/diy/ - Do It Yourself

>>1466111
this thread's digits brought to you by the venerable TDA6111 video amplifier

>>1466111
Are Chinese exporters not allowed to position the E "correctly"? Doesn't seem like they care too much about regulation anyhow

The

>>1466123
I think it's just their way of faking the CE

>>1466123
>>1466124
It's just bullshit. There's no "China export" mark.
There are plenty of Chinese who slap (correctly or incorrectly drawn) CE marks on their shitty and dangerous devices, though.

>>1466131
obviously it stands for Caveat Emptor

>>1466146
>Caveat Emptor

>nods head, good one...

>Resistance reflection rule: resistance looking into the base is (B + 1) times total emitter resistance.

What is meant by the phrase "looking into the base"?

>CE mark

Anyone who was even serious about this would use the correct distinction.

Also we get a lot of China junk with obvious engrish and design choices that definitely don't fall within the guidelines but bare the Conformité Européenne marking.

I mean, who is going to do anything about it?

>step on paper bag
>14 tin-plated copper pins pierce the top of the bag
>two enter my foot
>lift foot up and put bag away
One step closer to having a tidy workspace.

>>1466290
>What is meant by the phrase "looking into the base"?

say you were biasing the base with 2 resistors, the lower resistor would have an apparent resistance equal to (Re times Beta) in parallel with it, which you'd need to take into account.

How much power do you think this cunt needs? It's 3V and i wonder about the current to see if it would be at all feasible to power this from a battery without having to recharge it every day

>>1466368
Look at the datasheet for both ICs, it will say there somewhere. Good luck ctrl-f-ing your way through a few hundred pages of microcontroller addressing guides though. But by the looks of that thick trace it's more than a dozen mA.

>>1466370
well based on the shitty sheets it should be like 200ma when it uses wifi, but 0.5ma when it doesn't so theoretically a 9V battery with a buck converter should be able to power it forever since i only need wifi connection once every 10 minutes for one second

>>1466379
9V batteries are surprisingly low in capacity

>>1466379
9v a shit. 3xAA would be better.

How long do you want to run on a charge?

>>1466379
That's an average of 833µA. A small (~100mAh) lipo might be a better idea (~120 hours of life) than a 9V, depending on the q-current of the protection circuitry. Running it with a linear regulator might even happen to be more efficient. Add a monocrystalline solar to charge off ambient light and you've got yourself a near-immortal piece of kit. Those calculator solar panels will supply more than enough current for this thing to run (with a buffer) and will charge just fine off whatever lighting happens to be in the room.

You could probably even go for a supercapacitor, though you'd need a buck-boost circuit to get the right voltage to the ICs.

>>1466391
two weeks at least

>>1466543
For the supply I would use 3x AAA or AA cells (either AlMn or NiMH) and a 3.0V or 3.3V low drop regulator that has no minimum load current (which excludes the x17-type regulators) and not more than 0.1mA idle current.

This fucker broke down on me 5 days before a party, plugged my multimeter in to the coil thing's outputs and that seems to work, unplugged every connector and plugged it back in, absolutely nothing, doesn't even power on.

What the fuck could've happen to it overnight?

>>1466579
i only have lm317 at hand, but i would rather use buck converter instead of regulator since it doesn't waste precious ampere hours by making heat

>>1466671
You do not need to optimize the supply if you do not want to.

>>1466671
bear in mind that your typical buck converter can actually waste more power than a linear regulator when your load is small, so you may actually make things worse. there are special PFM bucks that fix this but you won't find those on breakout boards.

>>1466667
>What the fuck could've happen to it overnight?

everything dies eventually. but it should cause no worry whatsoever coz you'd previously gone to the thrift store and bought some old 250W stereos for $10 a piece which you can now use as backups in case of such an eventuality. right?

it's obvious you're in over your head, so you wont be able to repair it except by some miracle. anyway, check the fuse or fuses for continuity. (that's the ohm's scale; should read 0 ohms, just like when you short the probes together. make sure the power cord is unplugged when you do this.)

>>1466667
If the power rails are correct and there's no visible defects, whatever is wrong with it probably can't be diagnosed without a) an oscilloscope and b) a basic understanding of electronic circuits.

I've read that vacuum tubes operate similarly to JFETs. Could a constant current source as in pic related be made with a tube in place of a JFET?

Asking for a friend.

>>1466404
>whatever lighting happens to be in the room.
this would be a checkmate then
because the device i'm making will be a wifi module for a mouse trap which will be placed in pitch dark basements

i think that maybe i could completely skip a voltage regulator and assemble several batteries in series in such a way where their voltages give me exactly 3V
Is there any problem when connecting coin batteries with AA batteries in series?

Is this cascode diff amp properly biased? Output is single ended between Q3 and Q5. Is it supposed to be DC biased so high though? With 24k resistors instead of the active load formed by Q4 and Q5 I was able to get the output bias voltage to (VCC+Vtail)/2 (or about 13.5V) but I can't do it any more.

>>1466943
Well you might be able to get away with RF energy harvesting, if you so desire.

The problem with just having multiple batteries to make the desired voltage is that they have a range of voltages. AA cells for starters come in a variety of different chemistries, ranging from 1.2V to 1.5V, but as they discharge they will get down to 1.1V and 1.3V respectively, meaning the total voltage of 2 cells can vary from 3V to 2.6V. If your load draws higher current every now and then (which it will) than the voltage will drop here also, and a 0.2V drop for 0.5A would not be surprising. Using a supercapacitor to buffer this may help however.

I'd check the voltage ranges on your components to see how generous they are, but chances are they won't like dropping to 2.5V at all. Running it off 3 ~1.5V cells in series with a low-dropout linear regulator would hardly take up much more space, and would last a bit longer. A single lithium ion/polymer cell has a minimum voltage of 3.2V, which is also usable to this end.

>>1466925
It's possible, but not very common.

Any opinions on the T-12 clones on Ali? Seems to be a ton of them, but the 3.0 controller firmware seems to be the newest, so is that the one to get?

Also what is the difference between the 907 and the 9501 handle? Last one seems much longer for some reason.

also inb4: endless TS100 bitching.

>>1466975
The TS-80 is what's hot now-days, but personally I'd lean further towards one of those T-12s. Planned on making a spreadsheet about soldering iron tips, their prices and shapes, and the irons/stations that use them (maybe a pivot table), but I'm not exactly the most knowledgable person for the job. I've just got a cheap soldering iron with a potentiometer temperature knob on the handle, but I swear it's like 100° to cold since it won't reflow lead-free at 320°C. For my next iron I'll go for one with tips that have an integrated element, to reduce thermal resistance.

im trying to measure ac motor current using acs712 x05B (https://www.sparkfun.com/datasheets/BreakoutBoards/0712.pdf).). the current according to datasheet has to be ~3A, my multimeter also measures 3A.
for some reason, my acs712 shows me 6A peak (or ~4.3A rms).
wtf am i doing wrong, anons? i tried the same sensor with the exact same setup/ connections on a dc load and everything works (3A dc current registers as 3A). as soon as i switch to ac i get shit again.

>>1466671
Buck converter will waste much more into heat while the ESP is in sleep than an LDO will.

>>1467012
How are you measuring the output voltage?

Holy shit normies are so out of touch with reality.
>dads friends comes by
>he runs a small company that like helps old people and shit when they fallen and can't get up
>wants to expand his services by installing devices into his clients homes to detect if they fallen and can't get up
>got a quote from electronics company that does this kind of shit
>the number almost gave him multiple strokes
>"Hey anon you do like web design for a living?"
>i in fact like do like to do like a web design for a living
>here comes the boomer tech logic
>so since you know web design can you design, program and design ideal deployment for this fall'n'up system for cheap? apparently other companies use motion tracking for this or some shit
>...
>keeps looking at me
>...oh he wants me to get a design sketch, pricing and time tables right now of the top of my head...
>fuck it, i know what a resistor is, i can do this
>fire up aliexpress
>pic related
>motion sensor in each room sens signal over the wifi module to arduino when it detects movement, the arduino it self is connected to a wifi router, and if it doesn't detect input from any sensor within 60 minutes it DOESN'T do a GET request to a web server which is monitored by a dispatcher that sends the amberlamps there if no input is detected, whole system will cost like $20 from ping pong parts instead of $1000 from the retail jew
>boomer dude loves it, tells me to make a prototype happen, offers decent compensation
>agree to do it
so it should work in theory, but how reliable will it be? do you have any idea if this basic design can be improved? are there any flaws?
the downside here is that if the system fails i could become an unintentional serial killer over night

>>1467049
why not one wearable motion detector instead?

>>1467049
Wanna know why all the others are expensive? Because you cant guarantee that the system went fail, and when it does and some old geeser lies on the floor for 2 days your buddy is gonna get sued.

>>1467049
don't waste your time and money, he's never going to pay you. He was just trying to be polite.

How can I monitor the temperature of Peltier Modules while they are sandwiched between a massive heatsink and hotplate? Can I simply measure the voltage/amps and calculate based on the datasheet then use that the built a circuit? Or should I slip in some type of Resistive Temperature Detector between the hotplate and Peltier Module? All I really want it to use the Peltier Module to power its own heatsink fan and an alarm buzzer if it is getting too hot using the least amount of electronics possible. Any left over power will go to charging other stuff.

Specs on the Peltier Modules I'm looking at state, "50C," as their max temp.

>>1467049
Yeah, but you need to charge him $1k per unit, because of your expertise in designing and making it. If you don't, you are a fucking moron.

>>1467049
>cheapass asks for anything
It will bite in you in the ass hard.

Product safety, certificates, false positive handling, quality, reliability.

>do you have any idea if this basic design can be improved?
Do not use PIR sensors. For something like detecting a lying person it's complete and utter shit sensor for such task.

Best approach would be to use camera and visual identification. Identify floor, identify lying person, identify movement etc.
Or use like other anon >>1467053 suggested a wearable option.
Reason why from actual companies shit costs so much is because they have to review possible risks, events, applications and build a system that will be capable to deal with it as well as take responsibility for their product.

Now consider following scenarios
Geezer sits down in front of tv, falls asleep. Your sensor detects no motion, triggers alarm. People from that company call, nobody answers. They decide to come and check and they are OK. Whos responsible? You.
You install your system, you also install IP cams. How many geezers will accept IP cams? Will his company hire people to actually monitor IP cams every time your sensor goes off? He is expanding business? Wow, he will need more people for monitoring.
He goes to sleep on low level bed? Your semi advanced motion and position sensor tracks it as falling, trips alarm, geezer is sleeping, again nobody answers and his people come. Soon you get a called with death threats.
Seems like a nice way to lose balls.

My suggestion from experience in designing building monitoring systems - at least use some advanced sensors, like Omron D6T or OKAO Vision.
Other options would be to use RF for human tracking but pretty much anything reliable is outside of one man team scope

>>1467071
your thermistor doesn't have to be sandwiched between, it can just go on the heatsink.

>>1467100
I don't have a datasheet for the heatsinks and their transfer rate. so, I won't be able to calculate the hot side of the heatsink from the cold side temp.

>>1467097
>Will his company hire people to actually monitor IP cams
teleiophiles would work for free, or maybe even pay to spy on elderly people showering. start a side business "GILF cams" make millions.
just make sure granny signs the contract, and initials the disclaimer.

>>1467102
https://www.sparkfun.com/tutorials/314

>>1467117
Thanks a bunch. My biggest problem is that the hotplate will be exposed to random levels of wood-fueled fire at worst. So 600C at worst. I'll probably need to get a laser temp gun and do some lengthy experiments with the hotplate first.

>>1466975
I got that one. Case wasnt grounded, and power connector had broke off the board.

The screen was crooked.

Despite all that i was happy at first. Then i tried tp change the tip and it wouldn't recognize it

>>1467034
im using arduinos adc. i sample it every ~200us, 100 points (this roughly equates to 1 cycle, im on 50hz mains).
as mentioned earlier, i dont think the problem is with wiring or code, since everything works correctly with dc load.

>>1467153
So you're performing a true RMS calculation? You're subtracting the Vcc/2 offset before squaring, right?

>>1467178
yes, i want to get rms value (not exactly true, since the waveform is not an exact sinusoid). my calculations are basically the same as here: http://henrysbench.capnfatz.com/henrys-bench/arduino-current-measurements/acs712-arduino-ac-current-tutorial/
to be more exact: i take my measured values, find min and max values, then subtract 511 (acs712 gives 2.5V @0A). i then convert these to voltage (*5/1024). then i convert this to amps by dividing by acs712 sensitivity (/0.185) to get peak current in both directions. then i divide these by sqrt(2) to get rms values.
i tried doing the calculations as given in the example, same result (6A peak, 4.3A rms).

>>1467138
Buyer pays return shipping to China.

>testing a batch of chinese arduinos
>notice one of them has a crystal resonator without the cover

Every now and then I get one in a batch of 50 that will have this defect, yet it will work just fine. Why is it? Do they accidentaly lose the metal cover during the reflow soldering or something like that? Why do they still work?

>>1467186
>Why do they still work?

Because the interference the shield blocks isn't a problem in the manners you are using them.

I've got an old Kenwood amplifier and Yamaha amplified mixing table from the 80s which both happen to have a dead channel. How to I investigate which components are dead? None of them are obviously damaged.

>>1467049
What's wrong with the good old pendant button? Or Google voice API that sends for help when it senses the word "HELP"? At least microphones are slightly less creepy than IP cameras...

why don't inductors attract magnetic objects?

>>1467299
a) the magnetic field isn't strong enough
b) they usually have high-frequency AC current through them such that the magnetic field spends much time with lower amplitude than the peak
c) shielding

The AC part means that it sometimes won't attract a permanent magnet at all, depending on the net current. A microwave transformer is a pretty strong attractor of magnetic objects if you've ever messed about with one of those, but they've got maybe 10A through them and a fair few turns, compare that to a tiny inductor and it's no surprise it isn't acting as an effective electromagnet. The shape of inductors/their cores often deflects magnetic flux back around inside it, like toroids. Electromagnets on the other hand are built to have high flux leakage in a single direction.

>>1467186
they keep bits of dust, insects, and other mass off the crystal. still works, but no guarantees on frequency accuracy or stability should a fly take a crap on it

>>1467299
mine does

>>1467053
too prone to various failures
>>1467097
i am not trying to use machine learning to 3D track the body in space. I don't care if he is lying down, all the sensor has to do is detect movement in the room (including during night), thats it
I will look into the human detection sensors to see if they would be better for simple motion detection i need
>>1467291
yeah i am sure someone collapsed from a sudden stroke will be totally able to string together a single coherent thought

>>1467546
I think there's nothing wrong with a wearable motion detector; an accelerometer will do well. This was you can detect the difference between lying down comfortably and falling down with a shock; the sudden acceleration can be a good indicator. Best it's close to the core of the body with little room to swing around, like in a pocket. Though it would do well to be coupled with another method.

>>1467547
*this way

>>1467547
old people are stupid they would always forget to wear it, don't charge the battery or break it etc

>>1467551
Then it's on them, no insurance needed. Make the thing beep when low on battery. The pendant buttons like pic related certainly have some degree of success anyhow.

>>1467553
I am not a chinese factory though. To make that shit i would have to use an arduino with a wifi module, and to make the battery last at least a month would make that module pretty bulky and heavy

>>1467554
you might be surprised how cheap it can be to get a chinese factory to make your a thing for you in 3- and 4-digit quantities

>>1467554
>arduino with a wifi module
You literally posted an ESP8266 in your opening post, the thing has its own wifi. Just make it take hearing aid batteries or a coin cell. A 433MHz radio module might actually be more power efficient though. Anyway, the battery life will be really long since it will only need to transmit when something bad happens.

Alternatively, you could put an ultrasonic transducer in it such that the receiver can be anywhere in the house at a longer range than wifi.

>>1467559
>transmit when something bad happens
you've never designed anything high-reliability or to do with life safety in your life, have you?

>>1466988
>>1467138
Yeah its so cheap I just ordered one anyway, so we'll see in ~3 weeks how it turned out. Everything is better than my current analogue shitter that takes 10 min to heat up properly and literally squeals from the supply when i dial in anything under 400C.


Another question: They all seems to come with the 'K'-tip as default, is that a decent one? I've only used the more standard C and D types before, so the knife-like shape is new to me. I guess for smaller SMD stuff the thinner edge might be nice, but I mostly do throughhole sized stuff these days, so maybe somthing a bit more massive is better?

>>1467567
I've only heard good things about those chisel tips, you shouldn't be disappointed. Though personally I've never used one. With a conical tip I find I can only get heat into a join by using the slanted side, which since it's on a pretty shallow angle means my iron has to be pretty flat to the board. That chisel should mean you can get something closer to vertical for getting heat into a join, plus it's pretty thick looking so I imagine the heat transfer is pretty quick.

But I wonder how close the element is to the tip, which is a big factor in how quickly the iron's power supply reacts to jamming the iron onto a heat-sinking part.

>>1467567
K tips, maybe those would work well for QFN but I haven't tried
the screwdriver type is traditional and most generally useful
generally, choose the tip that is close to the size of the component or wire

>>1467012
>>1467153
>>1467181
bump.
i went over everything for 5th time probably, everything seems to be fine except for the values i get.

I hate electricity

>>1467600
I don't get
>then subtract 511 (acs712 gives 2.5V @0A)
You shouldn't have to do that, you're only interested in the peak to peak voltage.

>>1467605
>I hate electricity

It appears that you aren't too crazy about photography either.

>>1467623
Yeah was kinda bumpy trying to fuck with the relays at 60mph

>here! Take some spare fuses just in case!
>fuses don’t even fit

>>1467628
fuses are for sissy pussies, real men use thick wires

>>1467181
But if you're sampling that frequently, why not just calculate true RMS?

It's possible that there are transients meaning that the peak value is nonsense. These may not even be actual current transients, but noise coupled into the ADC's input.

>>1467628
Can you please stop using your trip? It doesn't add anything, and it takes me out of the flow of this board. I like reading electronics posts, not electronics posts with names.

>>1467646
What's your entitled snowflake victim index?

>>1467650
Nice buzzwords, old man

>>1467628
that sort of misfortune happens a lot to tripfags, I hear

>>1467650
ask the trip, not anon

>>1467617
subtracting 511 just gives me how far away from center the read value is. but the problem persists even if i use peak-to-peak voltage. both calculations yield the same result, just done in slightly different way.

>>1467681
I thought you might be getting some negative values. I'd try a non-inductive load on AC if you can manage that, otherwise a scope would be nice. Like the other anon said, using just the peak values might cause problems.

>>1467633
That’s a good idea. I just bought that wire assortment the other day, I finally have a use for it. I’ll rip all the fuses out and jam a little U-shaped length of wire in each one.

Thanks for the tip!

>>1467633
Sorry, one more question: will 14AWG be thick enough or should I go 10AWG?

http://www.ladyada.net/make/x0xb0x/
this is probably the most beautiful thing from an EE standpoint I've ever seen
a fully functional, open-source, near 1-to-1 clone of the roland TB-303 Bassline Synthesizer

I have an absolute hard on for this thing and I want to assemble one for myself.
The problem though is that since it tries to be 1:1, it uses all the deprecated extremely rare parts of the original

Apparently even specific models of transistors are specified, though I don't see how it would make a huge effect on the end result.

I just want to know:
1. is it possible for me to find common modern replacements for these parts? It's all generic parts, no custom ICs, save for the atmega chip
2. from a purely functional standpoint, will getting the parts "close enough" to spec really make that huge of a difference?
3. what differences would I expect? I know there might be changes in noise level, which doesn't bother me, but could it cause it to not even function even though the parts are roughly the same?
4. how do I find suitable substitutes for the parts? Scan data sheets and try to get them as close as possible?

I'm still relatively new to electronics engineering and I guess what I want to know is, when replicating a circuit, though not necessarily 1:1, will substituting parts radically change the behavior of the derivative circuit?

>>1467720
http://www.ladyada.net/make/x0xb0x/fab/parts.html

here's the BOM so you all can see what we're working with here

>>1467720
http://errozero.co.uk/acid-machine/

>>1467708
>awg
no idea. normal countries rate their wires for max current and don't use some arbitrary made up scale

>wireless charging seems cool
>order a charger and a charging pad from chinks
>attach to phone
>start charging
>check on it some hours later
>battery only at 50%
>why the fuck is it so slow
>grab the phone
>instantly drop it because it literally signed my fingers
what in the name of fuck, are wireless chargers supposed to generate this much heat?

>>1467733
it's the square-cube law.
The wireless charger works via induction. It's extremely inefficient, meaning it needs to output a *shitload* of power to get any appreciable charging effect in the battery. Where does all the power that isn't used to charge the battery go? It gets converted into heat.
it probably doesn't help that it's a chink charger too

>>1467734
so whats the alternative then?
surely you don't expect me to connect a cable every time i want to charge the phone like some lowly dirty potato farming serf

>>1467735
buy a name brand wireless charger, it's about the only thing you really can do

why would I use a buck converter to step down voltage rather than a boost converter to step it up (or vice versa)?

>>1467776
Often better efficiency due to lower switch currents.
Requires less current from higher voltage source = often easier to feed.
Lower ripple, because inductor is after the switch.
Load can be easily powered off.

>>1467642
>>1467686
yes, i get one positive and one negative peak current doing it my way, or one avareged out current doing the way shown in the link.
pic related is what the graph looks. ill look into calculating true rms from this, but i doubt it will make up the ~50% difference (3A shown on multimeter, while i get 4.3A).

>>1467567
K is my favourite. Also use D24 often.

I have an electret mic that's a tad quiet when plugged directly into my PC unless I jack up the gain in windows settings, which introduces a lot of noise. Idea here is to amplify without contributing any noise so that I don't have to increase the gain in windows settings.

Is an NPN circuit like pic related going to perform better or worse than an lm386 circuit for a mic preamp? Which will have the least noise?
The circuit would be inline with a 3.5mm mic with the output feeding directly into my PC and the circuit will be powered by 5V from USB.

I'd like to make a amp with either the 386 or NPN(s) (or maybe a 741); I don't have any mic amp ICs on hand and don't feel like ordering some for such a simple project.

Working on the lighthouse circuit from earlier thread. The idea was to light up some leds sequentially with a 4017, but linearly ramping the off->on & on->off transitions of the leds with a PWM signal.

So far I've managed to breadboard an oscillator circuit with a 555 + npn/pnp + op-amps that generates a very low-frequency triangle wave signal with an amplitude of ~0.1 to 3.5 V.
Now I'd need to make a PWM circuit with 555/556 to convert the triangle wave signal into a square wave with 0..100 % duty cycle.

Pic is a rough sketch of the circuit that I'm working so far.

>>1467720
1. Yeah
2. Yes
3. Pretty much nothing
4. Yeah

You might have to swap some other parts around to work with your replacements for the really weird stuff like transconductance amplifier. But for the most part, swapping an old transistor out for a new one isn't going to make any practical difference besides saving you money. Just make sure the specs are similar.

>>1467784
>~50%
you sure it's not sqrt(2)-1?

>>1467831
LM386 is a power amp, not a preamp
the 741 is the more correct choice. TL0xx would be better if you have them on hand

>>1467855
you don't need the buffer after the 555, the input impedance of the difference amp will be sufficient not to disturb the oscillator
I suggest two triangle oscillators, one at PWM frequency and one at LFO frequency. feed each to an input of a comparator and feed the comparator output to an LED

>>1467880
>LM386 is a power amp, not a preamp

Even incorrectly used as a preamp I'd argue it's still better than a single transistor solution, and way better than the single transistor solution in his pic.

>>1467831
Buddy, if you're gonna go this route you should at least use voltage divider bias instead of collector feedback bias and put an emitter degeneration resistor so you're gain isn't heavily dependent on the collector current and temperature. You'd be better off with a two transistor solution though, like an NPN PNP feedback pair or a similar pair instead with a N-JFET for the input and a BJT for the output. If you want low output impedance to drive your next stage then you'll probably want a third BJT connected as a source follower as well. If the preamp just feeds an op-amp or the aforementioned LM386, or really anything with an input impedance over 1M then you can get away without it.

On a transistor data sheet is hFE (DC current gain) the same as beta?

>>1467910
Yeah

>>1467880
Doesn't the 741 require +/- biases? I think the 386 has true-zero.

>>1467922
no op amp should need a true ground. their one job is to make both inputs the same
but, the 741 wants 20V minimum rail to rail. scratch that
you should be able to use an LM358/LM324 for this, it only needs 3V minimum rail to rail to run but headroom at that voltage a shit

>>1467943
Older transistor-based op-amps need .7-1.4V +/- headroom on the rails. I think some for a while had odd behavior near the rails (non linear or something), but now you can find plenty of rail-to-rail op-amps.

>>1467922
Just get a bunch of TL072/74s. FET input dual/quad op-amps. Higher input impedance than older bipolar front end op-amps, better slew rate, 13V/us as opposed to 1V/us or less for op-amps like the LM358 and LM741, and just generally faster, lower noise etc. They're cheap as dirt these days. I don't even know why people even bother to recommend stuff like LM358 which by now is practically as dated as the 741. The TL07x series op-amps are old too but at least they still perform well.

TLC2272 is decent if you need rail-to-rail. It's "relatively" cheap and good performance.

>>1467953
read RULE 4 and don't post here again until you understand why your response was non-responsive

>>1466111
Include something about well-lit, focused images in rule 3, otherwise looks pretty good. I don't personally mind the SQT-tier stuff here though.

Also what happens when you use isolation capacitors before op-amp inputs, like a buffer or non-inverting amplifier arrangement but with a capacitor in series with the non-inverting input?

>>1467956
I'm giving him suggestions for op-amps that are actually good. The TL07x series op-amps can do single supply operation but can't truly swing to zero without a dual supply which is why I suggested a TLC2272 if he absolutely NEEDS rail to rail capability. These are both better choices than a 741 or 358 and are also better for a preamp than a 386 which the other anon mentioned is a power amplifier. There was no violation of rule 4, you're just being an autist.

>>1467989
the TLC2272 was the only recommendation that didn't violate his problem statement
>the circuit will be powered by 5V from USB.
>don't feel like ordering some
agreed that just about anything but the usual jellybean op amps would be an improvement over what he plans to use, otoh I'm slightly unusual in that I keep 6001s in my jellybean jar

>>1467963
will do. also thinking of adding something about trips
>what happens when you use isolation capacitors before op-amp inputs
they're high-impedance so they'd charge up very slowly and cause dc drift in the output. but the output coupling cap should differentiate that into a small and decaying dc offset. I would recommend adding a resistor to (virtual) ground or divider to rails, to bias the input and set a definite corner frequency on the high-pass filter you just built

>>1468017
>adding something about trips
Just asking for a flame war, best leave it be. I've got trips blocked and it's really pretty noninvasive. Sometimes I click unhide and I'm quickly reminded why I have them blocked.

I was more wondering whether it would function as a ground rail if you couldn't be bothered using split rails/a buffered voltage divider. I guess it's unlikely since the impedances on the input would hardly be split 50/50 between Vcc and Vee.

>>1466111
>>1466123
>>1466124
The image is bullshit. Both of those will be treated as CE marks, but those don't mean shit anyway

>>1467880
>you sure it's not sqrt(2)-1?
why sqrt(2) - 1?
anyways, i tried calculating the true rms, which yields basically the same result: ~4.2A rms current.
now that i think about it, my cheap ass chineasuem multimeter definitely does not read true rms either, and it shows the correct current of 3A.

>>1468105
what, this? it'd work fine. for impedance analysis, voltage sources can be treated as short circuits. real power supplies try hard enough to be short circuits that they can be treated as such. the source would see R1||R4 across its output at ac, and high resistance at dc

>>1468179
it seemed more like 40% than 50%. just spitballing
anyway, you can only estimate rms from peak if the waveform is a proper sine. that waveform is too irregular. do the actual rms math instead. pretty sure you've got a hardware multiplier, if not, use a look-up table to calc the square

>>1468187
No the point of the capacitor in the original question was to un-reference it to any rails, you've added resistors R1 and R4 turning it into a high-pass filter but I meant to use the op-amps internal input impedance(s) for that instead. But that circuit does describe what I was attempting to achieve. Not sure if it's more efficient component-wise than having a single voltage divider as your ground rail and pulling both the anode of C1 and U1's pin 4 to it with resistors.

How do i wire this thing? Can i phase chop control this?
The common point of the coils goes to neutral, and one of the opposite end of the coil goes to line? Where do i put the starting capacitors?

Im looking to buy this motor, this is the only pics the guy has of it, and he does not know electronics to start it by himself. Apparently i dont know shit either since it feels like too much wires for me.

If i set a voltage regulator to say 5V and then feed to it variable input of 10-20V i always get 5V which is great
but it generates too much garbage heat
so i wonder, i know a buck converter doesn't make heat, but can it also take a variable input? i know that if i feed it stable 20V i can set it to get 10V.
But if i set the buck to 10V output and start changing around the input between say 15V and 30V, will i keep getting 10V on the input without ahving to readjust it every time i change the input voltage?

>>1468221
>i know a buck converter doesn't make heat, but can it also take a variable input?
of course

>>1468222
that is not what i asked, i want to know if i will get identical output voltage when changing the input voltage like in say lm317

>>1468200
>Can i phase chop control this
You mean use a triac-style dimmer on it? I wouldn't, inductive kickback might do some whacky shit. Use a VFD if you want to change the speed, otherwise it will want to run at a certain speed (some multiple of 50/60Hz (converted into rpm)) and will drop below this speed somewhat as you load it up/as it pulls more torque.

Now that does look like a fair few wires, and it doesn't seem to be for 120/240 conversion since it states 220V on the side. Now it might require a run capacitor not a start capacitor, so watch for that. Personally I'd buy it provided it's a decent price for the trust you have in it, since it's not not going to work. Otherwise there's not much to do without knowing chinese or seeing the wires coming out of the thing. Message the bastard, see what happens.

t. never used an induction motor

>>1468224
and the answer is still "of course". they use a pair of resistors to set the output voltage, not entirely unlike the LM317 but with a fourth terminal. a buck or boost converter that doesn't regulate its output would be practically useless. Pic related extracted from the datasheet of the MT3406, a buck converter IC

>>1468224
>>1468227

>>1467784
> pic related is what the graph looks. ill look into calculating true rms from this
That graph has a crest factor (peak/rms) of 1.765, compared to sqrt(2)=1.414 for a sine wave, so peak/sqrt(2) is a substantial over-estimate.

Using I=(n-512)*5.0/1024/0.185, I get peaks of -5.861 and 5.854 with a RMS value of 3.318.

>>1467831

i think you're barking up the wrong tree. if your mic is noisy when you crank up the gain in your sound card it's probably because THE WORLD is noisy when you turn up the gain loud enough to notice it. i.e. you're likely gonna hear the same background noise irrespective of where the pre-amp is, or the quality of your components.

>>1467855

and exactly why does the lighthouse vary in intensity? real light houses just spin the light. they dont dim.

>>1468263
what is the alternative? create and unleash a deadly pandemic which will wipe out 99.9% of all organic life on earth just so you have nothing left to cause noise in your recordings?

can i connect this 10w 12V led directly to 12V battery or do i need some other fuckery around it?

>>1468298
>what is the alternative?

to get a better SNR when you cant quiet the noise, you increase the signal by asking people to speak up, play louder, or sticking the mics right up in their faces and instruments.


>1468300
>do i need some other fuckery around it?

every LED needs fuckery. if it's not included, you gotta add it. in your case, you need a constant current supply rated for the device's current expectations.

>>1468300
No, you need a constant current driver.

>>1468303
so it won't work with just a battery? i find that hard to believe since a tiny 2v led i have works with just a battery a like 5 ohm resistor

>>1468313

sure it'll work with a battery, but if the battery voltage is just tad high, it'll cause irreparable damage, if not death, to your LED.

>>1468314
how about in series with battery lm317?

>>1468313
depends on battery, what internal resistance it has. If it's small and voltage higher than LEDs need it will draw excessive current

>>1468313
those tiny 2V LEDs consume 1/25W. your fuckery will have to be 250x bigger.
get a constant current driver

>>1468303
but the noise could be coming from INSIDE THE HOUSE

>>1467831
first check your mixer settings and see if it's something as simple as the drivers thinking it's a line-in or just not having mic bias switched on

>>1468315
probably not, an LM317 drops too much voltage
get a constant current driver

>>1468317
but lm317 can be used as a constant current driver. i don't mind if it drops some voltage the battery is like 2 volts above anyway.
I also have buck converter if that would be any better

>>1468320

this might work if you can keep the 317 cool enough. figure out the power consumption by multiplying LED rated current by the voltage drop. if i recall, the 317 needs something like a 2.5V differential in worst case, so it might go dim as soon as the battery drops a bit.

>>1468328
i am not very happy about the wasted heat tho, that will rape the battery pretty fast
maybe i could use pwm instead to average the voltage to 12v, that would produce no wastefull heat

>>1468334

that wouldnt work very well coz the average PWM voltage output would be a percentage of the battery voltage, and as the battery voltage changes, so would the average output voltage.

>>1468346
what if i used something like arduino to measure the voltage and adjust the pwm width accordingly?

>>1468352

you shouldnt need the services of a computer to light up an LED. if you're not gonna buy a CC source, then get a chip and make one. like the microchip HV9910B Universal High-Brightness LED Driver, which can take from 8V to 450V on the input.

>>1468383
i don't really want to buy stuff. i want to use what i have and i have lm137s, arduinos, resistors, caps, buck converters and some op amps, but i really hoped i could just use a resistor. i will try with the lm317 for now, it will waste some heat but it is very easy to make and i don't have to buy things

>>1468254
interesting. how come my cheap shitty multimeter shows exactly 3 amps though? its definitely not measuring true rms, like those fancy flukes.

>>1468228
That's a much nicer design than the switchers in Jim's AN25. Are they that easy to implement nowadays?

>>1468405
> its definitely not measuring true rms, like those fancy flukes.
True RMS isn't exactly rocket science. Particularly if it doesn't need to be accurate to 4sf. The thing is, AC voltage/current measurement is almost always going to be an approximation however you do it. Measuring peak and scaling by sqrt(2) is wrong for anything that's not a perfect sine wave, and you're rarely measuring perfect sine waves (mains certainly isn't that). So even a fairly crude "true RMS" measurement is going to be more accurate than peak/sqrt(2).

>>1468442
A lot's changed in 30 years... they are highly integrated and basically plug-n-play now.

>>1468226
Dug up the part number and found a chink datasheet sort of thing. Does that look like a start or a run capacitor? Anyway, ill try to get a VFD.

The guy selling it doesnt know shit, its attached to a machine and i want the whole thing together.

I'm trying to use a Hall sensor (US3881LUA-AAA-000-BU) and the output is being real shitty to me. It's supposed to be open drain. With a 5V supply and 11k pull down on the output, I get .26mV and .5V out. The IC is sinking current at almost the same amount for both of the switch states. The same is true of a 5k and .5k resistor. What gives? It seems like the leakage current is stupid high.

>>1468627
Sick. You know people are working on integrating magnetic materials into IC fab so we can have switchers on a chip, no external inductor?

>>1468386
Then use a buck converter. Modify it with an op-amp or something to measure current and use that as feedback instead of a voltage-based feedback.

>>1468628
Run capacitor I think? I can't see anything resembling a centrifugal switch (which switches off the capacitor when it's up to speed) in the diagram. But since it uses the same sized coil for the capacitor-shifted coil and the normal coil (since one can equally be the other when switched) and there's no common wire inside the motor that both capacitor phase-shifted arrangements go through that could have a centrifugal switch applied I think it's safe to say that it's a run capacitor. In any case, if you buy a run capacitor with the right capacitance and it turns out to need a start capacitor it will still run fine; start capacitors I think are just rated for lower continuous current.

VFDs are kind of expensive, and you might not need one depending on what you're using it for. But some VFDs can output two or three phases, with which you could run the motor without any capacitor, so if you went down that route you could likely get some extra power out of the thing.

>>1468673
I found a couple of chink VFD for 50USD or so, with good ratings. But the chinglish page of the thing sort of makes me not trust it.

>>1468678
Well if it can provide enough current it might be good enough, I'm surprised at the low cost though. Might want to check the input and output filtering personally. Extra cooling may also be needed for the transistors

>>1468320
Pic related shows a basic buck converter modified to constant current output
the same general idea applies to a boost module or any other four-terminal regulator: send the LED's - power return through a resistor chosen to drop the specified feedback voltage at the desired current (caution, if working from a limited stock you may need to connect multiple resistors in series and/or parallel to dissipate power)
for the LM317 solution, the wasted heat will "only" be 1-2W. in this case, the average buck module won't be much better because of the high specified feedback voltage
you need to understand that electronics is not an "I have junk, what do?" craft. a starter kit is not going to get you everywhere

>>1468647
did you try a larger resistor?
and open drain means you pull the output UP

>>1468647
I don't see the physics being favorable for that level of integration, except at the micropower level. caps are still hella expensive to put on a chip

>>1468673
>with an op-amp or something
just a resistor. Pic related

>>1468683
Oh neat. Gotta be a small resistor though I bet, and not exactly adjustable on the fly. Unless you can get current-shunt potentiometers/rheostats, which could be interesting I guess.

>order pcb
>export a list of the components
>add them to my cart
>double check everything
>one month later
>missing a part
literally every time

>>1468685
yes, you're looking at a few ohms at 1/2W-2W dissipation for most watt-rated LEDs, so it's most useful when you have a few of them series-connected. you can Ohm's Law the numbers
no, not adjustable on the fly. you could increase current by placing a larger trimpot in parallel to the sense resistor, as long as the sense resistor can handle the power
there are regulators designed for just this sort of low-side constant-current action with feedback voltages of 100mV or so, where a single 1/4W resistor is just fine for 350mA or 1A strings

>>1468688
>88
EVERY
SINGLE
TIME

>>1468692
Perhaps use a current shunt and amplify it with a variable-gain non-inverting amp, that way you're not wasting a bunch of power in the resistor.

>>1468696
that's a few more components, but that can work, if done with care. it would be wise to observe the output on a scope to ensure that the output is resistant to overshoot and oscillation

>>1468683
My bad, meant pull down. Didn't try larger but wouldn't that be even worse if the leakage current is abnormally high?

I'm currently planning on setting up a comparator to 100mV and getting a digital output from that.

>>1468707
pull up* I'm retarded

>>1468707
oh, nm, it's already digital
are you sure you're using the correct pole to trigger that device? are you using the correct pinout?

Hoping this is the right place to ask, but what's a good LED strip for programmable lighting for a bookshelf/display?

>>1468741
someone might know, but this thread is more geared to component level tinkering than towards modules like led strips so don't be surprised if you don't get an answer.

>>1468697
I think you'd need a rail-to-rail amp or otherwise a second, lower power rail. The latter of which could be easier if you're already splitting rails if the converter module can't handle the full LED voltage.

>>1468628
>Does that look like a start or a run capacitor?

neither. it's used strictly for direction control. you see this all the time in antenna rotator motors.

>>1468751
nah, LM324 and similar bipolar op amps have a common mode range including GND to approx. (V+ - 1.5). read the datasheet

>>1468741
a lot of it depends on how you want to program it. try the chinkshit general /csg/ over on >>>/g/

>>1468754
That's called a run capacitor. A capacitor in one of these motors is used to shift the phase of one of the windings such that it lags behind the non phase-shifted winding. This phase difference causes the magnetic field around the squirrel-cage to appear to rotate, which puts a torque on it. Once the motor is up to speed, all you really need is a single set of coils, and you'll be able to keep the motor spinning and even pull significant torque with it, but if you stall it you'll need to activate the phase-shifted coils again to get it moving. A motor with a run capacitor keeps this second coil running all the time, but a motor with a start capacitor has a centrifugal switch, which causes the coil with the capacitor to switch off once the motor is up to speed. This is often used because the capacitor is somewhat lossy and is in general less efficient than the coils of a proper 3-phase motor would be. Often the capacitor coil of a 2-phase motor like this is smaller and rated for less current if it is not there for direction switching (direction switching can otherwise be achieved just by swapping the polarity of one winding), especially for a motor with a start capacitor as allowing more than the continuous rated current for both this winding and the capacitor isn't a problem for the time it takes the motor to get up to speed.

>>1468386
lm317s and most linear regulators can operate in a constant-current mode. It drops power like normal, but if you can heatsink it it works fine (I was using mine for .5A constant current-ish). Check the datasheet, 20-30mA would be much smaller.

Recently I found a useful circuit, a transconductance amplifier or voltage-to-current op amp circuit. Useful if you want to use variable voltage for a linear mA output. Can be driven right from the op-amp.

https://www.allaboutcircuits.com/textbook/semiconductors/chpt-8/voltage-to-current-signal-conversion/

>>1468741
I mean, if you're going RGB customizable then anything with 3 analog outputs would work. For white you'd only need 1 analog. I know this is DIY, but they make remote-control RGB strips for like, $20 including remote and power supply. Not programmable but might still do what you like.

Depending on the application, you could get a digital-only device and use 3 digital to analog converters. I've seen spi/i2c with multiple channels, and most little things like arduinos and raspberry pi has spi/i2c support. I wouldn't know how fast a respective D2A could be operated, I think they can run on batch signalling (all outputs sent as one message). You'd probably want to use C if it's bit-oriented, but Python is often used for RasPi IO

>>1468754
capacitor is in place while it's running = it's a run capacitor. It's providing phase shift to the coil not driven directly through the switch.

>>1468773
>analog
go back, reddit

>>1468647
>so we can have switchers on a chip, no external inductor?
You can get regulator modules with integrated inductors. Many look like ICs, except that they tend to be somewhat thicker. The prices can be pretty low (for non-chink stuff), but they often come in obnoxious packages like LGAs and QFNs with multiple bottom pads. They typically need external capacitors.
Research prototypes of fully integrated switchers exist. Last time I checked they ran at insane frequencies and had quite low efficiencies.

>>1468791
>external capacitors
Shouldn't these be far easier to manufacture with silicon than an inductor?

>>1468804
Probably. Capacitance is proportional to area but inversely proportional to distance. Inductance requires volume. Also, inter-winding capacitance increases as distance decreases, which is a problem if you were planning on increasing switching frequency to compensate for the minuscule inductance.

>>1468819
Even more so, good dielectrics have been being made inside silicon dies for as long as MOSFETs have existed, but trying to cram magnetic alloys in an IC might as well just mean slapping an SMD inductor atop your IC and potting the whole thing, with wires from the inductor going down to the die. Sure it's a multistep process, and probably more expensive in the long run, but to me it feels like we're developing pointless technologies. The only real use for an inductor on the die would be for very high speed stuff where the distance from the IC to the inductor matters that much. Delay lines and RF filters, that sort of thing.

>>1468791
They have 3.3, 5, and 12V switching versions of the 78xx series, designated 78xxSR. They're more expensive and have lower current limits, but are much more efficient.

>>1468773
I forgot something obvious: Any chip with digital outputs can do PWM, but it'd depend on the max output frequency. Arduinos are normally real-time chips, so you can address the pins almost as fast the the cpu can go. The Pi has 2 hardware PWMs and can do a bunch of software PWMs. This gets more complicated if you want to have 4x rgb strings vs 1 long white string, and the power circuit also has some complexity, because you need to decide if you're supplying power in serial or parallel.

>>1468819
>>1468822
You two can drool over this:
https://www2.eecs.berkeley.edu/Pubs/TechRpts/2011/EECS-2011-94.pdf

>>1468773
>use analog
>get poorer linearity
>use more components
>waste more power
do you even lift?

>>1468804
inductors and transformers can be printed as spirals
capacitance is proportional to area. it's expensive and wasteful to put hundreds of pF on a chip, so we don't when MLCCs are so cheap

>>1468823
the ATmega*8 has six hardware PWM channels. why are you avoiding this

>>1468826
>why use analog
This isn't your grandfathers age. I had to look up D2As for a project and I think there were 8-channel i2c chips for like $3. For the open problem of "light brightness", 8 channels of 8-bit analog + 8 voltage-to-current devices could be cheaper (in the short term) and easier than an 8-channel PWM. Only downside I can see is power efficiency.
>m-muh ..00001% linearity
Go screw yourself it's a light.

>why are you avoiding my knowledge
I have different knowledge than you. Autist.

>>1468827
>open problem of light brightness
well, that's not the problem is it? light brightness using an unspecified number of parallel LEDs is the problem. RULE 4
>could be
so you don't even lift. ok
PCA9685 $2.33 in singles, 16-channel 12-bit PWM I2C with built-in oscillator + (power driver = a MOSFET to ground per channel, or a BJT and resistor if you swing that way, or if you're driving a single low-current LED, connect directly with a resistor) per channel
vs. $3 + an output transistor for each channel + a heatsink for any significant number of LEDs on the channel + an op-amp to control the V2I converter per channel + 40 or so passives, some of which may be power resistors
or, for that matter, there are WS2812B strips for about $3/m, though they are a bit fiddly to control and not economical for single-color-per-strip usage
>different knowledge than you
apparently limited to the last thing you read about. may I suggest you spend more time looking at how things are done in practice and reasoning out why they did it that way instead of power V-I converters
>autist
reddit

>>1468827
embarrassing. digital is the objectively superior choice here. if you're going to do analog then that's fine, just acknowledge that it's for shits and giggles.

>>1468838
Again, you're Autistic and assumed I knew what you knew about that component. Go around looking at 1000 LED power guides and you won't find that part.

>>1468925
If something works, it's not objectively worse. I'm used to extremes, trying to drive 1000V max pk-pk at tunable .5-1.5 MHz and 200->2000 ohms for 5-10ms or constant 1-4mA over 6-30 volts, constant and pulse. Sorry you don't understand people have different conventions and knowledge.

>Engineering thread
>why is everyone double Autistic as normal
>nvm

>>1468838
P.s.
>Rule #4 they're using parallel leds
>re-read the post
>they didn't say whether or not they're using the strips in parallel
>the absolute state of your bait

I constructed a 1:1 npn current mirror with BJTs with a reference current of 0.5 mA. When I increase the load resistance, the output current drops from 0.5 mA to 0.497 mA (this is in the active mode). I'm supposed to plot the output voltage against the output current to find the output impedance but my slope is negative. Is this plausible or did i fuck up?

>>1468947
Reality isn't terribly accurate. There's lots of non-idealizations we ignore because we usually avoid working at the edge of ideal.

>>1468721
It's latching so you need both poles to switch it. I'll double check the pinout I guess.

>>1468947
Increasing the load resistance reduces the voltage across the mirror, right?
So, the mirror current drops as the voltage across it drops. No negative resistance.

>>1468825
>C1 through C9 are implemented with off-chip capacitors

Meh

>>1468961
It was a prototype, and gave >80% efficiency across most power draw. There's literally no reason the capacitors can't be integrated into the silicon.

>>1468966
Ok, Vincent. Let me know when it makes it to market.

>>1467733
if your phone has a metal back you are shit out of luck no matter what. Even if you attach the coil to the outside there will still be flow into the metal. As soon as that happens you are basically doing inductive heating at 15W (or what ever your charger has) to the back of the phone.

>>1468300
you need the other fuckery. without it you will have a very dead led

>>1468983
>wtf why would you even make a prototype lol
Thanks, Babbage.

>>1468997
led is an inorganic object and was never alive to begin with

>>1469005
Shut the fuck up. Being pedantic about colloquial usage of terms like "dead" (example: dead batteries) isn't clever, it's annoying.

>>1469003
Are you somehow involved with that charge pump or its designer?

>>1468943
LED strips are a parallel-series array of LEDs; usually 1-3 LEDs in series and lots of those in parallel.

>>1469005
what if it's an OLED

does anyone know where I could find bulk bags of say, spst/spdt switches and 3.5mm TS jacks? I can only find them for individual purchase and they're expensive as fuck

you'd expect to be able to find bulk bags of these things on aliexpress but I'm turning up blanks

This basically a hydroelectric generator, will provide schematics in the next post

>>1469136
Here the schematics

>>1469136
I meant to say, do I need more energy? I'm using tap water, the distance between the magnet and the coil is too big?

anyone have guesses on what this connector is called, crawling through google images isnt working
white is female and out of focus but 10 pins standing down the middle
worst case I need to put in different connectors

>>1469017
Are you serious? It's an 80% efficient PROTOTYPE of a switched capacitor design. If you're not excited to fucking ditch inductors for switched-mode supplies AND get higher efficiency you're an idiot.

>>1469105
RGB LEDs will be in parallel, but the strips themselves my be in parallel or series.. He never said if he wanted each shelf addressable/programmable. There can also be power issues depending on how many you want. Each shelf on/off vs each shelf variable vs all shelfs same coloring are different, design wise.

>>1469113
Ebay, that's where I got all mine. Search for 10x or 5x.

E.g:
https://www.ebay.com/itm/US-10-x-Mini-6A-125V-AC-SPDT-MTS-102-3Pin-2-Position-On-on-Toggle-Switch-Practic/302241862850

https://www.ebay.com/itm/10x-3-5mm-1-8inch-Female-Stereo-Panel-Mount-Jack-Solder-Connector-Gold-Plated/163046619877?hash=item25f655fae5:g:6w8AAOSwdbZbaTuw

It can be painful shipping times, and even after 20 years ebay doesn't have a sort by shipping time. They'd make more money from if they did. I'm surprised amazon wouldn't have 10x/5x

https://www.amazon.com/3-5mm-Stereo-Panel-Connector-10Pack/dp/B01C773OP6/ref=sr_1_sc_3

If you add diy you'll get better results for the components.

Alibaba usually has quantity listings, since they deal with huge amounts sometimes.

>>1469153
B-but I just want to light up a 3v LED. I don't need time processing
But curious tho, it would be very difficult to factorize time on this thing...

And how about the LED? Too low voltage?

>>1469144
i looked on digikey to no avail. better pics might help.

>>1469157
?

LED strips are 12V normally, and have 3 little pads for the RGB and one more for the common ground.

>>1469113
I too have had a look for such things, but from what I gather the only bulk deals you'll get are those aimed at mass-production, i.e. mostly SMD parts, maybe some thru-hole, but nothing much in the way of panel-mount. I bought some bulk TRRS audio sockets for not too much, but they were thru-hole. Chassis mount switches and sockets have always been pretty pricy, perhaps look on eBay instead.

>>1469136
CHANGING flux, anon, it has to be CHANGING! Make it such that the total magnetic field lines through the coil change, typically flip from one direction to another. This is why most motors you'll see have coils not wrapped around the axis but around a radial direction or two. The magnet there should have multiple poles for this purpose, i.e. N,S,N,S pointing inward as you go around the perimeter, with the opposite poles pointing outward.

>>1469139
CHANGING

>>1469142
I wouldn't slap an LED on there right away, hook it up to your DMM or something that will give you an indicator of having a voltage with much less current, like a transistor or darlington powering an LED.

>>1469153
I was assuming he meant a single LED strip, but if there is more than one shelf I guess it's a question.

>>1466111
is calculus/higher end math a requirement for working in electronics?

>>1469190
Not for basic stuff and you can usually get away with plugging stuff into online formulas, but I do find myself using complex number algebra with basic calculus somewhat often. In particular, things with AC capacitors and/or inductors (filters, oscillators, delay lines) will tend to require complex numbers and probably calculus to analyse, while for digital stuff you should be fine without. Unless you get into FFTs, complex numbers are also used here. Really high frequency stuff turns anything into a capacitor or inductor.

>>1469190
What's electronics to you? Audio amplifiers and filters/synth boxes? Embedded programming? Radio-frequency? etc... etc...

For "work" you usually don't do anything than some algebra if you need to work out some annoying dependent values and want to look smart (trade-school electronics). If you're being experimental? Some complex numbers, laplace transforms, fourier transforms, some simplified 2d/3d integrations. Phyiscal research? How much can you learn?

>>1468951
>>1468960
My voltage increases but the load resistance increases a lot more, hence the slight drop in current. Pic related is what my plot looks like

>>1469214
Is .6% really outside spec for you? Use a current mirror ladder or whatever they call them if it's a problem. I think it reduces the error from like 1/(1+1(1/B)) to 1/(1+(1/B^2))

I mean by definition a current sink is a non-ohmic device. Are you sure you're understanding your problem correctly?

Also, negative resistance is a thing: https://en.wikipedia.org/wiki/Negative_resistance

If this is for a lab, report your results, period. Don't fudge things until they "should" be correct. That is not science.

>>1469214
>>1469217
I was thinking about it, there's a specific non-ideal effect due to voltages. https://en.wikipedia.org/wiki/Early_effect

Increasing the resistance of the load increases the voltage across the load as long as current is still supplied. Less voltage across the transistor reduces the shortening of channels due to a smaller static effect. With a slightly longer channel, resistance increases. There's an explanation for you.

>receiver ic wants 4.75-5.5v
>whatever it'll probably work fine at 4.5v
>it doesn't
tfw

>>1466368
It has different energy saving modes. If you just want to do periodic measurements, sent it to deep sleep after you published the data. If it has to operate continuously, you need to charge the battery all the time. These fuckers need a lot of energy

If i have two PSU from old computers and i want to use them to power the same device, do i need to connect their grounds?

>>1469113
did you try ali yet
>yes
oh
what kind and how bulk do you need? I buy 'em by the dozen or so all de tiem. it never hurts to ask a seller if they can get you a better price for 100 or whatever

>>1469265
yes
/ohm/ RULE 1

>>1469267
Is it safe though? Since the use transformers and are thus isolated wouldn't that mean they are going to fight each other?

>>1469231
>not having 5V anywhere

>>1469223
Or maybe just one of the transistors gets bit warmer than the other.

>>1469268
>Is it safe though
No it's a shitty idea. The two supplies will not be balanced and one will end up doing the majority of the work, instead go on aliexpress or ebay and buy a power supply that can handle the output power you need. Assuming you want the same voltage but more current out of it that is. On the other hand if you just want double the voltage then you can probably stack them in series without much trouble, though you may need to isolate one of the PSUs. Just severing its ground connection should do the trick. But it won't be as power efficient as a dedicated module, and it's a bit of a waste of two computer PSUs unless you're using it as a multipurpose power supply anyhow.

>>1469268
>are going to fight each other?

the grounds, no. the supply voltages, yes. the one that's a little bit higher voltage will try to back-feed current into the other, to prove that it is number one dogge.

but that should be of no concern coz the diodes inside will save the day.

anyway, the reality is that the grounds are already tied together through the power cord, coz PC supplies are ground referenced. also, if the cases are touching, they're grounded together that way, too.

>>1469265
If you put the supplies in parallel first you wanna put a small dummy load on each rail you intend to use to ground of a value between 10k-100k since switched mode power supplies will often have higher than normal output voltages when unloaded. Then trim the voltage of each PSU to be as close as possible. It'll probably only be possible to do this for one rail since feedback is usually taken off one. The add a small power resistor less than 1 ohm linking each voltage rail on each power supply. This will drop any voltage difference between supplies which will hopefully be very small to minimize power losses.

>>1469265
You can't reliably put PSUs in parallel. The one with the higher reference voltage will try to supply all of the current leaving the other one idle. In practice, the resistance of the power cables means that it's not quite that extreme, but you aren't going to get a neat 50:50 split. The main issue is that you don't know what that split will be; it might be 90:10, meaning that each PSU has to be capable of supplying almost the entire load, which defeats the point of using 2 PSUs.

Can anybody explain to me if average values matter?

If you look at the formulas for mean/RMS values for rectified DC for example:
https://meettechniek.info/compendium/average-effective.html

Full wave rectified sine
DC(mean) = U(max) * 2 / pi
DC(effective) = U(max) / sqrt(2)

Using effective makes sense when you think about AC, but then why do we have the mean values? I know they matter for Power because when there is a phase shift of 90 degress(impossible, I know, just theorhetical) we would have the mean value of the sine wich is 0, but does average matter in rectified DC voltage?

Will the LED driver part of LED bulbs get better?

The cheap chink ones I bought the driver part failed before the actual LED.

Symptoms
Flickering and then it went away and stayed on, minutes/hours later it began again.

I'm guessing a capacitor went bad.

>>1469265
>I have junk, what do?
Take it to the recycler.

>>1469335
the usual failure mode for me is one of the LEDs in the series string gone intermittent, like a bond wire coming sort of loose and moving around inside the COB depending on moisture
sadly I doubt drivers will get better. the drivers are meant to be disposable along with the lamp, no reason to optimize for reliability. personally, I've been considering refitting most of my lighting with local constant current drivers running from a 12V supply, locally or remotely

>>1469351
I am a preper and when the shit hits the fan this "junk" will be the only way to survive

>>1469331
Well, the mean of AC is zero. Mean of the absolute value isn't really important in most cases, peak-peak, RMS or amplitude of the fundamental are more interesting.

>>1469360
https://www.youtube.com/watch?v=jMn272A60MU

>>1469271
it's powered on-board by mains via a resistive dropper. it works fine when i power it from my bench supply. i'm going to increase the bulk cap values and go to the next standard zener voltage next time i order parts though. hopefully that'll be enough.

>>1469376
there is a big difference between junk and trash sweetie

>>1466391
Ya open one up and its just 6 AAAA batteries in series. Essentially the capacity of a single AAAA (worse than AAA).

>>1469381
add a 1N4001 in series with zener ofc

>>1469406
circuit

>>1469407
nvm i see what you mean. it's an smd board so i'm just going to wait for another sot-23.

>>1469408
if the bottom diode is plain silicon you could also get a few more millivolts by replacing it with Schottky
but really, why didn't you put that diode on the line side of the zener anyway, where its voltage drop would be immaterial? now you've just got a zener that conducts in the other direction for no good reason, no? (PFC? kek)

>>1469409
good question. you'd have to ask the author of the ST whitepaper i stole the circuit from.

>>1467049
The reason why these are so expensive is the liability associated with a potentially life saving device.

>>1469274
>o isolate one of the PSUs. Just severing its ground connection should do the trick
can you explain why?

>>1469413
then no worries mate, the liability lies with the company selling them, not the copper monkey who macgyvered them

>>1469331
The mean is the DC component, i.e. what you'll get if you feed it to a low-pass filter. Also the offset you'll get from blocking the DC component with a capacitor or transformer.

>>1469412
>you'd have to ask the author of the ST whitepaper i stole the circuit from.

no need to ask him, coz we know. he put the fuse on the neutral wire. that is just fucking retarded. so, the answer is: the dude is retarded. (and probably a smells of curry.)

>>1469407
Why not get a real power supply?

>>1469497
That's too big. My bunghole will complain.

I don't have much information at the moment cause I'm not there, but is there any reason why a room would have no power even though the fusebox says otherwise? There was power in a few outlets in the house, but some of them (even ran directly from the fusebox) didn't have power. It's really weird.

>>1469497
i used one on my last project but they're so big their gravity kept attracting the 0603s i was trying to solder

>>1469414
The negative rail may be connected to ground, so if you kept that connection in place you'd be unable to have the negative rail of one PSU 12V above that of the other PSU, you'd just make a short.

What is the purpose of that?

>For additional noise immunity, connect "M" to chassis ground even if not using sensor supply.

DC ground is already tied to earth at the power supply side.

https://support.industry.siemens.com/cs/mdm/91696622?c=60466742283&dl=ru&lc=en-US

>>1469406
>not using a capacitive dropper

>>1469568
Hello again, Vincent

>>1469573
no I'm just some guy copying big clive

Anyone has used NFC arduino modules? I need to know its reliability, its vital, final year project vital kek.

>>1469807
All the modules are like $20 can't you get a reimbursement for expensens>?

>>1467107
you are the future

>>1469568
>not using a proper offline switching converter ffs

120/240V to 5V is probably one of the most common configurations for step down flyback converters and you should be able to easily find schematics if you're too retarded to design one yourself. In the event that's too much work it should also be easy to find pre-made ones for cheap.

>>1469841
Chances are for very low power stuff that a dropper (or switching capacitor-based converter) will be more efficient for such a purpose. Even better if you can run it with the IC itself as the only load on the capacitor, but it's not that likely.

>>1469843
Scratch that, I know there are some pretty efficient switching buck/flyback converters out there. But if he wants value engineering it's still a method he can take, and it's at least better than a resistive dropper.

>>1469521
OP RULE 0

>>1469841
>flyback
>schematics
getting the magnetics is another story, especially if one doesn't want to wind them oneself. if one does wind them oneself, one can generate any voltage required and the topology of the circuit varies little to none as a result

>>1469559
I'm not 100% sure but I tink it has to do with electromagnetic interference caused by surrounding equipment on your ground conductor.

I was getting training on a xray-related device last week and the sensor only outputted 150nA, so it was really important to have good ground. If you use a chassis as extra grounding it adds to the overal grounding conductor thickness.

>>1469871
What sort of sensors do those x-ray machines use, photomultiplier tubes?

>>1469872
It wasn't a normal x-ray machine, it was meant for measuring thickness of stainless steel plates up to 1µm.

The xrays are caught in an ionisation chamber(rare gas and electrodes) where photons from the radiation "bump" the electrons creating current. It's sent to a transducer that first converts it to voltage and then bumps it up to a 0...10V signal and converts to digital values.

Funny part was converting electrical energy to xrays only had 2% efficiency, lmao, all the other energy is lost as heat.

It used High Voltage with a strange spring-loaded cone-shaped plug/socket.

>>1469841
it's crazy to replace a few cheap components with the 30 needed for a switcher when you don't have to. have you fuckers even designed an offline flyback before? it's not trivial. and buying one can double the cost of a cheap BOM.

A friend wants to drive a "gimbal motor".
Basicly its a BLDC with sensor. The catch is instead of the usual 3 hall sensors it uses a single AS5600 with PWM output. Does anyone here know of prebuilt driverboards for this type of motor? Everything I can find only has the inputs for the typical 3 hall sensors. Designing a custom driver board is out of scope for this by the way, if he's fucked he's fucked.

>>1470039
>AS5600
the IC's datasheet might show something

Grower here. I wanted to try the new full spectrum led grow lights since they are the new hot shit, but the complete sets are expensive, and i saw the sell the bare leds for just two bucks a pop so i got some, pic related, but now am not sure what to do with them. the seller says
>It must need a heatsink aluminum for cooling, do not light it without heatsink aluminum.

Can i just glue an old heatsink from CPU to it and connect wires from an outlet?

>>1470104
>glue
Provided that means thermal adhesive/compound then yes
>connect wires from an outlet
Since it appears to be a driverless chip I'd say so. You got 240V mains?

https://youtu.be/KKd2L9Exw0M
might give you some tips

>>1470128
It says it has a led driver built in

>>1470138
Driverless means you don't need an external driver.

>>1466146
Prime

>buy bunch of Schmitt triggers
>buy dip-16 sockets so I don't fuck them up
>Schmitt triggers are dip-14
ayy lmao

>>1470150
funposting aside is this the correct way to put an IC on veroboard?

>>1470152

if you have to ask...

>>1470152

you need a tool, or drill bit so pins 1 and 14 etc arent shorted together.

>>1470152
Buy better perfboard, like the mock-breadboard pattern, or just plain dotboard.

I'm involved in a project that involves distribution of low voltage, high current, ultra clean DC power.
I took a power electronics course in college but we only covered switching power supplies and didn't really get into noise considerations at all. Any recommendations for books I should read to know more about this?

>>1470166

fat wires and capacitors

>>1470167

>>1470169

lol sorry.

"fat wires and huge fucking caps", by Jerry Glouck, Hempstead Press, 1973

>>1470171
>>1470169
>>1470167

also "bigger the glob the better the job"

>>1470166
There's an equation for output ripple voltage, something on the order of V_ripple = V*C/(f*I).

>>1470166
Throw a capacitence multiplier on your output.

>>1470239
i worry that the filter ESL and parasitic capacitances will kill a cap multiplier's PSRR in the high tens to hundreds of khz that switchers run at. plus if you want a stable voltage you'll have to incorporate it into the switcher's feedback path which will either require an extra outer loop or will kill its frequency response. i think an LC + LDO would work out better, with a characterized PSRR and a lower dropout voltage (smaller heatsink) to boot.

>>1466368
Depends on what you're doing. On the nodemcu ones I have, flashing the rom and usually anything touching WiFi is too much for the low current of a traditional USB port, and causes it to reboot. I haven't tried with higher speed wifi though. Way back when I read docs for it, they said something about 802.11b/g had the longest range, but highest energy consumption, and the faster modes uses less power at the expense of range.

How useful are single sided, no through hole boards? My gut says they're too hard to route to be useful for anything reasonably complex, but I've never tried actually routing one.

>>1470311
you're basically right. a single-sided layout is usually not worth the effort unless
- you're patterning/etching your own, where the second side is a significant step up in process complexity and cost, or
- you are making enough of them that an SS-only but cheaper and easier-to-work substrate like CEM-1 or phenolic saves you time and money, even with a hundred wire jumpers topsides

>>1470311
The other anon pretty much covered the reasons why single-sided boards are used, but here's one extra reason to not use them: they're much more fragile than boards with plated-through holes. A decent whack can crack the solder joints of heavier components. This can be mitigated a bit by bending the component wires properly, but still.

>>1470165
>Buy better perfboard, like the mock-breadboard pattern, or just plain dotboard.
Not that guy, but I always found these striped boards (vero/stripboards) to be the most pleasant to work with by far. Its much easier to remove traces than having to create them yourself (with dotboards etc.). And its super easy to cut the traces with a simple box cutter or anything remotely sharp.

>>1470171
I saw the made-for-TV movie and I was suitably impressed

If a led says it's 30W, does that include heat?

>>1470478
It's the power you have to cool away. Photon efficiency is tiny.

>>1470502
that wouldn't make any sense, since that would mean only like 1% is used for light and 99% for making heat. leds are many times more efficient that other lights so that wouldn't make sense.

>>1470478
LED wattage can be thought of as more of a class rating than a spec

>>1470512
i basically need to know two things:
-how much heat i need to remove
-how many KWH will this fucker ruck up on my bill if i run it 24/7 for a month

>>1470515
>-how many KWH will this fucker

3.7

>>1470532
it can't be that small.
if we take say 12V led that is say 30W, that is 12*x = 30W = 2.5A = 30W
So, 31days*24hours*30watts/1000 = 22KWH.
How did you get such a small number?

>>1470509
>leds are many times more efficient that other lights
Consider it from other side - if leds are efficient, just how fucking inefficient other light sources are.

>>1470545
LEDs are efficient enough that I suspect even the "reputable" manufacturers take it into account when dimensioning heat sinks.

>>1466368
The max current is near 200mA, but using power saves modes it consumes way less than 1mA
Obviously you can't do anything when sleeping, you need to wake it up, do whatever you need to and put it to sleep.

>>1470478
Yes. 30W will be its power consumption. Assume that at least 70% of that will be heat (for comparison, an incandescent bulb is ~5% light, 95% heat). If you don't have actual specs on heat dissipation, size the heat sink to dissipate 30W.

>>1470612
how much heat does average gaming computer put out? i need some reference since i will be using cpu heat sinks

>>1470632

in the area of 70-120W. but that's with a fan, without it it would shut down in 2 mins.

>>1470648
bullshit, i could take the fan from my cpu heatsink right now and it would run hot but without any problems , only cheap tiny heatsinks can't handle shit

>>1470652

you make a good point. the fact that your experience contradicts that of millions of others, somehow invalidates them all coz you were put on earth by god to act as humanity's reference point.

Do I need a heat sink for the cree xlamp xp-e2? I'm just trying to put a small, bright, light in my costume and it seems like the most viable option. Also would a 9v battery be okay?

How strong are these?
I need to make a small robot arm that will push a reset button remotely. I cannot take the button out and add a relay so it has to be a robot arm.
The button is kind of hard to pres but not oo much kind of similar to the button to roll down a car window
Any ideas?

>>1470705

>>1470698

your LED is typically gonna take 0.7A to 1A. a 9V battery will last about 10-20 mins at that rate. that makes for a very expensive lighting solution. it would make much more sense to use 2 rechargeable Nicad or NiMh packs (keep one in reserve). better yet, punch a tiny hole in the wall to run some low voltage DC from a wall wart.

>>1470698

your LED is typically gonna take 0.7A to 1A. a 9V battery will last about 10-20 mins at that rate. that makes for a very expensive lighting solution. it would make much more sense to use 2 rechargeable Nicad or NiMh packs (keep one in reserve). better yet, punch a tiny hole in the wall to run some low voltage DC from a wall wart.

>>1470713
This is for a costume prop I'm gonna be carrying around for like 3 separate days, max. I don't mind using disposable batteries

I am trying to solder an extremely small 40 pin connector onto a board
if I have the iron on it for more than even a second the plastic will get melty and the pins will fall out

How can I do without $5000 in tools?
My idea when I get another connector is to try and tin or dip each pin then solder the tabs to the board to hold it in place then attempt to barely touch each pin just enough to melt the existing solder to the pad

pic semi related, about the size of the pins

>>1470743
my other idea is to get big balls of solder on each pad but not on the ground tabs, then solder the tabs while pushing down really hard so the pins will be under tension against the pad/ball but not actually solder them

>>1470743

this kind of thing is done way easier with solder paste and hot air. youtube has videos. a heat gun can be used, or an electric skillet. you can even use a soldering iron. the fact that the solder is already present, and already liquid, makes the contact time much shorter.

>>1470705
they're actually reasonably strong. you won't have any issue with pressing your button, but if you put any torque on them then you'd have to start worrying.

>>1470742
who's the suit maker and is it full or partial

>>1470705
what you need is a linear actuator
https://www.aliexpress.com/item/2PCS-DC-5V-10mm-Stroke-Length-10mm-s-Speed-Mini-Linear-Benis/32920971766.html
or a solenoid
https://www.aliexpress.com/item/DC-12V-300mA-Push-Pull-Type-Open-Frame-Solenoid-Benis-10mm-4N-0-9LB/32218777623.html

>>1470743
>for more than even a second
don't touch the pins with the iron. only touch the pad with the iron, let the solder convey the heat to the pin
and also use solder wick to clean the existing solder completely from the pads
and also use paste or liquid flux on the mating surfaces of pin and pad
and also use a fine tipped iron
and also work quickly

>>1470705
I second the solenoid, since it has no holding power at all once the power is cut off; it will go right back into a more neutral position once the power has been turned off instead of having to worry about return strokes, position feedback, etc. Though it will take a somewhat significant amount of power to use.

See how much force it takes to depress the button, multiply it by the distance the button moves, and multiply the two together (making sure to use coherent units) and you'll get the amount of work required to press it. Divide this by the time the motor will be on for and the voltage it runs at and you'll get a minimum current value. This value isn't useful at all, but hey mathematics is fun, right?

What's the project anyhow? Sounds like typical hoop-jumping from a college assignment.

How important is it to have active power factor correction in a power supply project (100W hopefully) where I want the circuit to be a hybrid power supply? (switched-mode tracking preregulator stage, linear output stage)

>>1471033
pretty unimportant at 100W
almost entirely unimportant if you don't care about the Caveat Emptor mark
not important at all if your utility doesn't care

>>1471035
Fair enough
Should I give up on electronics and just do meme worthy arduino shit? I feel like "wanting to design my own power supply" everyone would just tell me not to be stupid and get buy one since people do this shit for a living
Like, why even bother? Even figuring out the switching preregulator stage is hard enough. Let alone having to make it track the linear stage's output (and ensure it is ~5V higher than the output)
>inb4 just make completely linear LM317 power supply with 9 gorillion uF filter caps, pass transistor, and massive heatsinks

>>1470859
but isn't doing anything at all, such as pressing a button putting torque on them? since they have to do work

>>1471041
>Should I give up on electronics and just do meme worthy arduino shit?

generally people spend time on a hobby because it's fun. pretty much no one thinks an ''active power factor correction'' power supply is fun. but you're different. and that's to be celebrated. (until you go too far and they lock you up in the loonie bin.)

# # # # # # # # # # # PASSED BUMP LIMIT # # # # # # # # # # #
# # # # # # # # # # # NEED A FAG TO MAKE NEW THREAD # # # # # # # # # # #

>>1471021
>the solenoid
does not push, it only pulls unless modified. I would use the mechanical part of a suitable relay to push the button.

>>1471057
That's the point. You use the retraction force of the button to handle the reverse action. When you let off the power the solenoid, the button will push the solenoid back into a neutral position, no feedback or commutation required.

>>1471055
chill, it's only on page 5 and this is a slow board, not /g/

>>1471058
and, even if it isn't a completely neutral position because it's a tact switch buried inside some piece of consumer electronics, it doesn't matter because it's not applying force to the button anymore

>>1471057
>you might have to modify a solenoid
>so let's modify a whole relay instead
le_trollface.jpg

>>1471073
well neutral was relative to begin with

oh and RIP csg

>>1471048
I definitely enjoy the challenge of it (for the most part). Making an efficient and effective power supply and such
But I still feel like i'm just groping around in the dark.
There's barely any info on active pfc that isn't basically just "do the circuit in the application example"
Same too with some of the switched mode power supply stuff, namely selecting capacitors and inductors for a SMPS, and how to figure out phase margin shit for any sort of feedback system.

I also need an electronic load (preferably) to test this shit out with
Naively, I would run mosfets in their linear regime and with feedback to control desired load current, but certainly that's too simple. Esp since big boy electronic load boxes are hundreds of dollars

>>1471081
yeah, a lot of power stuff is pretty cookbook now because designers need some new voltage at the point of load almost every time they add a chip
>but certainly that's too simple
surprisingly, no, that's just what you want to do, if you have MOSFETs that can handle the voltage and the power dissipation, and you can get a low-tempco, high-power, low-value resistor for sensing. power paths are usually kept pretty simple, however complex the control circuitry might be.

>>1471021
>>1470986
the solenoid costs like 7 times more that the motor does, plus i already own the motor and an arduino to control it so i want to use them as it would cost me 0

>>1471095
>i already own the motor and an arduino to control it
Then why are you bothering everyone with it.
Just do it and see if it works.

>>1471095
then what you've got is a mechanical engineering problem, not an electrical engineering problem, and therefore you should consider asking in /qtddtot/ instead

>>1471085
>surprisingly, no, that's just what you want to do, if you have MOSFETs that can handle the voltage and the power dissipation, and you can get a low-tempco, high-power, low-value resistor for sensing. power paths are usually kept pretty simple, however complex the control circuitry might be.
You're right. I watched a teardown and an array of powerful and expensive mosfets do the loading. everything else is sense, safety, and operating system related stuff.
That's cool to know then. I could probably build a rudimentary one
I'm curious though - where would one find heatsinks that are shaped appropriately and large enough? They also seem to be expensive on digikey. I'm thinking of the eevblog rigol electronic load teardown. that heatsink tho

>>1471118
haven't seen it, but you could do worse than drilling/tapping PC CPU coolers according to their rated dissipation

>>1471126
>>1471126
>>1471126
>>1471126

>>1469190
yes, that's why it's in the degree

>>1467049
what if they fall slowly?