/diy/ - Do It Yourself

this thread's digits brought to you in part by the CSD17579Q3A itty bitty power MOSFET

Is it just me or is there like no electronic jobs at all these days? I'm almost done with my master's in microelectronics and I live in a medium sized city (metro area population almost half a million), yet I can't find even a single listing for electronic engineer work right now. Is that normal?

>>1377588

dont supposed you've notice that everything electronic has a ''made in china'' sticker on it nowanights. time to brush up on your Mandarin.

>>1377588
My buddy gets to freelance as an ECE tutor.
He teaches circuits/helps with homeworks and all that shit to our hard-learner neighbors $28 per hour

>>1377588
90% of the jobs aren't advertised. Talk to people and network.

>>1377579

I'm looking to make an adjustable high wattage low voltage power supply. I know nothing about actually building the board.

I need an output of 1-6v capable of a load of up to 250w in worst case. Total load resistance will vary from .15 to 1ohm.
It can be a dirty solution, like PWM if needed.
It will be powered by 120v ac. Can you give me any premade ideas? I thought of a pwm motor controller with a standard 12v power supply, 300w. Would this work?

>>1377778
At 6V and 0.15ohm you will only get 5.4W.

Why only such a low voltage? You will be looking at crazy high currents

>>1377779
1..6 volt 0.15 ohm 6.7..40 ampere 6.7..240 watt

>>1377783
True. But still why the low voltage? The high current will be hard to control/regulate

>>1377789
because super slow motion super heavy duty linear actuator

>>1377804
if it's super slow motion, why not use a geared motor?

>>1377778
obtw, E=IR, and 14AWG Cu wire resistance is 2.5 milliohms per foot

>>1377779
>>1377783
>>1377789
>>1377804
>>1377828
All wrong. What it drives is irrelevant, the load is .15 to 1ohm and requires adjustment of the input voltage from 1-6v.

I'm looking into using a standard 360w 12v PSU and a 1.2-(12max)v buck converter, 300w rated.

Is this the best way? Considering the input signal doesn't matter to the load at all, would PWM work just as well for cheaper?

>>1377852
What? I've been using an ohms law calc because I'm retarded, phyiscally and metaphysically. Good info on the wire, as I've got a 2 ft run to the load. Will need to keep that in mind.

>>1377588
I feel ya. My degree is in EE/CE but I got a programming job because it paid more. I just do electronics as a hobby now.

>>1377860
>Considering the input signal doesn't matter to the load at all, would PWM work just as well for cheaper?

If it's not a capacitive load, yes, PWM will be a more efficient (and, at that kind of load, cheaper) method of adjustment.

>>1377921
Inductive.
So, can you reccomend me either a search term or generic board with which I can control the output with? Remotely wired potentiometer preferred, but I guess I can learn to solder.

>>1377945
ah, PWM will be fine then, use the inductive load as in a buck converter i.e. add a freewheel diode
also endeavor to keep your high current runs as short and thicc as possible

>>1377945

Your problem is going to be that you need a whopping 40A at 6V. That's an unusual combination for a power supply. Off the top of my head, none of the cheap Chinese modules would work, meaning your SOL if you don't want to shell out the cash for a "real" supply.

I'm assuming you're trying to drive a motor with this. Motor controllers usually also have built-in current limiting to protect both itself and the motor.

Honestly, it sounds like the terms you really want to search for are "brushed ESC". RC speed controllers are about the only off-the-shelf solution for what you seem to be looking for.

So do I just solder the resistor to the positive on the LED and then attach the power wire? Standard USB charger 5V 1000mA.

>>1377970
should work, IF the output of the charger is voltage-regulated. some are shit and therefore are not

>>1377978
I have a few different chargers I can try if it doesn't work. I just want a small blue nightlight for one of my aquariums because the fish panic every time the light is turned on or off. Sucks because it is a breeding pair and if they do breed and then panic, they will eat their young.

>>1377986
i dig it
if it's not regulated it could burn your resistor or LED. I'd stick a meter on the charger output first to make sure it's within 5% either side of 5.00V, and recalculate the dropper resistor for the actual output voltage if not

>>1377993
The parts only cost me 2 bucks but it won't hurt to test. Now only if I could find my meter...

>>1377993
I found my meter but it is stuck on 1666 for some reason... not sure why.

>>1378020
Nevermind the hold button was pushed in for whatever reason.
Voltage on that charger reads 5.00. So I guess it is OK to use.

>>1378025
>satan trips on your meter
nice
yep should be fine for this

did you buy too big a bag of those XL6009 modules and can't find anything you want to boost with them? how about LEDs?

>>1378029
Pic related: rebuilding a magnifier desk lamp with half-decent LEDs and 12Vdc power supply instead of crap lowe's bulbs that don't last even their pussy-rated 15000 hours

Probably looks like shit

>>1378046
But it works

>>1378047
Might need to get another one. This LED was only 1 watt.

>>1378035
What type of LEDs are those?

>>1378049
>moonlight is for lovers
just for fun, measure the voltage across the resistor and the - terminal of the LED. the cable could be dropping voltage if it's very long

>>1378056
1W white LEDs off the chink boat, 50/$3 or something like that, color temperature looks like about 6000K

>>1378057
If I measure from the end of the resistor where the wire is connected to the negative terminal, the voltage reads 4.45. Is that what you meant?

>>1378057
>1W white LEDs off the chink boat, 50/$3 or something like that, color temperature looks like about 6000K
Wow that is a really good deal. I've been considering building some type of light with those cheap-o LEDs and then measuring the PAR to see if it stacks up to the big boy LEDs that cost possibly hundreds of dollars.

>>1378060
the cable appears to be dropping half a volt. you're not getting the full watt, maybe 1/4W or so. if you have an amp meter with a range exceeding 0.5A, you can check and see for sure
I advise you shorten your USB-power cable as much as you possibly can, or splice it at the plug end and use heavier gauge power cable from there to the LED/resistor

>>1378067
https://www.aliexpress.com/item/sharks-with-frickin-lasers/32698195986.html
https://www.aliexpress.com/item/obtw-stars-cost-extra/32596642167.html
I can't speak to CRI or any other spectral qualities, but if you're using them to power some plants you'll probably want mostly blues and reds anyway

>>1378068
The wire is quite thin. I probably have an old printer USB that has thicker wire around here somewhere.
This is the multimeter I have. I've never measured amperage with it so I'm not really sure how to do that.

>>1378070
Wow that is very cheap. Almost worth buying just to test them out with my PAR meter. How long did it take for your order to arrive?

>>1378071
measuring current:
>place probes in COM and 10ADC jacks
>turn selector to 10A
>disconnect red wire from resistor
>attach red probe to red wire
>attach black probe to where red wire was on resistor
>power up
>read
chargers and power managers don't mind the voltage drop but a simple unregulated current dropper like this would. the cable might say how thick the internal conductors are in AWG

>>1378075
that seller shipped within a couple of days. the package arrived in about 22 days during USA tax season.
>helpful tip when dealing with power bead LEDs: don't try to solder the rear pad, put some of that STARS-922 "heat sink plaster" on them instead

>>1378079
Damn, it's only reading .31! Since this charger is rated for 1000mA, I would assume it should read much closer to 1.0. So these wires are way too small. Shit. There is imprints on the wire but nothing that seems like the size of it. "NON-PVC" and "POWER" is pretty much all I can make out.

>>1378079
I found a cable that has 28AWG on the outside of it. I cut it open and the wires look much bigger than the one I had used.

>>1378082
it's a 1W LED, right? 0.31 * 4.3 is a little over 1W but still probably within spec (check datasheet to be sure). if it's a 3W LED, then you'll definitely need some thicker wire

>>1378092
Is the 4.3 from the resistor? What I did was go to this electronics parts store and bought the LED there. The owner asked me what I was using to power the LED and I told him a standard USB charger. He picked out the resistor and told me that it was close enough. But yes, the light is 1W. Do you think I should try the thicker wire anyways or just go with what I have on there already?

>>1378094
4.3 is actually a misestimate on my part. it's the forward voltage of the LED. in fact it's more like 3.8V for blue, in which case 0.31 * 3.8 is 1.17W, which is safe and reasonable. the LED is actually rated for current, you can give "1W"-class LEDs up to 350mA before exceeding their do-not-exceed rating
anyway, the setup you've got seems to be delivering, but if you were to set up another LED you would probably want to go with the thicker wire

>>1378101
You are right about the LED. I found it on their website (www.surplusgizmos.com). I just found this place recently and it was my first time in their shop. Really want to go back and spend like an hour or more in their store, they had so much cool stuff. Way better than what Radioshack used to be like. I do want to add another blue LED so I think I will just use the thicker wire anyways. Gives me a chance to get some much needed soldering practice.

>>1378101
If I add one more of the same LED, will I need to change the resistor? Is there a website that you know of that has basic info on wiring LEDs?

Ive got a weird project in mind. How high of a pressure can I get from a super tiny air pump, and how small can I get? I've got this 2 1/2" x 1 1/8" or so pump, but it's a bit larger than I'd like. I've tried searching a bit online, but it's one of those things where you can't really just browse to see what's out there, you have to already know where to get it.

>>1378117
The one in your pic is rated 90kpa I think. Just google miniature air pump.

>>1378110
woot. that shop seems like good people
don't forget about heat dissipation. the star backplate isn't necessarily enough to dissipate 1W without being firmly mounted to something
you'll probably want to use a separate resistor for each string of LEDs. if you change the supply voltage or the Vf of the LEDs in series, you'll need to change the resistor(s). or you can use a constant-current driving board.
I don't know of a good how-to-LED site unfortunately. if you just read anything you can get on the subject and then try to synthesize what they all agree on, you'll know the stuff pretty well

>>1377948
>>1377951
In that case he'll need a whopping output capacitor, particularly one with low series resistance (so not a normal supercap). V_ripple = duty_cycle*I/(f*C) more or less, but the main problem here will be ensuring you don't pull 40A out of the 12V PSU for long enough such that the

>>1378128
caps discharge or anything else bad happens. Pick a PWM frequency high enough that the voltage doesn't droop much, but not so high that you dissipate too much power in your switching FETs.

>>1378128
but the load itself is the inductor, which implies only average current regulation is needed. no output → no cap :D

>>1378117
I don't know what's the absolute minimum, but I have Koge Electronics' KPV08A-3A. It's around 40 x 18 x 8mm. That particular pump is for vacuum, but Koge has other similarly sized models.
Too bad they tend to have very limited total running time, measured in hours. That's mitigated by the fact that they're meant to be run only some seconds at a time, but still.

>>1378070
>$46 shipping for a $4 item
Fuck you Postnord

>>1378136
Well worst case scenario, the output caps discharge to close to 0V before the PWM pulse is up. The amount of charge you can pull out of the PSU is proportional to the capacitance. If I*duty_cycle/f << C (with some give for the flyback converter's own frequency) then there's no problem.

>>1378150
but if the load itself is the inductor that anon wants to power, where would the output cap go?

>>1378070
Went ahead and ordered some of the royal blue LEDs and stars. I have resistors. Found an easy LED calculator (http://led.linear1.org/led.wiz).). Can't wait for them to get here. Getting them from the US so it shouldn't take too long. I will have enough to set up on all of my aquariums.

>>1378154
bretty gud, anon. don't forget power supplies!

>>1378157
I have so many aquariums that I would need to get the power supplies for, I figured I would just use more USB chargers.
Gonna need to figure out the heatsinks too. I have a bunch of old CPU ones but it doesn't seem very practical.

>>1378159
or just use drivers
https://www.aliexpress.com/item/high-voltage-nipple-toys-kawaii/32756673848.html
the nice thing about constant current drivers is that you can string multiple LEDs in series, and it doesn't care how many you put in the string or how chintzy your wire is

>>1378161
Nice, I might have to do that because I was just thinking how shitty it would look to have 4 USB chargers all plugged into the same thing with wires going everywhere. My biggest cluster of aquariums has 4 right next to each other and I will need 7-10 of the LEDs to get the coverage I am thinking of.
So, do I add up the required voltage of all of the LEDs and then pick a driver based on that or how do I pick the right LED driver?

>>1378161
Wait, what if I wanted the kawaii nipple toy instead of the LED driver? lmao

>>1378163
>So, do I add up the required voltage of all of the LEDs and then pick a driver based on that
you got it. leave a volt or two of headroom if you're near the max.for "1W" LEDs you want the 300mA version

>>1378164
surely ali's got you covered. happy hunting :===DD

>>1378166
Easy enough. Is there a certain gauge of wire that you would recommend for this?

>>1378167
22AWG drops ~1V/170ft at 350mA, which you would add to your LED voltages according to the length of your loop. I'd try to keep a 22AWG loop under 60ft just to keep losses low
20AWG would be slightly more efficient, but not really enough to worry about. brightness is the same (as a function of current) in either case
>heatsinking
>>1378079 is the real shit. you can stick your stars to just about anything with this. you'll also want this stuff to build your stars. otherwise you would probably want to use screws to mount them to something
unless you stack like 10 of them in close quarters, the surface you mount them to suffices to keep them in safe temperature ranges. get some aluminum trim channel or angle at your local home improvement emporium, if you don't have some strips or sheets of scrap metal laying around

>>1378177
Alright, 22AWG it is. Not going very far at all, I estimate the entire loop will be less than 20ft. Good to know I can just mount it to whatever aluminum I find. I will have to check out the local stores and see what I can get that will fit my needs. I just ordered some of that plaster. So cheap from China. Are you going to use lenses on the LEDs in the light you are upgrading or just stick with the stock lens? I'm considering some lenses.

>>1378153
Before the PWM "switch", your simplification of the PSU as a battery is your model's shortcoming. Because as far as I know he's settled on a 360W 12V PSU with a PWM circuit afterwards.

Hello guys,

a few weeks back i wanted to build a "big" 7-segment-display. Just made my first segment prototype and then aborted that project, because i am a lazy fuck.
Now i think ill gonna take another try at it and double the amount of leds, because muh brightness.

Is my calculation correct regarding current drawn?
If i take those 120mA what that led array calculator gave me, multiply it by 7 (segments) and 4 (digits) this would give me 3.3A max. Is this correct? That would be nice, because i still have a 12v5A power supply lying around.

pic related

I don't suppose you can get breadboards with 3 bus-strips on each end, can you? I always have to have 2 or 3 5-socket rails for a ground voltage if I use the + and - for my ±15V supply for using opamps.

>>1378182
just six naked beads shining down on my workbench. it's done, wired into my bench 12Vdc bus, and it works wonderfully
you don't even need much, one 5g tube should do the slugs on the back of the LED to the star plate and tack the finished stars to your base material 100 times over. just drop a little dollop on the back and smear it around to cover the mating surface, and done

>>1378187
sounds good, maybe you should multiplex 1W LEDs for serious brightness, since that's what the rest of us cool kids are using :^)

>>1378188
you used to could get breadboards with separate strip-areas and bus-areas, and just snap them together in the combination that tickled your fancy. I have a breadboard here in the normal bus-strip-bus configuration. I haven't looked lately but maybe you still can

>>1378187
Yes those current calculations are correct, though you could under-drive the LEDs if you were worried. Something that bugs me is where the wizard got the values for the forward voltage of the LEDs. You're dropping 0.94V each across those resistors, meaning each LED is dropping 2.765V, which sounds like it could be a little low. If you measured that value yourself or got it off the datasheet then no worries.

>>1378150
>then there's no problem.

There are already large filtering capacitors on the output of a power supply for, more or less, that very reason. That doesn't help with the fact that 6V (whether DC or an average derived via PWM) will push out 40A against a 150mΩ load. Ripple is not really a concern here, the actual average current is.

However, again, if this is a motor we're talking about, you don't want to actually be drawing that much current in the first place. He's going to need to limit current somehow to prevent damage to the motor. It'd help if we knew exactly what he was trying to control and why, but he hasn't said. So...

>>1378194
>meaning each LED is dropping 2.765V, which sounds like it could be a little low. If you measured that value yourself or got it off the datasheet then no worries.
entered 2.8V myself. Just looked around and apperently 3.0-3.2V is more realistic for even brighter LEDs (3528 or 5050). Will looks a bit around, order a small batch and then test it i think

>>1378195
Well it's only 40A peak (provided the PWM is tuned really low with the freewheel diode in mind) with a very low duty cycle. Is it more likely that this peak current damages the PSU or the solenoid, provided the PWM device is overbuilt?

>>1378194
>ach LED is dropping 2.765V, which sounds like it could be a little low

That's a little high, if anything. The only LEDs I would expect to have a Vf above 2.4V would be blue, UV, or white.

Greens are usually ~2V, maybe a little less.

>>1378199
Well they could be green by phosphor mixing, not by band-gap. You know, for higher luminous efficacy.

>>1378197
Measure the voltage across one with a trimpot adjusted to give 20mA.

>>1378199
Most of the green ones on ebay say 2.8-2.9V (3528 housing).

Does the mcd value actually determine the brightness? Apperently the green ones seem to have a lot more mcd than red or blue. White has the most, but not sure if they are any good for outside

>>1378198

Depends. If the output filter caps on the PSU are poorly-specced and the PWM frequency is low, they may overheat, either shortening their lifespan or, in extreme cases, failing completely. Additionally, remember that that 40A won't be at a "very low" duty cycle on a power supply that only outputs 12V to begin with. Meaning, in ideal circumstances, you'll need the PSU to be rated for at least 20A to keep up.

But, fact remains, without further info on the part in question, 40A for any significant length of time on a device built for 6V can be assumed to be a bad idea. If the device is inductive, it wouldn't be much of a peak current, either; if the PWM frequency is anywhere near adequate, the peak current won't be much higher than the average.

>>1378138
Shit, they're rated that short of a lifespan? That may not work then. Are we talking at least tens of hours, or flat out hours? And thanks though, that at least gives me another brand to search.

>>1378201
A very interesting topic:
https://en.wikipedia.org/wiki/Luminosity_function
Have fun converting your luminous intensity to luminous flux to radiant flux, then dividing that by the electrical power to get a luminous efficiency, then convert the luminous watts back into luminous intensity again because that's what your eyes will see anyway, and divide by electrical power to get luminous efficacy. Oh yeah, have fun finding a CRI curve on your 1-page chinese LED datasheets, converting that into data points, and "integrating" over the thing to get your luminous efficiency. For bonus points, the radiant IR flux as the LED temperature stabilises can also be incorporated into its luminous efficiency.

>>1378200
>Well they could be green by phosphor mixing, not by band-gap. You know, for higher luminous efficacy

I guess, but all the shitty cheap ones I've gotten so far have been <2V. Granted, they're all low-power indicators, so...that might be part of it.

>Does the mcd value actually determine the brightness?

By Chinese eBay seller standards? Fuck no. Doesn't help that MCD isn't an indicator of total luminous flux in the first place, either.

>>1378205
>>1378207
>Doesn't help that MCD isn't an indicator of total luminous flux in the first place, either.
Oh I forgot, try to calculate the radiation pattern for the LED as a function of vertical and horizontal angle so you can convert luminous intensity to luminous flux in the first place. I think the integral is 1/r^2 dΩ.

>>1378205
>>1378207
>>1378212
lol okay, then i just have to build different prototypes and see what pleases me outside. Thanks anyways!

Is this battery a kill? I suppose I need to clear the brown stuff from the contacts if I want to put in a new battery?

>>1378230
>Is this battery a kill
I think it's a pretty good bet. Clean the battery contacts thoroughly and the thing should be fine, but since it looks like it might have been water damage you should try to prevent water ingress in the future.

>>1378234
It's inside a garden solar-powered glowing snail. The casing is cracked and I suppose that's where rain soaked through. I'll need to consider if repairing this thing is worth the effort.

>>1377579
Does anyone know how to solve this? or what theory is used?

>>1378247
mesh analysis? Kirchoff's second principle?

>>1378237
smother it with glue or something

>>1378247
>>1378254
more specifically, both circuits are closed loops, so you know that the current is the same through all elements (eg in circuit (a) the current through the 5k resistor and R is the same, and it's I=5mA). You also know that the sum of all voltage drops in a loop is 0 (kirchoff's second principle).

So for example, in circuit (b) you add together all voltage drops that "are facing the same way", so 16+6+12+8=42V. Voltage E is the only one facing the opposite way (look at the + and -), so it has to be equal to 42V for the total sum to be 0.
For R you just use ohm's law. The 16V 2k resistor tells you that the current through it is I=V/R=16/2000=8mA. That same current is going through the 12V R resistor because there are no other branches for it to go. Reapply ohm's law: R=V/I=12/0.008=1.5k.

For circuit (a) you use the same techniques, but with the unknown resistor you need to use the power formula, P=VI. However, you don't have the voltage drop! Only the current...

>>1378247
Ohm's law and common sense
4K 25 cw 1K5 42 ccw

Regarding the LEDs I purchased earlier (>>1378154), is there a way to use a dimmer on them?

>>1378395
inefficient and bulky: rheostat
inefficient: linear current controller ic / opamp and a transistor
cheap and easy: PWM, cheap chinese dimmers use a 555
efficient: switchmode current controller ic

>>1378395
you need a different driver than the ones I recommended last night. you don't want to screw with the output of those, as they'll keep turning up the voltage until they see the average forward current they expect. here
https://www.aliexpress.com/wholesale?catId=0&SearchText=constant+current+dimming+driver

>>1378269
>>1378270
But the resistances instead of adding are not subtracted?

>>1378412
>rheostat
I was thinking something like that but looks like I won't be able to.

>>1378417
That makes sense. I'm not sure that I will need to dim them or not, I was just sitting here daydreaming about what I can do when my package arrives.

>>1378269
>>1378270
but the resistances instead of adding are not subtracted?

>>1378435
resistances in series are added

>>1378430
if it's total light output you're interested in, you could just put a switch across some of the LEDs to bypass them. as long as you stay within the capabilities of the driver, it's as if the bypassed LEDs weren't there

>>1378441
at point b you seem to have to subtract them
20+E-20+25=0
20+E-45=0
E=-20+40

this has less to do with a project and more of a general curiosity:
learning that neutral is basically 0V, is there an actual difference between neutral and ground? (ie. on a standard american, german or japanese outlet)
it seems more like "ground is there because you can always have 0v before the live comes in" but a longer neutral prong would seem to be better here.
the only other thing i can think of is structural integrity of an outlet/plug.

>>1378448
voltage is potential difference.
there is no such think as 0v unless you measure it, you define it as 0v.
in your house
neutral exists as a return path for current to flow.
ground exists as a return path for current during a fault condition to allow overcurrent protection to operate.
the reason ground connects first is to protect you, to establish a path for fault current before the appliance is energized.

>>1378448
since wires have non-zero resistance, a wire carrying current will drop some voltage (E=IR), which could be dangerous if exposed parts of a device (e.g. the coax input or volume control shaft on old TVs) are held at "0V but not really". ground is there to provide a generally low-current path to "0V", so that devices won't have stray voltages on exposed parts

>>1378441
Why not just have the switch bypass a series resistor/connect a lower value one? Doing it the above way would increase the current through the connected LED’s wouldn’t it?

>>1378454
so would changing the resistor
current through d4,d5 will be 0

>>1378452
roger that, anon. thanks

>>1378451
>ground exists as a return path for current during a fault condition to allow overcurrent protection to operate.
okay so something like, if there were a live>neutral fault, the protection would trip and it would go to ground instead of being live across the break point.

>>1378448
Ground is a potential current return path that exists in case a hot live wire touches the case of the equipment, to allow the short circuit to trip the breaker. It's for safety / fire prevention.
It is not used as the current return path for normal equipment operation, neutral is.
Neutral won't be exactly at the same potential as ground because neutral is a current-carrying conductor, and you get a small voltage on that wire because of ohm's law.

>>1378454
yes, if the input is constant-voltage. not if the input is constant-current

>>1378457
>resistors
>for power LEDs
not recommended

>>1378458
no, the protection would trip and live would no longer be live, and would fall to neutral or ground potential (both approximately "0V" since there would be no other current flowing in that branch circuit)

>>1378451
>>1378458
wait shit, it would still be live>neutral at break point since otherwise live would be at neutral voltage, right?

So, I have a Philips TV, sound, ambient light, volume control, changing channels, all of that works, but there is no image, just some glitches on the display. It doesn't happen immediately when turning it on every time, some time it works well for a while and then it glitches.

I've read it could be an issue on the T-con board, anyone knows?

>>1378458
ok
lets say you have a toaster. a metal one.

lets say it develops a fault, the live wire inside comes loose and touches the case.

say the case isn't grounded at all, it was floating and now its live. you touch it, you get a shock, maybe you die. bad.
say the case is grounded. current flows unimpeded through the ground wire, because the case is grounded. because it flows unimpeded, not through the toaster element, not doing any work, very little resistance, the current that flows is HUGE. like, massive. so large that the fuse in your breaker panel melts. that is literally its job. which is good. you touch the toaster, no shock. no electricity because the fuse saved you.

ok so why not just connect the neutral to the case? a few reasons. sometimes plugs aren't polarised. i think in the US polarised plugs are fairly recent? so you connect the case to neutral, plug the plug in upside down the case is now live. oops, you are dead.
but plugs are polarised right?

how much current does it take to kill you? say 30mA because thats banded around a lot.
now the fuse in your panel is about 1000 times that, because things need current to work. so there is another device you might know as GFCI or RCD depending where you live.
say there is a fault in the toaster again where not the live wire come loose but a bit of marmalade causes a short between part of the heating element and the case. so in this example because the current has passed through some of the heating element already and the electricity has done work the current flowing in the short is also a little lower, its actually below the amount of current it would take to blow the fuse in the panel.
now this is where the other device and the ground/earth comes in. it measures the current in both the line and neutral wires and turns off the current if the imbalance is more than say 30mA. it only works if current can escape via ground and cause an imbalance in line/neutral current.

>>1378473
it could be because you touched yourself as a child.
hold on let me consult my crystal ball to diagnose your tv for you, phillips only make one model tv right?

>>1378479
It is a 40PFL7606D/77, if that helps.
Now I've read that the T-con board is the one that drives the display, so if everything else works it has to be that one, right?

>>1378487
could be the display itself.

Does anyone have any advice on some videos that I could watch to get a better hang of the physics side of electronics, particularly dealing with Kirchhoff's laws in circuits? I get the laws, but I suck at applying them.

>>1378507
>kirchoff
>videos
do problems on paper, faggot

>>1378507
>the physics side of electronics
if you think kirchoffs laws are physics i have some bad fucking news for you boyo.

pick any point on a circuit, literally anywhere you like. component, wire, junction, doesn't matter.
current in === current out.
thats part 1. you are welcome.

>>1378513
>current in === current out.
does not apply to caps in time domain

>>1378516
yes it does. its the law.

>>1378247
Point b. does not comply with the sign convention for passive components and sources, or are I being given voltage signs?

>>1378519
>capacitor charges
>current goes in
>stays in

>>1378530
then which terminal of the cap does the current go "into"?

>>1378555
The one that is most positive, if following conventional current

>>1378530
are you sure that is correct?
hint: no.

>>1378568
>what is conservation of charge

>>1378572
maybe it is your bed time?
current flows through capacitor as it charges.
no current flows when it is charged.
current is different when?
go do it, measure it. don't just listen to me. sim it if you want lol.

>>1378579
>DC current trough a capacitor
>what is time domain

>>1378588
maybe if you explained what you meant from the beginning I could more efficiently help you.
DC through a capacitor in the time domain is a problem for you? you meant something else? AC? the result is the same.
show me your results that differ. perhaps publish a paper on it.

>>1378593
>faggot didnt read properly
>gets butthurt

>>1378596
i still don't know what your point was.
you are avoiding explaining because you were embarrassed about being wrong? its anonymous here nobody knows who you are, no need to be ashamed.

>>1378599
>you are avoiding explaining because you were embarrassed about being wrong?
nah, it's ok, you're right, I was wrong

>>1378441
I was thinking about getting some type of switch so I could either turn on the blue LED or some white ones, or a mix of some type. I've seen some DIY moonlights that incorporate UV (<400nm) but I'm not sure that is safe. I'm probably going to end up ordering 1 of each type (1, 3, 5, 10, 30 W etc) of the white and testing those with my PAR meter.

>>1378648
hmm, you could use an SPDT switch, but the driver might not react well to an open circuit. if unsure, you'll want a make-before-break switch, or to switch them upstream of the driver
>1 of each type
but you'll need 1 of each type of driver, too
>UV
the sun does, why not you

>>1378655
I was thinking driver -> switch (probably SPDT) -> LEDs
Shit I keep forgetting about needing a damn driver for each type of light! They are cheap enough though, might take me a bit between all of my various projects to get them and test them, but it will all be worth it in the end for the data that I will be gathering.
UVA and B causes skin cancer and cataracts, that is why I am leery of using it. But maybe if I stay really close to 400nm it won't be so dangerous. This guy used 395nm (Min) 410nm (Max) LED along side a white LED on the outside and a blue LED in the middle. I guess it won't be a problem if I put the lights on a timer and don't keep them on all the time. I will just have to see how I like the Royal Blue LEDs I bought (should be the same as the one I have) and I will go from there.

>>1378660
sure, that switching arrangement will work great if the driver can handle an open circuit. some simple ones might not react well, or might send a strong current pulse through your LED string when the suddenly shorter string connects up
>UV
good point. trying for photoluminescent effects? you might get enough residual UV off of white or blue LEDs to serve your needs. if not, and one UV LED is too bright for you, if you are using a constant current driver you can switch a resistor in parallel with the LED (ballpark 22 ohms to start) to shunt some but not all of the current around it

>>1378676
Is there a way that I could test the driver once I get it? I don't really want it fail and either break the driver or my LEDs. Could I avoid the possible strong current pulse by unplugging the driver before using the switch? Sounds like a pain in the ass.
Not really going for any sort of photoluminescent effects. I was more concerned about algae growth from running the moonlight. Hopefully they won't be strong enough to grow anything. I'm not sure if algae utilizes light outside of PAR (400nm-700nm) to grow. I would think that green algae would have the same requirements as plants but there are other non-green algae that can grow in an aquarium.

>>1378681
>Is there a way that I could test the driver once I get it?
pretty much build that dual switched string with the shortest possible strings and watch for bright flashes when you switch between the two. or, maybe the seller would know whether their drivers can handle such treatment
"when in doubt, blow it out"
>Could I avoid the possible strong current pulse by unplugging the driver before using the switch
absolutely. in fact, you can use a center-off DPDT switch to do it for you, Pic related (center-off position not shown)
>UV
ahhh. I guess we'll all find out!

>>1378692
>center-off DPDT
Yeah, that's what I would use. All the fancy lights have something like that but on a rocker switch instead. One position controls the non-blue lights and the other controls the moonlights.
Currently going through and figuring out beam angles and crap for some lenses for these LEDs because I really don't want to fill the light with blue, just the aquariums. Not too hard to do, just tedious, especially because after I figure out which lens will work at what height, then I feel compelled to put it in Photoshop so I can dream about it more.
What program did you use to make that schematic? What is the purpose of having two switches on it?

>>1378599
>>1378604
What a shit argument. If you measure the current on either side of a capacitor, they're both the same at the same time, regardless of whether it's in DC. Now if you make your gaussian surface have a side between the capacitor plates then yes, a net current flux does flow through the surface and hence there is an imbalanced current, but a capacitor has 2 plates, and the charge accumulating on one side is equal and opposite to that accumulating on the other side, so the currents are equal. As far as electronics is concerned, or pretty much anything else, you never do calculations on each seperate plate of the capacitor.

I need to waterproof the electronics on one of those Chirp i2c soil moisture sensors, will any silicone epoxy work? That bitch is going to be fully underground (hopefully) semi-permanently and I need to figure this out today because I let a deadline get away from me.

>>1378697
I had another idea about these moonlights I am building. What if instead of using a switch, I used a photodiode? I think that is what they are called. When the normal lights go off, the moonlights would automatically kick on. I know they already have stuff like this out there, but they command a high price, almost $250!

>The photosensor of the Tunze 7095 Multicontroller is housed in clear acrylic. When the lights go out, this sensor automatically turns the LED on (in the black tube to the right) and vice versa. This assembly is less than 2 inches (5cm) long.

>>1378697
they're two poles of the same switch, so that the driver gets no input power for a moment while the output is being switched from one string to the other
I use KiCAD to draw the schematics. kicad-pcb . org

>>1378735
photodiodes are finicky. most brightness detectors use a simple photoresistor. you can find a ton of simple circuits to do light/dark switching around the internets, then just use an SPDT relay to switch strings (again, if the driver is okay with that, but relays will probably be quicker than a toggle switch), or, you can hack some dusk-to-dawn switches from the local home improvement store for an order of magnitude less money than $250

>>1378738
Ah I get it now. I reread your post about a dozen times before I put it all together.
>center-off position not shown
I should probably wait until I get all the components and test everything out before I add more and more complications to this project. Although it seems pretty simple. I only have a basic understanding of electrical components and can barely read a schematic. I did build one circuit two times though using a 555 timer but that was like 13 years ago.
I wrote the manufacturer of the driver to see if the driver will work, but they don't understand my question.

>>1378749
hah, yeah, candy store syndrome is definitely a thing
"Will the driver burn if the LEDs are disconnected?" should get the message across, or feel free to send them my schematic

>>1378735
LDR + resistor in series (trimpot for adjustment), MOSFET, job done. Using MOSFETS as shitty comparators is something I enjoy doing, but if you don't swing that way use a real comparator I guess. But since an LDR goes through a couple of orders of magnitude of resistance from day to night it really shouldn't pose a problem. A phototransistor probably also works, but is harder to tune, photodiode even more so.

>>1378749
It should just hit max voltage and stop unless it's particularly poorly designed. The problem would be its voltage ramping up until it kills the control circuitry, provided the switch can handle it. If it runs on mains then it should have circuitry to prevent such a failure, and one that runs off 12VDC or so should be able to handle that input voltage at its output anyways.

A link to that listing could help, if it's a common model or otherwise one for which relevant datasheets could be found.

>>1378757
Here is a link: https://www.aliexpress.com/item/LED-Driver-Adapter-Light-LED-Power-Supply-Lighting-Transformer-1W-3W-5W-10W-20W-30W-36W/32756673848.html
I bought the one in this screenshot. They still haven't got back to me with my question yet. Probably stealing that schematic I sent them as we speak.

>>1378770
Well worst case scenario you can always trace the circuit once it arrives and ensure it will be fine, but there's probably nothing to worry about. A make-before-break switch would work properly also, which you could make with some transistors if you have to.

Bump for this question

>>1378723
When set its fine.
When wet it's conductive.
Be careful

>>1378838
>When wet it's conductive.
Shit, really? I'll measure it next time I whip out any adhesives.

>>1378853
Never tested it myself but I heard a lot of stories from sparks that silicone an outside socket then test it right away and it fails insulation resistance with no visible cause. Head scratching ensues.

Hello guys,

i got a 12V RGB (common anode) strip and wanted to control it with an attiny13. Though i have no mosfets at home right now, but a few bipolar transistors.

Can I do pic related with any of those?

>>1378960
oh nice, i just googled a bit and apperently it just works with the bc547 quit nice. Webm related

>>1378979
that was just on a breadboard with buttons. Now just gotta attach a µC quick.

>>1378960
depends on how long of a strip you expect to run. each snippable section of strip ~= 20mA. mind the maximum collector current of your transistors (100mA for BC54x) and you'll be fine
(it's good practice to add base resistors to limit base current to roughly Ic(max) / hFE(min), so 1mA, so 3.9k or 2.2k depending on Vcc=5V resp. 3.3V)

>>1379009
>(it's good practice to add base resistors to limit base current to roughly Ic(max) / hFE(min), so 1mA, so 3.9k or 2.2k depending on Vcc=5V resp. 3.3V)
Thanks! Got 1k at the base atm, but will put a higher one in then.

>>1378960
As the other anon said. Multiply the maximum current each pin could source if all pins were on 100% by the minimum hFE of the transistor (datasheet, I use 50 to be safe), that will give you a figure for the maximum amount of current each transistor should pull. Needless to say, also do the P = 0.7*I calculation to figure out how much power each transistor will sink if you're putting at least a few hundred mA through.

Should get some FETs on order though.

How do I know how a DC motor is performing without having a performance chart available for it?


I've got a 200 RPM 'high Torque' 12V motor, operating at the full 12 V vertically with a light shaft and tiny bit of weight around the shaft.


If the motor is rated at 200 RPM, would it be obvious if the weight is slowing it down?

If it's torque is well above what's needed, does that mean the voltage is allowed to push it full speed? It has a rated torque, do I just have to ensure that the full speed for the weight is using less torque than the rated torque or is there some sort proportional rule?

What kind of toroidal core should I use for a diode ring mixer designed to handle signals between near DC and up to 1MHz or so? How many turns would I want on said core?

>>1379060
max torque for electric motors is at stall.
lowest torque is at the fastest speed, when unloaded.
http://lancet.mit.edu/motors/motors3.html
>If the motor is rated at 200 RPM, would it be obvious if the weight is slowing it down?
if the difference is big, you can hear it. do you have a way of measuring it? you can get creative and record video of it and count turns/sec if you need to.
>do I just have to ensure...
the question you need to ask is "does the motor do the job i require of it at the right speed without overheating?"
the motor won't spin at full speed if it's loaded, and the more it's loaded, the slower it'll spin.

>>1379106
Exactly this, the more you load the spindle the more the rotor opposes the field, the more current is drawn to drive it,the current causes heat.
Heat increases resistance and reduces current,cycle continues until motor can't run anymore.
If you can keep it cool then you don't have a problem.

>>1379106
>you can get creative and record video of it and count turns/sec if you need to
This or stick a magnet on it and record the audio signal from a nearby solenoid and FFT it on Audacity. Magnet + Hall-sensor + Hz measurement on a multimeter also works, or even just measuring AC current through the motor with a scope or DMM.

>>1379077
Well you can't go wrong with air-core as far as hysteresis losses, but it's not very ideal for such low frequencies. >>1369944 also leaves out this circumstance. Did you google it?

>>1379114
>Heat increases resistance and reduces current,cycle continues until motor can't run anymore.
The temperature coefficient of resistance for copper is insignificant, you're describing a runaway effect that doesn't happen.
Motors need to be cool so the wire insulation doesn't burn up, it's not because they are in danger of stalling from increased wire resistance.

>>1379116
Here's what I want out of a mixer
>High linearity
>Good LO to RF isolation
>Low noise figure
>High IP3/Low IMD
>As low a balun count as possible (ideally zero)
>works at low frequencies down close to DC

I don't care so much about conversion gain/loss and LO drive level which I am willing to make sacrifices on.

I have tried googling and looking at more mixer topologies than I can count but the problem I have is all of the really good balanced mixer topologies are designed for radio frequency use and don't necessarily work or are not designed for low frequency use. Those topologies that would be more suitable at low frequencies (single transistor mixers, etc) are ultimately not balanced and don't meet the above performance specs I am demanding so I'm having a real problem finding anything that will work.

>>1379137
Though this is second-hand knowledge, I think diode ring mixers do work with DC (at least on one side) because its required for modulating AM audio properly. But as your higher frequency gets close to your lower frequency your results really start to suffer, I personally wouldn't use one without an order of magnitude between the frequencies of the two mixed signals. I got to this by making a diode ring mixer in spice, but by all means give it a shot yourself. If you can characterise the mixer with an equation then even better, but the diode ring alone makes me shy away from that, let alone the transformers.

>>1379077
you should seriously consider a digital architecture. you're asking for one circuit for many decades, here
that said, does an NE592 or similar completely fail to work for your application? they're cheap, it's worth a try

>>1379178
I'm working on designing a swept tuned superheterodyne spectrum analyzer and going with a digital FFT based approach kinda defeats the purpose.

Can I build this circuit from the AD538 datasheet with just general purpose OP-amps and 2N3904s? The AD538 is about $80 a chip and I don't need the super high accuracy it's built for. Does the schematic give me an accurate depiction of everything going on in the chip?

>>1379191
>Can I build this circuit from the AD538 datasheet with just general purpose OP-amps and 2N3904s?
Almost certainly not, it's indicative of function only. But at least it doesn't have those whacky multi-collector NPNs or suspiciously unpolarised JFETs.

>>1379191
It's a simplified schematic diagram so it doesn't capture all the detail. Building it with discrete parts won't achieve close to the performance of the IC if it even works at all. Feel free to try though. If you need a log amp there are other ICs that will do it for more around the $10 to 20 range. Even if you want general purpose analog math ICs I think there are cheaper options.

How would I program an arduino to sequentially cycle through 5 different LEDs individually when a button is pressed?
Also, how would you have a buzzer make a sound on the 5th button push.

> I found the following code for something similar:

Demonstrates the use of an array to hold pin numbers.
Each time the push button is pressed, next LED in the
sequence will be turned on while the rest of LEDs are off.
*/
const int buttonPin = 2; // the pin number of the pushbutton input pin
int ledPins[] = {3, 4, 5, 6, 7};
// variables will change:
int buttonState = 0; // variable for reading the pushbutton status
int buttonPressCount = 0;
int numberOfLED = 5;
void setup() {
// initialize the LED pin as an output:
for (int i = 0; i < numberOfLED; i++) {
pinMode(ledPins[i], OUTPUT);
}
pinMode(buttonPin, INPUT);
}
void loop() {
// read the state of the pushbutton value:
buttonState = digitalRead(buttonPin);
// check if the pushbutton is pressed.
// if it is, the buttonState is HIGH:
if (buttonState == HIGH) {
for (int i = 0; i < numberOfLED; i++) {
if (buttonPressCount % numberOfLED == i) {
// turn LED on:
digitalWrite(ledPins[i], HIGH);
} else {
// turn LED off:
digitalWrite(ledPins[i], LOW);
}
}
buttonPressCount++;
delay(400);
}
}

>>1379191
the only way you might get something like this to work is if you use *perfectly* matched transistors or do a fuckton of trial and error with adjusting resistor values and shit, and even then, you probably won't even approach the accuracy of the single chip, if you can get it to work at all. And even if you do get it working, a minor change in temperature might throw off your "calibration." With ICs, identical transistors are inherently parameter matched and always operating at the same temperature, since they're embedded in the same chunk of silicon.

>>1379197
Toward the beginning, define the buzzer pin
>int buzzerPin = 8;

Initialize it as an output
>pinMode(buzzerPin, OUTPUT);

Then after the led loop(right before buttonPressCount++;), add a check to see if the count is divisible by num of led's evenly
>if (buttonPressCount % numberOfLED == numberOfLED) {
>....// turn buzzer on:
>....digitalWrite(buzzerPin, HIGH);
>} else {
>....// turn buzzer off:
>....digitalWrite(buzzerPin, LOW);
>}

>>1379197
>program an arduino to sequentially cycle through 5 different LEDs individually when a button is pressed?
You know there are single ICs that do this, right? I remember when some guy asked how to trigger stages in his rocket sequentially by sending his rocket a simple pulse (of RF I guess) for each stage.

Finally got an answer. I think.

>>1379316
DPDT switch it is then

>>1379318
Or make-before-break. A zener diode would also work, provided it could sink that much current.

what possible harm could befall a computer if I were to power a 12v case fan off of a 5v USB port? I'm more concerned about the fan trying to pull more amps than USB can supply

>>1379487
also, could I theoretically bridge two USB power and ground cables into one fan so I can power it with 10v?

>>1379487
worst case, your port doesn't have any protection whatsoever and fries. very unlikely though, because usb ports often have current protection and it's a case fan.
>>1379490
if the usb supplies in your pc float (are galvanically isolated) then it can work.
if the usb is not floating then you'll just short circuit it because putting them "in series" just short circuits 5V straight to a common ground.
almost guaranteed it won't work.

>>1379503
>almost guaranteed it won't work.
figured, after looking at the USB 2.0 spec sheet and considering how shitty the old PC is, I highly doubt they put in any kind of protection in the PSU

thanks for clearing it up

>>1379490
From one device, like a computer, no way. Two plug in phone chargers? Absolutely is possible. Maybe if your American you can use the split phase to your advantage

holy fucking shit. Webm related is the best thing that could happen to me today.
I never believed that this would work out, but i fucked with my atmega8a around and accidentally wrote the fuses wrong. So now I tried to do high voltage programming to recover it. According to this:
https://mightyohm.com/blog/2008/09/arduino-based-avr-high-voltage-programmer/

And what can I say: It fucking works. Holy shit.

>>1379561
pic related, my high voltage programmer lol

This thread excites me so much. I'm always
"repairing" broken electronics around my house (TV displays, cell displays, heaters, wall outlets, lights, etc.) but I don't actually know much about electronics. I have a firm grasp on the beginner stuff like what components do. I used to know a bunch of the equations for building circuits but I learned them like 15 years ago and forgot them 14 years ago...
Anyways, I see there are some good beginners books in the OP, I think I might pick one or two up. But my question is this: is there a good beginners electronic kit? I don't really care if it is geared towards children, although I would like it if it wasn't, but it would be nice to have something like that so I can ease myself back into making electronics. I already have a multimeter and soldering station, but those are the only electronic tools that I have.

>>1379565
Ok guys, more troubleshooting to come.
Apperently the microcontroller is working alright.

Yesterday I wanted to try out this module we made in uni in first semester and do something with it. As retarded as i am i didnt notice that it was laying like pic related on a metal plate and connected it to USB.
To that point it was still working and afterwards the function was okay. EXCEPT that the usb wasnt working anymore.
Thats when I tried reflashing the firmware and fucked up the fuses. Fuses should be alright now. Still the USB does not work. Are the zener diodes likely do go ded when shorted?
If that's not the case then i still might have some firmware problems.

Though, when I flash the atmega with a programmer everything else seems to work perfectly fine

>>1379587
Doubt it's the diodes, assuming the circuit is correct (I'm used to pic ics that don't have any components between pin and plug) easiest thing to kill is the ic. Just because everything else works doesn't mean its not fucked. New microcontroller will be only a few of your currency units and its in a socket, buy a new one and swap them, then you will have another one to play with anyway.

>>1379565
>anon just gained a priest level
congrats

>>1379573
>multimeter and soldering station
you're ahead of the pack already. you could just buy a solderless breadboard and some resistor/capacitor/transistor assortments from alibay, because you will almost always have a good use for those and they can come in quite handy. what components you buy in addition to that is best determined by the bill of materials of some project you would like to do. e.g. don't just buy one LM555, buy a few extra of whatever parts your projects use, so as to build your stock

>>1379587
you could have blown the I/O pins related to USB while everything else works fine

I'm making an outside shop's insigna with LED. 220V get to the inside wall, I plan to put a 220 AC => 24 DC transformer on the innerside of the wall, drill the wall, pass a cable through and step down to 5V DC with some waterproof converter fixed on the wall outside.

With about 1 meter between 2 devices, will it avoids voltage loss ?

>>1379646
>5V
y tho

https://www.youtube.com/watch?v=ixlzlna_BUs
Is he going to be okay

>>1379668
>Perpetual motion exists
o fugg he's losing it

>>1379668
no, he ded DD;;

I just printed a dual later pcb and forgot to specify some isolation for the bottom ground plate on this board.

Since I'm doing this for a project, feel like everybody (including the guy who printed it)will slaughter me if I have another one milled. How big of a deal is it if I'm careful with soldering? Only thing that could get iffy is the fact that I have some AC lights being controlled by relays

>>1379685
And for the record, that's all on top layer. Only things I have on the bottom are some 5V lines and signal paths. Like three traces.

This is my first pcb

>>1379687
>milled
you could take a pen knife, score/peel up some copper, and add your own extra clearance, and/or, you could just have your fabricator cut some isolation slots into the existing board?

>>1379690
Alright I can just do it myself just to make sure.

I guess I just forgot to specify some clearance on the bottom layer and didn't notice. It was auto milled, so it was hundred percent my fault. So it only has a few mil of clearance on the bottom

>>1379513
Provided each phone charger is isolated (flyback converter, ungrounded 0VDC pin) you don't even need split phase.

>>1379487
Why don't you have an extra rail on your PSU?

>>1379640
Are there any beginner projects that you could recommend? Or a textbook that has projects in it?

>>1379786
it's a disgustingly cheap mid 00's off-the-shelf prebuilt being used as a server in a micro-atx case with a micro-atx PSU, it doesn't have an extra rail

>>1379800
um, anon...

LED strips. I seem 5M of RGB strips for 10-20 bux on eBay. That's with the controler, and remote and crap. I'm guessing this stuff is fail early fail often?

If I'm right, where do I get some reliable strips?

>>1379807
I'd love to use some molex converters but this thing has a really strange PSU, it has no spare molex headers at all

>>1379809
>spare
that's what the Y adapter is for. piggyback off of one of the disk drives, which are almost surely "Molex"

>>1379808
it's more like signmakers, decorators, etc. buy miles of it so you get to enjoy some of the volume savings

>>1379810
disk drives haven't used molex in years.

>>1379811
>mid 00's off-the-shelf prebuilt
that should have been in the transition era, no?

>>1379820
maybe.

>>1379808
Should be okay for that price, unless they're advertising too good to be true numbers, like high CRI.

>>1379876
>high CRI
I don't think this applies to RGB, unless they split up the high-CRI phosphors between different LEDs. I wonder if you use yellow-emitting phosphors with a green LED?

>>1379892
Sorry, missed that it was RGB, so I guess the CRI thing applies doubly so.

bravo uk
https://www.youtube.com/watch?v=ixlzlna_BUs

>>1378604

>>1379197
microcontroller classes at ufsc?

I'm looking to do stereoscopic vision for my robot. However, the hardware for having two camera inputs doesn't seem to be quite there yet for the hobbyist, so I was wondering if I could instead use just one camera (eg with a RPi), and then use a solenoid to shift its position and take a second image to get that stereoscopic effect. The computer could then do its calculations to find an object's distance from the robot, by knowing how far the solenoid moves the camera.

The downside of this I imagine would be its slow speed, i guess because solenoids aren't super fast and they probably oscillate a little when they reach the extremities of their travel distance. I am not sure about this though as I have never used one before.
Anyway all this led me to this idea: would it be possible to use one camera and then some sort of piezoelectric actuator vibrating at very high speeds and time the camera exactly right so that it takes images at the full extension of the vibrating element, thus introducing this stereoscopic imaging effect with just one camera? Or would it just become a blurry mess and not really worthwhile over just buying the more expensive hardware for two cameras?

>>1379921
>the hardware for having two camera inputs doesn't seem to be quite there yet for the hobbyist
I think one of the more obscure Rpi models has 2 camera ports, but can you use a USB webcam as the second camera? There are camera port multiplexers, and if that fails you can always figure out how to run one off the GPIOs.

>>1379924
Yes the compute module has both ports broken out, but then you need the development board or whatever is called which is an extra 70 bones I believe. (Though i did find a chinese clone for 40 euros on ali). The disadvantage of the multiplexer is that it still connects to just one port so it's really just alternating the input between the two cameras, though I guess that would still have the benefit of doing away with the solenoid nonsense.
The USB webcam is always an option but I read that on the Pi that has the downside of not making use of the integrated GPU (which apparently only works through the CSI). So you'd be using your CPU to do the work of the GPU, though the former should theoretically be used to run your vision algorithms.
GPIOs was another alternative indeed, especially with those arducam models, but I remember reading it wasn't the best idea though I forgot the details of exactly why that is.

So yeah there's quite a few ways of going about this, I was just wondering if the solenoid/piezo thing could potentially be added to the list or if it's just a bad idea all things considered. Especially because I would imagine that with the speeds you'd have to work with to make it worthwhile (assuming the vibration is at least on the order of 10s of Hz), you'd probably need a pretty powerful chip or even a dedicated FPGA to process all those images being captured per second, which would get you back to square one (as in, at that point you might as well just spend more money to save yourself the trouble)

>>1379931
Perhaps having a second camera with a much lower resolution being run by via USB off the CPU wouldn't be an issue. And instead of moving the camera you could simply move a mirror in and out of the path of the camera, with another mirror fixed to reflect the reflected line of sight in parallel with the original line of sight. High speed vibration I think is a bad way to go, and while I'm unsure how shutter speeds work on those tiny cameras, it would be a pain to tune anyways. Multiplexing doesn't give you resultant data any worse than the solenoid, so it's just up to you whether you want the extra camera or extra mechanical element. But I wouldn't go more than 2Hz with the moving camera, while the multiplexer is good for frame-rate frequency switching. The frequency you need shouldn't be too high, but it will be dependant on the speed of robot.

So I see no reason not to go with the multiplexer other than price.

I've got a boost converter that I use to drive a 12v fan. It's a boost converter with a potentiometer so I can change the output voltage easily. Now a friend of mine says I could use PWM to control the speed of the fan.

Now my question is what the difference is. Variable voltage or PWM because as far as I know PWM basically does the same.

>>1380003
I forgot to add something: What's better?

>>1380003
>Now my question is what the difference is. Variable voltage or PWM because as far as I know PWM basically does the same.
PWM switches the device on and off, varying the on-time duty cycle to vary the amount of average power.
Variable voltage is just variable voltage.
They aren't really the same, PWM is simpler conceptually than a variable voltage power supply (especially a switching supply) because it is open-loop.
Better is usually what's cheaper and still does the same job.
For a fan it doesn't really matter, but for a variable-brightness LED backlight, variable current is preferable to PWM because there's no flicker.

>>1380013
Oh ok, so I won't notice any difference? I know how PWM works, when I said "basically does the same" I actually meant the result would pretty much be the same (when using a fan that is). Thanks for the explanation

>>1380019
>Oh ok, so I won't notice any difference?
Correct

>>1379639
>>1379640
I was just being a fucking retard.
There was one wire missing which tells the microcontontroller to go into usb mode instead of running the firmware. (the yellow one around the piezo)

Prof just gave me a spare one he had around and i was actually changing the diodes, but it didnt help. Then i was looking around and sometime i noticed this missing cable.
Pls kill me now. Thanks.

But atleast it works.

>>1380003
PWM on "PC" fans is usually implemented as a separate control input to the fan. I don't happen to know the electrical specs offhand, but you should be able to use an open collector to ground to stimulate it
if not, and you're trying to switch the power on and off quickly, at least you won't lose heat in the switching device as you would in a variable-voltage setup

>>1380066
derp happens, mein dude

>>1380066
werks now as a control for my led strip behind my desk.
But while this webm was rendering i was just thinking about how much overkill this is. Either ill rebuild this shit with 3 flipflops or ill add something else to the microcontroller. Will see, but its a start.

>>1380084
also (if i keep that µC) i gotta add some debounce logic, but thats the least problem I guess

>>1380003
pwm turns fully off and fully on, variable voltage should be smooth output.
Switching supply will have pwm somewhere inside it to drive the voltage converter

If the fan is brushed motor pwm should work fine.
If the fan is brushless there is a controller inside it that turns the fan coils on and off. The controller will not enjoy pwm turning the voltage on and off all the time.

>>1380013
Pwm is used all the time for led because they are driven at 10s to 100s of khz well beyond what your eyes care about.

Benefit of pwm is that when the power is on its at max voltage, in a motor that means you can lower the speed without losing torque. If it's smooth the max voltage is lower and so max torque lowers with speed.
Also with LED max voltage is really down to how cool you can keep the junction, with pwm led you can overdrive the led and get it brighter because it cools down during off time and doesn't overheat.

So yes absolutely different things for different applications.

>>1380090
Meant to say if it's a pc fan I'm assuming it's brushless and won't work with pwm at all which is why a pwm of fan (typically cpu fan but I increasingly case fans too) have 4pins, because the controller resets every time the pwm is 0v.

Saying that it's pretty common for 12v off fans to run slower at say 5v, maybe you can pwm between 5v and 12v so the controller is happy

>>1380084
woot
>USB mode
if you're feeling especially advanced, there's a project: control the lights (including PWM?) from your PC

>>1380085
debounce would be even more important with the flip flop solution, and there you can't just do it programmatically

>>1380013
>variable current is preferable to PWM
not (always) true. color output can vary with forward current

>>1380099
>if you're feeling especially advanced, there's a project: control the lights (including PWM?) from your PC
that might be a nice idea. Will think about it.

Not sure how pwm could benefit me with this setup. Got a simple rgb led strip with common anode and depending on what wire i put to ground with a transistor i get a different color (see >>1378979 ). Also pic related

>>1380104
>Not sure how pwm could benefit me with this setup
right now you have 3-bit color. with PWM you could get many more bits of color

>>1380109
>right now you have 3-bit color. with PWM you could get many more bits of color
Would I? Even if a have a common anode? I thought I'd need a common cathode strip for this

>>1380112
nope. your cathodes are still independent. it's just a wiring difference. CA (as opposed to CC) just makes the driving circuitry simpler, either can be fed with a PWM source and still produce variable brightness

>>1380131
Nice thanks. I'll try it out!

Has anyone any clue what a ~10kHz interference on literally every audio system in the house causes? I just started living here. It just comes up randomly for a couple seconds and then goes away again, and it randomly returns (although I didn't measure this yet, but on some days it's worse than others so it's not that regular.)

It seems like there is some signal modulated over the powerline? I unfortunately don't have measurement equipment available that I can just jam in the power socket. I am going to record the audio signal next week and try to decipher it.

I don't have any audio recording FFT images anymore, phone didn't really save it. But it starts at a carrier wave (?) that was around 9.9kHz but eventually it got stronger and by then it seemed like it got some sidelobes with 100Hz in between.

Anyone have any clue what the fuck can causes this? My guess it at some faulthy bullshit at some neighbour in the flat, but what would cause that interference on the power line? (my assumption it's on the powerline)? Better yet, what can I put in front of my gear to stop this headache noise?

>>1380169
>Has anyone any clue what a ~10kHz interference on literally every audio system in the house causes?

ahem

"is caused by"

*cough*

check your front yard. mexicans are there. it's a landscaping frequency that coordinates blowers and mowers.

maybe just ignore it.

>>1380169
what's the envelope of the wave? gradual attack? any decay?
same or varying amplitude per occurrence? same or varying length of pulse?

Is there a possibility this is legitimately supposed to be like this? It's from a chink keyboard (well, macropad) PCB.
I've never done SMD soldering before so if this needs to be fixed it's gonna be fun

Is there a possibility these top joints are legitimately supposed to be like this? It's from a chink keyboard (well, macropad) PCB.
I've never done SMD soldering before so if this needs to be fixed it's gonna be fun

>>1380174

I would hire you in a minute.

Are you in or near to the RTP in NC, or can you relocate there?

(Research Triangle Park, Carolina del Norte)

>>1380172
What? No there are no Mexicans here.

>>1380174
Can't give you that much info. Only did a spectrum analysis to find the frequency. The amplitude seems to vary per occurrence. Often a signal comes in pairs. (it stops shortly and than starts again, one louder than the other). The signal seems to be the highest at the end just before it stops.

>>1380182
Anon, YOU'RE the mexican

>>1380180

well I don't know

>>1380180
Those are resistor networks. Connecting all the other ends together might certainly be what the designer wanted.

>>1380185
Yeah. So why did you make a post then?

>>1380186
Good to know, thanks. The seller is supposedly pretty reputable and well-known, so I didn't expect there'd be anything wrong, but it just looked to weird to not question it. Thanks for the info.

On that note, what's a good way to learn to solder SMD, just out of interest? Just buy some $2 practice board off aliexpress and watch a youtube tutorial, or are there any better resources/tips you could recommend?

>>1380019
variable voltage = less radio noise
You're using the boost converter either way, one less switching transistor in the way means more efficiency. Though I'm unsure the sort of driver circuitry a regular brushless fan uses, it should be fine to somewhat low voltages, but you won't be going below 5V without a specifically buck-boost converter.

>>1380182
It's your neighbour's bootleg /ohm/-tier garage door remote control that can be operated at 5km and can be received within shipping containers. Yes, it is controlled with an arduino.

>>1380197
>It's your neighbour's bootleg /ohm/-tier garage door remote control that can be operated at 5km and can be received within shipping containers. Yes, it is controlled with an arduino.

I am indeed scared it is some bullshit project from one neighbour. Can only be fixed by either putting a huge filter on the power line here, or deliberately trying to fuck with by creating something that puts a 180deg out of phase signal on the line when it recognizes something in the hope he throw his shit away.

Hello I am a newbie when it comes to this stuff. What is the value of this current source in the schematic of the lm386?

>>1380201
To see if it's from the power line, try making yourself a high-pass filter with a 3.3nF cap and a 10k resistor, it will give a corner frequency of about 5kHz, and so attenuate the 50Hz mains by 40dB or a factor of 1/100: 1.2VAC. Then just shove your scope in there and safely see the waveform, provided you use the right ground reference.

I think it sounds like a high power switched-mode power supply, perhaps for a computer or electric car.

>>1380204
forgot pic

>>1380204
>>1380208
If you're trying to recreate an op-amp with transistors as in the diagram, think again. See the replies to >>1379191.

>>1380189
Yeah, just get parts and practice, it's not different, just smaller. The solder wants to do all the work anyways.

>>1380211
>The solder wants to do all the work anyways
Only provided you've got enough flux. Though I've seen some soldering stations that blow hot nitrogen gas out of the iron around the tip for better heating and also as a blanket of effectively inert gas to prevent oxidation, so in that case if your surface is clean you may not need any flux. Makes you want to try it.

>>1380204
Not a fixed value, follows supply voltage.

This is probably a very stupid question, but when building lipo battery packs, instead of doing 14 banks in series, and having to deal with balance leads, why not just do 1 bank and use a step up transformer to whatever voltage you need?

>>1380258
or is it with balance leads, at say 4P, I can only balance per bank of 4 cells, and if an individual cell in that bank is way out from the rest, I'm fucked. and with a 1S52P instead of a 13S4P, I still need to balance individual cells regardless of the topology?

I am trying to find a way to play wav files in NI Multisim 12.0. I found a page for a custom instrument but the download links were broken. Anyone know a source for a custom instrument or some other method of doing this?

>>1379197
>all of that code when you could have just 1 output line, a clock, and a multiplexer

>>1380265
Check to see if the page was archived on archive.org

>>1380258
it depends on the purpose. suppose your battery is rated for 10A. if you wanted to pull 10A at 14S (140W) from a bank at 1S voltage, you're pulling 140A (140W/1S) from your bank to the input of your "transformer" (a boost converter). even though your batteries themselves are rated for that discharge, suddenly your wiring needs to handle 200 times the power dissipation. cable losses depend only on current, so even though you're pulling the same power the current is way higher.

>>1380335
Or a clock, counter, and demux with no micro at all.

I'm not sure where else to ask so here goes: is it possible to leave feedback on aliexpress about the website itself? How do?

>>1380418
complain on >>>/g/csg/

Any difference between a level converter and a buck dc-dc step-down converter?

I have some of these "5V to 3.3V DC-DC Step-Down Power Supply Buck Module AMS1117 LDO 800MA" lying around.

What I currently want to do is connect a 16x2 LCD with an I2C backpack to my raspi. I've read online that level converters are required to protect the pi's GPIO's.

Are these buck converters cutting it? Or is there some Amp fuckery I'm missing on?

>>1380483
An AMS1117 is a linear voltage regulator intended for converting voltage of a power supply by dumping the excess voltage * current as heat.

A level converter is intended to handle logic signals, not power supply lines. For I2C it must also work in both directions, stepping up the signal one way and down the other way.

Sure you can't get an LCD backpack meant to run at 3.3V? If so no level conversion would be required, makes things easy.

Couple of questions.
Is it allowed for an I2C device to change its address during operation (not communication) by changing where its address pins are pulled?

Also, im designing an I2C device that has a low pin count (SO8), and i have only one pin left to set addresses. How would you feel if you had to set device address by placing different valued resistors instead of pulling pins low/high?

>>1380485
>For I2C it must also work in both directions, stepping up the signal one way and down the other way.
Which is easy to do with mosfet trickery

I'm using an AD2S1210 to convert a resolver signal from a motor into encoder pulses. However, I'm getting four index pulses for each full shaft rotation.

Could this have anything to do with the resolution setting? I currently have pin1 and pin48 pulled to GND, making it 10-bit

>>1380497
i don't know about the spec but i2c communication (above the protocol) is always specific to the slave device so if you want to change your address that won't cause any problems as long as the master is written to account for that.

i'm fine with using one resistor to set an address, but you're talking about a divider which brings with it increased quiescent current (uc adcs i know of don't like >10k source impedance) and a bloated BOM if you've got many devices.

you can set your slave's address on the fly by a write command if it has eeprom. if you don't mind having to configure it out of circuit or use jumpers, that is.

>>1380500
What kind of encoder are you using?

>>1380501
>uc adcs i know of don't like >10k source impedance
Never heard of this, any info?

Maybe ill just go with a higher pin count, not sure.

>>1380516
your datasheet is your best friend here, but microcontroller adcs have pin leakage and an internal sample capacitor. the pin leakage provides a static offset which varies between chips, and the sample capacitor forms an RC with your divider that takes time to charge. not all ucs have a software configurable sample time so this can result in a deflated reading. you can try to account for these factors in software.

>>1380418
There is a chat link near the bottom of the page.
https://helppage.aliexpress.com/buyercenter/index.htm

I going to graduate EE next year with a shit GPA (around C). What should I do?

>>1380540
Retake classes and bring your GPA up to a B average.

>>1380540
i graduated with a ~3.0 from a low tier university. i worked at a shit outsource company for two years and then used the connections there to get a good job at a fortune 100. accept that you fucked up. make up for it by being visibly good at your job and nobody will care what school you went to.

I'm planning to make a valve amp but I can't get the mains transformer any were so I'm gonna make it with a 700w core, the amp uses just 400, do you thing this may affect the general performance?
Pic related

>>1380504
I'm using the AD2S1210 as the encoder. It receiver the resolver signals and converts them to encoder pulses. It's a 2013 Leaf motor if anyone cares

Hmmm or if I use it in it's current state, will it just rotate at 1/4 speed?

>>1380548
>amp uses just 400
and the transformer has 250VA
good luck

>>1380540
>>1380543
Negative. C average is not adequate, you must raise it. You need a 3.0 gpa. Less than that literally get binned before any one reads them. You can't make it look good.

>multisim has the 555's internal latch set/reset dominance mixed up
thank god i checked

>>1380713
nevermind evidently there's a difference between the LM and NE555s. too bad my cmos 556 is the wrong kind so i can't abuse it as a nand gate.

What sounds easier? Writing a code to simplify boolean statements made from a truth table that includes "do not care" outputs into POS and SOP expressions, or instead using that same truth table to generate a tessellating array of N-dimensional Karnaugh maps and fitting shapes to them?

By N-dimensional I mean 1 dimension for each input instead of cramming the whole thing into 2 dimensions like we do on paper. Tessellating them so you'd get patterns that pass through edges. A different Karnaugh map for each output variable.

The truth table would be sourced from a CSV file, with the top row dedicated to "IN"/"OUT" and the second-top row dedicated to "A", "B", "C", "Q1", "Q2", etc. Time dependant variables "A(t-1)" should be natively supported.

>>1380776
The former is far simpler.

>>1380790
Yes the shape fitting is probably not something easily done, but I can't get my head around simplifying an algebraic expression by code.

>>1380803
Sorry, I'm an idiot and just thought of going from truth table to non-simplified expression, which is simple, but simplification is of course quite complicated.
I think the Karnaugh map route might be easier, but I'm having trouble coming up with an actually efficient way to do it. Perhaps something like BFS from a 1 to create the largest rectangular area of 1s, mark those 1s as 'counted', then repeat for uncounted 1s? Not sure, but maybe something along those lines.

>>1380810
Do you have an actual reason for trying to develop your own logic minimization algorithm, instead of implementing some known computer-friendly algorithm?
To put it differently, is this about testing your ideas, or do you want something useful?

>mfw LEDs, stars, and driver will be here tomorrow
>but lenses and thermal compound still in china
;_;

>>1380810
Best I can think of is define some sort of string-based (or arbitrary length list-based) variable type with a series of functions created to explicitly deal with logical expressions. Without "X"s it's pretty easy, but with them it's all got to be custom. But with the Karnaugh map method I think it shouldn't be too difficult to brute-force a pattern-finding algorithm, then subtract the variables accounted-for after each iteration from a list of them all. Though wrapping my head around a "2x2x2x2..." matrix might take a little work.

>>1380936
Well only in that making something yourself is nicer than using something someone else has made. Logisim sort of does this, but it requires a (legacy?) Java version that installing alongside my current install can cause some issues so I leave it uninstalled, and I've not searched about for existing software otherwise. Having the flexibility to make (or generate) a truth table in a spreadsheet program and export it is something that I assume doesn't come with the standard logical expression simplifier. Either way I guess it's a fun project?

>>1380950
Looking at if if the map is only 2 wide in each direction I don't need to bother with Gray code and symmetry, since the neighbour to the left is the same as the neighbour to the right; I only have one neighbour for each input variable. So for SOP I'll make a list of each Q=1, like [["A!", "B!", "C!"], ["A!", "B", "C!"], etc.], then find the neighbours and see if they're Q=1s or Q=Xs. The ones that have no neighbours for which Q=1 or Q=X will be put into the L1 pile. Then I'll put all the ones that do have neighbours that fit into that category and pair them up with their first 1/X neighbour. Then I'll find out which of those have no pair of neighbours they can match to and put them in the L2 pile, and match up ones that do work, etc. Then I can just combine the L1, L2, L3, etc. piles into the final already simplified expression. And do something similar, possibly simultaneously for the POS expression.

anyone know of a parallel only lithium charging board? most everything i find is for parallel charging for battery packs, but im trying to build a multi battery charger so it would be dumb to have to have every slot populated to charge. pic related is basically what im looking to do.

>>1381244
*everything i find is for SERIES charging

>>1381244
You don't need balancing lines, but rather just use a microcontroller to ensure you get the correct dV/dt at any one time. The custom controller should be able to infer the composite mAh capacity of any arbitrary parallel bank by seeing how quickly its cells change voltage with a known current. A few kb of reference data will probably be needed though. I'd look into I/V, V/t, and I/t curves for Li-Ion battery charging. I'm unsure if there are existing charge controllers for this purpose, but if there are they'll only use one line for the entire bank since their + and - should be commoned anyways.

>>1381252
thats waaay more involved that i was trying to go. i think ill just pick up a 10 pack of single cell chargers and just hook them up individually.

I want to have an LED's intensity match the decibel output of a speaker on the same unit (connected to a pi zero).
I haven't picked a speaker yet.
Should I just get a microphone or is there some way to have the LED get the analog output to the speaker and use the intensity from that?
Thanks, and sorry for being a huge noob.

>>1381273
wire the LED in parallel with the speaker and add a resistor in series with the LED? Or if the current throught the speaker is low enough you could even wire the LED directly in series with the speaker.

Never worked with speakers though so let's wait for someone who knows what they're taking about.

>>1381273
it's a three-step problem: turn the audio voltage across the speaker into a peak or average level, convert linear to log (optional?), drive a proportional current through LEDs
part 1 is easily enough accomplished with a resistive voltage divider and a half-wave rectifier charging a capacitor with a resistor across it
part 2 requires a log amp which can be found in most reasonably large op-amp cookbooks. in short, you put a transistor in the feedback loop of an op-amp. you don't need a high-accuracy version or special super-trimmed components for this, cheap stuff would be fine for this wide-tolerance application
part 3 depends on how many LEDs you want to follow the audio level, which could entail anything from just an LED and resistor up to one of thousands of available switch-mode LED driver chips that accept linear or PWM brightness input
OR
since it's a linux device and you can probably monitor the audio out using ALSA, just set up a PWM output on the pi, read buffers from the monitoring channel, get the maximum of all the samples, lin->log in code, follow with a simple weighted-sum low-pass filter, and update the PWM registers as frequently as you like, which output goes to the input of your choice of part 3 design

Will this work with the LED moonlight I am building soon or no?

>>1381316

>>1381317
>>1381316
If not, what about this?

>>1381318

>>1381320
that'd work

>>1381326
Even though the power supply is more than 12V? >>1378770 I'm still learning all this stuff...

I have a 12V led strip that pulls like 4A, 48W. I found only pic related thats around the power range it can support the leds, but the voltages are off on this one. And i cant seem to find 12V supplies in the 50W range. Can i use this one?

>>1381351
nope, those dimmers are meant for constant-voltage strips. you'd have the same trouble here as if you tried to switch the loads. you can get drivers that have dimming support though

>>1381364
LEDs need a current-limiting device of some sort to avoid overcurrent and destruction. that means EITHER a resistor and a constant-voltage "power supply", OR a constant-current "driver". trying to use both will just waste power in the resistors
LED strips have the resistors built-in so that they can be run from a constant-voltage power supply (e.g. a vehicle battery) so that you can connect as many as needed in parallel. drivers are designed to deliver a constant current to a series string of some number of a particular power class of LEDs
what you need is a 12V power supply. there are 42000 of them on ali. for example
https://www.aliexpress.com/item/100percent-tested-on-animals/32698541940.html

>>1381372
>LEDs need a current-limiting device
Right, i realized that im retarded and deleted the post since.

>>1381372
Sounds like I should have got a different power supply or maybe not, who knows. Everything but the lenses and plaster arrives tomorrow. Still need to get the aluminum to mount them to anyways. Will be a fun project but I am getting anxious and just want it all to be here.

>>1381378
that depends on whether coarse dimming by just bypassing lights will be enough, and if you'll really want dimming everywhere that can't be accomplished by e.g. interposing frosted acrylic or changing distance/angle of the lights
it's fair to consider these an experiment. you can always buy more later!

>>1381383
Good point. I might not need to dim the lights with the lenses anyways. In my design I will be using the proper angle of lens so there should be little to no spillover. I probably will buy some more in the near future. Those two things I saw I found on Wish. Just discovered that site, same with Aliexpress, and just feel like burning money lol.

>>1381386
financial leakage is a known hazard of this hobby, yes

>>1381389
Just like my other hobby, so they should go well together.

>>1381258
Chances are you don't need to do that, since most USB battery banks use multiple cells in parallel. It should be settable by changing the value of a (shunt?) resistor, provided you're using the same amount of cells each time. But have a look around to see if models that evaluate the dV/dt do exist.

>>1381386
>Wish
I keep getting ads for their bullshit, but do they have electronic components any better than Ali or ebay? I'd have to make an account to find out.

>>1381435
I think it is the same exact shit from Ali, but they use your global search history to target you with products you might want. Shipping might be slightly cheaper, but it all still comes from China from what I've seen.

>>1381455
I guess I'll stick to ali for bargains because I can't read chinese to go onto the real sites, and ebay for obscure geiger tubes or other obsolete components.

>>1381460
Wish is in English lol sometimes Chinglish on par with Ali

There are so many types of resistors and capacitors, but I just want to get into making circuits, so I don't know what specifically I need.
Should I just get something like pic related?
I'm at a bit of a loss on what to buy to get started.
Thanks.

>>1381506
that kit is only useful for learning the very basics of lolduino interfacing
>get into making circuits
electronics is an engineering discipline. you start with a goal and work your way to a solution
if you don't know what you need, you don't know what you're doing. choose a project, buy components according to its bill of materials (with some extras), build project. now you have some extras. pick another project and do the same.eventually you'll know what kinds of resistors/capacitors/other items you'll be using and can build a more comprehensive, prospective stock
meanwhile, read a book or two from the beginners' books section in the OP

>>1381489
I was more referring to the direct-from-china sites that apparently have better prices.

>>1381506
This kit is actually pretty good. You'll learn about shift registers which helps you in driving 7-segement displays. I'd buy one with an Arduino.

>>1381510
Not sure why I hadn't thought of that, thanks a bunch anon. Yeah, I'm about two thirds into Mims. Thanks again for the advice.

Hiya, /k/ here.
Rat ate my chickens, so I'm working on a rat trap, hoping you guys can help.
Basically I need something that can generate upwards of 300,000 volts DC, and a large bank of capacitors that can hold a 300,000 volt charge for extended periods of time (being constantly topped up)
I'm not sure how I feel about putting a bunch of lower voltage capacitors in series, as dumping all that current through them at once could damage them.
All help is appreciated, I'll make sure to post a video here for you guys once it's done.

>>1381521
Might just make a 6x10 grid of these or something, it'll work for tests at least

>>1381519
a point I missed: the extra stock you accumulate will be the stuff that *you* regularly use in your specialties. someone who does mostly analog audio stuff isn't likely to need shift registers in their inventory. someone who specializes in extra-low-power microcontroller platforms probably won't need high-current inductors

>>1381521
consider an electric fence low to the ground which you would just charge to a few thousand volts several times per second, using a circuit similar to this https://www.homemade-circuits.com/homemade-fence-charger-energizer/

>>1381522
>1000pF
not enough joules to do much

>>1381506
>Resistors
1/4W, 5%, with resistance ranging from ~100Ohms to 1M.
>Capacitors
Ceramic capacitors in pico to nanofarad range, film for nanofarad range (usually more expensive than ceramic), and aluminum capacitors for microfarad range. Technically the capacitor you want to get depends on what you will be using them for, but the ones mentioned are alright for general use.
>BJT transistors
2N3094 (NPN) and 2N3096 (PNP) are very common.
>Opamps
LM324 and LM741 are alright for general use.
>LEDs
5mm or 3mm are common for visual feedback.
>Electromechanical components
Potentiometers, 10k, 100k and 1M good to have. Trimmers, 1k, 10k and 100k. Buttons and switches.
>Microcontroller
Arduino, very popular and easy to use.
>Other
A proper breadboard, get multiple ones or a large ones if possible, its always nice to have some extra space. Also breadboard wires.

Naturally there is a lot more that is handy to have, but with this you should be able to get your hands dirty on the basics at least. This is from a noob, and I am sure there are someone out there reading this that have way better advice.

To be fair, >>1381510
is probably the best idea. Hoarding components is a curse.

>>1381525
Oh and I forgot diodes, 1N914's are very cheap and great to have lying around.

>>1381522
>$120 of tiny caps
Shouldn't need that much capacitance, but if you do you can always make your own capacitor. Alfoil + 1cm melted wax in a bunch of layers will work, 1cm will hold a minimum of 500,000V, possibly up to 800,000V. Cast the whole thing in a plastic container, preferably with square corners, making sure to leave 1cm of wax on all sides. Ideally you'd parallel the lot like how the plates of an SMD ceramic are arranged, but casting that and ensuring proper electrical contact might prove a little difficult, in which case leading the alternating layers of foil out each side and tying them together externally could work fine. Watch that the output terminals are greater than 10cm apart though, air isn't that great a dielectric. Calculated with the volume between the plates only, your cap would have an energy density of 24,000J/m^3 at max charge.

Thrift store for old candles.

>>1381525
>Hoarding components is a curse.
ehh...
>2N3094
you mean 2N3904/2N3906
>LM741
I have not found a circuit that made good use of the 741's compensation facilities. I would recommend LM358s over LM741s for all-around applications
>resistance ranging from ~100Ohms
I'd throw some 1.0 ohm 1/2W into the mix, maybe some sub-ohms 1/2W, which are handy for measuring power consumption from time to time
>breadboard wires
kek, get a wire stripper and a 100' spool of 24AWG single-core, and you'll always have just the right length

>>1381524
>charge to a few thousand volts several times per second
For my purposes I need something in the hundreds of thousands of volts, it needs to jump an air gap of around 4.5 inches. The idea being that I make the gap five inches, so if something over half an inch tall, like a mouse enters, it'll automatically discharge. If a rat enters, it'll do the same. I just wanna see them pop on a slo-motion camera.
>not enough joules to do much
I was thinking something more like pic related, but the higher the voltage I can find, the smaller the capacitor.
Five or so capacitors at the size in pic related would be good, but if I can only get them in 3kv then I'd need at least a hundred.
Thus, I'd instead use more capacitors like in >>1381522, in an array of, say, 6 in series, to get the voltage, but then run them 10-20x wide, to get the current/capacity.
>>1381528
Those numbers are sounding much better than what I was looking for, but I didn't think anything homemade would have even a fraction the capacity of anything commercial.

>>1381536
>I didn't think anything homemade would have even a fraction the capacity of anything commercial.
That's cause low voltage means really high surface area for the same energy density, which isn't really possible for a home-gamer. But when it's gaps of 1cm, and since the energy density is approximately a function solely of

>>1381537
cont.
solely of volume, dielectric strength, and dielectric constant of the chosen dielectric, it becomes quite plausible to make your own caps. I've been considering using it for cheaper than lead-acid in J/$ of storage, but it requires high voltages in a way that it's just a pain to run anything practical off them. Assuming having a 10^-4 times the energy density isn't a problem, that is. Paraffin wax probably isn't the best cheap dielectric either, but it is up there with polyethylene.

So I want to make a one key keyboard for my foot. I'll use it as mod key for i3wm.
Where do I get started? I'm familiar with c and have used a pressure sensor in a tiny electronics project. How do I connect it to my computer? How do I write the driver (just Linux will do)

what would be an adequate replacement for the PH3330L MOSFET? i need to replace the ones on my old gpu and there are none on aliexpress, so if anyone can recommend me one with the same characteristics i would be very thankful

Here's the 3330L datasheet:
http://prom-electric.ru/media/PH3330L.pdf

>>1381554
Pressure sensor? It just needs to be an on/off button, and not an analogue pressure sensor, right? You'll want something with low-force required, but a solid enough detent for your foot to feel whether it's on or off. No clue where to get them, but there are some old (bakelite?) microswitches with variable spring strengths at my local shop. If you can't find one of those then you might be able to whip up something with a variable lever-length on a microswitch, or otherwise a tuneable spring. If the fabrication is not up your alley, then there are plenty of pedal/foot switches out there, though the power switches (for a bandsaw or a sewing machine) probably require too much force, so look into guitar pedals.

As for the electronics side, easiest way is to get a broken/2nd hand joystick or keyboard, salvage the circuitry and bind a convenient key to it that you also hardwire the switch to. A joystick likely won't work natively with your distro, if my experience with macOS is anything to go by. If you don't want the project to look like a dog's breakfast, head over to /g/mkg/ and see what kind of custom keyboard drivers they use, probably the better choice regardless of whether you've an old keyboard lying about anyways.

t. thought about making a custom joystick once, certainly not an expert

>>1381536
>For my purposes I need something in the hundreds of thousands of volts, it needs to jump an air gap of around 4.5 inches. The idea being that I make the gap five inches, so if something over half an inch tall, like a mouse enters, it'll automatically discharge
RUBBER BOOTS IN MOTION

>>1381572
>guitar pedals
those switches are pretty stiff, if you're not wearing shoes or slippers with hard soles it won't be comfortable at all, not to mention the clunkyness

>>1381536
Is spraying an easily ionised gas with a motion sensor to get the spark voltage down a good idea? Once it starts arcing the voltage can drop significantly, so perhaps a small coil to cause a voltage spike could decrease your required maximum capacitor voltage.
>sprays helium at rat
>rat asphyxiates

>>1381577
Damn, I thought so.

>>1381583
I'm after automatic discharge without any mechanical warning whatsoever, the very first action in the order of operations being the spark forming, so on-the-spot gassing won't work. Anyway, flooding the chamber with hydrogen comes later, then the spark will be used as dual purpose.
All dead rats get potato-cannoned into a brick wall, just to be doubly sure they're dead.
Plus, then my chickens can eat pre-cooked rat. No Pathogens!

>>1381530
>you mean 2N3904/2N3906
Yes, my bad
>kek, get a wire stripper and a 100' spool of 24AWG single-core, and you'll always have just the right length
You are right, I failed to mention it

>>1381618
Your eggs will taste of wrathful blood.

Im starting in electronics and as a first proyect I want to make a
function signal generator
(http://ucdesignplay.blogspot.mx/2011/11/signal-generator-sa-icl8038ccpd.html)),
already ordered a integrated circuit ICL 8038. But there are some
things I dont understand abot the schematic.

1 (red).- Are those trimmers?
2 (blue).- What does that symbol mean?
3 (green).- Can I use a 1n4007 diode instead of 1n4148?

>>1381807
4.- I'm using a bridge rectifier 2w10 2A, should I change it?
5.- Is it necessary to use a l7912?

>>1381807
1. yes
2. ground (this circuit requires a split supply)
3. looks good

>>1381808
4. it's more than enough, you'll be fine
5. yes, that's the split supply

>>1381808
>Is it necessary to use a l7912
I'd say so, unless you have another negative 12V regulator lying about.

>>1381808
>5.- Is it necessary to use a l7912?

not really. the circuit will work essentially the same with or without regulators. that is, you can remove both the 7812 and 7912 as long as your rectified voltage is lower than +-18V, coz the 8038 has a max rating of 36 volts.

>>1381810
>>1381811
thanks, I'll post a picture when I finish.

>>1381831
he should do the math if he wants to try that. a high line voltage and low load could well result in voltages above 18, if the transformer is designed to give 12 in a similar circuit at max load and min line.

LED parts arrived today. I expected the driver to have a plug on it... Do I need to use the plaster to attach the LESs to the stars or just solder them?

>>1381872
LEDs*
Also, what should I use for the plug? PC power cord or something? Slightly irritated because I paid two shipping charges and it all arrived in one package.

>>1381874
Oh wait, I didn't, nevermind.

>>1381872
>solder
you probably don't have the machinery or supplies to solder the LED's center slug to the board. use a dab of plaster on the LED center slug, smear firmly to cover most of the center pad on the board and get as close to flat as practical, then solder the leads
>plug
that would work, better to solder it to an inlet taken from an old PC power supply or other office-type equipment e.g. Pic related

>>1381885
Right, I meant solder the leads not the center. Although I see they are metal and the star's center is metal too. Good idea with the inlet. I can harvest one from an old PSU I have. So with this driver, can I use it to just test one LED? It says output is DC36-66V, current 300mA (constant).

>>1381894
not a LED expert so defer to anyone with actual experience, but i am familiar with the typical LED driver topology. you can safely run just one LED. your only risk is that the driver's on-time gets so low that it loses stability and you don't get a consistent current.

>>1381894
the stars are typically made of an aluminum core with a thin sheet of copper-clad glass-epoxy on top
>one LED
nah, I'd use a diode tester on a multimeter to make sure the LEDs are alright (of course you'll get a dim glow instead of full brightness), then just wire the correct number of LEDs up for testing by tack-soldering thin wires onto the terminal pads

Has anyone worked on ancient electronics? I have a thing I'm repairing that has a 50s era cloth wrapped electrical cord. It just needs some minor work, but I'm not sure what this fuzzy material around the leads are. I'm worried I know exactly what this is and pondering if i should dump into the nearest incinerator and forget i left the oven on. I'll just rewire it with modern stuff instead of salvaging this cord if its that which shant be named. .

>>1381911
Is it safe to use less LEDs than the voltage produced by the driver, like one or five instead of 10?

>>1381960
Dispose of it properly if you think it is asbestos.

>>1381960
old fabric cords weren't particularly durable insulators. even if they are just nylon (which you can confirm via flame test), you should replace it with something more modern

>>1381966
well, >>1381897 says it should be okay but I personally wouldn't trust it to behave well at output voltages below design spec
>if only because I've toasted LED boost-mode drivers by running them at string voltages below Vin, derp

>>1381969
I don't have all the parts I need yet anyways so I might as well play the waiting game. Last time I ordered something from China I forgot I did until I actually received it like 2 months later. ePacket tracking is horrible.
>HuNanShengZhangShaShi guojieyuanqu Despatch from Sorting Center

I have a low frequency sine wave that I'd like to phase shift by 90 degrees. I thought capacitors were supposed to phase shift signals by 90 degrees but I see virtually no change in phase between the input and output. There is a very slight delay but no more than a degree or two.

I don't know of any other way to phase shift besides an inverting amplifier but those always phase shift by 180 degrees and I want 90.

What am I doing wrong? Did I not pick the right size capacitor or something?

>>1381807
>Can I use a 1n4007 diode instead of 1n4148?

I wouldn't recommend it. It's not going to destroy the circuit or anything but it may cause some weird problems. The 1N4007 is a power diode intended for low frequency mains power. As such it'll have a long reverse recovery time and a high junction capacitence. The recovery time for the 1N4007 isn't even spec'd on the datasheet though it's probably in 10s of microseconds. The 1N4148 is small signal switching diode with a recovery time of 4ns. It also has a junction capcitence about 4 times less than the 1N4007.

Okay, why does this all matter? You are building a signal generator and improper component choice can roll your frequency response off prematurely. I assume the IC probably does most everything but parameters are often set by components external to the chip and I don't know what role that diode plays in that. If it does turn out that that diode affects frequency response then by using a slow rectifier diode you will have a hard time generating square waves above a few kilohertz because the rise and fall times will be too fast for the diode to respond to. If the speed of the diode doesn't turn out to play a role in the overall frequency response of the circuit the capacitence might cause different undesirable effects such as ringing on any fast edges.

It could be you luck out and the diode selection is not critical to anything but you shouldn't assume that or you'll be pulling your hair out trying to figure out why your circuit isn't quite working right. Buy yourself some 1N4148s or 1N914s. They're jelly bean small signal diodes and they're super cheap. Get a few hundred, will run you like $5. And by all means put the 1N4007 in circuit and see how it behaves on different frequency ranges and waveform types. Swap it out with a 1N4148 and see if anything changes. That's how you learn shit.

>>1381977
>sitting on the tarmac at the airport waiting for a cargo flight with enough room for a container Despatch from Sorting Center
iktf

>>1382064
you'll need a PLL probably

>>1382077
looks like the diode is only there to keep the adjustment inputs off of the positive rail

NEW THREAD
>>1382101
>>1382101
>>1382101
NEW THREAD

>>1382064
You can do the phase shift with a capacitor, or more accurately, with an RC low pass filter. The thing is that you'll never get exactly 90 degree phase shift, but you'll get closer and closer as you lower the cut-off frequency. Yeah, this means increasing (and significant) attenuation as well.
If your frequency is constant, it's better idea to put two RC filters in series, each one shifting 45 degrees. You can use the normal f=1/(2*pi*R*C) to determine the proper values for R and C.

Both of these suck if your frequency varies. If you want to cover a reasonably wide range, you need fancier phase shifters. Pic related is one option (for audio, IIRC for somewhat limited range). You can also use LC filters (lead-lag filters, producing two outputs having a relative phase shift of 90 degrees over an octave or so), if you use simulated inductors for the coils.

Then there's that PLL approach the other anon mentioned and since it's 2018 now, you can also use DSP-like solutions.

>>1381572
Thanks, will look into them

>>1382745
Actually a mechanical keyboard switch may be the best route to go since they come in a variety of force ratings, provided you can mount it in a pedal enclosure properly.

>>1383059
That depends how he wants to use it, if he's looking for something like a pedal then "normal" keyboard switches will probably be way too light and hunting down rare super-heavy variants is probably more hassle than just finding a heavy microswitch.
On the other hand maybe a Box Navy could work, as it's on the heavy end for "normal" switches (80g) and remarkably tactile and clicky.

Also Kailh Big switches could possibly be a fitting meme

>>1383065
Depends on what kind of arm you use to depress it. Torque = F*r after all. It seems he wants to press it as a modifier key for GNU+linux stuff, so I imagine he want's something light but with a strong detent so he can use it often over a long time. A keyboard switch sounds fine for that, at least if a microswitch has not enough of a detent. Her hardly wants a guitar pedal or treadle switch.

>>1382745
also new thread:
>>1382103