/diy/ - Do It Yourself

Fuck the guberment

>>1841874
Whoopsie forgot image

I'm trying to migrate my arduino project to a more dedicated platform to learn more about electronics. My project takes input from a force resistive sensor, outputs a voltage, displays battery on a screen. Would MSP430 be a good chip choice? Trying to make it as minimal as possible.

Bought to you by Phoenix Contact 1841869, a big boi connector

>>1841885
Seconding this. I'd really like to see a list of specifically hobbyist-friendly MCUs for different purposes. There are so many MCU families I've never even heard of, so I just tend to stick to the buzzwordy ones that I'm familiar with, like the Atmel ATetceteras, the ESPs, and the STM32. And the Padauks, I guess. I've barely even heard of the MSP430.

what is the minimum safe spacing between bare pcb traces for 230VAC? Guide i found says 1.2mm but that seems like a lot.

>>1841919
STM32 is easily the most common one and will probably fit 100% of your needs. And you can buy a cheap debugger on Ebay. And you have the most freedom in choosing your development environment.

>>1841885
those seem a little overkill
why not use an ATtiny

Yesterday we talked about sound card oscilloscopes. I just found the following link, which improves the concept even further by subsampling the signal. Obviously this only works for repetitive signals, but it's a nice concept.
https://www.analog.com/en/analog-dialogue/articles/turning-pc-sound-card-into-sampling-oscilloscope.html

I bought an ALTERA USB Blaster clone from aliexpress
it has mcu but marking has removed.
what mcu is this?
i guess this is not PIC or STM32...
usb/power pin position is not same in known USB MCUs (avr,pic,stm32,cypress ...).

>>1842083
My USB Blaster clone uses a PIC18F14K50

>>1842053
They are overkill, but most people who make one-off hobby stuff don't care to be efficient. Bulk buy one microcontroller and use it for everything.
I've done this myself for simple things that I slapped together for personal use.

>>1841874
socket looks shocked

>>1841869

what is that on the chip in pic related under cpu. never seen something like that.

>>1842209
looks like someone botched an smd resistor on it.

>>1842183
>if only you knew how bad things really are

>>1842183
>only the dead can know peace from this evil

>>1842321
>and it shall come to pass
>that the slaves will defend
>their slavery

I want to implement DSP for an RTL-SDR so I can filter out the AM and FM bands entirely (passive components ain't doing it)
now I'll be honest, I don't know much about DSP other than what music production has taught me.
the Sampling theorem says I would need to be able to sample at twice the highest frequency the RTL-SDR is capable of receiving in order to be able to perfectly handle the incoming signals.
If I understand that right, since the RTL-SDR has a max frequency of 1.75 GHz, I would need to be able to sample at at least 3 GHz.

I don't know of any microcontroller that goes anywhere near that high in clock speed, let alone actual sampling ability.
Would I have to do this on an FPGA, then?

>>1842367
on a related note, if I wanted to instead implement this with active op-amp filters, something tells me I can't just get something for nothing. There needs to be a gotcha somewhere that would limit it.

I'm thinking this would be the highest speed at which the op-amp could switch at? If it can't switch as fast as the highest input frequency, it'll distort them, right?

would it be possible to perfectly filter out an offending signal in a radio by, say,
>getting a time-identical but 180-degrees-out-of-phase duplicate of the incoming radio signal
>bandpassing the frequencies I don't want
>combining this signal with the unaltered signal
so that the bandpassed unwanted signals destructively interfere with the same signals in the final output?

>>1842372

the way cable tv blocks out certain frequencies is to put a tuned circuit across the line that shorts out the band they wanna block.

>>1842407
is there a name for this kind of filter?

>>1842353
Fuck off racist

>>1842421
idk that comment seemed to poignant to be about niggers

>>1842372
Why would you want to do that instead of bandpassing the main signal?

What you're describing sounds sort of like a superheterodyn circuit, but not exactly.

>>1842468
If I were doing the just the main signal, it would just be a notch/band reject filter, right?
I figured the bandpass-and-destructively-cancel filter method might be a more complete way of filtering out the unwanted signals. By all means, I'll just use a band-reject filter if it's simpler to implement

but that just begs the question in >>1842369, in that I have no idea what affect the slew rate and bandwidth of an op-amp would have on the higher frequencies of the antenna signal

where do you put a "portfolio" of your electronics projects? Compsci brainlets have github, is there an EE brainchad equivalent?

>>1842498
buy a domain, set up AWS S3, and host them on your own website

>>1842372
You can put a bandpass filter in the negative feedback loop of a high open-loop-gain buffer, like how the opposite of how a wah-wah pedal works.

>>1842367
Look up how SDRs work. They use mixers to receive signals at far higher frequencies than their ADC's sampling rate. They have a certain bandwidth of frequencies that they can detect with each setting of the mixers' oscillator, and this bandwidth is just equal to the sampling rate of the ADC. Plus or minus a factor of two, idk. So the DSP only needs to be as fast as the ADC, so an STM32 of some sort should do the job.
If you're designing your own SDR, then A: you need to research more, and B: consider using a higher ADC resolution. Something like 24 bits. This way even if you have a hugely loud AM radio broadcast, you'll be able to still detect quieter broadcasts so long as the AM signal isn't hitting the ADC's rails. There was a video processing 24bit ~16MHz ADC that I had my eye on for this purpose, because most common SDRs only run 8 bits and much less than 16MHz.
If you're not making your own SDR, consider learning how to use GNUradio. You can put digital filters of any sort you want in the path of your signal, all on your computer. Programs like GQRX are basically just GUIs for GNUradio, IIRC.
The whole reason common SDRs don't go down to AM ranges is because they'd get swamped out by the strong signals. I suspect these existing SDRs have filters for this purpose. FMs are comparatively better, I think.

>op-amps
Op-amps are too slow to pass high-frequency signals, even if they're fast enough to filter out AM. You could technically make active filters with RF transistors or RF amplifiers, but it's more common to just cascade a bunch of RC filters (impedance matched ∏ or T filters), that's what I see being used for low-passes and high-passes at the very least.
Maybe a passive twin-T filter is a thing?

>go to junk shop (the same one I got a 10lb bridge rectifier from)
>notice half-full reels of SMD components on a shelf
>pull one out
>says "4R7" on the component
>says "WI322522-4R7KF" on the side of the reel
>check the next reel
>same
>check all 8 reels
>all 4R7
>think it's really disappointing that they're all just 4.7Ω, which is basically useless
>the packages aren't small so it's possible that they're ok for 0.1W or higher
>think maybe I could use them for current limiting LEDs or something
>check packet that two of the reels were in
>says 4.7µH inductor
>INDUCTOR
>grab reel with most components in it
>take up to checkout
>guy says he thought they were 16mm film
>$5
>deal
>get home
>write this
Oh boy if these are actually inductors that will be great, time to figure out how to test them. I suspect the good old LC band-pass filter should be the way to go, once I solder onto them.

>>1842526
Ok, it's 2.5mm wide, 3mm long, 2mm high, and sticks to a magnet. Its leads are also not like a standard resistor, so I'm feeling pretty lucky.

>>1842530
I took the inductor out of the breadboard and the ringing frequency didn't change from 25MHz. I'm going to have to adopt a different strategy.

>>1842540
Ok I broke the leg off and I can't be arsed doing any more of that before I eat, but I was getting inductive spikes so I think it's working. The ESR is really low, btw.

>>1842498
i use chadbucket because unlike the jewhub it offers private repos for free

>>1842083
it was removed for a reason
stop stealing other peoples hard work tyrone

>>1842498
can linkedin work for this?

>>1842418
>is there a name for this kind of filter?

Trap Filter
https://www.soontai.com/TRF.html

>>1842573
surely you don't want to do EE for a living? do you have any idea how tedious it is?

>>1842577
Better than being a lab jockey or doing customer service. Being able to do electronic engineering or even electronic assembly would be a dream for me and my BSc.

>>1842582
>electronic assembly
you mean being a 6 axis soldering arm? no thanks
or designing 100th iteration of some buck converter for some chinkshit? might as well neck yourself
start your own company and build the new Juicero but not shit, EE is only fun when working on fun and interesting projects

>>1842585
>start your own company and build the new Juicero but not shit
That's the long-term plan, yes.

>>1842498
You don't. What you do is to put best of your projects in pic related and go to your job interview. You then throw that bad boy on boss' desk like a chad and open the case with loud noise of hinges hitting the top of the case. You then proceed to empty contents of your case and start describing your projects. Boss then sees you're not some web project hosting service virgin but a chad who brings physical boards to the interview. Also don't forget to wear your lab coat.

>>1842586
just don't turn into one of those cringey pajeets or slavs who make a kickstarter for their project with a stop motion video and that royalsy free upbeat music that is in every single fucking kickstarter video

>>1842589
just be careful with how you present the assembly, when i last did this the boss looked at the soldered joints and he was instantly able to tell i didn't use TS100 for assembling that board and had the security escort me out of the building immediately

Never really learned analog electronics in school, how would I go about generating a trinangle wave where the aplitude, frequency, and offset can be controlled via changing passive component values?
Can I do this exclusively with some transistors and caps/resistors?

>>1841869
>ordered cheap PAM8403 as headphone amp
>hook it up, notice humm/ static noise
>Google the issue, find something about caps/ low-pass filters
>add 1000Mf cap across 5V input, gets rid of some of it, low-pass filter distorts audio
anything else i could try or keywords i can look up? I'm using it with a 5V powerbrick and audio coming from my pc's 3.5mm out. I'm a beginner when it comes to electronics

>>1842603
>5V powerbrick
There is high possibility that power brick has a switching regulator inside which will produce a lot of noise. You could try adding LC filters and chokes to your power input but the best thing would be to get a linear power supply.

>>1842590
I've seen enough kickstarter crap to know what not to do.
The aim is a promo video of a finished prototype (no "give me your money" with nothing concrete to go off) with high apparent production value, with editing synchronised to dank beats.

>>1842596
IIRC there was a dual op-amp oscillator circuit that outputs a constant-amplitude triangle wave that should be a good start. Amplitude and frequency should be alterable independently via potentiometers. Not sure what you mean by offset, phase? Relative to what?
>transistors
For analog electronics, especially low-frequency stuff like audio, op-amps reign king in terms of convenience.

Then there's the digital option, an MCU + DAC, or a DDS I guess.

>>1842603
What's the impedance of your headphones? Is that compatible with the output impedance/driving current of the amp IC?
You may also want to add a sizeable cap between the virtual ground (the upside-down T, or the audio output ground, or both) and each rail.
Assuming those L/R negative pins are connected together that is. If they're not, then it would be incompatible with a TRS common-ground headphone jack.

>>1842611
Plan it to have it powered by the pc's psu, its noisy on there too so i guess I'll have to come up with a linear power supply that i can throw into my pc case? I tried ferrit beads as choke, didn't do anything noticeable.
>>1842620
the headphones are 32ohm and yes the negative L/R on the output side are connected. So you suggest that i add 4 more caps? What is "sizeable" ? I have a bunch of caps lying around at home

>>1842620
I think I know what you're talking about. I looked up some IC that spits out a triangle wave on digikey, but it was obsolete.
For offset, I meant DC offset. As in able to pull the waveform completely above the zero cross. Guess I could just feed an extra source voltage for the output to ride on.
I have heard OPAMPs are easier to work with than say BJTs. But how would I go about doing something like that with OPAMPs exlusively? Maybe I need to do some reading on how OPAMPs can be used for such an application?
Kind of want to stay away from digital solutions and keep this purely analog.

>>1841869
Hello /diy/, I got inspired watching this crazy bastard (https://www.youtube.com/watch?v=8_T-YIvWNmQ)) and I got inspired to want to make a flashlight. No, not just a regular flashlight, but a really freaking bright flashlight, like a sun in your pocket bright. Any tuts or guides on putting together a 12V LED flashlight practically?

>>1842642
>Hello /diy/, I got inspired watching this crazy bastard (https://www.youtube.com/watch?v=8_T-YIvWNmQ)) [Embed] and I got inspired to want to make a flashlight. No, not just a regular flashlight, but a really freaking bright flashlight, like a sun in your pocket bright. Any tuts or guides on putting together a 12V LED flashlight practically?

>>1842645
any retard can build a flashlight all it needs is a battery, led, switch and a heatsink. And if the battery's internal resistance is not high enough then also resistor so the led wont sudoku itself

>>1842642
>tuts or guides
battery contact resistance is critical and hard to estimate, so be sure that whatever battery holder you come up with has good contacts. you'll be drawing at least 10 amps from a 3-4s pack and the milliohms really add up in that context to drop your low pack voltage even lower. if that isn't clear then start reading before you waste your money.

>>1842648
Should I buy the LEDs from YUJILEDS or Cree?

>>1842421
Nigger
>Also
Does anyone have a list of good off the shelf modules?
I know of dc-dc converters, power supplies, sensors for arduinos and battery chargers for li-ion with protection. How about some good pwm driving modules? In line polyfuses? What else exists and would simplify building projects?

>>1842642
>>1842651
Make this or you're a faggot
https://www.youtube.com/watch?v=LT5_-A0m8_U

i had issues with noise in my micro signal wires so i got a shielded cable that has two wires and those wires are wrapped in metalic foil
So wire 1 is +, wire 2 is ground, and what am i supposed to do with the shield foil? do i need to connect it to something too, or does it provide noise shielding even without being connected to anything?

I made a phone jammer and it works great. Low wattage, narrow frequency span, highly directional antenna. I use it to annoy loud people on their phones in the subway...

Not in the US luckily....

>>1842661
epic trole lol

>>1842634
>But how would I go about doing something like that with OPAMPs exlusively?
Make square wave generator and integrate the output with opamp integrator. Then you just use another opamp to add offset/adjust gain and amplitude like pic related.
>Kind of want to stay away from digital solutions and keep this purely analog.
Why? While analog solutions are fun, it's hard to achieve higher frequencies, using digital signal generator IC is easier and allows you to reach >100 MHz easily. For audio frequencies opamps would be ok.

>>1842589
I've done this in the past and it can be helpful but being a genuine, effective communicator is more important; people have been designing circuits for over 100 years now so it's nothing new. When they ask questions about your projects and see the passion it puts you in a better position regardless though, people love the enthusiasm.

>>1842634
Texas Instruments has an excellent book, SBOA092B. Google that number.

>>1842664
No, I need it for what's left of my sanity.
I need people to shut the fuck up.

I have a 14bit compass heading but the data is noisy and so I want to find an "average" of 5 measurements. However a regular average will fuck up if the measurements go across the 0 line. What do I use here and how do I implement it in C?

>>1842712

1. this is not an electronic question
2. normalize your data: change 15900, 0, 1 to 15900, 16384, 16285 or whatever
3. if you need us to write the C code then you should forget about working with code and paint warhammer models instead

>>1842712
Map your input data to pic related, then average x and y separately.
>but how do I map
1 bit of data gets an angle of 2pi/2^14 radians
so your input angle is
theta = <compass value>*2pi/2^14
which gives you
x = cos(theta)
y = sin(theta)
You then average x and y. Since there are no discontinuities (values from -1 to 1) you won't have same problem you have now. Then you get the angle of averaged values
theta_avg = arctan(y/x)
where you need to be careful to add 180° if x < 0 and 360° if y < 0. After you get your angle back you reverse first equation to get averaged compass value
<compass value> = theta_avg*2^14/(2pi)
but since you want your heading in degrees you'll probably omit this last step.

>>1842712
easy solution: store the last 5 readings in an array, sort them highest - lowest, and choose the median value (the middle one)

>>1842209
>>1842214
Its not botched
When this thing came out
"touching up" pcbs was pretty Common

>>1842712
Make an array of lowest to highest
If (array[0] < 500 && array[4] > 14000) {
While (array[i] < 14000) {
Array[i] +16000
i++
}
For (int x=0; x<5; x++) {
Sum += array[x]
}
Avg = sum /5
If (avg >16000) {
Avg - 16000
}
}
Assuming max is 16000. The other numbers are bullshit you can change. This code is messy but should work.

>>1841291

Finally! I have finished my beautiful work of routing art.

And now i have to actually manufacture this fucker.
This will be fun, since this is the first 2 sided pcb i will be making.

>>1842827
it's kind of funny how the regulation dictates at least 1.2 mm air gap between mains traces and contacts, so i did it for traces, but for contacts it's fucking impossible because pads for components like fets are only 0.5mm apart, so drm was giving me shit for that (red arrows) but eh fuck it, i used 0.5mm plenty of times before and never had any arcing

>>1842665
Damn, I feel dumb now because I actually learned about opamp configs in school. I need to brush up on this apparently.
I will play around with your circuit and will come back if I need more help.
>Why no digital?
I want to make a lightweight and tiny circuit, as well I don't need to be able to change the frequency on the fly. It will be fixed at a set frequency+dc offset, probably in the order of less than 1Hz so high frequency is irrelevant. Figured an analog solution would give me the cleanest slopes as well.

Do you factor in a batteries internal resistance into Ohm's law? It's not relevant to my project, but I'm at that stage where I have 50 tabs open, I'm more confused than when I started, but I've started to kinda form where I need to be starting for my project.

>>1842728
If hes using a microcontroller, needlessly calculating sin and cos is a big no no. Huge waste of memory and speed. Your solution is a waste of computing time anyway.

>>1842834
You factor in as many parameters as you see fit. If you feel that battery internal resistance will play a significant role in your circuits performance, include it in your calculations.

>>1842837
Sorry, guess I worded that poorly, I kinda meant it as a theoretical, not how you personally practice. Like, if I want my circuit to be 1Ω, and my battery is 0.01Ω, I use a .99Ω resistor, right?

>>1842836
>Your solution is a waste of computing time anyway.
It's too soon to start optimizing when you don't even know for what or where is he using that compass. I will give you better one to satisfy your optimization autism while laughing at your 8 MHz ATmega 328p:

int16_t read_compass_avg() {
// get initial reading from sensor
static int16_t p = read_compass();

// get additional values for averaging
int16_t delta = 0;

// read 6 values to get average delta error
for (int i = 0; i < 6; i++) {
delta += read_compass() - p;
}

// if white noise (equal + and - delta distribution) delta == 0, p is true value
// if p is wrong delta will correct it and come closer to average value
p += delta;

return p;
}

okay nerds rate my dumb fucking idea for a project
after watching a bunch of videos on DIY coilguns i came to a conclusion that electromagnets are an efficient method to obtain high velocity kinetic energy
so here's my proposal
airgun coilgun hybrid
or rather, electrically charged spring piston airgun
having some sort of an electromagnet/linear stepper motor compressing the spring
or maybe even completely replacing it
problem is, it will have to provide force of 100kg minimum or maybe even more than that and i have no idea what kind of coil and power supply size would that result in

>>1842859
That's already a thing

>>1842861
that's airsoft
im talking full power pneumatics, 7J and up

>>1842840
Again, it depends on how theoretically specific you want to get.
You can consider the wire resistance based on it's physical properties, the ambient room temperature the circuit is operating at, the temperature rise on the components+wires overtime as current flows in your circuit. All things can affect resistance, this is just what I can come up with. You can get pages and pages long of all the functions needed to control for all aspects of a circuit.
But yes, if you need theoretical resistance to be within 10mΩ accurate and your battery has 10mΩ of internal resistance, then consider factoring the battery resistance.

>>1842634
>some IC that spits out a triangle wave
Not what I was talking about. Pic related.

>>1842665
I was going to suggest that, but there's a slightly better circuit than that. IIRC, the circuit you posted has an amplitude that's dependant on frequency (because of the integrator), but there's a design that uses the integrator as a part of the oscillator that avoids this issue. Pic related.

Just add an inverting amplifier for variable gain and variable DC shift on the end.

Half a brain cell here, im trying to fashion a dpad where an anolog stick once was. I need to have the four inputs correspond to the four extremes on the x and y and have no input return immediately the center. If I only wire the four extremes to the dpad contacts it takes a full 2 seconds for the signal to return to neutral. When I tried to wire it in parallel I found I would need a switch that's closed when not pressed, but I can't do that with contact points to keep the dpad intact. So I thought I could use an hex inverter.
I'm new to electrical stuff, would my circuit drawn below theoretically work?
If it does my next problem is where on the board to draw the voltage to the ic?
I know there are modders that can add led lights to various controllers so it should be possible but I have no idea how.

>>1842827
nice acid traps there buddy

>>1842596
Look up how the korg ms-20 and other synthesizers do it. Basically:
sawtooth generator -> ac coupling -> through two diodes into an op amp. Use a potentiometer for amplitude.
>offset
dc offset?

>>1842888
So you need it to be 1.9V when neither button is pressed, one button pulls that to ground, and the other pulls it to 3.8V? Why not just use a 10k/10k voltage divider across a 3.8V source, with one button across each resistor?

>>1842828
There's always potting or conformal coating.

You're right, that's exactly what I need. I'm still learning electrical analysis and my brain kept getting stuck on solving it in parallel. My sideways brain cell thanks you.

I suck at calculus. I can't even do Laplace and Fourier. What do?
How do I calculus in couple weeks?

Can I do electronics without a background in advanced mathematics?

>>1842929
No. This is why I ask how do I calculus.

>>1842915
>How do I calculus?
Read a book lol. Half of my ECE degree was just calc courses

>>1842893
jokes on you i will be milling it

>>1842931
Any recommendation? I can't find a book for autists with down syndrome.

>>1842938
just invent some device capable of doing it for you... you can call it something like ... calc-ulator

>>1842929
>electronics without a background in advanced mathematics?

if you wanna design, build, and repair, you just need basic algebra and the most basic notion of what integrating and differentiating mean. if you want diplomas and certs, it's all about jumping through mathematical hoops you will likely never use IRL.

>>1842938
>Any recommendation?
Not really. I learned everything off the textbooks they forced us to buy in school and I have since sold them to recoop some of the lost money for them.
Any standard issue calculus textbook meant for university classrooms will work imho. You'll also need to look into finding a book on differential calculus too. Again, standard issue class text will work. Even better if you can get your hands on answer keys. Torrent is essentially the only way to get your hands on these things because the publishing companies are in bed with universities as to not release answer keys to anyone that isn't a lazy professor who can't make his own problems for you to solve.

>>1842954
Also this
Having the understanding for the complex calc is nice, but CAD tools do all the heavy lifting for you nowadays.

Noob here. I bought this:
https://www.amazon.com/gp/product/B078XV3RK2/
but I don't have a 9V, 1A DC power cable. The only one I could find was 12V, .75A, AC. Can I still use it, or will it not work?
I also have a Raspberry Pi, but I'd rather not use it for this.

>>1843018
I also found this.

>>1843018
Do you not see the 6.5 to 9v rating? You have options for PSUs. 12v could fry it

>>1843025
So then what about
>>1843023

The current is pretty low, but the voltage is correct. I'm just trying to avoid having to pay another $6 for a power supply.

>>1843018
those two black chips are most likely 3.3V and 5V regulators. All the excess voltage/power gets dumped off as heat.
If you're willing to take the risk, plug in a 12V and test them with your finger to see if they're not getting too hot (especially while your device is running)
Worst case, you'll burn em out and have to learn how to set up a 7805 (literally just a 10uF, 100nF, and the chip itself is fine) power supply.

>drench the fuck out of solder wick with flux
>apply the tip good and long to solder joint
>only a little solder wicks up
>the rest just gets real tacky/gummy and refuses to suck up
>wick then adheres with the force of 2000 suns to the solder joint
every fucking time
what am I doing wrong

>>1843095
freewire anything you expect to ever replace
rework is a bitch, my condolences

What's the simplest, most dependable way to put a 100ms delay on the output of a piezo sparker switch?

>>1843018
5V and 3.3V regulators are probably perfectly fine off 12V. Check the datasheets of the regulators themselves to be sure. But doing so will limit the amount of current you can draw through the rails to significantly less than with 6.5-9V (1/5-1/2 respectively).
The 9V 80mA will be fine as-is unless you want to turn on a bunch of LEDs, you may want to put a large cap across the output though.

Also you should have bought one with a micro/mini USB 5V input.

>>1843095
>what am I doing wrong
trying to desolder plated through-holes
manhattan or go home

>>1843118
Have a timer IC (555) or MCU (arduino) wait 100ms before using a solenoid to push the sparker switch. Solenoid driven by a MOSFET with a diode in antiparallel with the solenoid. Trying to switch the sparker switch's wires is a futile effort unless you have a dozen kV worth of transistors and capacitors.

A simpler, less mechanical option would be using an ignition coil or some other small transformer you've got lying about (like from an old CFL's ballast circuit or xenon flash globe circuit).

>>1843157
>Have a timer IC (555) or MCU (arduino) wait 100ms before using a solenoid to push the sparker switch. Solenoid driven by a MOSFET with a diode in antiparallel with the solenoid. Trying to switch the sparker switch's wires is a futile effort unless you have a dozen kV worth of transistors and capacitors.
>A simpler, less mechanical option would be using an ignition coil or some other small transformer you've got lying about (like from an old CFL's ballast circuit or xenon flash globe circuit).
Can I use a delay line of some sort?

I took apart a lapop battery and it's getting hot, what do? it's going to catch fire?

>>1843170
Actually, yes. Panic.

>>1843168
Arguably, but it's neither the most simple nor the cheapest method.
Now it's time for the section I like to call:
>Tell Me What Project You're Attempting so I Can Tell You You're Doing This The Wrong Way
Tell me what project you're attempting.

>>1843168
>Can I use a delay line of some sort?

yeah, inject signal at one end of 30M meter wire, and expect it to reach the other end 0.1s later.

the way analog delay lines work is something like this: convert electrical audio signal to electro-magnetic signal (similar to a speaker). send signal down a foot-long spring. reconvert to electrical. so you're delaying the speed of sound, not speed of light.

>>1843188
He's trying to delay a very high voltage impulse. Even if he can find the transducer to use an acoustic delay line, the voltage will have decayed somewhat and the wave's shape will be much more spread out and less instantaneous (a low-pass filter of sorts), making it likely useless for anything that you'd use a piezo sparker switch for.

>>1843186
I'm prototyping a new type of model aircraft engine. I need to delay the piezo spark by a few fractions of a second, from the previous power stroke, in order for it to function properly. No I can't use a battery because reasons.

>>1843192

reading comprehension fail.
the solution i gave involved millions of meters of copper wire, not a delay line.
the workings of a delay line were added as background infotainment, to explain why it wouldnt work.

>>1843196

oh yeah, i worked a few weeks on something similar: canoe shoes that let you walk on water. powered by a gas canister and piezo sparks for propulsion.

>>1843196
Is the piezo spark switch actuated by a cam? If you want to offset the spark time by a fixed 100ms instead of an amount proportional to the rotation period of the engine, you can't just change the shape of the cam or move the position of the piezo. While you could build a mechanical delay (basically the mechanical analog of an RC circuit) it would need to be very bulky in order to have the force to actuate the piezo (IIRC those things take a fair bit of force to actuate). An electric solution would be more compact and lighter than anything else I can think of.

If just moving the existing cam or piezo is an option however, I'd do so.

>>1843200
It uses constant volume combustion for better efficiency, so the piston is stationary at top dead center when it needs to spark. I'm not sure how to trigger the sparker when it's not moving.

I need more amps and I have a handful of boost converters. Can I just wire them in parallel?

>>1843201
>the piston is stationary at top dead center when it needs to spark
That's not a delay problem, but a timing problem. Pretty sure you just need to adjust the timing belt.

>>1843204
The whole camshaft stops and locks in place. There is no movement.

>>1843205
Yeah but what's locking the camshaft to the crankshaft?

>>1843203
>wire them in parallel?

no, use diodes in a wired-or config, and add small resistors in series to equalize the contributions.

>>1843208
Damn, I shoulda asked before I wired it up. What'll happen if I turn it on? I just need it for short bursts, and there are pots to adjust voltage/current. I don't have any diodes available currently.

>>1843210
>What'll happen if I turn it on? I

depends. if one unit is particularly strong, it might kill the weaker ones by injecting current backwards.

>>1843208
>resistors in series to equalize the contributions
Won't the diodes do that sufficiently? They have an I/V curve after all. Assuming they don't thermally run away, that is.

>>1843217
>They have an I/V curve after all.

a very sharp iv curve implies the equiv resistance is small.

>>1843221
So the resistance needs to make a bigger impact than individual differences in the buck converters, right? That could be measured.

>>1843230
>That could be measured.

or you could wing it. say you need 8A at 5V from 4 units and you want to lose at most 5%. so, what resistance drops a quarter volt at 2A? 0.125ohms, assuming all 4 units are similarly powerful.

>>1843243
That's a lower voltage drop than the diodes, but not by that much, so it should also be enough to counter thermal runaway. I'd use 5A+ diodes just to be safe though.

Is capacities dropper a god way to power arduino?
I need to power small arduino board from 230V mains, it will be encased so nobody can touch it. So i wonder how reliable cap droppers are as power source to supply up to 500ma?

>>1843258
Yeah, so long as you don't need it to be ground-referenced. I'd put a proper 5V linear regulator after the shunt regulator, maybe with a normal diode between the two so you can have a cap before the linreg without shunting it.

I've seen it's common to use an inrush-limiting resistor in series with the dropper capacitor, but I'm not sure how much good it actually does. Oh and I think your capacitor needs to be rated to twice the peak voltage (i.e. the peak-to-peak voltage), the diodes too.

>>1843263
>Oh and I think your capacitor needs to be rated to twice the peak voltage
why?

>>1843266
Because of the worst case scenario where it's unplugged on the peak of one half-cycle and replugged on an opposite-polarity peak. Actually it's not that big of a deal, I don't see commercial capacitive droppers taking this into account. Just I think it is a possibility.

shit man, when i buy PCBs from chinks they cost half of what they do here, and those are blank pcbs, but if i buy actually finished pcb with traces and solder mask and shit, it costs even half less a pop than blank pcbs from chinks, so 4 for a finished pcb, than for a blank one purchases here
how the fuck can they turn profit like that?
honestly if it wasn't for the $6 shipping cost and 18 days delivery time i wouldn't even bother with making pcbs at home at all, this is just crazy

>>1843272
>so 4 for a finished pcb
*4 times less

>>1843258
>500 mA
You'll need a switcher for that. You could maybe get this amount of current if you design it around full bridge rectifier but components will get hot. See
http://ww1.microchip.com/downloads/en/AppNotes/00954A.pdf
https://www.tdk-electronics.tdk.com/en/373562/tech-library/articles/applications---cases/applications---cases/everything-for-capacitive-power-supplies-from-a-single-source/1380514

>>1843272
I just buy offcuts for blank PCBs.
Even cheaper than chink ones.

>>1843170
>>1843173
...and then he was never heard from again.

I've gotten myself all fucked up and confused over power electronics thanks to a solar panel and a buck converter

I bought a cheap solar panel to fuck around with.
It's a 7 watt panel that nominally generates 18 volts but actually generates 23 here in the summer Florida sun
P=VI -> P/V=I. Given the specs, it can deliver .3-.38 amps.

1. I attached a buck converter to it, tuned to around 5.1 volts so I can power USB stuff from it.
Given the initial conditions of 18-23 volts, 7 watts, and .3-.38 amps, what the fuck is the buck converter actually able to supply in terms of power and current? Slapping a converter on there obviously won't magically bypass the original constraints of the source of electricity so something's gotta give. But what?

2. Can someone explain to me why it's valid that, given P/V=I, it would seem my solar panel would be capable of delivering ever more current the lower the voltage goes? Obviously this isn't the case, so what IS the real-world case?

>>1843412

The converter deals with energy, that it stores in a magnetic field. It can retrieve that energy without any regard to the volts or current that put it there. The main other factors are efficiency and ripple, but what you need to read up on is basic electricity. You don't need to fully understand the converter circuitry, just first semester EE.

>>1843412
>what the fuck is the buck converter actually able to supply in terms of power and current?
Power in = power out. If you give 7 watts to the converter you'll get 7 watts out of it (actually a bit less since it's not 100% efficient). You can have 7 watts at 5 volts and 1.4 amps, or you can have 7 watts at 18 volts and 0.39 amps it's the same amount of power. You can make an analogy with water and pipes. To move same amount of water in thin pipe you need to have high pressure (voltage) and if you have thick pipe you can move the same amount of water with low pressure. Buck converter is changing high pressure thin pipe to low pressure thick pipe.
>Obviously this isn't the case, so what IS the real-world case?
Solar cells have current(voltage) relationship as displayed on pic related. They can't provide more current than their short circuit current, and they can't provide more voltage than their open circuit voltage. For solar cells power is not a constant so P/V=I doesn't hold everywhere, but only in that specific point. If you try to lower your voltage you'll move left on pic related. As you can see current doesn't change (e.g. move from 10 to 5 V), but since you changed voltage you've lost available power. For that reason solar cells are usually used with maximum power point tracker that keeps voltage at a specific point to transfer maximum amount of power form the cell.

>>1842028
But it is chinktech.

>>1843412
your bucking to little and there are losses in transformers in the core and length of wire. bucking at such low voltages ( sub 100 volts) should yield any real noticeable increase in current due to the fact the losses in the core are present assuming your using a steel core one

bucking is almost always done with a generator that can produce around 100 - 200 volts or socket power.

make darlingtons and you can get more current

or if you make a monoastable multi vibrator (oscillator circuit) you can have that charge 2 caps (or cap banks) and discharge them. so you exchange constant dc power at lower current for pulses at higher current over time. it sounds more complicated than it is. so you make a simple inverter circuit type oscillator. 2 mosfets (or 2 bjts that can handle some current) 2 caps (to a large extent these control switching speed 10nf is good for inverters but 100nf would be fine for just getting more current) a cap for charging , 2 resistors 200 - 400 ohms and 2 inductors (inplace of the transformer 20mh is fine) . put the 2 small caps to connect the 2 n channel mosfets gates to each others drain (npn bjts bases to each others collector) the resistors connect to the gates (bases) and positive power connects to them and the inductors at the same point and the other side of the inductors goes to the drain (collector) and this is where it differs from a transformer connection on the source (emitter) put a cap across each one and then connect each one individually and then the wire off connect to 2 diodes and those connect facing each other off the wire from the caps and source / emitter. that goes to the load and then load to negative of the power. this will pulse by charging and discharging over and over. but the gains depend on a number of things like total copacitance of the charging cap and speed its charged and if it was charged to full. the change may not be to great and darlingtons may be better in your case since you get 0.3 amps

I am not knowledgeable with circuits or voltage or watts or amps or anything like that. But I have done some research, and I -think- I have everything I need to complete my project.

I want to power the air pump with the 12V battery at the bottom, and control the flow of power/air with the PWM circuit that came from HK. I'm using the PWM for minimum power loss. My concerns are this:
To connect the power source to the circuit in a way that is sturdy and safe. I'm want to attach a female end to the circuit so I can just plug in the cable w/o cutting it up.
To connect the air pump to the circuit. I know I have to cut that cable up, but again I want to have it reliable and safe. Is there a way to make a "plug and play" connection?
To extend the pot on the circuit 2 meters. Will any similar cable size work or do I need to calculate things to get the right cable size/length/type?

I also received 3 items from a friend, after he heard of my project: A pot, a package of 2SD882 transistors, and 2 resistors. Would these components achieve a similar effect, just with greater power loss? I have a side project in mind, and with the politics with HK and the US, I don't think I can get another PWM from there any time soon.

Also also, would a good move for me be to look up colleges near me and see if any staff members w/ electronics backgrounds set up a one on one meet to help me, or do colleges in America restrict their professors to semester classes?

>>1843157
>The 9V 80mA will be fine as-is unless you want to turn on a bunch of LEDs
I've just got one LED right now, but it's not lighting up. What am I doing wrong?
p-pls be gentle i am a noob

>>1843447
You're connecting:
+5V -> resistor -> LED -> +3.3 V
Red LED needs about 1.7 V to turn on, since you have 1.7 V across it (plus resistor) it won't turn on. Connect the white jumper to - (blue) rail.

>>1843428
Alright, it's making a lot more sense. I figured it had something to do with solar cells being nonlinear.

So, just to make sure I'm understanding right, regardless of what the current into the buck converter is, assuming favorable conditions, the buck converter will supply up to the same power that the solar panel is producing, less the losses from the conversion inefficiencies?

I notice that when I attach a load to the buck converter (such as a USB power bank) there's significant voltage sag (5.1 volts unloaded -> 4.6 volts loaded) at the buck converter's output, even if I bump the buck converter's voltage up. It always sags to 4.6-4.7 volts. I haven't yet checked if this is a USB standard thing, but I don't think so. I also haven't tested what current it's supplying yet.
Any insight as to what might be happening? Is this normal for a buck converter to sag that much? How might I go about stiffening the output? I was thinking of sticking a few chunky capacitors across the output but I don't know if that would work.

>>1843438
I'm going to look into using a monoastable multivibrator to charge those capacitors. I'll finally have something practical I can use my lm386s for

>>1843451
>regardless of what the current into the buck converter is, assuming favorable conditions, the buck converter will supply up to the same power that the solar panel is producing
Yes, as long as it's in the operating range of the converter.
>Any insight as to what might be happening? Is this normal for a buck converter to sag that much? How might I go about stiffening the output?
You should expect around 2 % voltage variation on the output. You have more than 9%, so it seems to me like voltage from the solar panel drops too much which limits your power output. If you want to get full 7 W out of your panel you'll need a maximum power point tracker. Such devices vary the effective load of the solar cell such that maximum power is transferred. Your buck converter (when load is constant) is some resistance that determines power point of your solar cell, which can be shown on characteristic as red line. You can see that input resistance of buck converter determines your power point (intersection). When you use MPPT that circuit changes its input resistance to force panel in maximum power point. Try measuring output voltage of your cell with load (power bank) attached.

>>1843447
you just fucked up connecting two components and ONE wire to a fucking breadboard
Congratulations, you are now on par with the American EE degree

Is microcontroller general dead. Is this the place to ask arduino questions again?

>>1843482
>Is microcontroller general dead. Is this the place to ask arduino questions again?

Why not start the thread? Anyone can. Or ask here.

>>1843443
>I'm want to attach a female end to the circuit so I can just plug in the cable w/o cutting it up.

just cut the cable. you already have too many connectors in series with each other, and you want to add more. if you did, you'd have to source a female with the right dimensions, which you're prob gonna screw up, and then solder wires, which you're gonna screw up, and get the polarity right, which you have a 50-50 chance of screwing up.

if you really really value your cable, you might be able to find a power brick with a connector that fits in your battery, and sacrifice its cable.

>To connect the air pump to the circuit.

cut the cable! you might find a cigarette lighter extension cord at the dollar store (i did last time i visited) but most likely you'll have to wait another week or month to get it delivered.
(if you have a soldering iron, heat shrink and skillz, any cut cables can be restored good as new)

>To extend the pot on the circuit 2 meters. Will any similar cable size work

any wire will do

>Would these components achieve a similar effect, just with greater power loss?

yes, but it's not a good solution. besides power loss, the motor will have little torque at low speed. see pic.

> look up colleges near me and see if any staff members

a bit delusional. no college prof is gonna coddle some noob. maybe you can find some retired engineer/tech you will answer your questions while he caresses your hair.

>>1843447
>>1843450
Resistor is also way too high. That's a 1MegaOhm resistor. You need something much smaller, like 220 ohms to start.

>>1843482
I never really got why they started that general. It's not like /ohm/ moves at lightning speed. Also microcontrollers have always been an EE topic, so I don't see why they get their own thread when this one does fine.
I'd say ask here, but maybe I'm biased as microcontrollers are my best EE subject.

>>1841869
Is it practical to remove the top layer of a metal backed PCB so you only leave the metal backing?

I have an LED panel that uses the PCB not only to power the LEDs but also as a structural piece to mount the heatsinks and LEDs to the housing
If I wanted to replace the single diode LEDs with lets say some cobs then I would have to make up an aluminum plate to replace what the PCB does and I'm trying to avoid that

I will never not be salty that creating a rectifier is piss easy but making an inverter is a lesson in electronic cock-and-ball torture

it's like fucking calculus all over again

>>1843579
so /ohm/ doesn't get flooded with code and blinkies

Best solar cell for DIY projects under a foot?

I need some advice guys.
A long time ago, I attached an 8 inch crt from a portable TV to a PC, pretty neat, except I had to use a VGA-to-composite and a RF modulator. Quality isn't great because of the modulation and demodulation process, so I want to get rid of it (at least, almost).

The TV main chip is a AN5151. It does the demodulation stuff, controlls deflection and stuff, and has the video output, great. The idea is, keep the modulator and demodulator (because the IC controlls all the CRT stuff and is synced), but I want to bypass the AV signal with the original one.

The main problem is the delay. I've measured something like 1,6us (pic related: up->original down->output of AN5151). I don't know if that time is really a problem. If it is, I need a way to get rid of that. I've got a few delay lines from old TVs, but I read somewhere those are 64us, but I'm not really sure.

>Inb4: Is the output signal the real problem? Yes, I checkd.

>>1843636
everything about pumping square waves through an inductor seems wrong, I don't understand why inverters insist on using that shit

This summer I'm going to finally learn how computers and electricity actually work, beyond the level of regurgitating definitions. I see NAND2Tetris shilled a lot from various communities, but I think Ben Eater's DIY 8-bit computer series (https://www.youtube.com/playlist?list=PLowKtXNTBypGqImE405J2565dvjafglHU)) looks a lot more hands on and fun (and therefore a better learning experience - plus it's more time away from the screen). If you haven't seen it, he walks you through constructing an 8bit computer from a breadboard.

Am I wrong in thinking that they will contain roughly the same content? I'm still waiting on some parts to arrive before my nosedive into this, along with clearing up and building a workspace, so I'm hoping to gather some resources and direction right now. I'm for sure buying that book Ben Eater mentions he got the design from. If you have any supplementary educational resources, I'd appreciate it. I've always wanted to read the famous CS:APP book and The Art of Electronics, so maybe I'll flip through those in conjunction with the project.

As an end goal, I'd like to make a compiler for a mainstream, standardized language like C, so I can run homemade C programs on my future 8bit machine. Is that an unreasonable ambition? I'd also like to be able to build some dope synths and get into FPGA programming down the road. I know I'm sort of all over the map and pie-in-the-sky sounding, just super excited right now and can't wait to get my space operable so I get moving.

>>1843565
>you might find a cigarette lighter extension cord at the dollar store
This, along with old barrel jack power bricks, is the sort of thing you should be going to a junk shop for. They get tossed all the time.
>pic related
pretty sure that transistor will eat power and explode

>>1843575
1M is usually enough to see a small glint, but if it's under reasonably bright light this can be hard to see.

>>1843579
It really got into the nitty gritty of MCU topics at times, like different architectures and such. Having the threads split made it a bit easier not to get lost in the huge range of topics.

>>1843641
>blinkies
Is this a derogatory term for arduino kiddies? Because I really like it if so.

>>1843675
I'd personally consider replacing that AN5151 to ditch the mod/demod steps, but it's up to you. No clue as to your question, sorry.

>>1843708
What do you think a buck converter or class-D amp does? They're all around us, embrace the slightly noisy future!

>>1843805
>NAND2Tetris

awful awful garbage: 1 million tangents, irritating accents, everything is virtual, completely useless end knowledge.

>Ben Eater's DIY 8-bit computer

brilliant, all real, nothing virtual, completely obsolete and useless end knowledge.

you want knowledge you can use, but these 2 series are 90%+ not applicable to current tech, so 90% wasted effort. what would be great would be if Ben did an equivalent series on a modern CPU or development system (say arduino or raspberry pi). i'm sure there are books and such if you look.

>>1843853
Yeah, the only useful guys on YT seem to be those dealing with parts of electronics that haven't really changed much in 20 years. Like analog and radio electronics. Or it's the guys doing arduino stuff that's technically modern but not actually very professional.

I think FPGAs or CPLDs are a bit more down Ben's alley, and I'd like to see a more theoretical approach to those.

Is there any reason why I should not use this: https://www.aliexpress.com/item/32965277185.html
instead of a classical adjustable transformer as variac?

>>1843884
It depends for what purpose you'll use it. It says it's a thyristor based regulator, you won't get nice sine on the output as you would with a variac. You'll get a chopped up sine like pic related.

>>1843884
I thought the same thing 5 years ago when I had no clue how TRIACs worked. At that price, it was too good to be true! It was only later that I learned this: >>1843887.

I implore you to learn the basics behind switched mode power supplies, both buck and boost, and flyback too I guess. Because using normal AC transformers is (for the most part) a thing of the past. There are some real cheap line switching controllers ICs on ali that you may be interested in. The TL431 is also an excellent part that may help you in your endeavours.

>>1843880
>technically modern but not actually very professional.
and this is where you fucked up buddy
by lifting your nose up on arduino stuff
arduino is the PERFECT entry point for getting into EE, dev boards for easy and fast prototyping and the best part is that if you make a prototype on say arduino uno, you can take the atmega chip powering it and use it in your fancy professional production while still being able to upload the same code to it with arduino ide

>>1843898
I mean, I lift my nose at it but I still do exactly what you're describing. But the arduino environment itself is a bottleneck, the hardware less so. There's a significant step between slapping code together in the arduino environment and making a proper project in MPLabX or any other dedicated MCU IDE.
If I wanted to use a lesser known MCU with no arduino support I'd be a fish out of water, both for the programming and compiling stage, and for the troubleshooting stage.

>>1843708
If you push a slow frequency square wave, then yes, its wrong. A 60 Hz square wave instead of a sine wave through a mains transformer is very bad.
SMPS use square waves, but in the 10's or some in the 100's of thousands of hertz, so the chance for total saturation while conducting is much lower.

>>1843840
>Is this a derogatory term for arduino kiddies?
Not just arduino kiddies, but they tend to be the worst.
Its usually the first "hello world" project where you blink an led at 1Hz. After the insanely hard studying and code writing in the arduino IDE, most are overcome with the joy of finally creating such a difficult design that they post videos and pictures about it online. /mcg/ was the net to catch them.

>>1843675
>>1843840
Turns out way easier than I thought. The IC doesn't get the sync directly from the demodulation process, but froom the demodulated signal, wich isn't (I think) what the block diagram from the datasheet says.
Anyway, I tried injecting the signal directly, and voilá, it works. Quality is way better, and it is a constant image without blurry and unstable sintonization.
The only problem I have is those white bars, kinda annoying. I don't really know the source, they arent from the original signal, but I'll inspect the signal anyway.
Do you guys know about these white bars?

>>1843898
>if you make a prototype on say arduino uno, you can take the atmega chip powering it and use it in your fancy professional production while still being able to upload the same code to it with arduino ide

The companies that do things this way are trash and make over-priced crap that typically fails.
The arduino system is good for learning, but not mass-production levels. Its terribly inefficient and not cost-conscious when producing large volumes.
Arduino code is bloated and not well optimized. Lets say you design a system and want to make 30k units. To maximize profits, you need to get the cheapest microcontroller you can to run your code. If the duino IDE dumps out a 2.5k hex file to load up, you'll need to get a uC with around 4k program memory (flash memory usually jumps from 2k to 4k). If you write the code in MplabX (Microchip IDE for Atmel chips), you can easily get the code under that 2k mark and go with a uC that has less flash memory, which will be cheaper. Lets say 5 cents cheaper, but at 30k units, that's a $1,500 savings. For one part, and also extends to the customer as a less expensive product.
You'll be limited to the chips that only Arduino IDE supports. If there is a cheaper and smaller chip not supported, you're out of luck.

Being an engineer doesn't only encompass slapping parts together and making some code. If you work for an actual company that sells real products, doing everything you can to make that product cost less is also engineering. For a few pieces, a $0.50 difference for more memory isn't a big deal. For mass production, that is a LOT.

>>1841869
What's the best budget soldering station I can get for my money? Around $50ish max.

I was considering
>https://www.aliexpress.com/item/4000444474862.html?spm=a2g0o.cart.0.0.7dc53c00ENSh6q&mp=1
Because ali's got their stupid summer sale and it's $30 but I don't know.

>>1843898
I'm the samefag who posted the Terry Davis image, basically my overarching goals are a) to learn how computers/electricity works on a fundemental, hands on level, b) get some applicable skills for automating stuff around the house, and c) I'd also like to build a eurorack diysynth to apply what I've learned.

Am I going down the right path with the 8bit stuff? I know 8bit architecture is obvioulsy outdated, but the computing concepts should be universal, right? But it sounds like maybe Arduino would get me where I want to go faster, maybe?

>>1843636
all these bucks shmucs. it is a fad. a cool new technology that will never mature. we will get back to linear power supplies eventually. they are still used when low level of noise is a must.

>>1844053
bump for interest, also what's the official recommended /ohm/ soldering station in general?

>>1844062
>c) I'd also like to build a eurorack diysynth to apply what I've learned.
Some modules are retard friendly, others require a bit of analog understanding, and some require some pretty heavy DSP.
Some modules, like VCO and delay/reverb, have ready-made packages you can just buy (that are used professionally).
What exactly do you plan on making?

>But it sounds like maybe Arduino would get me where I want to go faster, maybe?
Yes. Make blinky, make a pwm/CV sequencer, make a keyboard scanner/switchboard, interface with things like humidity sensors and a 2x16 display, etc.
>but the computing concepts should be universal, right?
pretty much

>>1844085
>>1844085
I don't know what I'm gonna make yet, just that I'm going to get started. I play bass and guitar, so I thought maybe some pedals would be nice since I don't own any, but I want to focus on getting a basic minimal modular Eurorack first. The heavy DSP is what I'm most excited for personally, I think it's really cool conceptually. Following this series for that the Eurorack build: https://www.youtube.com/watch?v=5DKU1m5_b_8&list=PLyE56WXw0_5Q5QGMEXWmskuhojKyRdA3T&index=1

I'll buy a couple Tarduino's today and toy around with those, I have shitloads of practical things I can use these for (homebrewing, indoor gardening, and a small business with lots of tedious labor that can hopefully be automated).

>>1843805
Look up "digital design and computer architecture" by harris and harris. It's on libgen and there's also a solution manual an there too. It's one of my favorite textbooks of all time.

>>1844062
arduino ide will allow you to make programs for chips in minutes which would take hours to make with anything else (at a price of a small overhead usually and a limited selection of chips of course since arduino ide doesn't support everything obviously, but the chips it does support can handle pretty much any project anyway).

>>1844111
>digital design and computer architecture
Thanks for the recommendation and digits. I wish my laser printer worked so I could print this stuff out and bind it.

>>1844111
The arm edition is probably best for me, right? It's also the most recent.

>>1843939
Replaced some caps, put bright to almost minimum and contrast high enough, and I got a pretty decent retro display.

The cut at the left is from the VGA-to-AV converter.

>>1844125
>>1844125
I haven't read it. I read the 2nd edition of the MIPS one. I think either one will work fine though. The concepts they teach about computer architecture apply to nearly all other architectures.
I managed to find a like-new copy for $30 once and I'm super happy about that, so keep your eyes peeled and you might be able to snag a cheap physical copy.

>>1844164
Crts are magic.
I have a couple of tiny ones myself. The smaller they are the cuter the look.

>>1844164
That's real cool looking, neat work.

>>1844164
>>1844219
retro technology is so aesthetic looking
>tfw ywn be an electrical engineer at a silicon valley engineering firm in the late 60's, spending your days taking advantage of the new microprocessor technology and your nights taking drugs with your hippie friends.
Why even live, bros? ;_;

Maybe this is a bit off topic, but this is today's Whomp! comic.

>>1844226
...and of course I lost the image somehow.

>>1844117
>arduino ide will allow you to make programs for chips in minutes which would take hours to make with anything else
One of the most retarded statements that I constantly hear. You'll save more time by using pre-built arduino hardware than making your own. On the software side, its not that time saving.
The IDE has a lot of built in libraries, so you don't have to toy around with settings. That is a learning curve issue. If you know how the SPI module works on a chip, you can set it up in seconds. The only difference is you need to learn it first as opposed to letting the IDE do it for you.

>>1844062
>I'd also like to build a eurorack diysynth to apply what I've learned
Audio isn't a walk in the park, just keep that in mind. If you want to build something that processes audio, you'll need speed and lots of memory - something the arduino lacks. You'll need to step it up to something with DSP abilities or excessive speed. Barring that, you can stick to external chips/circuits that process audio in hardware or are dedicated audio processors.

>Am I going down the right path with the 8bit stuff? I know 8bit architecture is obvioulsy outdated
8-bit is far from outdated. The bit depth depends on what you need. If you are building a timer to water a plant, there is no need to throw down a 32-bit controller. These are embedded microcontrollers because they are designed to get the job done without excessive processing power.
The best place to start is 8 bits. It is much easier to only deal with 8 bits at a time. If you need to jump up to a more data intense system, its usually a smooth transition.

How difficult would it be to design a VHF power amplifier that should be able to drive at least 20W into a 50 ohm antenna between 88MHz and 108MHz? What topology should I be looking at for the voltage gain section and power gain section? Obviously op-amps are out because awful bandwidth? Common base with a class A output driver? Maybe cascode instead?

>>1844253
please explain to me the appeal of jammers

>>1844240
Not him, but I have a audio question

I'm going to be wearing a face covering that muffles my voice. I want to be able to speak into a microphone in the covering, and have a speaker attached to the mask make what I'm saying clear. I also want to be able to have everything off by default unless I press a button, just like push to talk, only IRL. Otherwise if it's on all the time, it would pick up on breathing and other unwanted sounds.

>>1844271

you wanna look for bike helmet intercom systems. they let 2 people on a bike talk to each other. similar things exist for ski helmets.

>>1844271
>>1844240
So how complex will the needed circuit be?

>>1844267
I wouldn't know but that doesn't really answer my question.

>>1844240
In regards to the 8bit stuff: awesome, those were my thoughts exactly when I laid out my plans for this project.

In regards to the DSP stuff, I lack an understanding of what I'm trying to build, so what you said didn't really make sense. I'm trying to build an analog modular synth, I don't understand at all how memory comes into play since there's no 'computer' component, it's just voltages being passed around. Am I wrong?

>>1842596
you could do it by charging/discharging a capacitor through a constant current thing and some extra stuff to manage switching
you could probably do it really shittily without opamps by using a high resistor as a constant current element

reading the full reply chain, if youre gonna output 1hz charging a cap is probably a bad idea
the chink ones i have partly discharge in 1 second

>>1844344
You're describing a relaxation oscillator of sorts. By using a low threshold voltage relative to your voltage rails (i.e. a small amount of positive feedback) they typical exponential capacitive ramps will appear triangular. But these oscillators are usually made with a comparator/op-amp anyhow, the transistor-based ones are often much more finicky to get working and don't have a tuneable hysteresis. Using a >>1842880 circuit would give a much better result, I think using one comparator (for the slew rate) and one op-amp (for the linearity of the integration) would be best and still really simple.

Can someone explain what a 'dome' filter is?
I can find precious little information about what they are.

>>1844361
Looks like it's just an all-pass filter.

>>1844367
Is it just an all-pass filter which shifts the signal +/-90º ?

im gonna put my cum in a microwave,. what should i expect

>>1844377
Horrible smells

>>1844368
>an all-pass filter which shifts the signal +/-90º
I just looked that up and woah, never heard of those before.

>>1844380
I'm not entirely certain if that's actually what it is though.
If it is, what's the advantage of that over simply a unity+inverting amplifier, since that also creates two copies of the signal 180º out of phase?

>>1844383
Because you can make phaser audio effects

>>1844383
An op-amps inverting amplifier, well an ideal one anyway, will invert the signal 180 degrees at ALL frequencies. Look up an all-pass filter bode plot and look at the phase section. Notice how the phase shift changes with the applied frequency.

That's the difference.

>>1844385
That's more or less how I came across dome filters.
This is a block diagram for an all analog pitch shifter.
I just want to understand how it works.

>>1844386
Ah, I wondered if that was why it was needed.
The only source I can find that explains the circuit in detail is from electronotes, but they want $400 for the book, and they aren't even taking orders currently.

>>1844065
>>1844053
Bumpy Bump

I'm looking for a way of measuring force on a fixture. While the force should stay less than 1kN in normal operation (usually in the realm of 200N), it's highly likely that it will spike over 10kN during sudden impacts. While I don't actually need to measure these events, I do need a force sensor that won't be permanently deformed by such a thing happening. The force sensitive resistors I'm seeing on alibay are all either alimunium blocks with holes in them or those little stamped junk metal plates, which both look bad for this. Then there are those force-sensitive resistors that aren't mounted to anything, but they don't look easy to mount on a custom piece of barely elastic metal. The maximum force I actually want to measure would probably be ~2kN.

So I was wondering maybe if a linear encoder would be a better option, something like this:
https://cdn.ozdisan.com/ETicaret_Dosya/525365_671433.pdf
But I don't want the barely elastic metal to flex more than 2mm in normal operation, meaning I want good resolution around 200N (say, ±20N) while having a full scale of 2kN. So If 2kN > 2mm, then 20N > 20µm. I don't think this linear encoder can handle that, can it? I'm not good at reading its spec-sheet.

I'd consider using piezo discs, but I need a DC-coupled method, and I suspect piezos will leak/drift too much. Also maybe a bit too brittle for those impacts.

Anyone know a force-sensitive resistor or other force sensor that I can easily connect to a semi-elastic piece of metal?

>>1844435
this is 2020 nobody uses bulky bricks like that, it's all about ts100 these days

>>1843941
>The companies that do things this way are trash and make over-priced crap that typically fails.
grbl, the code used to run a big part of CNC machines all around the world, both in factories and smaller shops, is written in arduino. It is written by a guy with PhD in EE, but yeah he is a just retard making babbys first blinky who has no idea what hes doing, you could do much better job than him for sure by not using arduino ide

arduino ide is just like any other ide for MCUs, you can write both shit code and good code in it, the difference is that most goal functionality is much easier and faster to achieve with arduino ide. And your code will actually have smaller amount of bugs, because arduino functions have many safeguards and optimizations under the hood. This is a fact.

The ONLY reason to not use arduino ide is when you need to use a chip that isn't well supported by arduino ide and since chips supported by it can do almost anything today, the only scenario when this happens is when something like attiny13 (dirt cheap) is still too expensive and you want to use some 10 cent IC for mass production. For any small production run cost saving is not worth it (woohoo we saved $20 on ics total). In closing, any other argument against using arduino ide than cost saving for MASSIVE production runs I will take as ramblings of some hipster retard who hates on popular things because they are popular and thats it.

>>1844549
Does any "real" CNC actually run grbl?
I know some cheaper ones run LinuxCNC but I haven't seen grbl on anything but those tiny Chinese/diy CNC routers.

Certainly never seen an Arduino in a factory setting. They use Industrial PCs and PLCs - hardware specialised for industrial automation.

>>1844550
Arduino != Arduino IDE

>>1844551
grbl only runs on the Atmega on Arduinos.

>>1844515
https://en.wikipedia.org/wiki/Strain_gauge
https://www.hbm.com/en/0014/strain-gauges/
This is how they measure force of deflecting materials in industry. There should be something that'll fit your needs.

>>1844557
Those are the same force-sensitive resistors that don't seem to have a good mounting solution I was looking at on alibay. Interesting to see that they've got tempco compensation for different mounting materials. How do you mount them to things? The wikipedia page recommends CA glue for short-term and specialised epoxy for long-term, but how would I get the gauge in tension for mounting?

>>1844549
>grbl written in arduino
>An open source, embedded, high performance g-code-parser and CNC milling controller written in optimized C that will run on a straight Arduino
>written in optimized C
There are no arduino libraries used in this project. You can run it on arduino, but it's not using arduino software libraries.
>because arduino functions have many safeguards and optimizations under the hood. This is a fact.
Due to abstractions required to run on many different boards you also have a lot more code - lots of bugs.
https://github.com/arduino/ArduinoCore-avr/issues
>optimized
Writing to a register to turn on pin takes 1-2 clock cycles, using arduino digitalOut() takes 50 clock cycles. This is optimized how?
>The ONLY reason to not use arduino ide
It doesn't even have a debugger included with it. Pic related is how professional development environment looks like. See the registers, the breakpoints, call stack, instruction counter... Arduino IDE is just notepad with upload button, which is great for artists, makers and hobbyists, but will make your life very difficult in real world embedded programming.

>>1844567
>Writing to a register to turn on pin takes 1-2 clock cycles, using arduino digitalOut() takes 50 clock cycles
holy kek, how will blinkies ever recover?

>>1844563
>but how would I get the gauge in tension for mounting?
You don't need it to be in tension, glue it to your fixture in neutral position.
https://www.youtube.com/watch?v=s4Bq8MvwbyU

>>1844568
Nobody forces you to use the functions, you can easily write into registers directly if you so choose.
The functions take more cycles because they have safe guards to prevent problems like cross talk with analog reads and such. Not to mention that in 99% of projects a big part of the CPU resources is unused anyway so if you use these otherwise wasted clock cycles to do do some extra safety shit, it's a good thing.

>>1844567
If you learn arduino syntax (very easy) you can now write code for 1000 different chips ranging from tiny ics that cost under a dollar, to big multi-core ics with dozens of pins with built in wifi bluetooth and other shit

If you learn some chip specific IDE used just for that one chip you can now develop software for that one specific chip. And the leraning curve on top of that is usually much steeper than learning arduino. Want to use another chip? Have fun learning new IDE from scratch again for that new chip.

The only real downside of arduino ide is the lack of hardware debugging, but there are some extensions for certain chips that work with the ide which allow you to have certain hw debug functionality. But for simpler code you rarely need more than serial print or led blink for debuging anyway.

I'm what you might call a retard when it comes to electronics and have only done the bare minimum of soldering.
I picked up this metal detector on craigslist that the guy said broke after he put batteries in backwards. How plausible would it be to fix? Other than the circuit that turns it on/off, everything looks fine to me.

>>1844575
> Want to use another chip? Have fun learning new IDE from scratch again for that new chip.
Not true, most IDEs support multiple architectures and can be used with cmake/makefile based projects. And since most modern MCUs are ARM based it's already standardized across different manufacturers. Besides if you take time to learn how MCUs work at lowest level you'll find out that they're all basically the same regardless of the device. You control the hardware by writing to registers, different platforms just have different names for them. That's all what arduino does, it's just sets the correct bits in the registers based on the selected platform.
>but registers are scary and I don't want to read 800 page datasheets
Then use HAL/libraries provided by manufacturer. It shouldn't take you more than few hours to get familiar with it. Good programmer doesn't learn specific development environment like arduino and complains if shape of run button changes when switching to something else. You should learn concepts not environments and you can learn that by learning to program on lowest level using registers.

Arduino is great for people who don't care about how anything works (artists, makers) and just want to do something quickly. Once things start to get complicated they give up or buy a new device because the old one was "not powerful enough", even though it was perfectly able to do the required task. There is nothing wrong with that, but claiming that it can be efficiently used in production is just bullshit.

>>1844576
>it didn't have a reverse protection diode
fuck
Do you have the electronics know-how to trace the circuit and figure out what's what (assuming there's no circuit diagram online/in a booklet)? Because I'd want to do that to find all the routes for the reverse voltage to damage things, and to figure out from context what components are what and what values they should be for proper functionality. But even then it may be impossible to tell what component to replace a dead markless transistor with unless you have the service manual. Just hope that a generic BJT or MOSFET replacement works, I guess.

Desoldering SMD ICs is a real pain without hot air, to the point where if there's even one 8+ pin IC I'd give it up. Maybe some sort of solder-proof shim that you can slip between the solder pads and the IC's legs (aluminium soda can?) will work well enough, however. You absolutely must avoid situations where you delaminate the copper from the fibreglass board, because reworking that sort of problem on an SMD board is hell.

>>1844583
>That's all what arduino does, it's just sets the correct bits in the registers based on the selected platform.
If I was a real wiz who was comfortable in my IDE, I'd make my own addons/software to do this for every MCU I use. Doing it myself would really get rid of a lot of the uncertainty between porting code from one MCU to another that I currently get in the arduino environment. But before that I'll need to get an actual operating system.
>Arduino is great for people who don't care about how anything works (artists, makers) and just want to do something quickly
IIRC the remote control lot use arduino a fair bit, and for controlling a few IOs it's perfectly adequate. In fact, they'd probably be fine with the capabilities of a bottom-end padauk, but the arduino is just really convenient.

>>1844586
>assuming there's no circuit diagram online/in a booklet
electrical diagrams are readily available for it, for what that's worth.
It sounds like it would be a pain in the ass though from the rest of your post though, especially given how little background I have with this sort of thing.

I've a switch mode power supply that I had to replace caps on recently.
The switch driver's pwm output is driving a push pull pair of bjts which are driving a mosfet gate.
What would be the reason for doing this rather than driving the gate directly?

>>1844586
>>1844583
>>1844575
>>1844567

Feel free to continue these conversations here, or perhaps move to the new MCU general >>1844608

>>1844611
>please leave my dead thread with one post per day on average alone
you should be grateful we are keeping this fucker alive you pleb

>>1844598
mosfets take time to switch due to internal parasitic capacitors at the gate. the faster you switch them, the less power you waste. which isn't just about ``green'' energy, you also have to worry about overheating the mosfet from switching losses. so you want to dump as much peak current as possible into and out of that mosfet gate to charge it as quickly as possible and thus minimuze your switching losses. the bjt buffer helps accomplish that.

so after >>1843458 introduced me to the concept of MPPT, I've been trying to read up on it and figure out how to implement it with my shitty solar panel

the diagrams always show some kind of input to the buck (or boost) converter. What is the MPPT controller actually controlling on the buck converter?
I assume it's outputting a PWM signal to the base/gate of the buck switching transistor, which would vary the voltage and thus the load provided by the buck converter?

>>1844719
Yes, MPPT controller is changing the duty cycle ratio of boost controller, which in turn changes input impedance of boost controller (what solar panel sees). Voltage at the output of the converter on schematic won't be fixed, as you're using that controller to provide fixed input impedance so that panel is in maximum power point. After that converter you have to use another step down converter (because you boosted it to get correct input impedance) to get usable voltage.

>>1844722
Alright, that makes sense. I'll wind up doing the math soon enough, but are the gains from implementing MPPT worth the efficiency drop from having two converters on the circuit? I assume so since they're used in any actual PV system.
But this panel is only 7 watts, after all...

>>1844723
It depends on what you want from your panel and what you're using it for. Attaching low impedance load to it could easily half your available power so instead of 7 W you would only get 3.5 W, but if you had MPPT you would get 6.5 W (some loss due converter inefficiency) in same conditions. You can get pic related for less than $10 so implementing it yourself is not worth the effort unless you want to learn how it works.

>>1844593
how long do I need to study electronics to understand wtf is going on here

>>1844732
Few years from zero knowledge, more if you don't know algebra, complex numbers, Fourier/Laplace transforms. Here is a course that would give you a good base level knowledge (provided you aren't scared of equations)
https://wiki.analog.com/university/courses/electronics/text/electronics-toc

>>1844744
lol what? Fuck no. If you know basic op-amp theory you know 90% of what this circuit does. Just analyze each block individually then put it all together. You don't need any fancy calculus to figure out how this works on a basic level.

If I only have a 0.1uF and need a 0.05uF for this shitty project, is there a way to get around this? Like, can I do something to the 0.1uF to convert it to a 0.05uF?

>>1844753
Put 2 of them in series

>>1844764
Thanks, which formula is used here? Wouldn't putting them in series make it a 0.2uF?

>>1844765
Nah, capacitors act the opposite of resistors.
Capacitance gets bigger in parallel and smaller in series.

https://keisan.casio.com/exec/system/1258035746

>>1844768
Ah there we go, I used this. I didn't know that. In school they only thought us resistors... But what do I know, it's been a while hahaha

Thanks Anon

>>1844746
>You don't need any fancy calculus to figure out how this works on a basic level.
I don't know about you, but there are some 2nd order systems (U10B for example, capacitor in feedback C21 and to ground C22), which would require you to know about bode plots, which requires knowledge of fourier analysis to fully understand (at least designer needed to know that).

having a bit of a gamer moment right now. The input to this buck converter is 22 volts, +-.5V of noise.

Why, if the square wave is at 100% duty cycle, is the output voltage only 7 volts?

>>1844785
with the load resistor removed, it's at 8.5 volts which still doesn't sound right

>>1844785
That PMOS will turn on when Vsg > Vth. I don't know what is your input signal but it should turn on if you give it 0% duty cycle and not on 100%.

>>1844788
> 0% duty cycle and not on 100%.

And he should understand that neither of these is a square wave; both are DC inputs, so his circuit composed of inductors and capacitors is not going to do much of anything useful.

>>1844788
And I'm being retarded again, you're using NMOS, it should be PMOS (arrow up). It's confusing enough since they use 2 different symbols with different arrow directions.

>>1844791
that's a textbook buck converter

>>1844788
>>1844794
the input signal is a constant 22 volt DC signal, with .5 volts of noise.
I'll try playing with a p-channel mosfet

>>1844797
>that's a textbook buck converter

Fine. Now got to the index of your textbook and look for "duty cycle", and read about 0% and 100% and why they are not square waves, and therefore your textbook buck converter won't do shit unless you actually drive the fet with an actual square wave.

>>1844797
Problem with NMOS is that your gate voltage will need to be higher than output voltage for it to turn on, ICs have integrated level shifters that take care of this, but you have to do it yourself here. PMOS will work as long as your output voltage is higher than Vth, you can turn it off by getting gate higher than Vout - Vth, and turn it on by connecting it to ground. You can use either but your gate signal needs to be adjusted to reflect source voltage.

>>1844801
keep it in your pants smartass, it should be obvious that the circuit isn't working correctly for any value. Getting pedantic about the 2 infinitesimal values that happen to cause a PWM square wave to become steady state DC, just so you can suck yourself off, isn't helping anyone

>>1844802
Got it, thanks. I'm going to mess around with the PMOS and gate voltages like you said. I forgot to take those into account.

>>1844804
Pic related is how this looks like in IC, you need high side driver to get it to work, Q2 is there instead of diode. You can also use low side driving but then output won't be reference to ground. For simulations you can just adjust gate voltage to be greater (or less) than fet source pin voltage so that you don't waste time drawing a driver circuit.

Should I get the weller wes51 or the ts100? I think the wes51 is a lot sexier. SO basically tell me why I should't get the ts100 pls

>>1844271
That's a fairly difficult question to answer. If you want basic, no frills quality, you can use a microcontroller with a built in ADC, DAC, and DMA. The higher the bit count, the better the quality. 8 bits is pretty bad. The dsPIC33FJ64GP802 is a 16 bit controller with all that included in a dip package. You can find example code on microchips website for real time audio pass through.
But that's extreme levels of custom. You could hack a switch into some toy karaoke amplifier

There's no tarduino general so I'll post here.

I don't have a clock in my room so I sleep in too much esp in summer when it's sunny at 5AM so it's hard to tell when I've overslept. So I'm looking to get a wall mounted screen that displays the image above (from wttr.in) and a simple clock. Might be one screen or maybe two seperate screens, two might look better and be simpler. I thought it'd be fun to do with a raspberry pi or something and DIY it, is this as simple as just buying the components for the RPI and plugging it togther?

>>1844292
Computer is a loose term. An arduino is technically a computer, but really small. Any time you are messing around with audio, a lot is going on. There are pure analog synths, but going full analog is the opposite direction of what you are doing.
Analog is constantly variable voltages using things like capacitors, transistors, relays, ect. Voltages are being constantly manipulated by active electronics.
Digital is manipulating two voltage levels, in the arduino case, 5 volts and 0 volts. The 1's (5 volts) and 0's (0 volts) are insanely quickly used together in the arduino to make different results. These results are stored in memory and shuffled around at specific timings to create a "digitized" audio wave.
Analog and digital are the two major directions in electronics. They overlap, but if you want to go pure analog, the arduino is not the way.
I'd recommend more research and studying the basics like individual components and what they do.

>>1844772
wow it's a non-inverting integrator so complicated.
Unless you're the engineer designing the system any higher understanding of the circuit is 100% unnecessary. At no point when troubleshooting will you ever need to figure out frequency and phase response of that network. The engineers have already done that and if all the components are good and the op-amp is good then it just fucking works.

>>1844893
why is full analog the opposite direction? The analog project is totally seperate from the 8bit thing, fyi. For the analog stuff, I'm following this https://www.youtube.com/playlist?list=PLyE56WXw0_5Q5QGMEXWmskuhojKyRdA3T..

I think I'd like my synths to be full analog, does the playlist above look right for me? I'd still like to learn about digital electronics and arduino's of course too. I have a few immediate projects I'd like to do using arduino's, actually (tons of uses in homebrewing, for example).

>>1844902
I meant that if you wanted to make a true analog synth, the arduino isn't going to help. "Digital" electronics and "analog" electronics are two different animals.
I'll check out the playlist later when I get home.

How inconvenient is it to tell an attiny board to run on 3.3V?
Do I have to add a single line of code or can you just tick box in atmel studio that tells the mcu to run on 3.3V or maybe even lower?

>>1844898
Yes that is all true but given >>1844732 we want to understand what's going on in that circuit. We want to know why and how it works, not how to find the bad opamp.

>>1844907
Got it, yeah. I think the synth series is full analog, which I like, but I'm sort of all over the map with wanting to get into arduino and the 8bit breadboard so I see why it'd be confusing. My brother, an EE-type was just getting rid of a bunch of shit the other day and said I can have what I want, so I grabbed a surplus of electronics stuff and I'm itching to get the ball rolling before my schedule picks back up.

Thanks for taking the time to check out the playlist, I did some preliminary research earlier this year and that was the best and cheapest tutorial I found.

>>1844889
or just get a clock?
this screen shit will just cost money and will be useless in the end anyway

>>1844934
I don't know, this sounds cooler. Plus I could display my daily agenda and project tracker (which I use emacs for) on there and stuff so it's always a glance away.

>>1844910
If the chip physically supports running at 3.3v, you just apply the voltage to it, no coding necessary for changing Vdd.

>>1844974
Personally i use smart band for shit like weather notifications and alarms, so i have it everywhere with me and on top it tracks my sleep

I can't find a proper datasheet with voltage ratings for Deans connectors.

They're made for low voltage, high current RC stuff.
Would they also be safe to use for 220V AC application?

There's at least 2mm of plastic between both plugs and 5mm of air between the pins.

>>1845248
Looks fine voltage and current wise, but not necessarily safety wise. Just use one of the IEC 60320 standards, like C7/8, or C5/6 if you want it to be polarised. C17/18 might be worth looking into if you want non-grounded but polarised, for some reason. Personally I'd just go for based C14/15 on everything.

>>1845250
>Looks fine voltage and current wise, but not necessarily safety wise
What do you mean? I don't need to plug it in and out while the whole thing is powered, if that's your safety concern.

It's for an optional external switch on a power tool.
Which is also why I don't want it to be very big. C7 would be okay size-wise but it's current rating is too low.

>>1845251
>I don't need to plug it in and out while the whole thing is powered, if that's your safety concern
It isn't plugging in and out, but rather fingers contacting metal when placed near the live female connector. Proper mains female connectors have deliberately recessed contacts for this reason, but the Deans T connectors I've seen look to be only 1mm recessed or less. Also the male's prongs don't have insulation halfway down them, so if the connector isn't pushed in properly, a piece of wire or a coin or whatever could fall into the gap and short the connectors.
I'd safely push my tongue up against a 240V mains socket on my wall, but certainly not one of those sockets.

>C7 would be ok size wise
C9 or C11, if you can find it, isn't much larger and can handle more current. Why not use a relay to switch the power tool, fed by a low-current switch?

>>1845252
Ah thanks, now I get it.

I can't fit a relay inside the tool and don't want it dangling outside.
What if I used a shrouded connector like an XT60 instead and 3d print a cover for the exposed contacts when not in use?

>>1844925
>Thanks for taking the time to check out the playlist
I briefly jumped through the videos and they seem pretty good to run with. I'd strongly recommend spending a few days getting used to the different basic components you'll be using to build the synths (resistors, capacitors, transistors, ect...). He does a decent job explaining a bit about the parts, but leaves out some information that could be useful.
All in all, go for it, but for the love of Ohm, use some better wire management than him, your finished synth shouldn't look like spaghetti.

If you were an electronic part, what would you rather be? I feel like a capacitor. I think I am rarely charged to 100% and I feel drained most of the time.

>>1845281
Fuse, because people only notice me when they need me.

>>1845281
an op-amp
I'm constantly on edge and even the slightest deviations from the norm cause me to spiral into an anxiety attack

mira, es la creatura....
I finally got it working though I don't know how retarded the design is.
Time to work on the algorithm to control the PWM signal

>>1845299
one thing that concerns me about my circuit is that, if the duty cycle is dropped fast enough, there's a large negative voltage overshoot and I don't know why.
That's why I added the diode clipper, but sometimes it seems to conduct for far too long after it's triggered.

Considering the time scale is 75 nanoseconds, is this just standard inductor-capacitor ringing?

>>1845308
>is this just standard inductor-capacitor ringing?
Yes, if you short inductor to ground shortly after it was already discharged by capacitor (in off cycle), only thing that can happen is for inductor to flip voltage and transfer energy from capacitor, which then transfers back to capacitor. Only thing that is strange here is that period of oscillation (from valley to next valley) should be around 628 us (2*pi*sqrt(LC)). I'm not familiar with your circuit simulator so time scales are a bit confusing to me. If I simulate in ltspice I get close to calculated value. You should also consider working with realistic models, inductor will have some series resistance which will have a great effect on transient response (0.5 Ohm for example could eliminate negative voltage).

>>1845308
Yes, it's LC tank ringing, but try adding series resistance after inductor, it can prevent voltage from going negative. You should also consider how diode is modelled (ideal/not ideal). For example ltspice analysis confirms LC resonant frequency after abrupt change in duty cycle (pic related, calculated T=628 us) .

Friendly reminder from Major Armstrong that glass flows over time so make sure to rototwnikulate your vacuum tubes every spring so they remain aesthetically pleasing and don't lose vacuum. Retard over and out.

>>1845406
>glass flows over time
no it doesn't

>>1845407
Hey I plan on being alive in 3020, I'm only 200 years young!

>>1845294
Sounds like you need some negative feedback.

So far so good,ish. I have the air pump connected to the PWM, and the air flow can be changed. Before I go forward though, I did notice something. The circled areas, the switch on the board and the output of the battery got warm after a while of use. I need this thing to work continuously for 8+ hours. Should I be worried parts are getting warm?

>>1845516
derp

>>1845516
>>1845518
How 'warm' are we talking, and how much time is "after a while"?
They are heating up due to the amount of current passing through (amps).
It all depends on how many amps the parts are rated to handle vs how many amps are actually flowing through them.

>>1845580
It was about 20 min, it warm as in noticeably warm, like not burning, but smart phone warm after watching a full length movie.

>>1845516
>got warm after a while of use.

that's one of the reasons i suggested you cut cables to avoid connectors: every connector adds a bit of contact resistance which causes power loss thru heat.

the switch is obviously cheap, so it can be replaced or eliminated. to bypass it, just run a very short fat wire from red to red and another from black to black. this then makes the battery connector the power switch. so, to prevent it creating lil sparks when you plug it in or out, make sure motor speed is a zero before plugging/unplugging.

as for the battery wire, cant tell for sure, but the wires look to be very thin. you'd want 20awg wire there if possible, which can be rescued from some power bricks. i have a bunch of d-link 5V power supplies for routers which have that kind of thick firm meaty sexy wire.

>>1845516

just looked closely at battery wires and they are, in fact, pathetically thin in comparison to other wires.

>>1845647
>cut the switch
I don't need the switch on the board, it's redundant.
>get better cable
>>1845649
OK, then I'll cut the cable the battery came with.
Thanks, I'll do these tomorrow.

>>1844791
dc pulse acts like ac when its not giving positive voltage. proof of this is that dc pulse can cross a cw (cockcroft walton) but just normal dc cant. when its not pulsing in the circuit all the current that passed the source terminals decides to go back the way it came

i think its a simple switch mode power suply he made but it depends on the inductor value really. im not a expert but thats what it looks like he made. the fol;lowing isnt the same but its similar in how it would operate

https://www.youtube.com/watch?v=ldR8Ey-KnZw&list=PLVCnWm0AqwwQSWbhoO859qigDgv5GAVRa

>>1845651
>I don't need the switch on the board

just realized the switch is forward-reverse, not a simple on-off. so my advice to short the red-red, and blk-blk may be wrong. another poss. is you gotta short pin1 to pin3 and 2 to 4. you'll have to check with ohm-meter which pairs of wires are shorted when switch = forward.

>>1841869
I was cruising down the freeway today and some asshole was riding my tail. I really wanted to brake-check him but he was so close, I was worried he would rear-end me. That got me thinking: how about a brake-check with no actual braking?
Essentially I am looking for a way to make my brake lights come on without hitting the brakes. I imagine this should be fairly easy to do. A pushbutton switch somewhere, wired from the switch to the lights and the battery.
Does anyone have any ideas to help point me in the right direction? I was thinking of putting the switch somewhere on my steering column shroud so I don't accidentally hit it.

>>1845728
Most cars brake lights activate before the brakes actually engage. Besides, when someone does that, just letting your foot off the gas until they go around is usually a good way to handle it.

>>1845702
>keep in mind this setup only works with dc current
pretty sure it works on a brushed universal motor even when running on ac, and it would work on a single one of the coils of a 2-phase induction motor too. be a trivial jump to use it to swap two coils on a 3-phase induction motor also

Hi all, trying to power this microphone, an iSK CM-20Q, electret condenser (See: https://www.swamp.net.au/isk-omni-360-condensor-microphone-vocals)) through a phantom power circuit.
The specs ask for 9-52v Phantom Power (through balanced XLR cable), and I have thrown this (https://www.shure.com/en-GB/support/find-an-answer/how-do-i-build-a-phantom-power-supply)) together on a breadboard, but am getting a signal identical with and without power applied.
If I scream into the mic with the gain on max, I get a barely audible signal.
I understand fucking nothing about mics, plz halp

>>1845816
Did you measure voltage when you connect mic between pins 2 and 1? Your mic could draw more current than your designed power supply, and cause voltage on mic to drop below 9 V. If mic is drawing 5 mA, there will be 7.65 V drop across R3, R4 and R5, which when you have 12 V on the input gives the mic only 4.35 V. I would start with something simpler like pic related, your schematic includes low pass filters to make power supply more clean, but you should first start with a basic circuit without any filtering and add some if there is a noise problem.

>>1845825
Forgot to mention to replace 6k8 resistors with appropriate values for your input voltage (6.81 kΩ for 48 V, 1.2 kΩ for 24 V, and 680 Ω for 12 V).

>>1845816

this is what i did to power my mic and to make sure that nasty 48V stays away from my equipment.

>>1845831
>>1845827
>>1845825
Thanks for the support lads, turns out, as these things always seem to go, I just had the input and the output wires in backwards. Oopsie woopsie.

Not quite sure why the schematic I used looks so needlessly complicated, wish I'd ask you guys sooner seems much cleaner

>>1845862
>the schematic I used looks so needlessly complicated

monkey see, monkey do.
if you look at cartoons a lot, you end up drawing cartoons. if you look at Picassos, you end up drawing Picassos. if you do both, you get this.

>>1845741
If run a switch parallel to what's already there, will the lights come on when I push the switch?

So JFETs are basically a semiconductor version of the vacuum tubes? Same characteristic curves, same biasing techniques, same voltage controlled voltage/current source? Also, both have some kind of internal current limiting properties so technically you don't have to have a plate/drain load resistor and you won't fry them?

>>1845903
N-type FETs of any kind are pretty similar to a vacuum tube.
That was actually a selling point when transistors were just starting to replace tubes.

>>1845884
>will the lights come on when I push the switch?

obviously yes.
just as obvious is that this whole plan is gonna create a distracted driving situation thats gonna see you wrapped around a light pole. so, by all means, go for it.

>>1845884
just to be sure, i'd attach a mirror next to the brake lights and then another mirror that would reflect back to your rear view mirror. can't go wrong with such a debugging system of mirrors. you can tune your switch that way.

What is this?
It's from an optical barcode scanner and it has its own little window on the scanner's plastic shell to look outside the device. my theory is it is an ambient brightness sensor to help calibrate the barcode scanner but I need a second opinion.
The black box on this pcb is the size of the * symbol on your keyboard.

>>1846249

a brightness sensor wouldnt need like 20 wires. plus the label, U1, refers to a chip.

>>1846249
That's probably the actual CCD that reads the barcodes. Might be a CCD+image DSP all in one, like you get in optical computer mice.

what the fuck is stopping me from making an inverter by using a microcontroller to generate a sinewave then hooking it up to the gate of a power FET

>>1846327
A: mosfets aren't very linear devices so your sine wave would be hugely distorted without negative feedback.
B: the voltage dropped by the mosfet would be on the same order of magnitude as your output voltage, and with the same current. acting as a linear regulator your mosfet would overheat the moment you tried to power any modern appliance with it.
C: you need to start with HVDC in the first place, which many of the aforementioned modern appliances can run off without needing an inverter. making a HVDC boost converter is half the problem.

but if by "generate a sinewave" you meant "generate a sinusoidally varying duty-cycle pwm signal" alongside some big LC filters, then yes it would work to some extent provided you take care of C, but you'd want a high voltage half-bridge as opposed to a single mosfet, and you'd need a proper half-bridge driving circuit to go with it. while you're at it you might as well add feedback to keep the output sinusoidal in the event that a non-ohmic load starts pulling current. once you've done all that, congrats you've just made a HVDC class-D amplifier!

Looking for instrumentation amplifiers in handy packages and I found this:
>https://www.ti.com/lit/ds/symlink/ina122.pdf
But looking at the circuit diagram there's only two internal op-amps, which isn't a topology I've seen before. I want to amplify a signal relative to a reference voltage that I'll set with a trimpot, but the level that the output voltage is reference to isn't very critical since I'll just calibrate it with an initial ADC reading. I intend on running a relatively high gain, and it's on a 5V single supply system.
Thoughts?

Will it catch fire?

>>1846349
not with that attitude

>>1846345
Actually, then again, I can just buffer the reference trimmer with one op-amp, use a second op-amp as a non-inverting amplifier for the signal I care about and tie its reference/ground to the buffer's output, and set everything with just two resistors. Then I can calibrate the ADC with that buffered reference. So I'll do that with my existing dual op-amp ICs instead of buying new ICs. Might add some more resistors for bias current reasons, something which is new to me.

>>1846312
>That's probably the actual CCD that reads the barcodes
can't be. the green circles are where this mystery sensor looks outside, the red arrows point where the barcode scanner is looking.

>>1846345
this is a valid ina configuration. if i remember this version just has bad CMRR so if that's a problem you'll need to design your circuit to mitigate it.

>>1846312
>>1846374
Barcode scanners haven't used CCDs for about 15 years. They tend to use a row of phototransistors or a single phototransistor with a laser that sweeps back and forth with a galvo.

>>1846529
oh, neat

This will become the front light of my bike. Total of five 5 mm white LEDs and a 150R resistor network.

>>1842859
The additional step of the air just wastes your energy. Either make it compressed air only or electromagnetic only. I believe compressed air has higher energy density. Also please remember that EM coilguns by most hobbyists have poor efficiency since the coils aren't able to be timed accurately, and with things like ringing not corrected for entirely.

I.E just fucking use a barrel of air

>>1846582
For the streets? Perfectly fine if it's just for visibility. If you want to see the road by the light though, you'll want at least 10W of LED, probably closer to 30W. The ~3W LEDs are kind of a waste of time since they're not really bright enough to see the road by, and they're pinpoint enough to be potentially dazzling to drivers. I wonder if EL panels will take off for maximising illuminated surface areas any time soon. Even for automotive brake lights I'd think they'd be nice to look at than a handful of point-LEDs shining through some highly refractive plastic.

>>1846595
>you'll want at least 10W of LED, probably closer to 30W.

good advice if you're pedaling in the middle of the Black Forest at 4am.

>>1846374
Could be for IR communication when placed into a holder. I've dealt with handheld stuff before that uses an IR port to send/receive data from its main station. No plug/jack to get messed up or not properly lined up when put into its holder.

>>1841869
Need a 1.6Mhz oscillator
How do frequency scale a 16Mhz crystal?

>>1846687
frequency division is done by powers of two. if you insist on dividing by ten you'll either need a discrete counter and some logic gates to mix some of the counter taps, or you'll need a uc. at that frequency probably a uc with an nco peripheral too. i suggest not doing this.

>>1846595
Purely for visibility. I also built a similar rear lamp with red 3 mm & 5 mm LEDs.

>>1846749
I am adapting this design but using a 4017 which is a binary counter?

>>1846769
To detail, the 4*3 mm high-brightness red LEDs shine light on the left & right microprisms to spread the light to the sides, and 2*5 mm high-brightness red LEDs in the middle. Seems to work fine and is quite noticeable even during daylight.

>>1846781

4017 cant handle 16mhz. theres ttl chips that can tho: from memory, it seems the 7490 might do the job.

>>1846791
How do you tune frequency 1,7 Mhz

>>1846749
>you'll either need a discrete counter and some logic gates

Simplest solution is to build an AND gate with diodes. Connect diodes to every bit of the binary counter you want to be '1' when it resets and run that to the counter reset pin. You can clock divide by any integer you want that way.

>>1846814
Use a decade counter IC or dedicated programmable frequency divider IC, which might serve you a bit better at these higher frequencies.

I got a 100W laser, dont have any data on it.
Theres 3 diodes paralleled and a temp sensor, thats all.

Would it be feasible/possible to diy a drive circuit?
Other than losing an eye what should i be looking out for?

>>1841869
Planing on designing an animatronic face, I see tons of people using blue chinesium servos that sound like squealing hamsters.

Is there a small servo that isn't loud?

>>1846814
>>1846687
How accurate it needs to be?
If you cant get anything divided down to that frequency (+-tolarance) with logic, use a pll.

>>1842827
Good now import into TopoR. Your board doesnt look drunk enough yet.

>>1844644
I would like to add to your excelent explaination that there is more to the story. in most inverters there is a trade off between switching speeds and EMC. There are some 30nS 5-10A drivers available that will switch mighty fast but will create a great deal of EMC in the process.

Here's some aditional information for the anons that need to drive some FET on the cheap https://www.eetimes.com/power-tip-42-part-1-discrete-devices-a-good-alternative-to-integrated-mosfet-drivers/#

A TC4420 will also work but is more expensive than those jellybean transistors.

>>1846997
Yeah you can DIY a drive circuit, so long as you know the current to shoot for. Laser diodes need a strict constant-current source, and it would be a good idea to integrate that temperature sensor into the driver too. But if you're running it at 5% power or lower it's probably fine to just use a power resistor in series and ignore temperature.

You'll want it to be a switching driver for the 100W though, and high-power SMPSs aren't that diy friendly, especially in the department of inductors/transformers.

>>1847093
>TopoR
I do like the look of those curvy traces, but I'll be sticking to KiCAD.

>>1847103
access denied to that webpage

>>1847159
I think you can import kicad boards into later versions of topor. Use the autorouter and export back to kicad again. But im not 100% i know topor eats Eagle .BRD files ant Altrium designs at least

There is a similarly named video on youtube. For the discrete mosfet driver

>>1847192
>discrete mosfet driver
is it just a totem pole?

>>1847193
Yeah it is. However there are some really wonderful zetex high gain low VCEsat transistors if you want a really beefy discrete solution

>>1846814
>How do you tune frequency 1,7 Mhz

you just changed from 1.6 to 1.7. in case thats not a typo, and the frequency is not super important, then you can use a very popular crystal called the 3.579545 MHz Color Burst crystal. you just add a single flip-flop divide-by-2 to get 1.7897725Mhz. CMOS chips can easily handle 3.58Mhz, unlike 16Mhz.

>>1847199
I've messed about with maximising the drive current through the BJTs, but I find they get pretty hot and using darlingtons for improved gain doesn't really improve speed. I also have the issue that my signal comes from an open-collector comparator output; I need to use a pullup somewhere but that's a compromise between slew-rate and output impedance.
Those beefy zetexs sound good, but MOSFET drivers like the IR2183 have all the baggage needed for high-side driving, along with deadtime/anti-shoot-through if I need a half-bridge (also ~2A output). Also floating high-side for HV driving. Half-bridges are really versatile.

>>1846902

this is supposed to divide freq by 10 without ext components. havent tested or simulated.

>>1846902

this is supposed to divide freq by 10 without ext components. havent tested or simulated, but the datasheet confirms this in a sentence near the top.

>>1847216

This is the kind of transistor i'm referring to: https://www.diodes.com/assets/Datasheets/FZT1049A.pdf

These have the gain of a small signal transistor and the current capability of a beefy device. I lucked into two reels of those some time ago and have been using them primarily as relay drivers.

>>1847302
Oh yeah those are nice. But I would want complimentaries.

>>1847304
There are others that are also high gain low VCEsat that have PNP complements.

>>1847093
that routing looks terrible, as if a child with a sharpie painted it directly on the copper

while nice straight lines with 45 degree turns look sexy as hell

What is a good quality style of crimp and crimp tool for automotive electrical accessories?

>>1847325
REAL TRACES HAVE CURVES

does anyone know of any resources for inverter design?

>>1841869
How do you increase the precision of a 555 timer.

>>1847648
use a non-shit IC
a good old 32768 crystal and a clock divider is an acceptable way to go
or just a faster RC oscillator with a divider, in case your lack of precision is due to using a large and leaky capacitor for a high time constant
or an mcu, if you have need for its functions

>>1847774
>>1847774
>>1847774
>>1847774
>>1847774
>>1847774

>>1841869
Hi, I have a Zyxel GS-105B v2 that uses 9V DC 0.6A power.

I have a Huawei HW-120200E6W thats 12V 2.0A.

Will that power adapter break the switch or will it work just fine?

>>1848026

impossible to say. depends entirely on the robustness of the voltage regulator, and the design of the heat sink.

i wouldn't chance it. the extra heat will likely kill the unit in the long term. also, a capacitor may blow up.

if you have skills, cut one wire from 12V cable and put 4 diodes in series. that'll drop the voltage by 3V.

>>1842938
Basic calculus is either Thomas or Stewart's calculus. For engineering, Fourier or Laplace isn't too bad since you aren't a mathematician, so you can just read online guide and then use the tables and identities (or wolfram). If you want more then you will need a bit of complex analysis which you can omit if you want to only apply and not give a fuck.

Also the online Lamar or whatever it's called calculus stuff is pretty good. Usually first result in search.