/diy/ - Do It Yourself

>>1772647
this thread's digits brought to you by TIs Dual Synchronus Buck Regulator

>>1772647
What are those two big black plastic packages on that board? I can't imagine they're diodes.

>>1772702
>L1 L2
probably inductors https://pdf1.alldatasheet.com/datasheet-pdf/view/286558/CT/CTMC1812F-1R0J.html

>>1772717
Oh didn't notice the silkscreen. I'm guessing the antenna is on the other side of the board as a coiled trace? The 2 inductors, 6 caps, and 2 resistors match up and there's only those vias connected to the resistors.

Retard from /mcg/ here, I suspect I've got a hardware issue. I'm testing with an arduino nano and an SD card module in a breadboard, and even without touching anything I'm getting very intermittent results. Using the But I rely on the serial printouts for troubleshooting so I can't swap to a PCB just yet. I added a 100nF cap from Vcc to GND but that hasn't helped, are there any other things I can do to help my case?

>>1772862
*Using the* basic ReadWrite example sketch from the SDfat library.

Any STMbros on here today? I have firmware i wrote for an L496VGTx to test both availble CAN devices and I ported it to a devboard setup with a Nucleo L496ZG to test that both CAN nuses discover each other.
I hook up the nucleo via STLink usb for flashing and debug, and then use a TTL to USB serial cable to communicate using Putty as I run the firmware. It worked fine last week, but since Friday Ive not been getting any commands out of the pc into the board, while the TX line of the nucleo works fine as I can see output from the line into the laptop. What are the odds winbows has fucked with the serial port?
Things i have done so far:
>durrr reboot and retry
>changed serial cable
>changed ports between laptop and docking station (dock ports used to be the only ones that worked)
>scoped into Tx and Rx (referred to board) lines to find that Tx signal is visible, but Rx is idle no matter what keys i try to transmit from pc
Things I have not done:
>change uart device on board - code was written for a low power uart that doesnt change its pins in Cube
>try a different OS/laptop - although the same issue has been plaguing coworkers on the same board type but different firmware

Thanks in advance lads.

>>1772870
Forgive the typos, I was phoneposting in the cold outside.
I'm happy to take this to /mcg/ if it belongs there.

>>1772317
It's a balanced push-pull oscillator/generator configuration. If you need a Saint then pray to St.Royer. Pic: kit cheaper than parts.

>>1772870 and >>1772872 here.
issue is resolved, turns out the interrupt callback needed for receiving serial data was somehow removed from the code, works as intended now.

>>1772862
Haha are you me? Had similar problem, had GPS that introduced too much spikes. 470uF cap solved it for me.

>>1772910
I thought it looked familiar.
>kit
How many watts is it good for? Can it melt metal? Is the coil a pipe for water cooling?

>>1772928
search alibay

Rate my captcha. Also google shittyboard if you wanna try

>>1772978
9/10, fits the theme of the page, gets a smile out of me as well. Doubt it'd be hard to circumvent, but then again the page it protects won't really be under attack seriously.

>>1773004
Well, google images suck at recognizing my hand-made symbols, plus I can distort things more. But yeah, nobody will ever try to bypass it.

>>1773010
Hey, there's a admirability in being able to recongize and accept "good enough". It's something a lot of people seem to have issues with their personal projects. As I said, I like it.

>>1773030
True

Got a bit of a problem:
I want to directly interface with my I2C bus via USB. The first chip I stumbled across that seems to do this well is the MCP2221. I'm using KiCad and they don't have a footprint for it and when looking for one online I only find a random old commit in a old depreciated kicad version (v.4) .
My questions therefore:
>Is the MCP2221 a decent choice or are there I2C to USB chips that are better?
>If it is, I know it's a longshot, but does anyone know of a ressource to get the footprint? I never created one myself. If it fails I guess I'll have to though.

>>1773040
Nvm on the second part of the question, using another search engine gave me a souce. Still unsure about how good it is though.

Possibly a dumb question: If I have a datasheet that specifies 2x22uF as a capacitor value, is that 22uF or something else? Why is the 2x there? Do you add some in series/parallel?

>>1773070

2 parallel caps, obviously, totalling 44uF

Is it okay to store ICs in Styrofoam sheets?

>>1773081
Thanks, I'm new to the notation.

>>1773082
>Is it okay to store ICs in Styrofoam sheets?

analog and TTL are fine, it's risky with Cmos coz it generates static.

This just keeps getting harder and harder.
If I control it from the voltage source, the FETs get fried, I think I can't use phase control too, since it's high power and high speed.

the sim game must go on

>>1773087
Thanks.
I have 74series stuff mostly plus some microcontrollers and UV eeproms, i dont think i have any cmos.

>>1772978
Get more obtuse so people have to look them up. Each post is a learning experience.

is it 4chan x compatible?

>>1773113
>74series
74LS or 74HC or something more obscure?

>>1773128
>is it 4chan x compatible?
Nope. It is not.
> Each post is a learning experience.
First 4 or so. Then it disables captcha, utill next faggot tries to wreck the site

>>1773095
there is something fundamentally very, very wrong with square inverters.
Also your rectifier looks like it's gonna build up voltage

>>1773169
Speaking about inverter, how do you design a three phase motor controller. I give up. Remember shitposts about diy e-bike controller? I ended up blewing up all transistors I found in trash trying to run an hdd motor.

Where is (hobby) electronics headed? In 20 years, will everything be so miniaturized and integrated that hobbyists can only fuck around with "vintage" items?

Due to size issues, I wanted to make a custom power supply and found pic related(in german print Trafo), which seems to do an all in one job of converting 230V to 6V. My question is how reliable are those, or are there any hooks I should keep an eye on(something like adding another condensator for safety)?

>>1773178
Well in the 2010s, microcontrollers replaced control circuits and PCB modules replaced a lot of other common circuits (ex: sensors).
I am guessing that SBCs will replace MCUs (this is already slowly happening) and more stuff will focus on wireless communication.

>>1773175
You were shorting the high and low side FETs in the intermediate points I gather? Pretty sure you need some comparators, positive feedback would be good too.

>>1773189
It's a transformer, do you want 6V AC?

>>1773193
Yep. How do I make a fet driver with dead-time delay. (getting IC is not an option, just like pre-made controller, since coronavirus nigger).

>>1773215
Buy it from a local seller, or on ali but which ships from local warehouses?

Anyhow, assuming you're using NMOS and PMOS FETs, you need to ensure that it operates in a break-before-make operation, so I'd have three totem-poles. One for each FET, and a third current-limited one that feeds a couple of comparators. Use base resistors and an emitter cap to ground to set this third one up to produce a ramp from Vcc to Vee (or at least well within the gate threshold voltage of these rails). Then have your (schmitt trigger?) comparators trigger at points further apart from one another than the gate threshold voltages of your FETs, and feed these outputs into the totem-poles that feed the FET gates. Pic related, plus whatever diodes and what to optimise the output totem-poles. Then you tune the RC constant until you don't get switching overlap (due to gate capacitance) but still get as little down-time as possible when operating at maximum frequency. Ideally, you'll have the ramp time much smaller than the fastest period of the driving square wave, but slightly larger than the FET's ON time.

Note that I've never used an H-bridge in my life, so take this with a grain of salt.

>>1773277
>Buy it from a local seller, or on ali but which ships from local warehouses?
You can't find anything here locally. Only basic crap.
I ended up salvaging optocouplers for bootstrap circuit and lower fet, lol.
At least my Arduino didn't explode,lol. I think it wouldn't have exploded, if I used pure AVR C...
>pic
But I also need floating option, and pic related don't seem to feature this...

>>1773277
Oh and that design is beta dependant, might want to do the current limiting with a resistor on the emitters, not the bases, though it will give less of as straight ramp. Might also want pullup and pulldown resistors on the FET gates respectively to ensure they stay in their off-state when starting up, alongside potentially some zeners in parallel with the totem-poles' bases to keep the gate-source voltage from getting too high, in case that's something you have to worry about with your voltage rails.
Might want to do something else about the voltage references than a simple divider too, like a pair of TL431s.

>>1773283
>floating
What do you mean by that? What's preventing you from slapping an opto after the each comparator?

>tv has shitty speakers
>use parts-bin 3-band 3.5mm to composite A/V cable to hook audio source up to a stereo
>if I put the 3.5mm in the stereo half-way, it sounds fine, just in mono
>if I push it in all the way, I get stereo but with a fuckton of noise even though composite video isn't plugged in
>the noise goes away if I short the unused composite video inner metal piece to the outer metal piece
Does anyone know why this is happening?

>>1772978
>captchas that require research or knowledge to complete
Could this be the key to an effective retard barrier?

I live next to a big intersection and occasionally retards with really loud speakers wait for the traffic lights outside my window and make a right racket. I've heard tales of devices that wirelessly project noise into nearby speaker systems, does anyone here have experience with that? I gather it would need to be rather directional, but all directional antenna topologies I know of use extra reflectors or active elements that are 0.5/0.25/0.75 wavelengths away from the central emitter, and that's just not possible with my favourite music: 1kHz Sine Wave. Can I string together a bunch of small loop antennae in an array and control the signal phase going to each one as a function of displacement from the central loop for directionality? Or would a parabolic dish work even if the dish is smaller than the wavelength?

Or perhaps could I transmit noise at/near the class-D amplifier's PWM frequency and somehow interrupt it like that? But that would require a ham license, transmitting at 1kHz doesn't.

new faggot here. Any reason why this wouldn't work?

>>1773128
Both LS and HC, random gates, counters, fliflops and stuff. I keep the rare, obsolete stuff in conductive foam. I dont have enough foam for all the common stuff

>>1773362
Don't drive motors through resistors unless they're really small motors. If it's a 10W motor, worst case you'd dissipate 2.5W in the variable resistor. Normal size thruhole resistors are good for 0.25W, and most potentiometers won't be too much better than that, 1W at most. You'd need to buy specialty power potentiometers.

What real circuits do is called PWM, or Pulse Width Modulation. Essentially, by switching the motor on and off dozens or thousands (or millions) of times per second, you approximate the motor being driven by a lower constant voltage. This uses a switching MOSFET, a triangle oscillator, an adjustment potentiometer, a comparator to compare the triangle-wave with the pot's voltage, and a handful of passives. Such an oscillator can be made with a comparator, so a quad comparator IC would work for your purpose. Or you could use 3 fucking NE555's like some kinda retard.

>>1773368
they're one inch salvaged motors with a very small fixed load.

>>1773369
One inch probably isn't small enough, could be anywhere from 2W to 20W. What's the winding resistance? The current drawn at maximum speed?

I have a chip powered by DC and the chip takes AC wave as input on one of its pins (audio wave). the wave can be at most 350mv peak to peak (no dc bias).
What is the simplest possible way I can protect the chip from overvoltage, if for example some retard tries to for example crank up the volume and the audio becomes 500mv peak to peak?
I need a way to either clamp it to 350mv max OR to cut it off fast enough to prevent damage to the chip, either of those options is fine
I am looking for the simplest solution possible here

>>1773376
Resistor in series, then antiparallel Schottky diodes to 0V, they should clamp the voltage to below 0.3V, though it will depend on the size of the resistor. Too large of a resistor may make an impact on the received signal due to the input impedance of the chip. What chip is it? It's unusual to have an input range smaller than the supply rails, unless it's an ADC or mixer or something.

>>1773371
don't know about winding resistance but it one motor pulls .015 amps at 12 volts. So .18 watts?

>>1773377
https://www.sparkfun.com/datasheets/Components/General/MSGEQ7.pdf

>>1773303
https://help.longtailproducts.com/hc/en-us/articles/207970396-Smartphone-Headset-Standards-Apple-iPhone-AHJ-CTIA-OMTP
>>>/g/

>>1773331
just shoot 'em
>>>/k/

>>1773189
I think you're looking for Pic related

>>1773178
hacking on consumer products isn't the sum total of hobby electronics. just lern2smd and everything will be fire. it's not that hard, you just need finer tools than your fingers and grandpa's soldering iron. oh and better vision than your grandpa, get a "USB" microscope to get started

>>1773082
it's ill-advised. conductive foam bug rugs are awesome, get some

>>1773040
you've probably got the footprint already. what you need is the symbol. it's pretty easy to create a new symbol: draw box, add pins, fill out info, save, use

I wanna make an IR detector so I can tell if I'm being recorded.
I want it to beep when it detects IR light from a camera, what would be the cheapest and easiest way to go about this?

>>1773378
Oh that's surprisingly small, an effective resistance of 800Ω. Yeah you could get away with a variable resistor, but your minimum power output will just be an 800Ωload with a 100Ω resistor in series, so it won't be much different. You'd be better using a 1k or 10k potentiometer, or having the motor be the centre tap with +12V and 0V being tied to either end of the potentiometer, or both.

>>1773379
Oh I saw the datasheet of that when it was posted a few days ago. What's the output signal actually like? A bunch of DC voltages that vary depending on the filtered input amplitude? Sounds neat, but how are you meant to use that data? Just use it as-is and rely on persistence of vision?

>>1773395
Photodiode + light filter + bunch of analog and/or digital signal processing. Some dedicated IR photodiodes come with a filter on them (they look black), and there are probably camera filters that work like this, but remember the sun outputs IR too, so you'd need a really narrow filter to get the SNR high enough. Plus not many cameras use IR LEDs at all, only cheapo "night-vision" security cameras. And things other than cameras use IR LEDs, like remote controls and
You might even have better luck listening to the electrical noise to find whatever a digital camera sounds like, I imagine you'd get a peak at the framerate. But that would also be a real pain, as interesting as it could be.

>>1773395
>so I can tell if I'm being recorded.
why?
do you have something to hide fren?

Look frens, this arrived today from my fren Chong. He is probably dead by now, but luckily he managed to ship it just in time.
Are you jealous? Are you?? Yes you are.

Not really electronics related per se, but it's the closest I'll get...

I've built a synthesizer a couple of weeks ago. I then altered the design yesterday. Would you consider the altered design to be rev 1 or rev 2?

>>1773458
when did he ship it? covid-19 can survive on surfaces for 9 days

>>1773479
it's china buddy, minimum shipping time is over a month

>>1773458
How do you mod these things to test at higher frequencies? 1kHz is dogshit for testing inductors

>>1773487
It's not intended to replace tools like a multimeter, it's for quick component identification

>>1773489
That's bullshit. It works fine for capacitors, resistors, and large inductors. I've built resonant circuits out of them and measured them with a scope, and they're fine.
The only shortcomings i've found are hFE measurements and <10uH inductors

>>1773482
you can always pay like 16€ to get it delivered in 4 days

>>1773497
why would i pay extra to get infected?
slow postage saves lives

>>1773498
because it would be

>>1773380
Thanks, I created the part for the footprint. Since then I also found out that you can just download the footprint & part from mouser, which makes things a lot easier.

>>1773418
Obviously it woildnt be perfect (the camera detector) but it would work at night at least.
>>1773436
>this argument again
Yea, and guns should be banned and freedom of speech is stupid

>>1773531
>>1773418
https://www.adafruit.com/product/1980
Would this work for detecting IR light or is the IR diode only present for discrimination purposes?

>>1773533
>That means you can separately measure infrared, full-spectrum or human-visible light.

That's from the product description.

Hey everyone,

Maybe some one has encountered impedance measurements? I'm thinking of making a device that can measure impedance. Any ICs suggested? I already checked few Analog Devices chips. Plus, maybe I should give a try for more hardware options, like reactive Wheatstone bridge? Looking forward for advice

>>1773549
don't some digital meters measure the V-I angle?

>>1773566
They do, but I want to make a device that can measure it

I wanted to use the MCP2221A chip as a USB to I2C bridge. I read the datasheet (https://ww1.microchip.com/downloads/en/DeviceDoc/20005565B.pdf)) completely and the only information on wiring in it is on page 17, section 1.6.2.2.

>In the pic on the left, the datasheet states that C1 must be no more than 10uF. Do I just control that by limiting trace length? Why did they draw it in the circuit?

> What exactly are the white triangles used for? Do they represent ground? Why not use the normal ground sign?

I found that Adafruit makes a chip that uses the board, in the pic on the right. Can anyone give me ID on part 3 (472 written on it) and 4 (G3P written on it)? Either the chip name or just what they do? There's no mention of any need for these in the datasheet.

Thanks for taking the time.

>>1773585
Capacitors are for so-called bypasing. It acts as both low-pass filter and assures more stable voltage supply, if for example chip's current consumption would rise rapidly, voltage could possibly decrease, capacitor ensures that this effect may be as negligible as possible. While talking about noise, it is possible that sometimes power supply may introduce some of it - if it is a switch mode power supply, there are plenty of noise, etc.

Yes, triangles are ground. Often used as a digital ground (in constrast to analog ground). For example, you have both digital and analog circuitry in your schematic. Like, you have an audio signal coming to an op-amp (analog) and a analog-to-digital converter (digital). You may probably use different power supplies in order to minimize noise (digital circuits are noisy in comparison to analog in this situation), so different grounds are used as there are two different power supplies.

>>1773589
First of all, thanks for the fast answer.

Am I right in assuming that they want me to add a capacitor, just one that has a capacitance of less than 10uF? Do you have some experience as a rule of thumb what you'd use?

Triangles for a digital ground makes sense. Thanks for that.

Lastly do you have any idea about the parts?
The Adafruit page states that one of the features is a onboard 3V regulator so I assume that's the Chip Nr. 4 (G3P). I have no idea what the other one does though. There's no mention on their page about it.


Again, thanks for the quick answer

On a unrelated note, maybe someone can get a laugh from my highly sophisticated reverse engineering technique. As it turns out this is probably a 3 layer board and thus this technique is pretty limited in usefulness.

>>1773574
idk of any IC's to do exactly what you want. But maybe you can still measure the angle like I mentioned, with a hall sensor for current and voltage divider for voltage, then count zero crossings and time delta between I zero and V zero. Though it feels like there should be an easier solution so maybe also wait for someone with more knowledge to reply.

>>1773591
I doubt that something bad would happen if you even would add a capacitor of 100uF. I guess Microchip simply tested a chip under condition when 10uF SMT capacitor was attached to circuit, that's why they encourage to use no bigger value. Talking about rule of thumb, yeah, while talking about generic SMT capacitors, I would always put some near power supply pins near to the chip. It is a common practice in PCB design (I see that a lot while playing with old hardware PCBs). Although I only made like three PCB's but I always tend to do this - add 10 uF capacitor near to power pins of a IC. Adding more different capacitors of different types in parallel (preferably smaller, like in nF range) would be even better. I don't want to go into details as it takes time, but it simply ensures as low ESR (Equivalent series resistance, a quality of capacitors) as possible, they provide excellent filtering.
It is stated perfectly here why you should do it:
https://electronics.stackexchange.com/questions/239764/why-multiple-capacitors-in-parallel

Part 3 is a resistor array (472 translates for 4.72kOhm as 47 is a number and 2 is number that has to be a power factor of 10. Resistance is a multiplication of these two - 47*100). It may look unusual, but it is sometimes used.

Yes, part 4 is a regulator (LDO), probably from Diodes Incorporated.

>>1773596
>Adding more different capacitors of different types in parallel (preferably smaller, like in nF range) would be even better. I don't want to go into details as it takes time, but it simply ensures as low ESR (Equivalent series resistance, a quality of capacitors) as possible, they provide excellent filtering.

Yeah I watched a video by EEVblog about it and why you'd want it. Still found it weird they'd specifically mentiona "maximum" capacitance they want.

Thanks a lot for the answers and the ID on the chip, this really helped me out!

My refrigerator can't maintain the temperature if the room temp in my kitchen suddenly drops. It got really cold overnight and now my milk is frozen solid. It's happened before so it's definitely the cause. I have an electric heater in the room but I wasn't paying attention to the weather forecast.

Is this normal or is the fridge on its way out? My house is old and not very well insulated.

>>1773606
>Is this normal or is the fridge on its way out? My house is old and not very well insulated.

That's not normal. Your refridgerator should have a control circuit that adjusts the cooling power according to the measured temperature inside itself. Your best guess is probably checking if the thermometer works, if you set the settings correctly or if there's any visible damages to electronic parts (unlikely). Also iirc appliance repair isn't really in the scope of /ohm/, but there's not any threads up for that either.
Is it a very old fridge? That might influence the complexity of the control circuit

>>1773611
I'm not sure by how much but yeah it's getting kinda old. There's a temperature setting outside (9 small lights and you hold down a button to choose) but it doesn't stay on the one I select. I can't get to any electronic parts at the moment. The last two times it happened it eventually went back to normal but it hasn't happened yet, it's colder outside at the moment.

Yeah I wasn't sure where to post but this seemed like the best bet. I was hoping someone knew and could just be like "yep it's fucked".

>>1773640
Well chances are the issue is a damaged thermometer or a damaged control circuit.

Since either of those aren't easy fixes and a fridge that needs a heater next to it in cold weather isn't what I'd call functional "It's fucked" might be a good call here.

Okay lads I need your fucking helps 'cause I'm a fuckin moron. I need to replace 4 driver transistors in a 40y old amp and I can't find a replacement. I can only order from farnell or mouser. The OG T's are Panasonics and 2SA0794A-R. I found a through hole equivalent for one polarity but for some reason they ran out of stock for the other and of course no one makes these anymore so I'm fucked. I'm pretty new to all this so I want to know what characteristics I can change and by how much. I've been told I should match the switching frequency in these amps because a high rating may cause unexpected behavior and wreck the output transistors.
How close should I be to the original Hfe to get away with a different value ? Can I exceed the collector current ? If THT equivalents don't exist can I get some sort of SMD adaptor on these sites ? Thank you.

>>1773418
>You'd be better using a 1k or 10k potentiometer, or having the motor be the centre tap with +12V and 0V being tied to either end of the potentiometer, or both.
k, I figured 100 ohm because the highest ohm resistor on the device i salvaged from was 68 ohm

>>1773331
My amp occasionally picks up police comms when the pass in front of my window. Son it should be doable with your average baofeng. My guess is that blasting this kind of EM noise will get you fined real quick though.

>>1773684
Forgot to post the other polarity original transistor. It's 2SC1567-R.

>>1773687
>My guess is that blasting this kind of EM noise will get you fined real quick though
Yeah that's why I want to do it at audio frequencies, not at one of the regulated frequencies. I should make a massive resonant driver circuit like this >>1773095 but tune it for 1kHz.

>>1773525
oh yeah, forgot to mention the meta-librarians. they are kinda useful for a starting point but you will probably want to edit the symbols to be more usefully organized e.g. with power and/or functional units separated out and left-to-right signal flow. in my limited experience with them they usually just give you a box with pins mimicking package pin layout

>>1773395
method 1 is good
https://www.wikihow.com/Build-a-Hidden-Camera-Detector
also listen for the periodic clicking noises of a cheap autofocus going to home. always fun to grab a custom fw for a chinktablet only to hear that a few hours later

>>1773436
if you don't, maybe you're not paying attention

>>1773461
usually they start with A and bump each time the design is sent to manufacturing

>>1773491
ok so it happens to work respectably well within its ambit

>>1773498
topkek

>>1773549
are you looking for something called a network analyzer

>>1773599
the on-board LDO for Vusb may have a small transistor not sized to handle a long inrush pulse into a larger cap

>>1773684
it depends on the circuit design. if you keep hFE and fT close to the original and meet or exceed all absolute maxima, you should be ok

>>1773699
how much power do you think will be transferred to a 1/150000-wave antenna
actually nm you're still well within the near field lel
also the inverse square law called and said it hates you

>>1773814
>a 1/150000-wave antenna
Doesn't coiled wire count as a longer antenna? I wasn't exactly planning to use a monopole.
>near field
What do you mean by this?
>inverse square law
Yeah that's why I have to make it directional.

>>1773418
>but how are you meant to use that data? Just use it as-is and rely on persistence of vision?
is is meant to mostly be used with a micro that processes that signal and then does stuff based on that
i actually am not sure how exactly the signal coming out of it looks as i was swamped and even though i have the circuit hooked up, i had no time to power it up and look at it with my scope yet, i hope to do it tommorrow

>>1773828
It would be very strange if it communicates to an MCU in a way that needs an ADC on the receiving end, and still pretty odd if it used PWM or differential PPM or something.

>>1773816
Wait, I could use a spark gap and two 1mF 400V cap to make some wicked arcs, right? Doing that fast enough would produce a bunch of electrical infetterence. Doing so would make me a complete asshole to the local hams, but I'm wondering how easy it would be, just for curiosity's sake.

>>1773836
the output is definitely analog and not digital so you need ADC to read it, the chip isn't very smart, it is just 7 various pass filters in a row and you select which of those pass filters you want on the output
But beggars can't be choosers, i was unable to find pretty much any chips at all that provide at least 5 band sound filtering and aren't too expensive, or unavailable

>>1773858
>the chip isn't very smart
Yeah but it could at least output a variable pulse width instead of a variable pulse amplitude, it would just require some internal comparators and a capacitor ramp. Doing so would mean you could use a much simpler MCU.

I assume it's not designed to be fed into an MCU, but rather some sort of voltage dependant analog-switch multiplexed VU meter, perhaps through an op-amp gain stage. Because any MCU with its own ADC can probably be used as a VU meter without the need of another IC.

On that note, are there any analog VU meter circuits that don't need a transistor for each LED?

>>1773869
>analog VU meter circuits that don't need a transistor for each LED
The more I think about it the more difficult it sounds. Assuming instead a digital method, charlieplexing would be pretty efficient, but there's some degrees of freedom there that I'm not in need of, specifically that no LED beneath an on LED will ever be off, and I'd rather have all LEDs on at once rather than rely on persistence of vision. I smell linear algebra.

>>1772647
Help understand cable impedance. Why aren't coax cables rated for a Zo at a given frequency and are instead rated as only 75, 50, whatever Ohms?

>>1773877
IIRC it's because the impedance is independant of frequency, the frequency component from the inductance and capacitance cancel out. Have you learnt the concept of impedance matching?

>>1773884
Yes. I "know" hot to use that but I was always curious why Z0 didn't change with frequency. So I guess the limit of a coax cable is when it starts behaving more like a wave guide and less like a cable.

>>1773884
How does a coax cable make the inductive and capacitive frequency components cancel eachother out?
Or does this happen at a previous stage?

>>1773885
>>1773886
Wanna do an integral?

>>1773887
I'm only the first quoted, but yes. I was good at calculus.

>>1773816
I meant the car
>near field
https://en.wikipedia.org/wiki/Near_and_far_field

>>1774046
I see

Do these multiplexer ICs work how I think they work?
For example I bought one of these https://www.ti.com/lit/ds/symlink/cd74hc4067.pdf
to power 15 LEDs by multiplexing

After further studying the datasheets these pins I1 to I15 are inputs
My plan was to use a 4bit counter with a high enough frequency to make the 15 LEDs appear to be on simultaneously.
Now I don't think it would work with this one. I'm not even sure what the common input output pin means.
What would be an alternative? Preferably without a microcontroller

>>1774068
>What would be an alternative?
74HC595 SIPO register, pretty easy to set up
> Preferably without a microcontroller
idk

>>1774068
What arrangement are the LEDs? Just one after the other such that when the counter is at "12" the 12th LED is lit? Then just use a 4-bit binary to 1-of-16 decoder.

>>1773287
>What do you mean by that? What's preventing you from slapping an opto after the each comparator?
Both fets are turned off, so wheel can spin freely.

>>1773306
No, since I've made it, lol.
I expected it to sort out furries, anime faggots, but some still made it though.
It must be in reverse.
"Find image containing 1234567 and describe what is on the picture in one word" or something like that.

>>1774150
needs to be more obscure, you should have used the US de facto standard AND gate

>>1774068
multiplexers act as a signal switch, much like a rotary switch on an old audio receiver, but without the rotary part. key word, "signal" not "power", they aren't usually large or low-resistance enough to pass the milliamps required by an LED to be seen under normal room lighting conditions
maybe try 1/2 of a 74xx393 for the counting part, followed by two 74xx138 to drive the LEDs

>>1774169
>needs to be more obscure, you should have used the US de facto standard AND gate
It has both ISO (ones I grew up with) and shitty american penises gates, which I can't remember to be honest.

>>1774149
Well either you have a tiny bit of all-off time, or a tiny bit of all-on time, which means exploding your transistors. There's no way around it. With any luck it will only be a few ns so it's not like it's a significant issue. You could implement a braking function that activates both/all three low or high-side FETs of an H-bridge, and that would be pretty easy to do as my circuit doesn't include the side-to-side timing circuitry.

How the hell do BBD chips like the MN3007 contain over a thousandth capacitors? I'm thinking they must be extremely small; less than a picofarad. I'm thinking of making a BBD out of discrete transistors and if that's the case I can just use the capacitance between the traces and ground plane.

>>1774148
Yes that's exactly it. I'll look into it

>>1774169
I'm using low current LEDs
I'm planning to run each LED on 3mA but I'll check out those ICs out

>>1774223
Can someone teach me how this circuit works? What's the deal with the transistors connected to Vcc? How does that strange output stage work, and why do some circuits couple to two different capacitors? Why does the end connect to Vdd? Why is there a Vcc and a Vdd?

>>1774224
Looks like the 4514 or 4515 are good choices

I need some specialized switches and am considering printing contacts directly on the PCB and having some springy contacts slide over them like how this multimeter works. Is there a name for this technique? I'd like to do some research on how to do it good. My main concern is PCB pads weren't meant for this and will wear away quickly.

>>1773929
haha fucking nerd

>>1774223
the caps are quite small and quite thin. see for yourself https://zeptobars.com/en/read/MN3207-1024-stage-analog-delay-line-CCD
bear in mind you'll also need to handle stray capacitance, or it will destroy your signal before it reaches the end

>>1774252
I'd only do that for an analog circuit with potential overvoltage situations and/or an importance for high linearity. Otherwise I'd use buttons/toggles/rotary encoders along with analog switch ICs.

You need to plate those kind of contacts with hard gold, a type of gold alloy found on headphone jacks and USB sockets, which is different from the soft gold plating that you get on solder pads. I think some consumer PCB fab houses do hard gold, but I'm unsure whether you could specify different platings for different parts of the board.

>>1774223
MOS transistors can be seen as capacitors too..

>>1774223
>>1774236
Oh and also how do they stop the charge just getting spread between each pair of capacitors basically halving the amplitude after each clock pulse, and how do they stop each voltage from being affected by the voltage that came before it?

Guys, first time on /ohm/. I was adding a usb port to my Vita and I fucked up. I think I damaged pic related, there's no continuity anymore. How the fuck you desolder it?

>>1774292
And btw, what is? A diode? Total noob here

>>1774276
There are two FETs per stage. One is functioning as an analog switch. The other is functioning as a voltage follower.The switch isn't connecting a capacitor directly to the previous one in the sequence; it's connecting to a buffer of its voltage.

>>1774292
Did it used to say "SB" on it like that other component? Continuity alone usually isn't enough to measure a component, I'd test for resistance, maybe capacitance too.

>>1774294
Yes I thought they needed to be buffered, but doesn't a FET voltage follower need to be a gate-source thing, not a drain-source thing?

>>1774292
>>1774293
Ok I looked it up, looks like SB stands for "solder bridge", in other words, a 0Ω resistor or short-circuit. To desolder it I'd advise cutting/crushing it in half (being careful not to damage the board underneath) and desoldering each half separately with relative ease. Unless you happen to have two soldering irons. Replace it with a short length of solid-core copper I guess.

Also that looks like a solder blob on the multi-pin connector to the right.

>>1774276
>1774276
>>1774271
they run on two or more internal clocks.
I know the circuit is cringy, but I'm traveling on a bad computer without a mouse so give me a break.

>>1774304
Shit one of the clocks looked like it was connected but wasn't. That is why the signal is fucked up.

>>1774304
No that circuit is perfectly fine. Was just about to do it myself. Though your one is the one without the common-gate intermediate stages. There a reason that the capacitors are from source-to-gate not from source-to-ground?

>>1774306
>There a reason that the capacitors are from source-to-gate not from source-to-ground?
To avoid
>>1774276

>>1774307
Also because it is literally sampling a signal, the anti-aliasing and reconstruction filters are needed.
Here the "real" topology is presented
http://dafx.de/paper-archive/2010/DAFx10/RaffelSmith_DAFx10_P42.pdf

Is there a reason there aren't a million jelly bean components that are charge pump + FET driver? I have found a handful but they all seem to be fairly expensive, niche products. I was thinking about trying to trick one of those OR-ing controllers into doing it since they have charge pumps and gate drivers in them. Any search or chip recommendations?

>>1774307
Oh it helps the charge to flow, like lifting the left-bucket and lowering the right-bucket. That's clever, I was wondering if it would work in the opposite direction.

>>1774308
Yes it did distinctly remind me of a nyquist quantised sampling situation. So if I use one as a delay, I need to ensure that the input frequency is no more than half the clock frequency, and the signal moves two buckets per clock pulse. So with a 1kHz sine wave, I'd need to run the clock at 2kHz or greater, and with 1000 buckets it would move through all of them in 500 clock pulses. So the maximum delay in such a situation would be 0.25s. Wouldn't this be pretty shit at delays for higher frequencies or anything with a reasonable number of harmonics? If I go for a clock rate of 40kHz it's basically useless, but I'm not sure what frequencies common instruments go up to, in harmonics or in fundamentals.

>>1774315
Charge pumps are usually low enough power that they can get away with internal FETs, right?

>>1774317
Human hearing only captures things until a little bit above 20khz

>>1774317
>what frequencies common instruments go up to, in harmonics or in fundamentals.
Here's a guide that gives a rough idea. Just assume everything has harmonics
https://i.imgur.com/jr3NP7i.jpg

>>1774236
Isn't it basically just a sample and hold circuit? Two desynced jfets on either side control input, then output. Common gate amplifier jfet just takes a sip and stores voltage in the capacitor. I guess the Vdd to Out jfet pairs just average the samples or something.

>>1774317
>I was wondering if it would work in the opposite direction.
It's essentially quadrature logic. If you simply flip a switch "on-off," you lose information about direction. Voltage flows both forward and backward.
But if you have two clocks at a 90 degree phase shift, then CLK1on-CLK2on-CLK1off-CLK2off is seen as clockwise (forward) and CLK1on-CLK2off-CLK1off-CLK2on as ccw (backward)

>>1774303
Thanks mate! But basically if I solder the two metal parts with a metal blob current should flow? Just to avoid damaging the board

>>1774367
Not the guy you asked, but yes. Make sure not to burn the board or damage other parts. You can bridge the two pins either with copper wire, direct solder or any wire that's conductive. To make sure you aren't damaging the board, don't heat it too much, that's pretty much it. Use a lot of flux, you can just wipe it off later. Using just solder to bridge the pins might be harder than laying a bit of exposed wire inbetween the pins, soldering that on and cutting any excess.

I need quality BNC to RCA adapters. Anything I should look out for?

>>1774374
Thanks. I really should start using flux.. Btw I will try. I read the function of SB and I finally realized that yes, it a fucking BRIDGE. English is not my primary language and it really jumped over me the meaning of the word

>>1774388
No problem, language barriers with technical terms can be a bitch.

>I really should start using flux

You absolutely should, and apply it liberally. A decent thin solder wire that has some flux in it as well as a external way of applying flux makes an insane difference when it comes to soldering.

Does anybody have a link to cradked Eagle for Windows?

>>1774320
Yes, which is why you need to sample with a sampling frequency at least twice 20kHz. IIRC audio stuff uses something around 48kHz. It's the Nyquist sampling theorem.

>>1774342
Thanks.

>>1774348
Incorrect. It isn't a 90° phase shift, it's a 180° phase shift. The directionality comes instead from the buckets themselves being lifted/lowered by the clocks to force the charge to flow in one direction, see the image >>1774223. By connecting those capacitors to the opposite clocks and with no other changes you'd get data flowing in the other direction, assuming the ignoring and input stages on either end.

>>1774344
Well the common-gate amplifiers don't act like buffers, there's no way for the output impedance to be any much than the input impedance. But it is a small amount of gain that likely counterracts the resistive loss between each bucket movement. I guess you could tune the gain of the circuit by adjusting Vgg/Vcc. And the fact that the capacitors are tied to the clock lines instead of to ground is also a very important distinction to make.
Though the output stage at >>1774236 still doesn't make too much sense to me, is it meant to average two buckets? Or to show one bucket at one clock position and the other bucket at the other clock position? I also don't get the fact that the circuit ends at Vdd.

>>1774436
KiCAD is better than Eagle

>>1774453
Woah, I just realised I can use the multiple outputs of a BBD for some sort of analog DSP algorithm, like with those f(z), f(z-1) things that I did once

>>1774436
>Asking for fucking Eagle in 2020
>Can't even be assed to google it
Oof

>>1774315
IRF/Infineon has a bunch of options that are relatively cheap and quasi-jellybeanish enough to be sold by a bunch of alibay sellers, see IR21xx series for a variety of options, from single no-frills high (IR2117) side to self-oscillating half-bridge (IRS2153) to 3-phase full-bridge (IR2130), under a buck on ali

>>1774436
KiCAD loads Eagle layouts/schematics to some extent

>>1774453
>is it meant to average two buckets?
helps hold down the output with one bucket while the other is being "lifted", reduces output noise by a lot
>ends at Vdd
it's a reasonable place to dump that ~1pC of charge that has served its purpose. decouple it well

>>1774512
>helps hold down the output with one bucket while the other is being "lifted"
That's about what I assumed, I guess you could connect the 2 outputs of the MN3007 to have this same effect, which is actually done in the datasheet's example circuit.
>dump that ~1pC of charge
I'd have thought that leaving it floating with a 100nF cap (with parallel discharge resistor?) would do the trick, or just sink it into the input impedance of whatever device it's connected to. In the example circuit it's grounded, which isn't really what I think of when I hear Vdd, but you know, jfets.

The MN3007 also recommends an MN3101 BBD driver IC, which outputs the double clocks (presumably with a little dead-time) and also provides a Vgg of 14/15 times the 3101's input voltage. It's designed to run off a 15V supply (but can run from 8V to 16V) so this would output 14V, I'm not really sure why this is. The two clock outputs give at least 10V highs, which I imagine is to give a little headroom between the output and the V+ rail.

I'll probably add BBDs to my collection of of semi-obscure analogue ICs, alongside PLLs, mixers, TCAs, and optoFETs. Might get some cheap ADCs and DACs too, just to mess about with digital logic and audio. Maybe even some of those ultrasonic delay lines, but I think they're a little fast for most analog stuff I do/ Unless somehow mixing my signal to a higher frequency, delaying it, them mixing the result audible again would do anything.
Any suggestions?

>latest kreosan video:
>using a vacuum tube to make x-rays
>enough x-rays to be visible on a camera
>enough x-rays to make a geiger counter go ape from 3m away
>guy uses a sheet of X-ray reacting paper/phosphor to image the inside of a clamp-meter
>guy does the same thing to roughly see his hand bones
this is cool but also fucking scary

>>1773306
Retard barriers are fundamentally incompatible with anonymous posting. The only way to block retards is to associate them with a real person (with photo id e.g. drivers license) so if someone fucks up you can ban them and not have them immediately rejoin with an alt. Unfortunately this means manual verification of all applications so it's not scalable, but for communities where the rate of people joining is very low, it works well enough. E.g. old BBS systems.

Fucking engineers, why can't follow one fucking standard.

>>1774669
>he doesn't use coax for everything
balanced is a meme

Is BEAM robotics still a thing these days?
Virtually everything I can find is at least a decade old and the only store that even mentions it is a small shop in Calgary with a somewhat broken website.

I'm looking at powerbanks specifically outputting 5v 2.4A.
I am using the powerbank to power my camera and external monitor which require 7.4v.

Seeing as 99% of powerbanks use the 3.7v lipo batteries I am now confused as to why a majority of the powerbanks say 3A rather than 2.4.

Is this simply a marketing ploy or a thing they have to write for some safety reason?

I've tried the usb>9v dc adapters but they were outputting too much.
This usb>7.5v dc converter was much better.
It requires 2.4A 5v specifically though, hence I'm worried a 3A powerbank will have the same unwanted effect as the 9v converters.

>>1774812
for most electronics "V" is constant and "I" is variable in V=IR. your power bank puts out a fixed 5V and can put out anywhere from 0 to 3A depending on the R of your device. 3A is the maximum current your power bank can put out before it either melts or just isn't able to maintain a 5V output. it'll work fine with a 2.4A load.

>>1774826
Oh, so I'll be ok with a max 3A powerbank then. That opens up a lot more options.

Thanks, anon!

I want to ditch veroboard and start using PCBs, how hard can it be to learn to draw PCBs from existing veroboard layouts. Keep in mind I know nothing except some simple prototyping even though I've been building for years.

>>1774940
KiCAD (or any other full-featured PCB CAD software) lets you design your circuit diagram, assign footprints to the components, and export it to a PCB layout. Revisions to the diagram can easily be brought across to the layout. All you have to do is place the components where you want to and it will tell you which pads need to be connected to which other pads. You can draw the traces between them manually, or I think you can use autorouting. Ground fills and vias are easy, and the software also has provisions for making custom footprints and custom electrical symbols. You can use custom trace widths, custom clearance distances, custom via sizes. It has a 3D viewer, assuming you've installed the package full of 3D models for your footprints. It even has a microwave circuit design option, and there's a spice simulator built into the schematic builder. Best of all, it's free as in freedom.
Just use a circuit diagram instead of a veroboard layout.

I only have a couple of gripes with it, and they're probably just because I'm new to it. Firstly, I've not been able to set default footprints to components like resistors, or otherwise set a shortlist of recommended footprints. Since I mainly deal with THTs, there's really not much variety that I'll ever need in terms of footprints. Secondly, I can't get the 3D libraries to install properly. Whatever I do and however I do it, the 3D viewer just shows a flat piece of copperclad with my traces on it.

I'm using a little i^2c oled display with an stm32.
The... frame rate? Refresh rate? Display rate? is hilariously slow. I have it iterating through the alphabet and it refreshes at about 250ms per refresh, slow enough for me to count along with it while it shows the alphabet.

Obviously there's an SCL pin for clock purposes. Is there a way to speed up the SCL signal without overclocking/boosting the clock of the STM32 itself?

>>1774538
sounds pretty complete as long as you have some fairly decent low-drift caps, well, vacuum triodes can be made to work at 12V or so

>>1774951
you should try doing it the old way, assigning footprints en masse with CvPcb
if you can multi-select you can use the F hotkey in eeschema
>mainly deal in THTs
>current year
#include <judging/you.h>
>>1774960
data rate, yeah, I2C standard mode maxes out at like 10kB/s, fast mode at 40kB/s. keep in mind that the slave device can stretch the clock and regulate bit and byte flow in or out. find the I2C interface setup code in the program and set the timings to a faster speed (see the user manual for register details), then hope for the best

>>1774960
The clock speed is probably significantly lower than the STM32's clock, but rather limited by whatever internal I2C hardware it's got. You can likely program the I2C to go faster, especially if you're using an I2C OLED library of some sort, or possibly bit-bang it faster even if you're not, though I'm unsure if I2C has standardised clock speeds or not. Naturally the I2C receiving IC on the display itself will have some maximum clock rate on it, so I'd look up its datasheet. I2C isn't an ideal protocol for communicating with a display in the first place, I'd consider buying a display without the I2C interface. I2C is really only useful for low-to-mid speed comms with minimal pin-counts; you'll see that the displays on tablets and phones and the like all use like 15-40 pins on them. If you have the pins to spare on your STM32, this is definitely preferable.

>>1775013
>vacuum triodes
What do they have to do with BBDs?
>#include <judging/you.h>
They're what I already own for prototyping, and they make home-etching a more forgiving process. Not like I need my PCBs to be 12mm across. Only if I use enough of a component to buy two batches of it will I consider buying both THT and SMD versions of it.

>>1775014
oh, I thought you were looking for general component advice for screwing around with sound
BBDs are kinda rare these days, I think CoolAudio makes some clones of the old Matsushitas under similar part numbers
1206 passives and 1.27mm SOIC are perfectly useful even with such large trace/space as 0.5mm. also you don't have to drill as much. 1206 component kits on ali are cheap, or write a script to construct a bunch of order numbers from standard value series and upload them to LCSC's add BOM function

>>1775030
>general component advice for screwing around with sound
Oh yeah, but vacuum tubes aren't that cheap and they're a bit of a pain to work with/store/transport for fear of breaking them. But I'll definitely consider it. Diode tubes also sound interesting to have if their I/V graph is different to that of a silicon diode.
I'm more of a low-fi guy than a hi-fi guy, there any reason to go for low-drift caps? I don't see how ±15% would matter to audio stuff, except for making oscillators, or maybe fixed notch-filters. I'd have thought that not being microphonic would matter more. I've also seen people recommend to steer away from voltage-dependant capacitance, but I can't see that mattering much either.
>also you don't have to drill as much
That is a big reason why I would like to transition to SMD. Who knows, I might fold all my component leads at right-angles and use them as SMDs. Say hello to DIIC-14 and TOT-92!

>>1774292
>>1774303

Don’t do this. You won’t be able to break it in half, and you’ll damage the board trying.

Touch your iron to one end of the component and hold it there. Use your other hand to grip the component with tweezers and pull gently upward on it. The component itself will conduct heat through to the other pad. After a few seconds, the solder will melt at both ends and the the component will pop off.

Obviously not something you’d want to try unless the component was already fried anyway.

>>1775055
>You won’t be able to break it in half
Because it's a solid wire in there instead of a metal/carbon film? It seemed to work with the couple of resistors I needed to use it on. For SOICs I just cut the legs off with my side-cutters.
I'll be taking your technique though.

>>1773593
PCB's usually (99.9% of the time) have even layer numbers. If you really think this one is a 3 layer board then it's probably a 4 layer one with one of the layers not being used (but still having copper on it)

>>1775095
If I might interject for a moment, I would like to bring to your attention that you have just used an apostrophe incorrectly. I don't want this to interfere with the purpose of your post and in no way think that this small grammatical error degrades the validity of your message, but I kindly ask that you longer use an apostrophe on a pluralised word of any kind, be it an acronym or not. Apostrophes are to be limited to the indication of possession (Dupont's trademark) and the indication of a contraction (He didn't understand). With the notable exception of the word "its", which while it can be possessive, should never have an apostrophe, since the apostrophe is always used instead to denote the contractive form of "it is" for the sake of eliminating ambiguity. There's little sense in having a pluralised "it" to confuse the possessive "its" with.

By convention, most acronyms are written in upper-case, hence the lower-case letter "s" following the acronym's lettering should alone be sufficient to unambiguously denote pluralisation. Acronyms that do not follow this convention are typically proper nouns that make no sense to pluralise in the first place (entity names, technologies, et cetera), so the times where you'll need to pluralise a mixed-case or lower-case acronym should be exceedingly rare and dealt on a case-by-case basis as exceptions to the rule.

Thank you for your time.

I have an inverter welder with an analog pedal with two pots controlling the welder's amp setting. But it's a chinkshit welder and something inside isn't quite right, the pedal's control voltage range should be 0 to 19V, however when it gets near to 19V, the voltage starts to drop, and it's impossible to reach this end of the range, which is a problem since 19V is the 0%, so basically I can't set the welder to less than 40A, even though it can go down to 5A.
How do I fix this? I've thought about using an external power supply to give the control circuit more current to work with.

If i supply power to a stepper motor it turns. If i turn the stepper motor it generates power.
But how efficient is it?
What sort of motors do those huge wind turbines use?
What i am asking basically if using a stepper motor as a dynamo is comparable to efficiency of other motors you can use as a dynamo

>>1775192
Your main problem is that it generates high frequency AC instead of DC, but it is also less efficient. You could've gotten a comprehensive answer from google.

>>1775200
>frequency AC instead of DC
just slap on some diode rectifiers and a cap

>>1774623
>https://en.wikipedia.org/wiki/Goi%C3%A2nia_accident
my homestate ;)

>>1775192
>What sort of motors do those huge wind turbines use?
Most I've seen use synchronous motors, some use induction motors (they are a bit weird to use when in the network).

>>1775217
They are a bit weird to use. (anywhere really). If you want a easy generator just pick up a microwave motor (the one that spins the plate, not the fan. The fan is usually a shaded pole inductor motor). Because it's so tiny and has so many poles and a very very large gear ratio any little turn you give it can generate like 300V of high frequency AC (open circuit voltage)

>>1775219
Jesus fucking christ phone posting is hell.
Induction motors working as generators are a bit weird to use and not that common. Most windmills usually just use a synchronous motor with permanent magnets and do the good old AC-DC-AC conversion. Why you might ask? Well, they are synchronous machines, so their speed is directly linked to the electrical frequency they make. But unlike a hydro plant, you can't really control the wind (you can angle the blades but that is just for "throttle", not for frequency control). So it is easier to just convert whatever frequency it is giving to DC then properly alternate it for the network.

>>1775221
It is also common to use a mechanical interface to control the speed and make sure it always gives 60 or 50 Hz. If you want I can try explaining why induction generators are weird but it is just a brainfuck.

Wanting to get my feet wet with FPGAs, does anyone have any recommendations for dev boards to start with?

>>1775192
>What sort of motors do those huge wind turbines use?

It varies with brand.

>(you can angle the blades but that is just for "throttle", not for frequency control)

Funny you'd mention that. I have a few of picture related on my bench right now. It's the rotor current controller from a 660 kW Vestas wind turbine. It uses PWM and a GIANT 75 kW power resistor to precisely control the power output and rotor speed of the generator by shunting power as necessary.

>>1775095
Good to know, thanks. I actually just thought it was "more than 2 layers" and since not many traces were missing I thought 3 would be cheapest. If even numbers are what's normal then it's likely a 4 layer one as you said.

how does ups protect against overvoltage? is it just a bunch of 10 cent varistors that i am overpaying $50 for?

>>1775251
>rotor speed
but why would you waste all that energy as shunted heat instead of feeding it into the actual energy storage or grid?

>>1775262
Because these old turbines don't use inverters (as far as I'm aware; if they had inverters, I would have seen one on my bench by now), so the generator has to spin synchronously with the grid, within extremely tight tolerances. Newer turbines just use inverters instead, so rotor speed isn't as important.

>>1775259
google over protection circuit? It's that easy.

>>1775262
Not all energy is cost-efficient to capture and convert.

What is the output impedance of a MOSFET common-drain amplifier (used as a voltage buffer)?

Google suggests that the answer is "very low" or 1/transconductance-of-device. Is this accurate?

>>1775251
>
Funny you'd mention that. I have a few of picture related on my bench right now. It's the rotor current controller from a 660 kW Vestas wind turbine. It uses PWM and a GIANT 75 kW power resistor to precisely control the power output and rotor speed of the generator by shunting power as necessary.
Yes yes you do that do. But when the wind gets too strong, vibrations start getting above the max allowed you have to angle the blades.

Tektronix 1240 Logic Analyzer for $30, yay or nay?

>>1775318
If you like old timey stuff for that price its pretty good. Even for parts.
You can get a good 16ch usb LA for 60$ or so.

>>1775318
Yay
Yay
Yay

>>1775322
>>1775318
you can get a shitty USB logic analyzer equivalent to the low-end Saleae for about $5 but they call them "CY7C68013 breakout boards". it's plenty good enough for serial bus monitoring/hacking within its 24MHz sampling capability

>>1775347
Yeah they're great, I've got one and it's real useful, but if you're worried about insufficient input protection/grounding issues frying your computer I'd consider a standalone one.

>>1775316
What? Pitch angle is the main mode of speed adjustment (duh), the electronic control is only there for fine adjustment that needs to be much quicker than the pitch adjustment can allow for.

Hey so I recently moved and these LEDs I use for light in my aquarium are now strobing. Any idea what may be causing this?

>>1775423

>>1775423
>>1775424
Check the power supply, see if it's getting too hot.

Hi guys, what rookie mistakes have I made here? (Relay coils will be 5v)

>>1775581
add freewheel diodes to relay coils
add ceramic caps across all your ICs' power pins

>>1775594
Thanks for the advice.
I believe the Darlington array has diodes built in but should I add more to be safe?

>>1775595
Check the datasheet, the diodes probably won't be in the right position unless your power supply can handle the negative current spike. Even if it can handle the spike, you'll still be putting a ripple on the power rail which your more sensitive devices won't like if you're using that 5V rail for logic as well. Diode should go across the coil, not across the transistor.

>>1775596
Yeah makes sense, thanks.

>>1775099
n'o

>>1775581

470uF is prob too small. it implies the circuit uses less than 0.5A which may not be true if several relays are on simultaneously.

Quick question about wire gauges & ampacity

Say I have a 0.75mm dia (or 21AWG wire) wound up in a solenoid coil, and I'm pulsing 8.5A peak with a duration of 1sec ON & 1sec OFF, would the coil be ok heat wise?

>>1775605
Depends on the size of the solenoid. The ampacity of a wire is calculated by with the rate of heat dissipation, which is a function of surface area. Since the effective surface area for removing heat decreases if you wrap the wire up against itself, the effective ampacity will also decrease. Calculate the average power dissipation and do some rough mathematics to figure out by what factor the surface area has decreased from a straight wire. Then compare that to a standard ampacity metric, or a similarly sized heat-sink, or something else with thermal data you can nab for reference.

>>1775608
Hmm probably easier to do some bench tests

Here are the specs for coil if anyone is interested:
ID 13mm
OD 26mm
L 54mm

>>1775609
surface area in the realm of 100,000mm^2 (that's not the european comma)
how long is the wire?

So i hooked up one of those 12v rc plane lipo batteries to a breadboard and smoke started coming from the holes. It just had some ics etc. What happened?

>>1775613
You shorted something. Is the lipo hot? Do you understand the breadboard's pin layout?

>>1775616
Yeah i didn't short anything

>>1775618
Then where did the energy for the heat that caused the smoke come from? Smoke was coming out of the holes, meaning it's probably the plastic or a coating on the metal inserts that was releasing the smoke. Unless it was only coming out of holes next to a component, and even then I find it hard to believe that smoke got projected downwards and came out the holes.
Were you using any 7400 series ICs? They're 5V only AFAIK. Same applies to any other low-voltage ICs or transistors. Some MOSFETs may have a maximum VGS as low as the maximum lipo voltage. It's also possible that you got the polarity around the wrong way or otherwise miswired it.

>>1775618
maybe you got some pinouts wrong? just brainstorming

Hey brehs how the fuck do I tell if an output transistor is blown ? I did the good old diode test and they all came out good. I noticed that all 4 of them make a strange ringing sound when shaken. Is this normal ? These are old style TO3 transistors btw. The amp has an issues where the gain is pitiful and it has to be turned up to eleven to work. I was thinking the transistors gain might have gone to shit.

>>1775610
Used an online calculator and it said ~30m

>>1775771
measure its hfe

It's this alright for a power supply?
The mosfet on the capacitance multiplier would be an IRF530, I'm mainly worried about the ripple from the buck regulator.

>>1775844
I forgot to add, the Voltmeter/Ammeter consumes about 20mA, so I think it is alright to drop 12V with the LM317.

Hi. I have an expensive SACD player that got it's transformer burnt to a crisp in a storm. I don't want to throw it away so I considered rewinding it. After investigation I narrowed the fault to one of the secondaries because only part of the device got power. I managed to find a replacement part but it is for US voltage. Since getting a new transformer is out of the question I've been thinking that getting a US voltage trafo and rewinding the primary will be a gorrilion times easier than rewinding the secondaries. Am I on the right track or am I retarded ? If rewinding is the only option do you have any tips or ressources that could help ? Thanks.

>>1775853
Also forgot to mention that the trafo is a Kitamura Kiden R type fancy transformer expensive audio equipment oblige. Getting it apart will probably drive me mad.

>>1775844
You might be dropping 2-4V across that PFET due to its VGS-threshold, sure it's not going to overheat?

>>1775853
If it's a step-down transformer, there will be more windings on the primary than the secondary, so it may well be easier to rewind the secondary, even if there's more than one secondary you need to rewind. Are you able to take them apart and just slip the primary off one and onto the other?

>>1775853

there is, or was, a company in my city that did custom xformers and would rewind one for you for about $25. dont have a name coz it was long ago, but presumably google will locate one near you.

>>1775853
why not just use a commodity power trafo of equivalent spec and suitable shielding? it's a power trafo, it probably won't ruin your tone

>>1775853
>but it is for US voltage
There are 2:1 step-down transformers for that.

>>1776011
>a commodity power trafo
Not him but I'm assuming it has like 6 windings on it

is there any reason to own complimentary jfets? do they exist?

>>1776067
>I'm assuming it has like 6 windings on it

what a moron! it has only five.

yes! my puss bois full of corona from chinks are finally here
i can finally pretend to be luis rossman

>>1776098
oh shit i might be legit fucked, this thing is even more dangerous than corona

>>1776099
>reproductive ham
Nice

>>1776011
It has multiple windings and the OG trafo is a low profile ultra low noise R type transformer.
Even if it doesn't ruin the sound an equivalent EI type won't fit in the enclosure.
>>1776063
If I already had one I would just go for that. Problem is these inject a lot of garbage everywhere and finding one that won't will probably cost me as much as having it rewound by a pro.
>>1775976
I'll check for these. Thanks.
>>1775862
R type transformers are the easiest to rewind due to how they are constructed.
https://youtu.be/hY0b3emLGIE
The primary is on one side the secondaries on another.

>>1776098
>luis rossman
Do you really want to be him?
He is basically a glorified flux monkey.
Look at the repairs he does, he basically puts a board under a microscope, looks for rusted or damaged parts, then desolders them, takes identical board form another mac, takes that identical chip from it and solders it on the original board. Done. That is 99% of his repairs. he sometimes multimeters the power rails to narrow down which part of the mac is borked, like the display, or the cpu etc, but thats it. He has actually no idea how the machine works, or how to design a computer like that.
I am not shitting on him, i am just saying his job is shot and no the fun EE adventure people think it is. The best EE shit you can get into is actually designing circuits

>>1776112
Yeah he makes solid dosh doing easy repairs for faggot macbook owning new yorkers

>>1776129
That is not disputed here, he makes a bundle, like $250 per 20 minute board repair that is bill gates tier income, but that doesn¨t change the fact the work it self is tedious and boring
personally i always prefer having a fun job i enjoy for less money, than making a lot but fearing the send of the alarm clock in the morning every mon - fri

ive had 3 amazon packages missing. whats the cheapest way I can put a device inside a dummy box that will give me its location on my phone?

>>1776130
>do it for a year
>put that money in a good ETF
>enjoy the rest of your life

Hey, I need some ideas for measuring linear displacement.
Basically I have a shaft inside a pipe. A shaft is able to slide freely to a certain position. I need to measure that displacement (it is in centimeter range).
When I checked possible methods of measurements, I came to two most easily apprehended - LVDT and linear potentiometer. Heck, but those are quite expensive, especially LVDT.
I need at least five methods to compare and pick one who's relatively accurate and cheap to make it alive.
Other methods that I am thinking of are using multiple Hall Sensors with a magnet attached to an end of shaft.
Is that it? There probably are more great methods. I dug into ResearchGate and all I can find is overkill methods, such as laser or ultrasonic.
Give your advice or methods if you have any, thanks!

>>1776211
Laser ranger sensors cost like 5 bucks, is that really overkill? I've used some with around mm accuracy out to a meter that were not expensive at all.

how can I test more than one combination of bits? So how to transmit multiple codes, instead of only one code with GNURadio and HackRF?

>>1776213
I meant laser sensors who work on triangulation principle, I guess you are suggesting Time-of-flight type sensors, like VL53L0X. Actually thanks, I'll take a dive into them, but I still need some more advice.

Fuck it, comment’s lost, so I’ll type this quick.
Basically I need to create a converter for my trailer (sold as a two to three tail light converter) that takes two inputs, representing left turn, right turn, and when both combined, brake, and split it into three outputs, left, right, and brake.
I thought the easiest way to do this would be one relay on each signal line, and when both relays are active, they would allow current to pass through them both, creating the third signal.
Additionally, I need to make this without soldering, lest I buy a bunch of soldering equipping, so crimping connectors to automotive relays is preferable.
It’s shit but if it works I’m happy. Any thoughts?

>>1776222
I've actually only done retard tier stuff with this in robotics applications, so I know nothing more than looking at sensor spec sheets and just ordering the one that does what I need it to. I was talking about ToF sensors though, cheap and works.

>>1776211
what precision/resolution?

>>1776227
Centimeter-subcentimeter, while resolution in subcentimeter range would be great.

How do I actually implement chassis/frame ground in a PCB design? Just run a wire to the frame? Should I cut a section out of the ground plane and tie that to frame, then run components to it with vias like a normal ground plane?

>>1776233
I hardly imagine why on earth you would need that. Obviously you would need to design a cutout section somewhere in the corner of PCB

>>1776227
yo, have any ideas?

>>1776241
For USB, shield goes to ground and data is tied to frame/signal ground through TVS/zener diodes. I'm just going off the FTDI spec sheet, obviously they won't come to my house and shoot my dog if I just ignore it but I can't see any reason to since this is just a lazy hobby I have. The USB shield is just going to a ground plane, my question is more what benefits there would be for a separate ground plane tied to frame vs. just running a wire from a pad or via to frame like a truck or something. The USB and the signal traces in general are close to the edge of the PCB anyway so I could not care less about routing, I'm just wondering if it matters either way since I've never messed with something like this before.

>>1776244
I guess USB shielding is just for EMC. I thought at first that you are doing ground (earth) plane for ground wire (AC).
I doubt that there is a need for a separate ground then. The voltage induced in shield will either very small or electrostatic (low energy, low current), so I can't see any meaning for separate ground. Although I'm no pro in EMC, so maybe someone who knows better can give better advice, but I doubt it is necessary.

How can I learn using GNURadio? I think it seems complex but it's easier than it seems.

>>1776217
u wot

>>1776223
You need to ensure that when both are on, the indicators are off, so you can't just use that sort of AND gate while ignoring the indicators themselves. Worst case it would take like 3-4 relays, which might get a little messy. Do the indicator lines go on and off at ~1.5Hz, or do they just go flat on and the oscillator is mounted in the trailer? Either way you'll want to draw up a truth table. Since this is 12V we're dealing with, I suspect you can use diode logic without an issue.
I'm a little worried at the lack of a ground wire on the output of that image though.

>>1776230
Yeah it sounds like you're just within the range that a laser one will work. The time-digital-converters I looked at on alibay had resolutions of ~7mm, so the higher spec ones built-into dedicated linear distance sensors should work even better.

>>1776315
Have you asked on the /ham/ thread? I think you can get more info over there

Favorite/Least Favorite part about Electronics/Electrical Engineering?

Not sure if this is the right thread, but I bought a damaged DMG-01 gameboy recently (it was physically in decent shape, but it wouldn't power on). After I fixed it up, I decided to start modding it. First mod was a rechargeable battery. Put a nice fat 3000mAh battery in it, and rewired the system's power jack to charge the battery.

Been thinking about what to do with the LEDs. I don't want to modify the case (hence, rewiring the power jack instead of just cutting a hole for a USB port), so I don't want to cut a hole for a charging LED or anything. I also don't know if there's any simple way for me to set up a low battery LED, since I don't think the original model of gameboy actually detects low battery in the first place--it just dies.

Also not sure if I want to do a backlight mod. The pins on the display's ribbon cable are already slightly corroded and extremely fragile to begin with, so I don't know if I have the gonads to desolder it.

>>1776525
Are you using a proper charge controller and protection circuit? Those often have their own pins to attach an LED indicator to. If you're not using such protection and charge circuitry, what the fuck are you doing

>>1776498
Favorite:Building things and seeing them work
Least favorite: realizing that a undergrad student is in the lowest tier ever, I still need 1,5 years to graduate.

>>1776498
Favourite: starting somewhere really obtuse or difficult and slowly working through the problems and feeling that progress. Also the initial planning phase, because I never finish anything.
Least Favourite: being wholly unqualified to actually get a job with electronics, not having enough money to attempt a project, not having a panavice. Maybe I should buy one of those chinky knockoff bench vices with terrible tolerances, I've heard they're decent if you shim their sliding bits.

>>1776577
>being wholly unqualified to actually get a job with electronics,
I feel the same way even though I have a degree in mechatronics

>>1776548
Of course I am. The circuit I'm using even came with LEDs. The problem isn't wiring the LEDs. It's wiring them in a way that looks nice.

is it 240 MOV? cant find marking anywhere and cant test it myself

bros help me i feel dumb:

say I want to heat up a room, heat power is R*I^2. So the higher the current the more heat I provide. In the US, we have 110V, but in europe they have 220V. So at equal power, 110V systems give twice the current therefore 4 times the heat. Is that right?
But heat power is also (V^2)/R. So what now?

>>1776650
equal power is equal power

>>1776650
if power is same it ll heat equaly. p=iu=i2r, i=u/r so in order to have same power heater you would have to change both current and heater resistance. if you use 120w heater in 230 it ll draw more current and possibly overheat and other way around

I need to have few micocontrollers remotely around.
What device should I get when I value:
>should be small
>long battery life (button cell battery)
>cheap
Looking maybe BLE, wifi or infrared might even do it but not sure.

>>1776684
>whatever + wireless modules from aliexpress
My first priority is getting one you have the tools and capability to program.

>>1776767
>you have the tools
assuming they are most likely arm so gcc will do fine. If device without usb port is the best solution getting needed cables and peripherals to flash shouldn't be too expensive consider I'm getting multiple devices anyways?
esp32 doesn't seem to bad.

>>1776622
Ah, I thought you had LEDs built-in to the case already and were wondering how to incorporate them with your lithium ion circuit.
Why not just have them shining out through the plastic? You could sand a thin spot for them, or make a little box to prevent the light bleeding out too far.

>>1776628
>cant test it myself
yes you can, put a 1M resistor in series and slap it across 240V mains, then measure the voltage across the resistor.

>>1776650
Wrong. A 1kW heater made for 100V mains will draw 10A of current, and hence be a 10Ω heater. This 10Ω is the constant here. If you put this heater across 200V instead, it will draw I = 200/10 = 20A of current, and P = 200*20 = 4kW. Double the voltage, and for the same resistance you'll get 4 times the power.
What's important is knowing what variables are dependant and what variables are independant. My answer also ignores the thermal switch that heaters typically have to prevent overheating, which would drop the power output from 4kW to something closer to 1kW.

I'm a little new to this type of circuitry so forgive any dumb things I'm about to say. I bought two PCBs, one's a MCP73871 board and the other is a 4A boost converter board that claims to boost to 5V on the low end.

I've got everything wired up and the output power of the controller board is around ~2.9V and I've trimmed the pot down as far as possible but the output of the boost converter is still high at ~8.62V. I know the input voltage is not *exactly* in the input range but I can't figure out why this converter is hanging high. Is it because there's no load? The MCP datasheet is a little unclear about what kind of voltage output, too.

MCP board: https://www.ebay.com/itm/Battery-Charging-Board-MCP73871-USB-Solar-Li-Ion-Lithium-Lon-Polymer-DC-5V-1A/202357242346?ssPageName=STRK%3AMEBIDX%3AIT&var=502341962684&_trksid=p2057872.m2749.l2649

boost converter: https://www.ebay.com/itm/4A-Boost-Converter-DC-DC-3V-32V-to-5V-35V-6V-12V-24V-Step-Up-Power-Supply-Module/173744213303?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2649

>>1776932
Put a 10k or even 1k resistor load across it and see if that changes it? Not the best solution for a battery-powered solution, but you could at least check if it works. Else you could desolder the pot and replace it with more extreme discrete resistor values, I've done that a couple of times before, though I can't say I got no adverse effects.

>>1776936
Yeah I'm considering adding a load in there to see, but actually I might have a bigger problem. The battery doesn't actually charge if the converter is connected, I suspect because of the idle power consumption of the converter is too high.

That's a little unfortunate because I bought this charge controller board for the load balancing -- the idea was that I could just plug in the power and the chip would route power to the converter and charge the batteries instead, as a laptop does.

>>1776938
Yeah, I too would be a bit miffed if I got some sort of fancy BMS only to find out it can't charge and discharge at the same time. Sounds like you'd be in the market for a "USB power bank controller"-style of circuit, though whether you can get one of them up to 4A is another issue.

How big is your cell/battery anyhow? You could replace it with two seperate cells/batteries and hotswap them. Or increase your cell voltage so you don't need a boost converter, just a buck (which typically have a lower standby current).

>>1776941
The MCP chip is pretty complicated so I'm gonna read the datasheet really carefully to hopefully find some solution. I'm trying to power an SBC, so the power consumption on these is pretty nominal. I've seen benchmarks that give only ~1.3A at full power, so a 2A cap is probably fine. I want to add a small screen (like 4-7") though.

It isn't imperative that it manage up to 4A. It's been around six weeks since I designed this circuit and I looked at power bank circuits but can't remember why I didn't do it. I think some anon in a previous /ohm/ said that I couldn't trust those circuits to seamlessly change the supply for my application.

I haven't fully specced the batteries yet. Of course I have this single-cell MCP chip and a boost so I would prefer any 1xN configuration, but I'm not married to it if a 2xN configuration works better. I only spent $4 on the MCP and boost anyhow, the batteries were salvaged.

>>1776947
4-7" could potentially draw more than 1A depending on how strong the backlight is, I'd look for some potential specs before you assume what will and won't fit within your power budget. There might also be power bank circuits that can handle more than 2A, especially these days with USB C pulling I think up to 3A at 5V.
>some anon in a previous /ohm/ said that I couldn't trust those circuits to seamlessly change the supply
I'm no expert on this but it sounds possible, USB power bank circuits are designed to charge things that have their own battery, so they don't necessarily have to provide continuous power during the swapover. I'd buy one or two anyhow and test with a scope, see how long the downtime is, if there is one at all.

>>1776949
I did some current measurements and I'm pretty puzzled by how this thing works. With the battery plugged in the boost converter draws 120mA. That's a lot, and the floor voltage is what I said above.

If I disconnect the battery then I can trim it down to 5V, actually all the way down to 3.3V I think. The current draw then is 30mA, which is closer to what the specs say for idle current draw.

If I disconnect the boost and reconnect the battery, the battery appears to only draw 20mA, which I don't think is quite right.

The wall plug I'm using can deliver 2A so that isn't the problem. At the 30mA idle I should easily be able to charge the batteries too, the datasheet says the upper limit is ~1800mA. But there's clearly some kind of weird interaction; the datasheet says it does not regulate output voltage so I expect the boost voltage to spike when I connect the battery up, too, but why the power draw goes up by so much is a mystery. For what it's worth it does charge the battery just fine if the boost isn't plugged in.

There must be some kind of fuckery with the on-board SMD resistors. But I don't know what just yet.

>>1776958
What's the wall-plug voltage? Sounds like it's a 9V adapter that's losing 0.3V through a (schottky) diode. But that would be a silly mistake, and I don't want to insult you by implying that you did so.

>>1776964
I'm not that good at electronics so it's entirely likely I made some dumb mistake. It's a USB plug that supplies 5V/2A or 9V/1.67A.

Turns out there's a "precondition" phase where it trickle charges the batteries, but I have yet to figure out the set thresholds.

>>1776965
>or 9V/1.67A
That sounds a bit unusual. Is there some sort of switching mechanism there? Did you measure the voltage before the boost converter when it's plugged in and with the batteries installed?

>>1776966
>a bit unusual
I agree, it is visually a totally ordinary USB plug. Absolutely no indication how you could get 9V out of it. I only noticed it when I checked to make sure it supplied 2A.

The voltage at the output of the MCP board with the batteries installed is 4.56V, that's without the boost connected. If I connect the boost and measure at the same points I get 3.17V. The chip is not a voltage regulator AFAIK.

>>1776968
The boost converter can run as low as 3V, right? If it was in constant brownout that wouldn't be good.

>>1776970
Yeah the input voltage says it can run as low as 3V.

I just confirmed that it isn't the battery protection PCB I got, too. The thing is I know this chip can be used for this purpose because the original part that I was being stingy and not buying uses it:

https://www.adafruit.com/product/2465

So if Adafruit is to be trusted then it is possible to get at least 1.8A out of it. But the difference is I got my MCP board from some guy in China and who knows who designed this one.

>>1776976
I guess its possible that it's using out-of-spec resistors? But even then this is quite the discrepancy. Have any power resistors you can load the MCB board with in place of the booster?

goddamnit recaptcha stop making me train your robot dogs

>>1776965
the precondition charge is preset to 10% of the fast charge current, datasheet 4.4
make sure SEL is high, see the datasheet section 3.4, "Input Source Type Selection (SEL)"
>9V/1.67A
there may be a quick charge protocol that you can activate by placing resistors between either data lead and VBUS and/or GND. on the 73871's end see

>>1776947
you might want to power manage your SBC carefully, eg shut down cores and peripherals you aren't currently using. a stiff 4.5V should be enough to power a carefully managed system
have you checked out some of the benchmarq/TI chips? bq2419(5|6) is pretty nice, has a higher output current capability. rare or out of print on chinkboards afaict

>>1776684
wi-fi is a high-power protocol, can be a bit much for a coin battery
look into the nRF51822 or some of the other Nordic BLE micros. even there, battery life is mostly a function of how aggressively you manage power in software. and also of clock frequency

>>1773533
>>1773537
>>1773395
Ok so I bought this sensor and tried it out but it's not nearly sensitive enough on the wavelengths i need, 950nm, and the IR diode measures all the way down into plain old visible light, I need something with a much tighter spectrum, 850-1000 would be ideal, this sensor is entirely useless because I'm trying to detect half second pulses of Ir light and it has a hard time even refreshing that fast and it's not sensitive at all to the 940nm light when the leds are more than a couple feet away.
I'm really not sure where to go from here, I either need an IR photodiode or IR phototransistor but every IR phototransistor I can find has built in 38khz filters to eliminate interference and I need something that detects straight up DC output with no modulation.

>>1777004
Buy a dedicated IR-receiving photodiode. They're designed to receive signals emitted by IR LEDs, so I assume they have a reasonably tight band-pass IR filter on them. Probably says so in the datasheet.
Like this random one I found on Mouser, it has a nice spectral-sensitivity/wavelength graph:
https://nz.mouser.com/datasheet/2/445/1540051EC3590-1714684.pdf

>>1777004
You can use any IR LED as a receiver, you'll need to use a opamp tho. Or you can buy one of those TCRT5000 emitter receiver pairs (and take the receiver) they use for limit switches and shit.

>>1777009
Also I use IR sensors all the time (either diodes or phototransistors as the TCRT5000) and I never ever heard of a filter that was put in place but not mentioned in the datasheet

>>1777004
And you bought a sensor for measuring light intensity (lux) so why would you expect it to work with what you are doing? I wish I had this much disposable income.
My monkey brazilian genes are hardwired to be a cheap shit so:
>get broken remote from the trash
>get IR LED
>use said LED as a photoreceiver in one of the many configurations possible
>????
Cutting the top rounded part of the LED to make a flat surface, or using a lens of some sort might help.

>>1777009
btw, I used those to make a PPG thing and it worked like a charm, so it works well with low frequencies.

>>1777011
>I wish I had this much disposable income.
You wish you had 6 dollars?
that's kinda weird but ok
I spend that much on a coffee every day... wtf man

>>1777008
I'm not sure if a photodiode is what I'm after or a phototransistor, I want something with really high sensitivity to detect an IR led from 50-100ft away in darkness, the burst of IR light is around half a second long and I will like have about 6 or 8 IR diodes/transistors aiming in a circle to receive from 360 degrees around me.

>>1777009
Yes but IR LEDs don't have any filter on them, they'll pick up anything from ~1000nm all the way up through UV. Not like solar panels need to have photodiodes tuned to 50 different wavelengths to get the full spectrum, that's not how the photovoltaic effect works. You need an optical filter to get rid of the visible spectrum, hence why dedicated IR photodiodes often look dark-blue or black, compared to IR LEDs which often do not.

>>1777014
>6 dollars
you forgot the $20 per package that it takes to ship to his backwards country to stop the capoeira gangs from zoom-zoom-zooming his ass

>>1777015
There's fuck-all difference between a photodiode and a phototransistor, the latter effectively acts as a photodiode with a current gain stage in there (like a BJT), for all intents and purposes. In reality there's some distinct differences but they don't matter to someone using them like us. Phototransistors are probably better as binary sensors since they saturate more easily, while phototransistors are probably better as analog sensors. Unless you really need the gain for, perhaps, a pulse oximeter where you need to receive light through a person's finger. Photodiodes are probably better for higher frequencies.
Anyhow, chuck your photodiode in reverse bias with a pullup resistor. Resistor can be anywhere from 1k to 10M, really depends on how much gain you need and how much noise you're willing to put up with. For any one application you should be able to find a sweet-spot where the phototransistor is dropping 1/2 of the voltage in steady-state, and is near its maximum signal to noise ratio. This takes a bit of fiddling around. Having it away from both rails also makes it easier for an ordinary comparator to trigger off it, but leaving it a bit lopsided to fit within an MCU's ADC reference voltage is fine too.

Also google:
"nbp940"
https://www.aliexpress.com/item/32903871066.html

>>1777017
This is based but some older cameras still use 850nm LEDs that glow dim red
Guess imma have to order some components and fiddle around with this bitch

>>1777014
That's like a lot in brazil nowadays thanks to bolsonero.
>>1777016
>you forgot the $20 per package that it takes to ship to his backwards country to stop the capoeira gangs from zoom-zoom-zooming his ass
Yes. And the 3$ I have to pay the mail system for them not to take my package hostage (I'm not kidding btw)
>Yes but IR LEDs don't have any filter on them, they'll pick up anything from ~1000nm all the way up through UV. Not like solar panels need to have photodiodes tuned to 50 different wavelengths to get the full spectrum, that's not how the photovoltaic effect works. You need an optical filter to get rid of the visible spectrum, hence why dedicated IR photodiodes often look dark-blue or black, compared to IR LEDs which often do not.
yeah yeah I know but common it's a better idea than a lux sensor.

>>1777015
>I want something with really high sensitivity to detect an IR led from 50-100ft
You want a IR camera then, because you gonna need a lens for that.

>>1777021
Well, it's an array of like 30 IR leds, not just one.

>>1777021
also it's not completely unrealistic to expect a TV remote to work from 50ft away and that's a single low intensity IR led, albeit it's using a 38khz carrier to block interference but there wont be any stray light really if using this at night on a trail.
I don't think it's unrealistic.

>>1777019
>850nm LEDs that glow dim red
Oh I've seen those around the security cameras in the airport shuttles. When we go into a tunnel they're visible. Didn't know those were meant to be IR, but I suppose it makes sense. Makes me wonder if staring at them for too long could damage my eyes.

They have nbp850s too, which is good. But no nbp905s for me to make a laser jammer with.

>>1777020
>a better idea than a lux sensor
I think that lux sensor just had 4 photodiodes in there of different semiconductor materials, and they'd use differential measurement to figure out how much of each there is. One of which was a standard IR photodiode. It's how I'd do it at the least.

>>1777021
Not necessarily, I doubt a CCD will have a higher SNR than a single much larger photodiode die. Not like directionality matters. Remember, the whole point of the IR LEDs on the cameras is to illuminate him enough to be visible on camera, so if he's illuminated enough to be visible by a security camera, the photodiodes shouldn't have a tough time picking it up. He also plans to use it at night, which means the small amount of visible light that passes through the filter shouldn't matter either, even with a default filter like >>1777008.

>>1777023
Ah, it makes it easier. Well If it were me I'd just buy the phototransistors or whatever grab a camera and test it until it worked.
>>1777024
Yeah but the problem is birds, people and most other shit also give out IR.

After managing to repair my old dishwasher I want to try and repair an old HiFi amplifier from the 90s, pic related, that heavily distorts the sound. I hoped to find some obvious damage, but upon visual inspection I didn't see anything obviously burnt out. My knowledge of electronics was enough to troubleshoot a simple electromechanical machine with a cheap multimeter, but an amplifier is outside my current knowledge and instrumentation. Can you recommend me a textbook to fill the gap? Also, what tools would I need for the job?

>>1777030
>Yeah but the problem is birds, people and most other shit also give out IR.
That's at 10000nm, mid IR
he cares about 1000nm, near IR
there's a big difference
the only thing that will pick up on mid IR is a passive IR sensor, which literally uses the energy radiating off the heat source to heat up a tiny piece of pyroelectric material and measure the change in charge across this material. this method is used both by PIR motion detectors and IR thermometers.
there simply aren't any semiconductors with a band-gap that small, if there were we'd be using 0.5V TTL
the photovoltaic effect is quite similar to the photoelectric effect, they're both fundamentally quantum phenomena that require a certain threshold energy for a single photon to liberate a single electron. the electron's threshold energy is about 1eV for silicon and this corresponds to somewhere around 1000nm. doesn't matter how many of them you have, if you throw trillions of 2000nm photons at a silicon photodiode it ain't budging.

Hey, I need some advice on accelerometers. I plan on measuring linear displacement with accelerometer with Kalman filtering. Measurement has to be as fast as possible (tens of milimeters). Displacement should be around 10 cm. Some vibrations in the system from outside may be induced, so that should be taken into account. How to choose accelerometer knowing this information? Thanks

>>1777031
>looks at your post
>looks at OP
>sees "Geier, How to Diagnose & Fix Everything Electronic" is still there
>looks back at your post
hmm

I doubt anything will be burnt out if it's still working but distorted, check if the electrolytic caps are bulging. It's possible that your thermal goop has gone bad or something, but I've never seen it.
Also a class-D a 10th the size of that could output the same amount of power, so don't expect anything stunning compared to 2020. But wanting to fix something is always good.

>>1777033
>inertial navigation
Look for the ones with the tightest tempco and other tolerances. You'll also need to measure at a faster sample-rate than twice that of the maximum vibration frequency. Putting the sensor itself on a damped springloaded bed is a probably a good idea.
But I strongly advise you to consider an ultrasonic, laser, or surface optical/magnetic linear encoder instead. Even a geared system's backlash would be easier to handle than the drift coming from your finite integral. You might be able to calculate an estimation of the drift as a function of quantisation error.

>>1777036
I will use laser and magnetic methods too, but I need to do a research on all cheap methods possible, one of them being accelerometer with Kalman filtering. What about g rating of accelerometer? How it relates to sensitivity of accelerometer?

what is the difference between oscilloscope and spectrum analyzer

>>1777034
It's more for the feel good sensation of having fixed something, beside I'm not an audiophile so I don't really care about performance as long as it works properly.

>>1777042
Same difference as watching music video and listening to music separately. One shows waveform, time domain, other shows frequency content, on frequency axis, it shows signal composition.

>>1777054
i think most scopes do frequency domain as well now, no?

>>1777030
>Well If it were me I'd just buy the phototransistors or whatever grab a camera and test it until it worked.
I bought a trail cam at walmart for testing purposes, it was only 30 dollars surprisingly, the image quality is kinda ass but it'll be ok for testing.
This image I was about 50ft away
If it can light me up this good that far away, I'm sure that it can make some results with a photodiode or transistor, I'm just too retarded to figure it out on my own kek
Thank GOD I have 4chan helping me :3

>>1777056
Not all of them, plus they do calculate spectrum with FFT, do not measure it directly like analysers do

>>1777058
Ok, BR anon here. My experience with photodiodes/transistors was with a Pulse oxymeter and detecting stuff at small ranges. But here are my final 2 cents
>don't overthink this, just get the transducer that seems to fit.
Near infrared? Ok, get a near infrared IR detector (by detector I mean the transducer, not a "sensor" from whatever that you have no idea of what it does).
>you'll likely need to amplify and filter whatever you receive. I don't think that'll find a lot of IR noise in the conditions you have described, so you can amplify a lot. This means that if you hook up everything and it doesn't work, try it with a closer light and figure the gain you need
Good luck and drop the trip

>>1777109
If everything fails you could see if the camera triggers PIR motion detectors and just use those.

>>1777109
Huh? I have never used a trip, this is a name.
Are you blind?

Can someone explain superheterodyne magic to me? The principle as I understand it is you generate a sine wave (local oscillator) then mix it with an incoming signal. The result of this is a new signal containing two sets of frequencies: One is local oscillator frequency + the input frequency. The other is local oscillator frequency - input frequency.

What I don't understand is if this only works with input sine waves. For example if my input were a 10 khz triangle wave and the local oscillator were 1khz would the output be an 11khz triangle wave + a 9khz triangle wave? Also if A-B were below 0Hz would that mean the output only contains A+B?

What I want to do is take an arbitrary input waveform and increase its frequency without effecting its shape and I'm trying to figure out if a superheterodyn mixer can do that.

my country literally ran out of lm386 and 555 chips
what's a good store for components that ships around europe?

Poorfag here. What's the cheapest viable oscilloscope for someone getting into electronics?

>>1777158
lots of people suggest the rigol ds1054z for entry level scopes. It's 400 gold packs though.
Less than that you're looking at chinesium pos or usb scopes

>>1777163
That looks real nice, but I was looking for something a little cheaper, like an entire order of magnitude cheaper. I'm sure the chinesium will work just fine for me.

>>1777166
supposedly you can get $50-60 scopes on ebay, usually ~20MHz bw and analog
There's also a cheap little dso138 kit if you only care about audio stuff. Caps out at 200kHz though, which is kind of shit

>>1777155
where?

>>1777119
This is just what I learned from watching Technology Connections' video on it.
Basically, you couldn't amplify an AM signal at 500kHz - 1.7MHz with the primitive amplifiers of the time, they'd be limited by their gain-bandwidth-product. But their oscillators lacked the stability and the accuracy to tune it directly to the carrier frequency to directly demodulate the signal through a mixer. So they mixed the signal down to approximately 100kHz or wherever and putting it through a band-pass filter, the signal now at a point that their amplifier could amplify without issue. Doing this doesn't require an oscillator to be nearly as accurate as the direct-mixing method. They'd then feed this through a common diode detector + audio filter circuit.

These days we have solutions to both of these problems. Both amplifiers that can function up past 2MHz (basically every modern op-amp) and oscillators that can lock onto the carrier frequency (PLLs), meaning you could go for either mix the signal right down to audible and amplify it with an audio amp with a PLL as your oscillator, or amplify the RF signal before demixing it or using a diode detector on it. Diode detectors working on unamplified signals don't work too well due to the impedances and voltage drops involved (see: crystal radio), but these days you could ignore all those issues with an op-amp ideal diode. There are probably some more clever envelope-following circuits out there using op-amps.

>>1777158
Avoid chinesium scopes, USB scopes are kinda ok but still somewhat expensive. I got one of those DSO138s and while it helped a little and was fun to solder, I constantly found myself second-guessing my measurements and it only has 1 channel. 1 channel is simply insufficient for basically everything you'd ever do on a scope. Go second-hand. Ask around, keep an eye on local auctions, etc. CRT scopes aren't outdated, especially if you nab a digital storage CRT scope. I'd only upgrade when I can get a 4 channel.

>>1777119
Linearity.. Signals can be seem as a infinite sum of sines, for some reason god agreed with fourier on that one so a triangle wave is just a bunch of them in a specific pattern. Grab a introducton on linear systems. They usually explain modualtion on those texts.

So have some nixie tubes and I've got them working with a rasberry pi... What should I have the numbers display?

>>1777227

>>1777227
a counter for the random junk in your drawer

>>1777231
I have 7, but one of them the 4 digit is not working.

>>1777233
I'm going to need more nixie tubes...

>>1777234
And here is my messy as desk with a single nixie tube connected the pi.

>>1777236
>170V through a breadboard
thanks but no thanks

>>1777238
No magic blue smoke yet and haven been shocked.

>>1777238
Also its 178 volts...

>>1777236
>>1777243

>>1777194
>Avoid chinesium scopes
Damn, I just found pic related on banggood for 40€ shipped, which is pretty much how much I wanted to spend (and the size is a plus for me): I hoped that was just enough to play around and learn the basics as I go.
>USB scopes are kinda ok but still somewhat expensive.
I have no idea what those even look like.
>Go second-hand. Ask around, keep an eye on local auctions, etc. CRT scopes aren't outdated, especially if you nab a digital storage CRT scope
Thanks, I'll keep an eye out. Is there a rule of thumb to know if a scope is decent or not? How do I tell a CRT scope from a digital storage CRT scope?

>I have no idea what those even look like
I could have googled better before posting. I assume you're referring to things like pic related.

>>1777243
can you just use rectified burgermains?

>>1777248
>40€
That's definitely the kind of range you can get an analog CRT scope for. The digital ones will have extra buttons on them that say things like "hold", "save", etc. Like my Hitachi VC-6545 that I got from my university for free. It has a button to swap from sending the analog signal straight to the CRT to sending the analog signal into an ADC, sending that to RAM, and then sending it through a DAC to the CRT, allowing me to pause it and to see what happens before a trigger.

>>1777251
Yes that's right, they use your computer's screen and signal processing/analysing ability and this cuts down on requiring all those things to be installed in the scope itself. Never used one myself but I prefer having an oscilloscope that's electrically isolated from my computer. Dave EEVblog has a video on beginner oscilloscopes where he recommends going for a CRT scope over a USB scope.

>>1777227
Frequency counter

>>1777258
All right, after mining ebay I found pic related. Starting price: 55€ (+12€ shipping). Is that a good price for a scope like this? Sorry if I bother you, but I have no idea what are the rates these things go for.

>>1777260
How do you make those anyhow? Just use a dedicated IC? It's a neat thing to do with nixies as it's somewhat useful and related to the hobby, while not being as complicated as an RCL meter and not a fucking clock. Not sure about him but both my scope and function gen have very accurate frequency counters in them, negating any purpose for me to make such a thing. Anything else I can think of measuring and displaying with nixies would better be suited with a 2D graph (frequency response, transistor hFE, power factor, total harmonic distortion, etc.)
But I just thought of this:
propagation delay measurement
could also measure slew rate of a comparator I guess

>>1777275
That one looks pretty digital, and the specs are probably pretty good, but the UI looks a little unintuitive. Note that the digital side is only 20MS/s, meaning you can only go up to 60MHz in analog mode, and 10MHz in digital. My scope is closer to 500MS/s, or at least I can't see any time-domain quantisation error when running it at 50ns/div (probably 256 steps across the 10 divisions). Not that it will probably matter to you at all, the only thing I've used the higher frequencies for is checking that a 16MHz crystal actually works, and for hunting down noise that I probably don't need to worry about.

It isn't a bad price at all, but starting price isn't necessarily very indicative of how much it will go for in the end, it's quite possible that it will end up going for 4 times that. Looking at "buy-now" prices will give you a better idea of the market. Going for a digital CRT might be pointless to you, especially if you get one of those 24MHz $5 logic analysers. I'd lean more towards a straight-analog scope especially if you've barely/never used a scope before, since the analog knobs are more intuitive. What do you intend on using the scope for?

>>1777287
>What do you intend on using the scope for?
For now, fixing an old HiFi amplifier and learning about electronics in the process, but in my experience buying tools always turned out beneficial in the long run: just recently I spent 30€ on a good (enough for my needs) multimeter to check for continuity on a small soldering job I had to do on a small quadcopter, then a few weeks later the dishwasher broke down and thanks to that multimeter I fixed it, instead of wasting hundreds on a new one. I like tools.

>>1777290
that's a pretty decent tool for €55

>>1777287
it's just a counter that periodically copies its value to the display and resets itself. raspberry pi = too many layers. an MCU of the bluepill tier would be a better choice

>>1777290
Yeah you shouldn't need a digital storage oscilloscope, a basic 10-20MHz analog twin-channel scope should serve you well. The importance of two channels is that you'll set one to a clock or an input signal and set the other to an output signal, such that you can see the relationship between them. Look for distortion, lag, missed bits, etc. Analog CRT scopes also have the advantage over digital scopes that there's virtually no lag between what happens on the circuit and what happens on the screen.

>>1777297
>that's a pretty decent tool for €55
If he can get it for €55.
>a counter that periodically copies its value to the display and resets itself
How many clock cycles does it take to increment a value, and how many clock cycles does it take to reset it and print to the display? Sounds like a bunch of for any MCU, latency that needs to be compensated for. Compared to using an array of discrete synchronous counters or a dedicated IC which will run exactly at the clock frequency. Fine for anything slower than ~1/50th clock speed, but when you get up to 1/5th you'll really start to see the distortion.

>>1777227
probably a number anon

>>1777216
I understand that a triangle wave or any other waveform is a sum of sines but what I'm not getting is how a superheterodyne mixer can possibly produce the A+B frequency component with arbitrary waveforms.

Imagine you're feeding it with a triangle wave. After half the waveform has gone in it would look no different from a sawtooth wave. Now somehow from that the superheterodyne mixer should have already output a triangle wave with a frequency equal to that of input triangle plus the frequency of the local oscillator. I don't see how that can be possible without the circuit magically being able to see the future. To produce a waveform at a higher frequency it needs to know the shape of the waveform, but the higher frequency waveform needs to be output before the original one is complete.

>>1777327
You're not questioning how a superhet works, but how a mixer works, and how AM in general works.

Don't think about it in terms of adding frequencies, mixers are better thought about as multiplying two waveforms. Graph:
>y = (2+cos(2x))*cos(50x)
alongside:
>y = 2+cos(2x)
and:
>y = cos(50x)
The result should be intuitive. A mixer just multiplies the two signals. Note the DC component added to the audio wave for practical AM radio, this ensures that the envelope doesn't cross through 0V, and hence will be recoverable though a diode detector. In the frequency domain this is seen as a peak at f = 50 (the carrier wave) and two smaller peaks (the audio sidebands) at 48 and 52. SSB radio only transmits one of these sidebands, for some reason.

Mathematically speaking:
>cos(a)*cos(b) = 0.5*cos(a-b) + 0.5*cos(a+b)
Say a is much larger than b, the two frequencies (a-b) and (a+b) will be very close to one another, hence you'll get beat-frequencies between the two which is seen as the amplitude modulation. This equation is for modulation. When you want to demodulate it, say you're using a direct-mixing demodulator since it's simpler. The equation would hence be:
>output = cos(a) * (cos(a)*cos(b))
>output = cos(a) * (0.5*cos(a-b) + 0.5*cos(a+b))
>output = 0.5*cos(a)*cos(a-b) + 0.5*cos(a)*cos(a+b)
>output = 0.25*cos(a-(a-b)) + 0.25*cos(a+(a-b)) + 0.25*cos(a-(a+b)) + 0.25*cos(a+(a+b))
>output = 0.25*cos(b) + 0.25*cos(2a-b) + 0.25*cos(-b) + 0.25*cos(2a+b)
The (±b) frequencies will be far lower than the 2a±b frequencies, hence they can be easily isolated with a low-pass filter, and you're just left with the audio itself. cos(-b) = cos(b) after all.

>>1777327
If a signal is made of frequencies A and B and you multiply it by a higher frequency C. Then you'll have a signal that is just a copy of A and B but instead of being around in zero, it'll be around C. You really are thinking of frequency domain stuff in with time domain on your head. Also learn what convolution is.

>>1777376
>Also learn what convolution is.
Not him, but what relation does the convolution have? I tried to learn what the convolution was once, didn't end well.

>>1777384
Convolution is that whole operation.
Because of math, multiplication in the time domain is convolution in the frequency domain and vice versa

>>1777394
So if you are modulating a AM signal in the time domain
>sin(carrier) * signal
In the frequency domain it'll be the convolution of those things (but transformed. So sin(carrier) is a impulse)

>>1777394
>>1777395
Ah that's interesting to know. Using my intuition of the fourier transform is probably an easier way to get my head around the convolution than going at it directly. So is there a discrete time version of the convolution like there is for the fourier transform? If I wanted to do convolutions via DSP, for example.

>>1777397
Yes. It's the same-ish thing. It is usually in the exercises of DSP and Linear system books because you can do convolutions graphically very easily.

>>1777305
the timer-counter works at up to 36MHz iirc. you can use the event system to interconnect timers. for example, consider two 16-bit timer units where the first is your timebase clocked by the system clock, and the second is a counter clocked by the squared and leveled input signal. the event system can patch the output of a condition sensor on timer 1 to, among other places, timer 2's capture and clear inputs and a core interrupt, with a fixed latency (which is usually preferable to a low but spiky latency), and could patch the overflow output on timer 2 into timer 3's counter clock input to carry
since you're not dealing with floats and burning next to no CPU time on the measurement, you can update any display or UART output pretty quickly. if your meter has at least 5 digits, this is unlikely to be a great problem
if you're feeling especially enterprising, you could use DMA from some variable in your heap to the SPI port so you can stack up drivers on it

>>1777330
>SSB radio only transmits one of these sidebands, for some reason.
makes room for another channel. radio bandwidth with a very long reach is a scarce resource. when you get up into VHF and beyond you start to encounter more FM usage because there's room for it and it's easier to AFC

>>1777408
>makes room for another channel
I guess that makes sense, but from what I've seen SSB is harder to demodulate. Since you don't have a carrier wave to latch onto, both sides need to have well calibrated oscillators and know what frequency to look at. Though a somewhat clever scanner should be able to figure out the original carrier frequency of an SSB transmission.

>>1777305
Would you advise against buying a broken one and fixing it?

inb4 you need an osci to fix an oscilloscope

I did this, but it doesn't work.

As you can see, this "Sequences_vectors" python embedded block that I created, has a list of bit sequences and it uses a for cycle to select each sequence and then return them to the next block (Repeat block) to go on in the flowgraph.
I did this to simulate a "Vector Source" block(that takes a _single_ list of bits), but that takes more than a single sequence of bits (in my case, as you can see in the code, there are three sequences).

I'd like to transmit each of these three sequences of bits. How can I do?

If I start the graph, it reports errors, I think because I can't return list but only single int.

In fact, even if it started without errors, I guess that the python embedded block wouldn't behave as expected. I think it would return one list and stops, since the "return" only works one time and not more than one (so the other two sequences can't be "returned"). So I think my code wouldn't work.

>>1777458
what ´program makes those flowcharts?

Can i just charge some 18650 batteries using a power bank or i'll simply set my house on fire? In particular i have a lipo powerbank and i want to charge in parallel 2 18650x2. All the batteries are discharged

>>1777520
USB power banks output 5V, since that's the USB specification, unless they feature Quick Charge, which means they can output up to 9V. A 18650 battery can go up to 4.2V when fully charged: giving them 5-9V will surely damage them, but they won't necessarily burn your house down.

>>1777471
GNURadio

Are ICs like counters, timers and gates still used these days when you can replace them all with just a microcontroller?

>>1777556
don't know about others but I find coding and prototyping with mcus to be extremely frustrating so I try to avoid mcus

Has anyone tried this? https://gerbolyze.jaseg.net/

I can't seem to get it working. Either the image doesn't upload at all and I get an internal error, or it uploads but doesn't display the whole PCB, just a piece of it.

I am using "Vector Source" block. As you know, it's used with a list of ints. Of course, if I try to use more than a list, so if I use a list of lists, it gives me errors.

I also tried writing "[x for x in sequence]" in "Vector Source", but it doesn't work neither (same error of before).

Obviously if I write "sequence[0]" to select only the first sequence, it works, with no errors.


So my question is, how can I use more lists with "Vector Source" block?

>>1777455
>Would you advise against buying a broken one and fixing it?
Definitely. Fixing one with just a $5 logic analyser is probably possible, but even with a scope it's a very complicated circuit unless it's from the 50s or whenever. Got high-voltages to watch out for also. Plus if the CRT itself is busted you're not easily going to be able to get yourself a replacement. I'd only do so if you already have a scope and the broken one is going for particularly cheap. Or the fault is particularly minor and can just be fixed with a drop-in replacement. Heck, my scope needs a replacement coin cell for its memory.

>>1777624
Understood. Thanks again.

>>1777034
Naw man, class d sucks imo. Sure it has its advantages, being small and power efficient, but it processes signals to a higher degree using A/D converters/PWM. Fine for power control or motors where it's fucking fantastic but the goal of a great amplifier is to reproduce an input signal accurately with minimal changes. This means no d/a or pwm conversion, good snr and low distortion (imd/thd/timd). Tubes, transistors, opamps...whatever, they control an analog signal and are purer, when built for the purpose. The LM3876 for example is a ridiculously good part for the price and soundspretty good to boot.
>>1777031
Check the power supply. Electrolytic capacitors may look ok but test bad. Usually people replace them as a matter of course in older gear as they are prone to drying out as they age and can destroy rectifiers or even the transformer, norwithstanding frying other shit first.

>>1777855
>the goal of a great amplifier is to reproduce an input signal accurately with minimal changes
You act as if there's no analog feedback loop between a class-D's input and output

What does a cap labeled with an asterisk (*) mean in a schematic?

>>1777414
in practice there is a vestigial carrier that could be located and locked onto in the frequency domain. but it is also true in practice that 10ppm on each side can be rough on voice intelligibility even as low as the 30m band

>>1777556
>counters, timers
a 555 circuit in quantity can cost just a few pennies and be producible on saved-from-the-scrapyard equipment. the infamous 3-cent micros do seem to have it for anything much more complex than a back-up alarm on a piece of heavy equipment
>gates
sure, especially now that they come in single-gate packages. but more than a dozen gates or so tend to be absorbed into a CPLD or FPGA. even some newer low-mid-range 8-bit micros have a few instances of 3-bit LUTs with a limited fabric to connect them and the pins/fw

>>1777595
https://docs.python.org/3/library/itertools.html#itertools.chain.from_iterable
or try
[val for sub in sequence for val in sub]

>>1777888
failget
it means whatever the guy who drew it wanted it to mean. look for notes on the drawing or in associated text

I ordered some 5557 ATX terminals to do some custom work for my PC's wiring. How do I order the right crimping tool (from aliexpress since budget)?

>>1777862
Furthermore, even if there wasn't an analog feedback loop in a class-D, you could simply add your own with an op-amp on the output, possibly with a small ceramic delay circuit in there to compensate for the small amount of delay that comes from running it through a low-pass filter.

>>1777891
10ppm shouldn't be an issue with modern crystals, right?.

>>1777915
typical tolerance is 50ppm, 10ppm is a slightly extra cost option but not hard at all to find. 0.5ppm TCXOs are easy to find for under half a buck in China, some accept a trim voltage to adjust over a few ppm or so, which you could also discipline with a GPS receiver if you were into that sort of thing and had that kind of accurate timebase available

>>1777855
Yeah, so if something uses AD and has low distortion, good SNR it still bad because it is a class D? What the fuck I though this place was my safespace from audiophiles.
>"analog signals are purer"
I don't even know what that means but ok

>>1777943
DDS frequency generators still need a crystal for timing, right? I've no clue how DDS works so I always just think about PLLs with programmable clock dividers, but I doubt anyone actually does that these days. But maybe PLLs' VCOs make sine waves and DDSs don't so PLLs are still preferable when making a signal for heterodyning?

Is this a decent way of making a VCO or I'm just stupid?

>>1777998

>>1777999
use a jfet

>>1777855
>Electrolytic capacitors may look ok but test bad.
I got a cheap multi tester that also measure ESR, maybe I won't even need a scope for this job.

>>1778045
One thing you can do is measure the AC voltage on your DC voltage rails, that should tell you how much ripple you've got on your rails. But ripple on the rails shouldn't matter much even if it's half a volt or so, all stages of an amplifier should be designed to be rather resistant to power supply ripple. If a power supply cap is failed you'll see a ripple much larger than 0.5V, and if the rectifier is gone you might notice uneven voltage rails or even the absence of one rail at all; a ground hard up against a voltage rail could definitely produce distortion. If, however, there are electrolytic caps in audio stages as opposed to the power supply, the failure of them would also cause distortion. Assuming your rails are acting fine, probe the analog voltages through the circuit while it's running (a schematic will help) and find where the distortion begins. A scope isn't necessary here, you could just use a little piezo disc and listen for the difference, or else use an op-amp buffer that feeds a small speaker. Floating, of course. Using your computer's sound card (either directly through a wired connection or just by microphone) would be a more quantitative way to compare waveforms.

Definitely search for a service manual or schematic. If it's from the 90's it should have one that's somewhat available, can't say the same for most 2010s stuff.

Need advice on babys first smd board.
My plan is to do reflow soldering on a hotplate.
Should I be using thermal reliefs or not?
My thought process is that I heat up the entire board anyways so I won't need them but I figured I'd ask here.

http://datasheet.octopart.com/CR-77-EDSYN-datasheet-62136948.pdf

Fuck you edsyn this is not a datasheet it's a fucking commercial leaflet
How am I supposed to find out this paste's temperature curves

>>1778121
It's better to have them otherwise the copper plane might suck up all the heat and your part will never reach the final temperature it's supposed to

>>1778121
>Should I be using thermal reliefs

a person with a normally functioning brain with do some with, and some without, and learn from the results, instead of asking hypothetical questions.

>>1778196
How about we take the rest of my post into consideration?

Is getting a job in embedded background demanding? I am very interested in the topic and i want to specialize on it. The problem being i'm studying E.E (4th out of 8 semesters also not from the U.S, lineal program) wich is heavily power line-electric machines based with only a few classes about electronics and i doubt they cover embedded systems at all. My original plan was to get the degree and then apply for a Masters based on it, while trying to rack up on certs on the side and doing projects on my own, given that right now i basically know nothing about the whole thing. Alternatively i could apply for a job on the field with just the certs i get and my E.E degree, however that brings me back to the first question. Now i'm not sure if my E.E degree could be a mayor setback in applying for a Masters on the field, but the possibility of changing to an Electronics based program is also present, with the downside of having to start from, say, 3th semester, as not all classes are transferable.

>>1778189
you fucking niggers destroyed my diode

>>1778189
>>1778318

>>1777977
the Si5351A-A programmable clock generator is still very popular in homebrew radio applications even with its square-wave outputs. for <$1 at Digi-Key it offers three programmable frequency outputs that can be set in quadrature at 3.2MHz and up. the B-A version even offers a VCO input. worst-case output jitter is a respectable 155ps pk-pk. they're just a bit harder to program than the AD98xx DDS chips and you need an LC filter (or several, depending on output frequency range) if you want a sine output
you can usually use a logic-level clock in place of a crystal by introducing a few passives to match the clock input's expectations

>>1777999
pants are not entirely on head, but make that discharge tranny an npn (or n-channel MOSFET). there are many places in this circuit where the effects of process/voltage/temperature and temperature gradients could manifest, e.g. changes in the thresholds of the Schmitt trigger as the buffer drives the base/gate of the discharge tranny and heats up the Schmitt IC; input transistor Vbe voltage shift and mismatch
component selection and construction/packaging is paramount. you might even think about building an oven around it

>>1778121
I would leave the thermal reliefs in place, in case I need to rework it later

>>1778318
I agree, it is kind of a shitty leaflet
Sn96,5 Ag Cu0,5 is most likely good old SAC305, I'd take the standard IPC/JEDEC J-STD-020 lead-free reflow profile as a guide, use a thermocouple on the board to avoid overcooking anything

>>1777896
they look like weatherpack terminals to me

what to look for in a heat gun for home electronics repair?

>>1778456
>Si5351A
Oh that is neat looking. I see that it uses internal PLLs and whatever the hell a multisynth is.
>square-wave outputs
could you just feed the output into another PLL to get a sine? Might be cheaper than the filtering to squash that 3rd harmonic

can we make a new thread so I can post a question that has a chance to get answered?

>>1778499
it will probably be on page 10 for another day or so

>>1778500
thanks I'll post it Monday.

>>1778458
they're not

>>1777896
non-insulated open type crimps are very roughly interchangeable. SN-28B is adequate and does both wire and insulation crimps at once, but may need dressing with pliers after crimping for some connector series. Engineer PA-20 will do a much neater and more reliable job but costs 2-3x https://www.youtube.com/watch?v=3Cw9YFLDZeo

>>1778498
yeah, MultiSynth is their fractional PLL. AN619 is a much more detailed read into the setup. their Windows desktop clock planner is kinda shite. but the chip is bretty gud
otoh inductors with respectable Q are fairly cheap these days. a 5th-order LC LPF won't break the bank. you could build a few 5th-order filter banks and insert the best choice with a mux as desired. the Tayloe detector is perfectly happy with a digital clock so no big deal

>>1778468
replacement heating elements
corona-chan gona be gud

>>1777896
>>1778521
Thanks, I ordered an SN28B.

>>1778521
>SN-28B
you could totally crimp that in a weather pack crimper

>>1778521
>a 5th-order LC LPF won't break the bank
You mean a discrete filter, or one of those enclosed filters?
>Tayloe detector
Not just yet, I'm sticking to analog until I accrue the RF experience to just homebrew a full-on SDR.

On this note, after graduating from babby's first CD4046 (1.6MHz max), is there a common PLL setup to go for? From what I've seen they don't come all integrated like the 4046, so I'd need a seperate VCO and phase comparator and filters. Do people actually build their own VCOs, or do they get VCO ICs? A DDS or sinusoidal equivalent to that Si5351A would be fine too, tentatively.

I imagine the SA612's 500MHz will last me a bit longer, though perhaps there are other features I'm missing out on.