside from visual inspection (or thermal camera inspection) the only troubleshooting method that will help will be to hook up a logic analyser to the row/column contacts and see what makes P different from other keys. Using a scope would (and would be much better for finding non-digital faults) work but could take longer, and using a DMM might possibly work if you have one with min/max features and a lot of time. Getting the probes clipped on nicely might be a pain.

>>2660973
thanks for the post, fren <3
i googled a bit about keyboards after i posted and i found matrix keyboard diagrams, and i realized that there were some other keys that were fucked. i traced it back to a shorted resistor and replaced it, and now those keys work. i got lucky, it could have easily been the FET above it. also, i cant read SMT resistors, but im like 98% certain the chinks just put the wrong resistor on the board. some poor sap probably got the lemon and just threw it away to goodwill.
unfortunately now the space bar doesnt work for some fuckin reason so i have to fix that now.

>>2660980
fixed it, i bent one of pins for the switch when i was reassembling. we're so back.

>>2660980
>>2660989
cool

Will a power diode bridge save two unevenly charged batteries from excessive current when wired in parallel?
Would they both need an individual charger?

>>2661042
>save two unevenly charged batteries

no.
think of it from the point of view of the power source.
the loads are in parallel, so it just sees one load.
and dumps as much current as is demanded.
if one of the loads is sucking it all in, source has no way of knowing.

>individual charger?

yes, or a charger that monitors individual cells, and can do cell balancing.

>>2661042
>save two unevenly charged batteries

no.
think of it from the point of view of the power source.
the loads are in parallel, so it just sees one load.
and dumps as much current as is demanded.
if one of the loads is sucking it all in, source has no way of knowing.

>individual charger?

yes, or better yet, a charger that monitors individual cells, and can do cell balancing.

Seriously WTF, is this diagram wrong?

I'm trying to set up a DRV8825 for a stepper motor. I accidentally connected the ground wire to the FLT pin instead of the GND pin when I first set things up, and got a voltage of 0.59 when I tested the ground against the potentiometer. Then I realized my mistake and moved the ground wire to the GND pin -- and I get zero voltage flowing across the chip no matter what I do. Moving the ground wire back to the FLT gives me a voltage reading of 0.59 off the potentiometer again.

Is this diagram mislabeled? Am I supposed to connect the +3.3V supply voltage somewhere other than the two pins (labeled SLP and RST, shown with orange "wire") on the left of the diagram? Or am I supposed to hook up the motor power to VMOT and GMOT (top right two pins) before testing?

Here's the tutorial the image is from:
https://www.rototron.info/raspberry-pi-stepper-motor-tutorial/

>>2661173
Try connecting the EN pin to 3.3V.

>>2661173
I don’t know too much about this, but I do know that it’s a bad idea to try and let the raspberry pi power other devices…. It barely has enough power or itself.

>>2661173
https://www.researchcell.com/electronics/how-to-use-drv8825-stepper-motor-driver

>>2661181
Thanks. So it's supposed to be pin 9 on the corner . . . why am I not getting ANY voltage when I do it right?

>>2661176
Are you sure or are you trolling? This guy's >>2661181 webpage says that the EN pin is "enable" and that setting it high (putting voltage to it) disables the board. Quote:

> Enable Pin (Setting it high will disable the board)

>>2661179
>bad idea
Might not be ideal, but the Pi is literally designed to power other things. This is just to power the chip, the motor-driver voltage is a separate power supply. running at 24V.

>>2661189
>why am I not getting ANY voltage when I do it right?
In the diagram here >>2661173 you have the reset and sleep pins pulled high.

>>2661189
>why am I not getting ANY voltage when I do it right?

it looks like the chip is powered from the 12V motor supply.
would make sense if it's expected to power the motor with 12V.
do you even have that connected?
$1,000,0000 says you dont.

Hello anons, I am modifying controller board for my 3d printer, kingroon kp3s pro.

signal pin on jst connector for chipset stm32f103 clone, GD32f303, and
>pic related

in order for me to use serial I have to remove R33, then bridge, and given this schematic, is that all I need to do?

>>2661230
Just solder a jumper wire and leave the resistor.

>>2661235
Thanks anon.

>>2661239
No problem.

>>2660861
What about IGBTs or new fangled SiC fets?
>Also they are literally like 2-3 inches diameter.
This is not that much big of a problem for home security device, or even car security device.
>>2660883
Even if I mount it on a tripod?
>>2660887
It is a nail gun, officer, eco-friendly powder-free prototype.

>>2661189
> pi to power other things
I was having problems because my pi wasn’t even booting sometimes.
I was using a dollar store 8 foot cable to connect it.
I switched to a tiny 4 inch in-pocket charge cable and the problem went away.
Never figured out exactly why, but I figured the Pi was at nearly at it’s limit and booting maybe took a bit more power… just enough to cause the boot to fail (and after that it was off, by the way… it wasn’t just an error)

>>2661228
>picrelated mfw
Thanks. :-(

>>2661268
Yeah, they do that. But literally right now the Pi is sitting idle and acting as a dumb power supply, all I was doing was trying to set the potentiometer for the correct voltage (which is how you set the maximum current delivered to the motors).

BTW, the cheap shit Chinesium potentiometer screw is stripping out even with a good screwdriver. I'm just gonna have to hope the pots are set to a reasonable value already, because there's no way I'm going to be able to change the setting. :-/

>>2661289
You could wire an external pot to adjust, then desolder the pot and solder a fixed resistor of the same value.

Calculated a black hole collision (526 dBm) into a 50 ohm resistor and got: 4.461542169214018e+25 Volts. That's a huge current. Could it destroy the Earth through a 50 ohm resistor? On the hand, the air resistivity is pretty high, and the break down voltage from 1cm is 21.1 KV. Would 4.461542169214018e+25 Volts be sufficient to break down air all the way from the black hole and what happens in the vacuum?

>>2661042
If you're leaving them in parallel and they're the same chemistry and voltage, then you don't need to worry about balancing. Even if the batteries have different capacities. But you need to worry about it when connecting them together in the first place, a big spike of current can flow between the cells. I'd put a power resistor array between them so they equalise at a controlled current. Say, if one battery is 12V and the other is 12.5V, and they can handle 10A, you could pick a 5W 0.1Ω resistor to let 5A between them until it's mostly equalised. Then you can leave them connected together. If your batteries have balance lines, you also want to connect these in parallel.

>>2661173
If the enable pin doesn't have a pull-down resistor, you may want to wire it directly to GND. Also check that your stepper current is set to something other than 0. I'd also check that your GPIO toggling is working properly.

Apart from that, I guess it is possible that tying the GND line to FLT could have fried something. The DRV8825's FLT pin is the open collector of a FET, so assuming it floated high (or is pulled high by a resistor on the PCB), connecting it to the pi's 0V pin would mean that all other signals (STEP, DIR, M0, etc.) would go 3.3V above the DRV8825's internal 3.3V rail. These higher voltages may well have damaged the DRV8825. There's nothing in the datasheet that mentions protection or TVS diodes on its pins, but I imagine they have something along those lines. In that case, it's also technically possible that you damaged the digital outputs of the Pi as well/instead, so definitely check the voltage levels of those GPIOs in particular, both when connected to the DRV and when disconnected.

>>2661330

>>2661334
Can't see a pullup resistor on FLT. Also it looks pretty easy to measure the trimpot set voltage.

>>2658021
Hello. I have been directed here by this post >>2660547 which is in response to this. >>2659975
Basically I have a FRS walkie-talkie radio which in order to broadcast you hold down a button. You can only broadcast by holding down the button. If you need to broadcast hands-free, you need to clamp down the button with a rubber band or a clamp.
BUT it has a programmed in failsafe in the software that keeps it from broadcasting too long. Even if you hold the button down with your finger or clamp it down, it beeps and shuts itself off after about 1:30 to 2 minutes. Which is screwing over my tests.
So someone told me I could possibly hook up some other kind of switch that would work around this software timer. He said this:
>digital switch like a “4066”
So, I'm ignorant to circuits and wires. I'd be willing to open the radio up to make this work, I just don't know how. What is this switch that anon is talking about and how can I use it to get around the software limit?

>>2661341
why you want make radio jam, anon? why you knotty boi?

>>2661343
Not a jammer, no. I don't think I've learned enough about it yet to make jammers.
I've tested a bunch of FRS radios by having a friend just sit around and wait for me to radio back and have him tell me how I sound as I ride a bike around the city. But he got annoyed by that and I got the idea to place a song on loop on a phone next to the radio while I ride away on my bike. The continuous music signal made for a MUCH better test. I could just passively listen to the music while I rode by different blocks. It worked great! I could practically make an exact propagation map of my city this way. But, only for two about minutes before the radio back home just shut itself off. Which is the issue I'm truing to solve.

I realize this is more of a electromechanics and power electronics topic, but can someone tell me how I can safely, reliably and cheaply close a circuit at the ms range without much arcing happening ? Can a solenoid achieve that ?

>>2661341
>I'd be willing to open the radio up to make this work

that'd be a mistake for a noob.
it's more than 100% certain you'd break something if you tried it.
especially if it's a water-resistance model.
first thing i'd do is use a diff FRS radio.
I have a few and some allow for continuous transmission in the settings, even without holding the PTT button.
if that's a negative, consider designing some kinda cammed motor or solenoid-based gadget that will press button for 1 min then release for 1 second.
(the 4066 solution is feasible, but requires the kind of special skillz that only me and my retarded twin sister have)

>>2661353

oh, yeah, another solution is a cell phone and Zello.

>>2661341
To actually figure out if you can bypass the timeout feature, you'll need a schematic. I'd want to find one online, reverse engineering is not trivial. Anyhow, chances are it's impossible and you'll just have to switch it off again for a fraction of a second once per minute. That kind of timeout will be embedded in the firmware, so a fix to that would be obtained by flashing custom firmware. That's a hardware hacking question for /mcg/ or somewhere on /g/. Just finding the programming header may or may not require a schematic, but you'd need to open it up, identify the chipset, and look for JTAG or whatever headers.

>>2661350
why not solid state?

>>2661347
Use the output from a tarduino with blinky code modified for 60 seconds on, 1 second off to toggle an optocoupler that switches the transmit button electrically through the solder contacts.

I'm making one of pic's motherfuckers, but with these motherfuckers instead of optical sensors or potentiometers: https://www.melexis.com/en/product/MLX90393/Triaxis-Micropower-Magnetometer

It should actually be pretty straightforward, thanks to the magnetometers working just fine as 2d joysticks (though with limited throw). What I need is some design inspiration for what to make this magentometer six degrees of freedom input device look like.

>>2659160
you mean like this?
>+12V -> original fuse ---+----> new fuse -> radio
>.......................................|
>.......................................|---> original circuit
or like this?
>+12V ----+--> original fuse -> original circuit
>.............. |--->new fuse - >radio
if it's in the first case, that's not a good way of doing it. Since now the original 15A fuse must provide not only for the original circuit, but also for your radio. This means the 15A fuse before the 7.5A fuse will constantly blow when the radio is in use. Increasing the original fuse to 22.5A is also a dangerous idea since it no longer protects the original circuit and can potentially be a fire hazard upstream if the wires are not large enough for the current. If you want a simple way to install your radio, the way you did it might be ok if the fuse isn't tripping, but if you want the right way of doing it, try to check if there are some open slots in your fuse box, if it does, wire some 12V directly from the battery to the fuse box, then use your 7.5V fuse in the unused slot. From there take that 12V and use it to power your radio. To prevent the radio from being on when the car is off, you could have a relay that turns the radio on only when the ignition is on, and off when it's off. The relay itself is controlled via a mosfet that has it's gate tied to the ignition circuit switch negative side. When the switch is on, that node becomes 5V, which turns on the nmos. If you attach the relay coil to +12V from the fuse box and the to the drain of the mosfet, you now have it energized whenever the ignition is on, which will power the radio. LMK if you need circuit diagram and parts.

>>2661441
Can that move in 6 DoF? 3 translation and 3 rotation? If so, the orientation and relative strength of the sensor should give you enough DoF, but you'd have to arrange the geometry so they're sufficiently independant. The mathematics to extract both translation and rotation from the sensor data might not be easy. Having three of the triaxis magnetometers spaced 120° around the magnet's idle position would definitely work, but is probably overkill.

Don't forget a calibration/auto zero function.

I don't get ground as a return path. How does the charge carrier have enough oomph to return the wave across such a huge resistance (not-very-good-conductor surface-of-earth materials)?

How does resistance drop with increase surface area? Is it because of the geometrical advantage of a spherical charge's field-effect on the charges around it making an additive wavefront?

>get led lamp with fixed brightness
>want to make the brightness adjustable since it can be a tad to bright
>figure out the chip on it is an LM3414
>it has a pin specifically for PWM dimming
>painstakingly solder a tiny wire onto the solder pad peeking out from underneath the WSON package
>solder more wires for power and ground to spare pads
>start repurposing an old MCU board I made
>hook the MCU up to my programmer and start thinking about the code
>get back to double check the connections before I solder the dim pin to the microcontroller
>dim pin is showing 12V, not the 0V or ~5.4V I'd expect it to present
>look at board
>wire snapped off and is sitting in a +12V via
>tiny pad is gone, no copper of that node remaining
Well RIP that idea I guess, shoulda super-glued it in place.
The only other method is to feed current into the current set resistor pin to change the current amplified by the current mirror. That or replace the current-setting resistor with my own pot+resistor combo. Which I can do, but to get the right curve is proving tough.

On that note, anyone know how to get a concave resistance curve from a potentiometer and fixed resistors? Everything I've tried with one pot and one resistor has resulted in a convex curve. No, the traditional "log pot from linear pot + resistor" method doesn't apply here.
Not sure what it will do if I add a potentiometer contact to ground.

>>2661534
>enough oomph to return the wave across such a huge resistance

pretty much everywhere, if you dig you hit water.
water makes earth a reasonable conductor when you stick a pole a few feet deep.

>How does resistance drop with increase surface area?

super obvious.
electrons repel; they hate being in groups.
if you give them room, more will spread out and join the flow.
same way a larger magnet will have more field lines.

>>2661549
"reasonable" in what ohmage?
Electrons don't "hate" shit. Stop anthropomorphizing them.

>>2661548
>how to get a concave resistance curve

https://www.amplifiedparts.com/tech-articles/potentiometer-taper-modifications

>>2661551
>"reasonable" in what ohmage?

very variable.
but low enough that in some rural areas, a ''single wire earth return'' system is feasible.

>>2661554
What makes it feasible or not? Low resistance? The higher the resistance in the ground, the more losses in power, right? Someone kept saying "oh, but there are many concentric rings of return path, which lowers resistance." Literally how?

>>2661555
>Literally how?

earth is massive and so is capable of providing a near-infinite number of return paths.
these paths, being in parallel, reduce overall resistance, even if only a tiny bit as you get further away from point of contact.

>>2661552
I already said:
>the traditional "log pot from linear pot + resistor" method doesn't apply here.
I need a logarithmically variable resistance, not a logarithmically variable voltage divider. You see, if you have a series potentiometer + resistor circuit, the total resistance is equal to:
>R_pot*(1-x) + R_pot*x || R_fixed
>R_pot*(1-x) + (R_pot*x*R_fixed) / (R_pot*x + R_fixed)
Which graphs as so:

>>2661555
Parallel paths make for a lower resistance. A 1mm*1mm*10km length of dirt has a million times more resistance from one end to the other, compared to a 1km*1km*10km length of dirt. Assuming you can make good contact to the entire square kilometre of surface area at either end.

>oomph
Ultimately what matters is the effective ratio of resistances. Current will flow the entire way around a circuit regardless of its resistance, it's the amount of current that flows (and the resultant voltage across your load) that changes. If your load is effectively 10Ω and your return path is 1000Ω, then you only get ~1% of the power deposited in your load, the rest being dissipated by your shitty return path. 99% of the voltage is dropped via that return path, and your load only sees 1% of it. On the other hand, your load is effectively 100Ω and your return path is only 1Ω, then the return path will only drop 1% of the voltage.

>>2661560
EVERYTHING has near-infinite return paths. The air has infinite return paths. That makes no sense.
Those paths can't reduce resistance in parallel until you tell me how the fuck it works. From the electron doing the pushing, how is the resistance lower? The electron doesn't take every path simultaneously.

>>2661421
>why not solid state?
Not cheap enough for the 1kV 500A range

https://www.europe-solarstore.com/hoppecke-12-opzv-solar-power-1500-48v.html

Why waste thousands of euros on this when a few car batteries and an inverter do the same job?

>>2661565
Sounds like you don't really understand Ohm's and Kirchhoff's laws.

Ohm's law's vector form is particularly useful when dealing with bulk resistive solids:
>J = ρE
Where J is the current density vector (amps per square metre), ρ is the resistivity of the solid (ohm metres), and E is the electric field strength vector (volts per metre). The J and E vectors are both going in the same direction, and since the lines of constant voltage are going to be perpendicular to the E field, they're also perpendicular to the current density. So you end up with graphs like this, where the current flows easily through the bulk of the material. By flowing easily, I mean the electric field strength is low, meaning little voltage is dropped per unit length. But close to where the points contact the surface, there's less and less area for current to flow, and so there's more and more voltage drop.

>>2661566
Preventing 500A arcing is really difficult. If you can guess at the system's total effective inductance, you can get an estimate for the arcing energy, and decide what sized snubber network you'd need to absorb energy before it would significantly arc. There are some options for arc mitigation, like the arc dampening vanes seen in circuit breakers, strange gases or liquids to put between the contacts, applying a strong external magnetic field to force the arc to move via lorentz force, using liquid metal contacts, or even just opening the contacts extremely fast. This last one could be done by forcing an extreme spike of current into the soldenoid of a contactor. High dI/dt through an inductor necessitates a high voltage, so I think you could use something along the lines of a flyback transformer or ignition coil or whatever to produce that.

You may want to consider valves/tubes like thyratrons, though I'm not sure if they can be turned off.

>>2661484
>Can that move in 6 DoF? 3 translation and 3 rotation?
Yes, the physical travel is very small but you do get true 6dof motion on space mice. This kind of project has been executed before using three 2-axis pot-driving thumbsticks before, the only innovation here is swapping the cludgy interface around the thumbsticks with the three magnetometers, each reading off one paired magnet. You actually have a lot of usable translation space available, it looks like you can use a full 5mm radius travel zone around the magnetometers and still get good tracking.

Why do chinese call all transistors a tube?

I'm looking for a connector to attach wires to a 1/4" aluminum rod, preferably in a way that is removable and has a good contact, and requires no solder. At first I was thinking of using spring clamps (Picrel) but they don't make them in aluminum and I'm worried about galvanic corrosion over time. Preferable if it's removable, but a single time fitting would be good. Anyone know anything that fits the bill? Basically my idea was if I could get something like picrel in aluminum and solder the wires to it, then you can just clip it around the aluminum rod. I might just be searching for the wrong type of electrical connector. Any help on what to look for?

>>2661784
>removable and has a good contact, and requires no solder.

you can buy or make all kinds of little clamps from any material you like.
but surest way to connect anything temporarily is to drill a hole, then use a nut and a bolt or two.
use nylon screws if worried about metal-to-metal interactions.

>>2661791
Oh, that's a good idea. I didn't think about that. I have a drill press, I guess I was looking for the easiest way without having to drill each one, a clip is quick and easy. A shame they don't make aluminum versions. This will all be sealed, so I'm not sure it's even that big of a problem between aluminum and zinc

>>2661768
supposedly jfets behave almost the same way, just at lower voltage obviously

>>2661830
yeah, but does it make it legal to sell class-D amp with mosfets on the output as a tube amp?

>>2661837
no lol but you get what you pay for dealing with chinks

Has anyone itt ever built a DAC for high impedance headphone use?
I'm thinking of taking this on as my first 'real' project that isn't just fixing/breaking something but I do probably intend to use chinkshit electronics

>>2661616
>>2661484
>The mathematics to extract both translation and rotation from the sensor data might not be easy.
That seems to have been addressed by others but it was the first serious question I asked myself. You can do matrix math with a known body and how it moves across three planes to calculate roll, pitch, and yaw. I'm pretty sure that's what the triple thumbstick solution used as well: https://www.youtube.com/watch?v=Vs6Xte4gOxk

>>2661996
>You can do matrix math with a known body and how it moves across three planes
That works when your potentiometers are nice and linear (e.g. 60% voltage = 60% of the travel), but when your magnetic field strength is a function of the 3 dimensional arbitrary distribution of magnetic field lines coming out of the magnet, it may not be trivial at all. Well, worst case you have to do a bunch of trial and error to build up a lookup table or approximate equation.

>>2662001
>when your magnetic field strength is a function of the 3 dimensional arbitrary distribution of magnetic field lines coming out of the magnet
That's why you use three separate magnets, and calibrate each matching sensor to their relative fields.

>>2662013
>picrel
Are you just going to assume that the magnetic field strength as a function of angle is a trig function? Because my point is that may not be the case. You have the compounding effects of distance AND angle, so you may well need to map out a phase space of magnetic field strength with respect to those two variables. For each sensor. The best method would probably be just to simulate the field strength from a simple radiation pattern and make a lookup table or equation from that. Tweak the values until it feels sufficiently linear.

Anyone with experience in analog electronics? I have a circuit where I have a variable voltage A, and a variable voltage B which is always less than A. I need the result to be (B/A), where the output is basically "what percent of A is B". I don't mind if the output is PWM or a varying voltage.

Thought of using a triangle wave generator that stops increasing when it reaches voltage A, and a comparator that compares B to the triangle wave. I've also seen other designs using a voltage multiplier, and another using variable gain.

>>2662059
The original solution is using log and antilog amplifiers to realize function
>exp(log(B) - log(A)) = B/A
Depending on how much bandwidth you need you also have an option of using MCU with ADC and DAC as ICs are so cheap nobody bothers doing everything in analog, because analog is hard to do right and it's also expensive. $0.70 STM part can do what you want on multiple channels and can do a lot more than just B/A.

>>2662059
The triangle wave idea is pretty neat, especially if PWM is fine for you. All you'd need to do is ensure your triangle wave is nice and sharp, and that your voltage thresholds are well defined, which isn't trivial but is fairly easy.
If you need a variable voltage then you'd have to filter the PWM, necessarily making a gap between the response frequency and the triangle wave frequency.

The alternative is using an operational transconductance amplifier, mixer, or cascaded log and exp amplifiers to perform some sort of multiplicative operation. These will be higher fidelity, but generally these are comparatively expensive and hard to tune to be nice and linear throughout the entire range. Log/exp amps are going to be the cheapest but they're not trivial to tune, and can suffer from issues as the intermediate signal strength varies as a function of temperature. OTAs like the LM13700 are easier to work with and not too pricy, though I've never used one so I couldn't say how easy they are to get linear. You might actually need a 1/x curve on your gain control input, not sure.

What's it for?

>>2662086
Trying to make a simple pen-plotter using only TTL chips and linear ICs found in the 70s, for an electronics challenge. The idea goes that:

>send two 15-bit numbers over serial (X & Y) plus a "pen up/down" bit.
>digitally compare the new coordinates with the previous coordinates to determine directionality
>Subtract the smaller number from the bigger number to find net movement magnitude
>Do the division thing to determine how fast one of the motors go in comparison to the other one
>plot the line, and count the motor rotations until it reaches the magnitude limit
>ideally the other axis will have travelled the correct distance as a ratio of the bigger axis

This is why I dont mind PWM. I'll be driving the motors with it anyway. I thought of the triangle-wave method, but with lower magnitudes the PWM frequency would increase, which isnt ideal for motors. Thought of just low-pass filtering the earlier PWM output into a stable 20khz triangle wave PWM generator.
>inb4 just do the division digitally
Fuck no, space is at a premium for ICs, and I'm not smart enough to make a good hardware divider for 15-bit inputs.

>>2662095
>send two 15-bit numbers over serial (X & Y) plus a "pen up/down" bit.
How do you read and write this signals? What parts do you use? Just asking out of curiosity

>>2662095
Does it matter if the speed at which the motors turn is proportional to the distance they travel? As in, each linear move would take the same amount of time, regardless of distance. If so, you could feed the ∆X and ∆Y values into R2R DACs and use that for each motor's PWM for a fixed amount of time. Though that may not be precise enough to be repeatable.

Do you intend on doing closed-loop distance control? If so I'd want to use stepper motors, though a phase-locked-loop using a brushed DC motor and rotary encoder as its VCO would probably work too. You'd just need to produce a series of step pulses, whose frequency is proportional to that ∆X and ∆Y values, which you could do with a variable threshold triangle oscillator. And instead of a fixed time, you count each pulse on both axes, only shutting off each oscillator once the counter is back at zero. There might be some slight misalignment at the end of each line (e.g. a diagonal has a small straight line at its end to get to the right coordinate) but there wouldn't be any long-term coordinate drift.

The real chad project is to make a fully analogue CNC machine, where the analogue G-code is stored on audio tape cassettes. You'd probably frequency modulate each axis and use PLLs to lock onto each frequency. Clock synchronisation wouldn't be an issue since the tape playback speed would be tied to your master frequency source.
PLLs are cool.

>>2662059
>>2662086
>>2662106
Can someone make a flow chart of this idea?
I don't get it.

can anyone run a quick sim to see if this circuit oscillates? Or just check it by inspection, i think it should oscillate, but I always have problem simulating oscillators on kicad or ltspice. Kicad giving me the timestep too small error, and I don't know how to fix it, seems like the numerical method used to solve the circuit uses a dynamic timestep and at some point becomes too small to compute.

>>2662444
For errors like that, try putting "startup" in your spice command. Adding parasitic elements can also help.

Also isn't that totem-pole just shorting itself out? I think you have Q9 and Q8 swapped.

>>2662449
>Also isn't that totem-pole just shorting itself out
I'm not sure I understand the point you're making. The totempole takes the output from Q11, which is a simple positive feedback latch. Assume Q4 starts empty (0V). Q10 is off, which means Q11B is at 5V (minus R10 drop), turning Q11 on and Q11C ~0V. This is fed back to the totem pole where Q9B gets shorted to ground via Q11, turning it on, allowing current to flow Q9->R9->C4, which raises the voltage. Once voltage is above 0.7V, Q10 turns on, shorting Q11B to ground via Q10, making Q11C ~5V, fed to the totem pole, turning off Q9 and turning on Q8, which shorts R9/C4 to ground, discharging C4. Once C4 is discharged, Q10 turns off and cycle repeats, unless I screwed up somewhere and those BJTs end up in active region and reach some steady state. I'll give the startup command a shot, as for parasitics, what exactly should I add to the circuit and where? I'm relatively new to this, I just managed to get an oscilloscope so I started making slightly more complex circuits. I wanted to get a BJT oscillator that can be tuned with only 1 capacitor, so the astable vibrator is no good. Tried to come up with my own circuit based on my own understanding, so I'd rather not cheat, if the circuit doesn't oscillate, i'll go back to the drawing board.

>>2662455
>>2662449
perhaps I should remove R11 and put it in the feedback branch? That way Q9 E-B doesn't get shorted to ground with a 5V like pic rel?

>>2662459
wait, this is also wrong, now Q8 is floating when Q11 is off. I'll have to look into it a bit more.

>>2662463
>>2662459
I think this should do it?

>>2658021
I have no experience whatsoever in PCB design, which should be obvious. JLCpcb says this is a 2 layer board but PCBway says it's a 4 layer board... I think if it weren't for the 10x10 square this could be a 2 layer board, is that correct? I'm not planning to sell these I just want to make a few for myself because they're out of stock and I'm not paying $200 for some chink knockoff.

https://github.com/omahacommsys/KFDtool/tree/master/hw/KFDShield/gerber/revA

>>2662516
Idk, it has 4 layers in the gerber zip, but I can't really see it being a 4 layer board, probably a kikecad moment. Kinda retarded file output ngl.
>I think if it weren't for the 10x10 square this could be a 2 layer board, is that correct
Wdym?

>>2662530
>Wdym
I just realized that using PCBway's gerber viewer I couldn't flip the board in the layer view and the traces that were confusing me are part of the bottom layer. I'm wondering what to tell JLCpcb, I just got picrel in an email form them. The only reason I did this knowing that I don't know shit about it is due to the absurdly low price after all the discounts about $20 for 5 boards.

>>2662539
This is the email picrel

>>2662541
I see my email address in the filename but it's too late now. Oh well, one of many.

>>2662530
looks like a 2 layer file to me, you have F_Cu that stands for "front copper" and B_cu which stands for Back copper, I don't see any other layers. Maybe you didn't select the right pcb settings when ordering? Also, try allpcb, i think they give free samples if it's under 10x10 or 15x15, don't remember. Check with them tosee how many layer it displays.

>>2662543
wait a minute and try to delete it. Anyway, those files are for the front and back layer. You must include both.

>>2662545
see >>2662541
Front and back or bottom are what I was thinking as well, something seems to be confusing JLCpcb as you can see in the other post. I know I picked 2 layer in the configuration.

>>2662547
>wait a minute and try to delete it
Anyone using 4chanx will still be able to see it deleting it would be pointless, but it's a burner anyway.

>>2662547
>Anyway, those files are for the front and back layer. You must include both.
I agree so why is JLCpcb so confused about it, it can't be the first time they've seen it done that way.

>>2662553
did you accidentally zip the file twice? so there's a second copy? Also if you generated the gerbers yourself from kicad, look at this
https://jlcpcb.com/help/article/16-How-to-generate-Gerber-and-Drill-files-in-KiCad-6

Make sure your files are named like in the example. Also, from what I remember, when you order on JLCPCB, i think they ask you how many layers you want, perhaps you chose the 4 layer board instead of 2?

>>2662555
>did you accidentally zip the file twice? so there's a second copy? Also if you generated the gerbers yourself from kicad
I just used the zip as it was included in the git clone, I don't even have kicad.

>perhaps you chose the 4 layer board instead of 2
I definitely chose 2 layers I just attempted the quote process again and 2 was what it automatically picked and I didn't change it. I also just tried allpcb and their quote app says it's 2 layers, so pcbway is the only outlier thinking it's a 4 layer board for some reason. I also notice that the files aren't in alphabetical order in the pic the JCLpcb rep sent me and that make them look much more confusing I'm not sure why they did that.

>>2662559
I see, from a past file i uploaded, it looks like the second part of your image is not required. Delete all the .gbr files in the second circle and see if it looks good or not. It looks like jlcpcb ignores those extra gbr while pcbway doesn't

>>2662564
>>2662541
this one, try removing everything with .gbr in the second circle, keep everything above the kvlshield-edge_cuts.gbr and delete everything below including the edge_cuts.gbr that is in the second circle