/diy/ - Do It Yourself

here's a less boring pic for next thread.

Should I just avoid using anything I can't solder with an iron?

>>2006465
no, boaden your horizon. A lot can be accomplished with a 20$ cooking plate and some aluminium as heatspreader.

>>2006482
what's a cooking plate

>>2006482
Also for $20 I could already get a toaster oven. Not the convection type though.

>>2006465
other anon isn't wrong but imo yes, soics > tssops >>> qfns simply as a matter of convenience.

>>2006446
That is fairly sloppy wire wrap work, poor layout with little thought of maintenance. Little good I can say of it. Well done wire wrap is a wonderful thing to see.

I would give it 5/10 if it works and then snip a few of the blue wires which are tucked under the white and hope that he learns his lesson regarding layout and forethought towards future maintenance as he spends hours fixing it.

>>2006465
Look up "diy reflow oven" o whatever name english speaking people use. Small things are not scary, just tedious at times. Hot air got way cheaper in recent years too.

>spend hours on various manufacturer's websites looking for potential parts
>pick out a few
>they're all too obscure so no one stocks them
fuck

>>2006553
Did you also get a barbed-wire tattoo around your bicep in the 90s so you could be unique and stand out in a crowd?

>>2006556
I didn't exist in the 90s

>>2006557
Stop being gay.

I bought a 24V-20A power supply on aliexpress but like the nigger I am I forgot that I needed a 12V one. What are my options to step down the voltage? Complete retard btw but I can probably follow instructions.

>>2006566
Buck converters, regulators, or modifying the voltage divider for Vsense feedback in the PSU.

what happened to "thread brought to you by..."?

>>2006573
Sponsors pulled out because not enough trannies involved in micromanagement.

>>2006571
Just to be sure, the fact that the PSU is "24v 20A" means that the MAX amount of current that can be drawn is 20A right? Not that it erogates 20A constantly. I'm pretty sure but I might as well ask, sorry for the acute retardedness

>>2006587
Yes, 20A peak. That means you can't expect 20A continuous without killing the power supply. Always leave headroom.

>>2006588
Thanks!

>>2006590
no problem fren

>>2006557
based zoomer

>>2006557
>there are actually people born AFTER the year 2000
holy fuck my mind is blown

I got a part mismatch error on a PCB order. I'm pretty new to PCB design and I'm just wondering what the size of this LED footprint is? Apparently there's 1.5mm center to center but SMD sizing standards seem to reference edge to edge. Does this look like 0603 imperial? Many thanks in advance.

>>2006588
>That means you can't expect 20A continuous without killing the power supply.

Bullshit. He didn't say peak. If a supply is rated for 20 amps you can draw 20 amps full time unless it has a specified duty cycle.

>>2006605
You're being a pedantic faggot, anon. Go run your power supplies at 100% duty cycle.

Still the original 24v psu retard, I assume that it would be enough to power a MPCNC with an arduino mega and a ramps board + 5 steppers, right?

>>2006604
leds frequently ignore the standard footprints for passives. that doesn't look like a standard 0603. the solution is just to buy 0603 leds when possible. you can also fit sod-323 diodes onto an 0603 pad; i just use the same footprint for everything.

>>2006605
just like you could technically run an electrolytic at 105C and 100% rated voltage, but you'd regret it. i'm sure the chinks that design these supplies aren't properly debiting the power tranny junction temperatures. i wouldn't expect a reasonable MBTF on anything but the most bespoke PSUs running at their limit.

>>2006613
MTBF*

>>2006613
> that doesn't look like a standard 0603.
> the solution is just to buy 0603 leds when possible.

These two sentences seem to contradict each other. Will an 0603 not fit here then? Do I just have to hunt through datasheets to find a correct fit?

>>2006618

>>2006618
that pad doesn't look like a standard 0603 pad, pic rel is more or less what you'll get from supplier recs. you could solder an 0603 to that pad you have pictured though. when you're hand soldering it doesn't really matter as long as you're in the ballpark.

>Do I just have to hunt through datasheets to find a correct fit?
if you're not sure, yes.

>>2006620
Length and width for 0603 look decent. Pad size is a little questionable though... worth trying out do you think?

>>2006629
Sure, why not? You can't win if you don't play.

>>2006627
>>2006632

Thanks chiefs

>>2006638
You're welcome, Hoss.

>finally finished a project
>now I just feel empty inside
why is this? do I need to do another project again to quell this feeling?

>>2006643
stop now while you can

>>2006643

>>2006643
>Have several projects on hold and several new project ideas
>still waiting for the parts for any of them to arrive.
Ahhhh I just want to BOOILD.

>>2006643
You can sate your emptiness while building a portfolio now!
Aren't you glad :)?

>>2006694
>portofolio
it's doesn't matter, I am a software developer

I'm quite new to this, is it safe to use multimeter in a extension cord box? (unplugged ofcourse). Just want to try out the "continuity knob" in the multimeter.

>>2006698
You could become an embedded dev

>>2006700
>become an embedded dev
I am not sure, I'm still too dumb for low level stuffs

lmao I just downloaded kicad and I have no idea what's going on

>>2006725
Watch a tutorial on youtube, you'll pick it up sooner or later.

>>2006725
easy, but convoluted af
wish it was easy to import LTspice schematics, so I don't have to replot everything

>>2006741
Pretty sure you can import netlists, but there's a lot of mismatch between the parts used. And of course netlists don't store position data or anything.
Have you tried using KiCAD as a sim?

What kind of experiment could bring collaboration between /ohm/, /mcg/, /pol/, /k/ and /g/? How complex can hobbyists get with serious real world shit?

>>2006725
>this should help smoke those rats out

>>2006747
Distributed fully secure communication systems with tons of heterogeneous microcontroller nodes

>>2006751
Secure, unimpeded comms is probably the most important tool to have at your disposal. It's going to require an entire re-thinking of current systems and infrastructure.

>>2006751
What would /pol/'s contribution be?
Spam BBC porn?

>>2006747
smartguns, with AI-assisted user judgement

>>2006754
Philosophical autism's attention to detail is no joke.

>>2006754
ensuring that it's immune to censorship by posting things on there which would otherwise get censored by the big-tech-communists

>>2006756
You don't need extra autism, it's already a 4chan project.

>>2006756
>implying /pol/ is anything more than schizos

>>2006755
It's a nice idea and it actually works to an extent. Some university was working on smart rounds a few years ago.
>>2006758
The more the merrier. lmao
>>2006759
Schizo's are a small percentage. There's more commie shills there at any time, between Bidenfags and chinks.

>>2006747
big-data AR cyber-infantry
cameras and rf sensors and microphones situated about the battlefield provide their information to a central server, which processes the data efficiently with machine learning techniques, and feeds it to the AR display of the soldiers. in their display they'd see thermal overlays where enemies are, even through walls and obstructions, and it would keep track of how many there are. the soldier would have a minimap in the corner of his vision, showing the battlefield from a live arial drone footage, with thermal and audio layers. it would also monitor where the enemies are looking and pointing their weapons, so you can get a solid estimate of the total enemy field of vision, in order to skirt around it and choose the best order for eliminating them. it would coordinate the vision fields of allies too, in order not to leave any blind spots. aim assist too, allowing a user to shoot down a moving target, be it a running insurgent or an rpg.
for radio silence, have a portable version where all the sensors are mounted to the helmet of the infantry, and the computer is on the back.

also replace /pol/ with /sci/

>>2006755
yes, with a scope-cam feeding a low-latency stream to one eye, so you can fire around corners and such more accurately.

>>2006764
What about something like a swarm of autonomous mechanical bees with an exploding thorax full of cyanide gas?

>>2006747
autoturrets with racial recognition

>>2006764
>also replace /pol/ with /sci/
/sci/ has become more and more like /pol/ over the past couple years (and that's a good thing!)

>>2006769
https://www.forbes.com/sites/davidhambling/2020/08/24/mexican-drug-cartel-carries-out-drone-strikes-in-gang-war/?sh=7358aee89432
We're a lot closer to that sort of thing than people realize.
See, the real problem is that this new big tech/china/(((globalist))) alliance has both the drone production (from china) and the CV/motion planning software (from big tech) necessary to make massive swarms of explosive drones.
Any potential civil/world wars are going to be fucking wild.

>>2006779
Dangerously based

>>2006725
is there any reason to use kicad over pirated altuim designer rather than
> look at me im le supporting open source software xd ?

>>2006782
pls don't van me fbi, but this sort of bomb drone seems like it would be incredibly cost effective for attacking the electrical grid

>>2006796
you know the answer, there is no good reason not to pirate altium

>>2006796
not having to deal with cracks and dubious torrent sites?

>>2006798
AAAAAAAAAAAAAAAAAAAA
CHAFF DRONES JUST FLEW OVER MUH POWER SUBSTATION AND 5G TOWER

>>2006798
>>2006796
>> look at me im le supporting open source software xd ?
Completely serious, you should kill yourself.

>>2006804
how the fuck can you post on imageboards and not know how to detect safe torrents
>cracks
most cracks nowadays are just copy paste x folder to your program directory

>>2006807
not an argument

>>2006798
>you know the answer, there is no good reason not to pirate altium
Doesn't run on Linux.

>>2006798
Oh shit, sorry bro, didn't mean to reply to you in >>2006807
I was not telling you to kill yourself.

>>2006810
Not him, but if you realized how antiviruses detect malicious programs you would realize that your sense of security is entirely artificial.

>>2006798
>not dropping methylmercury in the water supply instead

>>2006814
piracy is too convenient so the best solution imo is just an airgap. i don't do any of my banking or shopping on this shitbox.

>>2006811
>linux trannies
>>>/g/

>>2006818
>piracy is too convenient
>so the best solution imo is just an airgap.
I do this for muh gaymez and let me tell you, it is anything but convenient.
Airgapping is a fucking PITA
also yeah, >>>/g/

>>2006820
Windows is complete crap cobbled together by pajeet hacks and runs like shit

>>2006820
>microsoft windows is for cha-

>>2006825
>linux is for incels
>windows is for bugmen
osx is the only option for the white man

>>2006826
>osx is the only option for the white man
osx is for faggots tho

>>2006826
What a retarded post
I expected nothing less from an itoddler

>>2006826
>the only option for the white man
*blocks your path*

based retard derailing this general

>>2006833
This guy fucks

>>2006835
Sneed

>>2006835
It's my fault for mentioning /g/. Sorry.

>tfw about to dive into an over 30 min eevblog video
wish me luck bros

>>2006643
going into my abstract merchants folder

is there a way to make an adjustable gain differential opamp without using dual track pots
i'm trying to measure my gpu and cpu power draw with an arduino
or would cutting the wires and adding a shunt resistor be better, wouldnt require adjustable gain amps but soldering copper cables that thick is hell

>>2006907
>is there a way to make an adjustable gain differential opamp
there's a popular instrumentation amp circuit topology with gain controlled by one resistor. you shouldn't need to do that though, post more project details.

>testing flashlight project
>too bright
>cover it with a notebook for 10 seconds
rip notebook

>>2006907
First of all, if you’re measuring current draw with an MCU, a potentiometer to control the gain is the last thing you want. You want accurate fixed resistors to set your gain to a precise level, and calibrate your MCU to that. If you change the gain as you’re making measurements, you’ll have to recalibrate it each time, which would be a pain in the ass.
Secondly, I take it your differential amplifier is to measure the voltage across a current sense resistor? The voltage will be so close to one of the rails that getting a working circuit, even with rail-to-rail amplifiers, will be tricky. I’d highly advise that you watch Andreas Spiess’s video on current measurement modules. He covers a few solutions, some with internal ADCs and some with analog outputs, all of which are designed to be able to snug right up to a rail and have an integrated amplifier. Could either buy an assembled such module, or buy a bag of the ICs.
Thirdly, don’t computers often have ways to measure their own current?

>>2006956
sounds like a good way to blind+burn yourself in the woods lmao

Should I wrap MOSFETs in heat shrink or leave the heat sink bare?

Thoughts on this list of resources? Seems useful.
https://github.com/kitspace/awesome-electronics

>>2007017
is it close to any other metal that it could short against? if not then who cares

>>2007023
What I'm trying to ask is, would it be detrimental to the component if I were to heat shrink it?

>>2007036
you need a specific type of insulating pad if you want to isolate the heat sink from the mosfet. just using heat shrink will reduce the sink's performance, probably drastically.

>>2007043
That's what I thought, thanks.

I am building a phono preamp (p06 by rod elliot recommended by an anon here) and I accidentally ordered capacitors that are comically oversized (physical dimensions, not capacitance) because they had better tolerance. Should I mount them above the other components and bend the leads so they fit in the case? Or should I mount them off the pcb and connect the leads via hook up wire? Is there a disadvantage to either of these? It's a phono preamp so it has relatively high gain and any added noise will be noticeable

>>2007060
Solder them to a piece of perfboard, then connect the perfboard to the main board with the shortest length of wire possible.

Similar to how you can find "old" well-built linear power supplies that outclass chinese switching power supplies, is there an equivalent to these? How did/do engineers measure ac power usage?

I've come across those hp power meters, like he 436A, but I don't think they do the same thing - at least not easily.

>>2007078
>How did
with a cool electro-mechano-magnetic multiplier and accumulator mechanism. there's a good video out there of a vintage power meter teardown.
>How do
for mains they use a current transformer and a microcontroller. ac/dc lab equipment will use a current sense resistor instead most likely.

>>2007089
>current transformer and microcontroller
Oh, like using a clamp meter/module and hooking it up to an arduino to log the data?

That older method sounds pretty neat, but not very practical for a novice to quickly hook up.

>>2007101
>like using a clamp meter/module and hooking it up to an arduino to log the data
yes. just be sure you bias it right so the arduino can read it.

I bought this EMF reader.
can you point me some course or readings on how to use it wisely to measure the EMF around ?

>>2007120

>>2006446
How do I find out what voltage and amperage a device needs from a wall wart? I've got a Chinese arcade console but no power supply and no markings on the housing to show what it needs. I assume it needs 12v 2a based off similar devices sold on aliexpress, but I do not know for sure. The other thing I need to figure out is if the tip of the wart is supposed to be positive or negative. How do i figure this out without shorting the unit out?

>>2007184
Open it and look.

>>2007184
>How do I find out what voltage and amperage a device needs from a wall wart?

- google replacement power supplies, zoom in to find voltage current and polarity
- use a variable supply and increase voltage until it works. to determine polarity, use LED+resistor, and compare with a power supply of known polarity.

>>2007184
>How do I find out what voltage and amperage a device needs from a wall wart?

- google replacement power supplies, zoom in to find voltage current and polarity
- use a variable supply and increase voltage until it works. for polarity, assume positive inside, which is 95% of modern units. hope there's reverse polarity protection if it's in the 5%.

>>2007192
>>2007211
P15 12V

lead from the center of the barrel is what I assume to be tip heads off to components, side of the barrel doesn't lead to anything, so I assume that to be ground. Any other clues I can look for to figure out amperage? I don't have a tester, other than a harbor freight multimeter.

>>2007223
I found a disconnected wire inside that seems to go to an external speaker. I'll see if I can get it working as is and worry about that next time.

>>2007223
>other clues I can look for to figure out amperage?

- if you can find a fuse in series with 12V line, supply current should be about 75% of fuse rating.
- there seem to be two 470uF caps on the 12V line. that implies about a load of around 1A.
- you can always use a HIGHER rated xformer without worries, so aim high, and you're sure to hit the target.

>>2007236
1.5A is working so far. I am asking /v/ for advise regarding the software side now.
I got the damn thing powered on and it has 2222 versions of questionably similar fighting game(s).

Pic related.

>>2007018
OP here. That's a really good resource, putting it on my paste now!

>>2007089
>ac/dc lab equipment will use a current sense resistor instead most likely
More specialist high-current stuff (solar power) may use hall-effect current sensing. In a fixed enclosure, it can be pretty low-drift.

>>2006705
It's not as hard as it used to be. An ARM Cortex-M4 part probably has some ludicrous amount of memory (for embedded), so you can fit a decent-sized C++ program on one. The days of everything needing to fit in an 8k "luxury-sized" ROM on some 8051 are over.

>>2007268
256 words is all we ever needed. SI prefixes were a terrible mistake.

Do you guys like Metal-Film or Carbon Composite Resistors more?

>>2007321
well, the cheaper ones are worse
but, the worse ones are cheaper

>>2007323
buy nice or buy twice

>>2007341
they don't somehow have a worse livespan, just worse tolerances and drift, right?

>>2007343
Metal Film advantages-Low Tolerance, Low cost, Less noise than Carbon, and the resistance does not vary so much with temperature.

>>2007344
none of those really matter to me other than noise
is it thermal noise?

>>2007347

the only people who realistically need to worry about resistor thermal noise are those trying to catch neutrinos 1000 ft down in a cave.

>>2007374
Or radiotelescope guys, and a few other kinds of very low energy instrumentation. I think it's more of an issue with high resistances, like 10MΩ or more, which may be required in some very extremely low current measuring circuits.

Anyone know if the current drawn by an LED at a fixed voltage increases or decreases over the life of the LED (or conversely if the voltage increases or decreases over the lifetime of an LED fed with a fixed current)?

>>2007374
wrong. IC designers need to take it into account, precise measurements also need to.

>>2007120
You bought a tool you had no use for? are you daft?

>>2007223
button arrays tend to be low power, so 500mA should be plenty (they usually draw 100mA or less)

>>2007321
metal films are like $0.01 apiece, there is no reason not to buy them

>>2007442
he is a CONSOOMER

It seems LCSC only really stock chinese parts? I found an Analog chip for less (like half the price) there, but everything else is rather lacking.

>>2007223
Just beep test the contacts inside the jack to the big ground plane, geez.
Almost all coaxial plug power supplies these days are positive-center or AC. The snowflake exceptions I've seen are "toy" guitar amps and P-touch label makers, both 9V. (The former I just rewired the fucking jack, but next time I'd put a bridge rectifier, the latter I tried once and it had a crazy circuit to interrupt the battery, but I didn't have enough fucks left to figure it out.)
Amps is less certain, just try it and see if it gets enough. 12V bricks tend to give 1.5-3 amps anyhow, but cheap 5V USB bricks can barely do 1 amp. I once had something which seemed to work with a 5V1A brick but it really needed an extra amp to go beyond the initial "blink some LEDs". It turned out the official brick was 3A (it had a lot of LEDs). I was trying to use a USB lead to have one less brick to carry around. I eventually found a multi-outlet USB with 4A and I could use the other plugs to power low-amp toys.
The real annoyance is when a weird coaxial plug size is used for a common voltage, much harder to bodge cleanly.
>>2007237
Well it's a step up from 9999 hacks of NES Contra.
>>2007268
It's RAM size that tends to be the problem with embedded MCUs now. Especially when you use a FAT filesystem library that wants 20K of buffers, with code that keeps config data for 60+ external devices including port names. (btdt)

>>2007467
>It seems LCSC only really stock chinese parts?
not really, there are plenty of non-chink stuffs, but mainly yes

I am trying to check a charging cable to see if it is just a cable or if it has and smart stuff in it and with ground wire i just get a continuity beep so that is a direct connection, but with the + wire my continuity tester just does like a 0.2s long beep and then stops, and then if i reverse the probes i get another short 0.2 beep and it stops and i can keep reversing the probes to get these short beeps
what sort of thing could be inside of that cable to cause this?

>>2007523
Maybe a capacitor to make up for voltage drop in the cable.

>>2007548
if the cap was in series with the wire then the cable wouldn't work at all with DC and if the cap was across wires that wouldn't have any effect on continuity test on one wire

>>2007548
>>2007562
If the cap is in parallel with V+ and Ground, it will appear as a short when discharged. Performing a continuity check will charge the capacitor, because the multimeter has voltage flowing between the test leads in that mode.
Once the capacitor charges, it appears as an open to the meter, so it stops beeping.
Reversing the polarity of the test leads also reverses the voltage across the capacitor, so the cap discharges to 0 volts, appears as a short again for a fraction of a second, then charges back up to appear as an open.

You can verify this by testing the wire like normal, and once the multimeter stops beeping, short the V+ and ground wires together, which will discharge the capacitor. Check the continuity again, and you'll have the quick beep once more.

>>2007523
Is there a boost circuit molded into the plug? Did you test the output while it was plugged in?

Can anyone tell me what the fuck happened to this PCB? The gameboy still works fine but i opened it up to check on some components and it looks like a whole area was doused in corroded acid

>>2007582
Could be spilled soda, leaking batteries or leaking caps. You need to clean it up and neutralize the acid if you intend on repairing it. A UV light will help you see more of the damage.

>>2007585
>neutralize the acid
I meant neutralize the alkaline corrosion with acid (like vinegar).

>>2007586
>>2007585
Interesting, its in an area too far from the batteries and too concealed for stray liquids to spill into it, and it works fine from what i can tell except for possible bad caps on the audio lines

>>2007599
Corrosion spreads like mold, and it only takes a little bit to make a big problem (one leaky cap or coin cell battery is enough). It will eventually stop working unless you deal with it now.
Maybe it was sitting for an extended time in a humid area with batteries installed?

>>2007603
good point, ill take a q tip and some vinegar (or 91% iso alcohol?) and clean it up. the most affect components are the 4 diodes and a 47pf cap and the lining of the pcb board around the cutout for the LCD screen
hopefully i dont need to replace the diodes (DAN215) doesnt seem like anyone is making them anymore

>>2007608
Use a soft bristled toothbrush and vinegar. Scrub the whole board well on both sides. Then rinse thoroughly with isopropyl. Let it air dry for a day before you power it up.

>>2007608
You should also do continuity and resistance checks on traces that look suspect after everything is clean and dry.

>>2007608
>DAN215
for reference

>>2007582
>>2007585
>>2007599
>>2007603
>>2007608
Why does everything have to be a bad capacitor or leaky battery with you people?
Its from old crappy flux cleaner or some other pcb cleaner that probably wasn't acid free in the chink factory that built it.
You can see the brush strokes around DA1-DA4.

>>2007673
the cap problem is unrelated to the acid leak. The audio on the gameboy emits a high pitched noise that doesnt go away when the volume is turned down. im no circuit expert but people have speculated a bad cap in the audio line could be causing buzzing

>>2007673
>you people?
fuk u raycis ass cracka wypipo
nytenda ain't use rosin flux on dey sheeit
i'll brush stroke yo lily huwhite ass byotch

>>2007442
I experience strange things in my house and want to eliminate hypothesis one by one. Here is a graph of vibrations in my basement.
My goal with this tool is to check for Low Frequency magnetic fields. Also just experiencing can be fun.

>>2007710

>>2007710
>>2007715
What's your best guess right now?

>>2007718
Either the city's watering pipes going all way through the street,
Either My neighbor's bakery, 81kVA of power. Many high amperage semi-industrial/cooking oven, fans, fridges, centrifuge devices. And due to the cost, before moving some expensive Lab-guy for Solid Noise/Seismic/Electrical problem, and Hire some expensive RUDY20 lawyer, I'd like to be confident with the cause of the trouble.

>>2007729
Maybe I'm a brainlet but can't seem to understand how the EMF reader can help you reveal the source of soil vibrations. Also what kind of sensor are you using for measuring the vibrations, I mean you already have multiple sensors you could calculate the direction of origin

>>2007737
If the source is from heavy machinery then mechanical vibrations & EMI spikes would overlap in his logs.

>>2007737
>Also what kind of sensor are you using for measuring the vibrations,
It's some 3$ piezzo electric sensor sold on Sparkfun. I'm gonna buy something more expensive, from PCB Electronics to confront datas. I saw there is accelerometers to fish into those PS4 joypad as well. One of my nephew destroyed many of them. Some spare parts to dig there.
>I mean you already have multiple sensors you could calculate the direction of origin
Like disposing them on a grid layer style and seeing how the direction is given? I should do this maybe yes. I monitor my desk room, my sleeping room, my living room, my basement, my shop and my warehouse.
ATM I've been focused mostly on building a neat solution. some small C nano-timestamping collector, some bunch or PERL script to feed an Influxdb database. Then used by Grafana for fancy graphics.
I also plan to set up some good microphone to produce same kind of graphics.
I also have some T° & Pressure sensors, enough to see it doesn't behave linear to those values.

I was looking at getting a charger/converter to work with a 3.7v 2500mAh li-ion pouch battery and a Raspi, but those shits are $20+shipping from Adafruit. Maybe I'm being stupid, but couldn't I just tear apart a dollar store 18650 backup battery and take the controller from that?

>>2007769
Yes you could. Another option is TP4056 modules which seem to be standard fare.

>>2007775
>TP4056 module
Oh, damn... Thanks. There's quite an assortment on Amazon too, I might even be able to find one that'll slim down the form factor of my build!

Is there any advantage on using an LED driver IC against a MOSFET + Driver + PWM controller?

>>2007779
YW. Get the ones with a DW01 protection IC onboard if you go that route.

>>2007785
>DW01 protection
Is that so it doesn't get drained too low? The battery I have is https://www.adafruit.com/product/328 which I believe has that built in, unless that's something else? I also noticed that it says to disconnect the load while charging, do you happen to know if there's any that allow it to run while the battery charges?

>>2007797
Adafruit sells Powerboost 1000 modules for an exorbitant price ($20 USD).

>>2007799
Yeah, that was specifically what I was trying to avoid. I was hoping I could find a knockoff on Amazon for pocket change.

I did just realize though that those cheap powerbank controllers require disconnecting the device while charging too, so those wouldn't have worked for me anyway, I need more of a UPS, because between the board and screen, 2,500 mAh won't last particularly long.

>>2007805
You could use a switched barrel jack so when the male plug is inserted, the battery output is cut off while the battery charger and power supply get power.

>>2007815
>while the battery charger and
*mainboard get power from the power supply.

>>2007784
cost. also led switchers are current mode, if you're using a discrete fet you'd still prefer to drive it with a LED PWM controller like the HV9930 anyway.

any one here know if you can give a op amp the non inverting input reference voltage from vcc if your doing a inverting input? this is with non inverting having a different power source than vcc.

assuming you do both inputs as a different power source than vcc to power it you can. crystal radios can be bosted with a opamp like this

most references to op amps talk of comparisons to the inverting and non inverting then tie non inverting to ground on the inverting circuit. seems like if im powering it with say 6v vcc and want near 6v output i need to tie it to vcc and have the non inverting for the input so when there is nothing coming in the non inverting line it gives a near 6v output

>>2007815
>switched barrel jack
That would at least take a step out of the process, but I was looking more for something like how a cellphone can charge while it's in use.

>>2007826
post schematic

>>2007842
the 40 hz signal is with a different power source and from ground it connects to both power sources negative side and from output it does this as well to allow for anything to cross if need be

would it work or do i need to tie the non inverting to ground?

>>2007856
that op amp is wired as a comparator, not as an inverter, because it has no feedback. post schematic including the wiring of your power sources for more feedback. i can't parse your post.

>>2007856
That circuit as pictured won't work the way you want it to. You can't have an amplifier circuit without feedback. Even with a Gain of 1, you still need it.

What you've made there is a comparator. You've tied the non-inverting input to your positive rail, and are feeding your input to the inverting input. As such, the op-amp will sum your incoming signal to the 6V supply. At best, you'll have an output signal that sits at your rail voltage, and periodically dips on the low side of your input signal (assuming it ever goes negative).

>>2007868
well thats the intent. its a op amp part but i just need it to not go when given power then go and give about the input voltage as the signal when it doesnt. im using like a comparitor but those have a different circuit symbol

its single supply no negative feed. so in this example 0v - 6v. so if it works as intended i get 0v- 5.7v ish for output signal . no feedback doesnt seem like its needed for this operation unless im missing something

>>2007871
>but i just need it to not go when given power then go and give about the input voltage as the signal when it doesnt.
I know all those words, but that sentence doesn't make any sense.
If I'm understanding correctly, you want your comparator to output your rail voltage when no 40hz signal is present? And then to pass the 40hz through when it is? Do I have that right?

>>2007856
You'd be best off just using a rail to rail op-amp if you need rail to rail operation.
Also I want to add that by putting a resistor in series with the op-amp's power rail you're significantly limiting the output current it can source. If you had a dead short load and you assume the loss in the op-amp's output stage is negligible then you can only source 6mA max. With an actual load that some voltage needs to drop across like an LED you can pull even less current and probably wouldn't even manage to light it. Maybe it's not an issue in your application if you're driving a high impedance load but since it seems like you don't know what you're doing you should be aware of it.

>>2007710
FFT it

this might be a stupid qustion
say I have three 3.7v Lithium cells that I have to use to power a 9v supply
What's the difference between wiring them up in series and using a buck-boost converter to regulate its voltage, and wiring them in parallel and using just a boost converter to boost the voltage from 3.7v to 9v?

>>2007970
3 batteries in series will have the capacity of one battery. 3 batteries in parallel will have the capacity of 3 batteries. Both configurations will suffer losses in the buck-boost converter, and you'll need a BMS to safely manage the batteries.

>>2007970
Series with a buck converter will typically be more efficient than parallel with a boost converter. It will also be better for the cells, as you're guaranteed to have equal current draw. Connecting cells in parallel is best avoided.

>>2007986
Only if you use "current capacity" (amps) rather than "capacity" (mah)

LED PWM question. I'm using this driver board (3A output) with a 12V power supply:
https://www.digikey.com/short/zn29m1

It takes a PWM signal up to 20 kHz. I'm using this PWM circuit (pic related) with pot = 10k and R1/R3 = 10k:
https://www.instructables.com/Yet-Another-Simple-Pot-controlled-555-PWM-generato/

It appears not to work, and I'm back to the drawing board. Would I need anything like a MOSFET in between the 555 timer's output pin and the PWM input on the driver board?

>>2007986
>>2008023
Capacity (measurement of charge, usually in amp-hours) will remain the same when connected in series, but voltage will triple. Capacity will triple when in parallel, but voltage will remain the same. In either case, the stored energy (capacity x nominal voltage, usually in watt-hours) will remain the same. Unless you’re using a linear regulator, in which case capacity is the more useful measurement.

As for switching losses, buck converters might be better than buck-boosts, and they should cover three cells in series unless you want to run them somewhat low. It’s not uncommon for cells to be declared drained when at or below 3.2V. Don’t forget to take into account converter dropout.

Bucks are usually more efficient than boosts, so I’d lean towards series cells. That said, you will need balanced charging if you’ve got three cells in series, which is fine if you’ve got a balancing charger to plug it into, but a pain if you need to wire up the balancing protection circuit directly in the device. A parallel cell circuit will still need some protection circuitry, but not as much, and it might be included in the cells themselves. 5V USB power banks usually use parallel cells with a booster, but I can’t speak for the higher voltage USB-C ones.

>>2008032
Cap to ground on CV might help. So might a transistor inverter on the output, to boost the 3.3V output to 5V. You’ve got the grounds connected, right?

>>2008036
Thanks. I see what you mean re: cap to gnd on pin 5. Pic related looks like a better version of the same circuit:
https://www.electroschematics.com/pulse-generator-with-555/

From the driver IC manual:
>A logic input PWM signal to the enable (EN) pin is applied to adjust the LED current. The brightness of LED is proportional to the duty cycle of the PWM signal.

It looks like I don't need anything else. But I'm a newb and don't know how to make a 0~90% duty cycle at a steady 20 kHz with a 12V power source. There has to be some kind of resistor math I'm missing. Would I need any kind of crystal or component to set the 20 kHz frequency?

The application is an array of high-power LEDs for optical use (microscope). Having no flickering is very important, as is dimming the LED at without distorting the wavelength.

>>2008071
555 is not a great chip for that. You can't easily adjust the duty cycle and frequency independently. The simplest solution is one 555 astable circuit producing pulses of a constant frequency and trigging a second 555 monostable that goes high for a variable time.

Is there an easy way to figure out what neon bulb this is? figure it might be an NE-2 but no schematic or model number on this GE flip clock
clock looks like this: http://4.bp.blogspot.com/-VRa-Ck_7hSY/Tl7q_BtskVI/AAAAAAAAAOw/6NnKXgG-yZM/s400/+flip+GE+rectangular.JPG

>>2007962
This. If there are 60/50Hz spikes, or harmonics of it, that tells you that something synchronous is vibrating. Like transformers in a substation. If you see something a bit below your mains frequency (or a harmonic of it), then it could be an asynchronous motor. If it's a multiple of 7Hz then it's schumann resonance.

>>2008071
Did you calculate the PWM frequency properly? It might be too low or high. You can try simulating if you want, I know LTSpice has a 555.

>>2008081
>You can't easily adjust the duty cycle and frequency independently
You can if you're using a potentiometer for it, I think. Not if you're feeding a signal to be modulated into the CV input.
Another method that will work is using a dual-comparator IC, using one as a schmitt oscillator to make an approximate triangle wave, and the other to compare that wave to a pot.

If there's flickering you can also always add an LC filter after the driver, to smooth out the PWM.

>>2008089
will add that it has a 47kohm resistor in series right now

>>2008089
>>2008095
What voltage is it connected to? Pretty sure most neon indicators of the same size will want the same voltage, but I couldn't say for sure. At the least they'll be able to stand the same amount of power, which you could calculate out yourself.
In a pinch you can just buy a similar looking neon, use a 200k resistor, and iteratively decrease the resistance until it looks bright enough.

>>2008100
well i though it was connected to the mains but im only reading 6.5V on one leg and 11.2V at the other leg where the AC cord and neon light legs are soldered together..

>>2008089

def some variation of the NE-2. the small resistor implies it's probably a high-brightness variation like the NE-2U.

>>2008102
>only reading 6.5V on one leg and 11.2V at the other leg

sounds like user error. you measure ACROSS the bulb.

>>2008107
thanks now im getting about 108V across it
>>2008105
this is helpful, definitely a high brightness i think. how can i check the mA of the bulb to be sure?

>>2008112
>how can i check the mA of the bulb to be sure?

you can search youtube for more details, but essentially:
- set meter to AC mA
- cut circuit
- add meter in series across the cut

>>2008114

or better, measure voltage across resistor, and do ohms law: current = voltage / resistance

>>2008033
>Capacity (measurement of charge, usually in amp-hours)
Right, basically, a given number of electrons. It doesn't change regardless of configuration.
What you're describing is maximum current per cell, which in series is the same current through all cells, while in parallel each cells current is additive to the total.

3 gallons of water is still 3 gallons regardless of whether they're stacked or side by side

>>2008114
im getting 0 mA across the bulb with the MM so maybe i did something wrong.

voltage drops about 1v from one side of the resistor to the other, 108.8 on one side 107.8 on the other
ohms law tells me i'd get around 2.3mA which is way higher than anything on that list

>>2008123
>voltage drops about 1v from one side of the resistor to the other, 108.8 on one side 107.8 on the other

this is more user error. you measure ACROSS resistor.

>im getting 0 mA across the bulb with the MM

more user error. to measure current you place MM in series, not across. pls watch introductory MM videos.

Can spraypainted wire double as enameled wire? I salvaged a transformer from a microwave oven and I want to use it as a spot welder. I managed to turn that thick solder cable a couple turns around the secondary but the voltage I get is too low.
I have no space to give it more turns so my only option is to reduce the cable thickness and I'm short in isolator options.

>>2008125
>Can spraypainted wire double as enameled wire?

sure can, but it wont be any good. 99% sure you'll get shorts.

>I have no space to give it more turns

if you have house wiring cable, like Romex, you can short all 3 wires to give you one thick conductor inside a vinyl jacket. or else, cut off the wires from your car's booster cables.

>>2008129
>you can short all 3 wires
Didn't think of that. I'll give it a try, thank you.

>>2008124
sorry i sound like an idiot but i swear ive been doing electronic repair work for a while, just never dealing with these particular things like current.

so if i have it right now, i measure 62mV across the resistor

>>2008142
>so if i have it right now, i measure 62mV across the resistor

very very wrong. if you're operating from 120Vac, then you should see 120-108 = 12Vac across resistor, which results in 12/47000 = 0.255mA

>>2008144
I have a watt/volt meter tool that you plug things into (kill a watt):
it reads ~109V, .02A being drawn on this little clock at the outlet
My MM only reads 108.8V at the point where the Ac cord meets the leads for the neon light and the transformer. I dont understand what's wrong here. even in the picture now im only getting 40mV across the resistor

>>2008081
>>2008093
Thanks. I spent some time looking around if there was a premade solution. I found this but it seems inappropriate:
https://learn.adafruit.com/tlc5947-tlc59711-pwm-led-driver-breakout

Basically I'm looking for any info that can help me meet these specs. This is the signal I must produce:
>0.5V to turn on the LED driver board PWM input
>the power would come from the same 12V source (parallel to the LED driver board)
>4 such circuits, each controlled by a sensitive 10 kOhm potentiometer
>duty cycle range: off to ~90%
>frequency: 20 kHz constant (the max supported by the driver)

I'm also looking for any info about how to make the circuit more robust. The LEDs will populate the round PCB sitting on top of a heatsink. The controller case will be fan cooled from the power supply fan output.

Is there anything I can do protect the LEDs, either on the main controller or the remote LED wafer? Would putting generic diodes after each bulb (about 20 total) be useful to prevent reverse current? The LEDs are arranged like this, each greentext being in parallel. So here there would be 1x3x2 LED series in parallel, each wavelength with its own dropping resistor and the whole thing being 3A max current:
>dropping resistor + 1x 660nm LED
>dropping resistor + 3x 740nm LED
>dropping resistor + 2x 850nm LED

>>2008147

oh boy! lets start over:
- make sure you know where every wire leads to
- measure voltage at point A
- measure voltage at point B
- subtract, which gives voltage across resistor
- divide by 47000 to get current in lamp

>>2008147

oh boy! lets start over:
- make sure you know where every wire leads to
- measure voltage at point A
- measure voltage at point B
- subtract, which gives voltage across resistor
- divide by 47000 to get current in lamp

(assumes lamp is working. if it's off then it's all meaningless)

>>2008162
lamp doesnt work which is why im trying to figure out the specs so i can replace it

>>2008163

ok. makes sense now. you cant figure out specs of a dead lamp, unless there's another one thats working.

>>2008093
>You can if you're using a potentiometer for it,

No. Look at how the 555 works internally. It's swinging the capacitor voltage between 1/3vcc and 2/3vcc. The two resistors set the discharge and charge periods but not in a 1:1 ratio. Pin 3 goes high during capacitor charge. Adjusting either resistor effects both the frequency and duty cycle. Changing duty cycle without effecting frequency requires changing both resistors simultaneously following the equation: R1 + 2R2 = constant . You can't do that with just a potentiometer.

>>2008164
sorry for the confusion.
according to >>2008105 my best best is either NE-2H or 2U

>>2007673
>You can see the brush strokes around DA1-DA4.
Yeah or you know, OP might have tried to brush it off gently.

>>2008166
How badly does the frequency drop with the resistance? I got a mockup in CircuitJS for the 555 timer PWM dimming method, dunno if the DISCH pin is connected right.

If it doesn't drop below the 100 Hz threshold at 0% duty cycle, that's okay. estimated usage per channel is either off or 50% ++ for the most part.

The light will be used at 1000x magnif so having a bright one is usually more important that getting a perfect flickerless light at 10% intensity

>>2008148
Is there any documentation? What voltage does the thing want the PWM signal to be? Read the part number of the 8-pin IC on the board and look up its datasheet.
I'd run your frequency a little lower, maybe 17kHz, since part drift in either device might make it unacceptable after some time.

As for your LEDs, by the sounds of things they're A: designed to run off a normal constant-voltage power supply, B: nowhere near 3A worth, C: quite possibly not as bright as you might want, and D: they're all red and infrared.

Read the datasheets and figure out how to drive your LEDs (without destroying them) without PWM. Then figure out how to PWM them properly.

>>2008166
You need to design a circuit such that as the charge time increases, the discharge time decreases. Using a diode (or two?) you can have the charge time proportional to one resistor, and the discharge time proportional to another. Not forgetting an extra resistor, pic related as the 1k, where the 5k/5k are both sides of a pot. Your duty-cycle is limited by the two 1ks though. There also might be astable circuits using the output for something like >>2008175, might even work better too.

How do you approach component selection?
Just pick whatever seems to fit from your favorite manufacturer? Or do you visit vendors first and see which is the most available? Or do you look for the most popular part that might fit your purpose?

>>2008212
I go vendor-first, and hope my favourite vendor has what I'm looking for. If not, I see if they have a good replacement, before checking other stores. Having to go through a bunch of different sellers with my shopping list is too tedious, and the seperate shipping is a pain.

>>2008197
Thanks. I'll make this circuit in CircuitJS and start playing with it.

LED driver docs:
https://www.issi.com/WW/pdf/32LT3953.pdf
>Logic input for enable and PWM dimming. Pull up above 1.4V to enable and below 0.4V to disable. Input a 100Hz~20kHz PWM signal to dim the LED brightness.

>As for your LEDs
There are 4 mini LED arrays with 3-9 bulbs each. Each array gets 12V DC power and each driver board can deliver up to 3A.

The individual bulbs are plenty bright, and with the exception of the blue bulbs, all draw at least 700mA each (avg is maybe 1.0--1.5A per bulb). Here's the list of all the bulbs:

UV band:
>3x 365 nm @ I=1; V=3.8
>2x 395nm @ I=1.4; V=3.5

Blue band:
>3x 450nm @ I=0.15; V=2.7
>3x 470nm @ I=0.05; V=3
>3x 505nm @ I=0.02; V=3.3

Red/IR band:
>1x 660nm @ I=3; V=2.3
>3x 740nm @ I=0.7; V=1.9
>2x 850nm @ I=1.5; V=1.5

White band:
>3x 6500K @ I=1.4; V=3.05

Anyway, I'll be playing on the circuit simulator until the mail comes. Cuz I already popped one of the driver board IC's and one of the 470nm bulbs.

Can you use WD-40 on electronics (to remove water residue)?

>>2008475
then you have wd40 residue. if that isn't a problem, sure. better to use isopropyl, zero residue

>>2008480
>if that isn't a problem, sure.
That's what I want to know. Can it attack something on the PCB? Does it conduct?

IPA is less convenient and is expensive here.

>>2008484
it isn't great for rubber. not good if you have any trimmer caps or old paper caps. not sure how it does with carbon track pots... might attack might not. oh, don't get it on any rubber pad type switches, gonna muck those up and block conductivity. doesn't conduct but is flammable.

>>2008391
>link
Yep, it's constant-current. It's designed to drive a string of LEDs efficiently without any resistors in series. You might be able to use it with your arrays, maybe, but it kinda defeats the point of having such an efficient driver anyhow. You still need a high-power DC power supply in the first place.
>Each array gets 12V DC power
So, you don't need a driver board to power the LED arrays at all, right? If you want to PWM your LEDs, just chuck a MOSFET on them and drive that. Be it one per array or one for the whole lot. Pic related, for the UV band.

>at least 700mA each
That's a fair bit, what package are they in? I was guessing based off the footprints in the round PCB shown >>2008148, which look like normal 5mm LED footprints. The resistors are also going to need to be high power.

>>2006454
HOW

>>2008584
I have to redesign the circuit, probably to add as many LEDs as I can fit. Most of them are teeny tiny microchips and some of the blue ones are bigger and 5mm.

The current limiting resistors need to be on the LED board along with some TVS diodes. So clearly I want to get as much Vf for the blues (UV, blue, and 65K) and step down to 6V for the IR/red.

This PWM circuit looks like it works, and I need simple pushbutton switches to turn on/off each band. Currently revising the circuit to use these protection measures and add more LEDs:
https://www.digikey.com/en/articles/coordinated-circuit-protection-for-led-lighting

The thing is a microscope light but it should be usable for light therapy at ~5-10min sessions at high intensity

>>2008676
>current limiting resistors
* voltage dropping resistors

>>2008676
>This PWM circuit looks like it works
Replace the 2k with a 1k. Or the 1k with a 2k. Either way you’ll have a 2-4% limit that stops you getting to 0% or 100%. A circuit that uses the output pin instead of the discharge pin might be better in this respect.
If you want a more efficient circuit, make a Schmitt trigger oscillator, and run comparators on its “triangle wave”. Three channels would only need one DIP14 this way, while with 556 timers you’d need two DIP14s. Also no duty-cycle limitations.

>I have to redesign the circuit
If you want to use that constant-current driver, you’ll probably want to only run a single series string off it, with no resistors. If you have a bunch of identical LEDs you could probably put them in a few parallel strings, though with slightly limited maximum current. Personally, I’d get a bunch of white LEDs in 1-4 series strings at a reasonably high voltage (say, 10S) and drive them with a single driver board. Should be more than bright enough. That said, you’ll also need a ~36-40V PSU. If you absolutely need multiple colours (as you conceivably might for better analysing biological stains), then I’d stick to the existing arrays with resistors.

Either way, I’d first do a test to see if the existing arrays with resistors are bright enough at 100% power, from a normal 12V power source.

Again, you don’t need current limiting resistors if you’re using the driver for one type of LED, it does current limiting specifically for the purpose of protecting LEDs.

is this a good and affordable way to charge a ev or will the Chinesium burn down my house?

aliexpress.com/item/1005001504089294.html?spm=a2g0o.productlist.0.0.41773709AnXNs8&algo_pvid=b427c9ba-3639-4950-be5a-c1204f4917c6&algo_expid=b427c9ba-3639-4950-be5a-c1204f4917c6-2&btsid=2100bdf016111858263983282e9fd8&ws_ab_test=searchweb0_0,searchweb201602_,searchweb201603_

>>2008634
>HOW

if you're Russian enough, and have the right tools, you can make magic happen.

>>2008720
>Schmitt trigger oscillator
Thx. Looking at this (555 section). I'd attach that schematic to the out pin and not use disch? ABout to have another CircuitJS session and clean up / update the project repo.
https://www.allaboutcircuits.com/technical-articles/how-to-design-schmitt-trigger-oscillators/

>multiple colours
>analysing biological stains
This is exactly the purpose and why it needs to be many specific wavelengths. I can use the 450/470/505nm blue/cyan channel to excite a wide range of fluorescent proteins, and filter all but the emissions around 530nm green. Same with IR (only pass 900+nm to a modded camera) and UV (only pass yellow+). Blend in cold white as needed or use it alone if stock halogen doesn't satisfy.

I think the problem was trying to copy the specs of a scientific diode/collimating lens that sells for $1200. Instead I need to work with my 12V and cram as many LEDs as the Vf allows, to reduce the bulk of the dropping resistors. Because they are essentially wasted efficiency. On a similar note, I need to step-down the red/IR channel to 6V *before* the signal leaves the controller (the heatsink/resistor/TVS diode wafer will be RJ45 connected).

Heat management is the main concern. I have the smallest heatsink possible due to space, and the LED board has a gnd bus around the outside. I tied figure-8 knots to 3 gnd wires, strung them through the sink screw holes, and will solder them to the gnd bus. I will also wrap bare solid cable around the sink core and solder that to the gnd bus.

Then the other protections such as TVS diodes to dump current in case of voltage spikes, and some kind of PolyZen things (encased zener diode) and PolySwitch resettable fuses

is there something like a relay or optocoupler that can retain the open/closed state without being energized?

>inb4 a switch controlled by a finger attached to a motor

>>2008752

latching relays. it's what they use in some thermostats, so the lil batteries can last a year.

>>2008755
thanks! i've been learning electronics since quarantine, didn't know these existed

>>2008752
There’s also magnetic cores and floating gate transistors. Pretty sure you can make your own magnetic core circuits with cores on eBay, saw some guy make a non-volatile flip-flop or latch with one. But I don’t think you can buy floating gate transistors.

Anyone know what IGBT (lesbian gay bi trans) modules do apart from participate in sexual depravity on /b/?

>>2008815
One broke in my AC 2 years ago.
I replaced it but then something else broke.

In my analog circuits class, my prof. said that in a class A amp, by assuming a separate source for biasing the base voltage and the supply voltage, the base voltage is offset at Vcc(R2/R1+R2), however, I think this shouldn't be the case since Ib exists, by my own analysis, I'm correct, however I would like confirmation if possible.

>>2008848
nvm, I see now.

>>2008815
Higher current than MOSFETs, easier to drive than BJTs. IIRC they’re like a glizzy-pair but with one FET and one BJT; not a new kind of tranny but a composite one like a darlington.

>>2008745
>allaboutcircuits link
I meant more something like this image. But with added resistors on either side of the pot to make the pot's range more useful. Never used a schmitt inverter for this before, since comparators are just things I use in circuits often anyhow.

>This is exactly the purpose and why it needs to be many specific wavelengths
Neat, gonna be interesting to see how it performs. Another method of getting this performance is via a broadband light source (halogen) and some filters. Pretty sure a variable thin-film interference filter is possible and would work well here, but I'm not sure if you can buy them. Diffraction by prism or grating could also work, maybe that's what the $1200 thing you mentioned does. But an array of LEDs is probably the most convenient solution unless you need something really narrow-band or in a region where LEDs aren't commonly found.

Anyhow, do you need to vary the effective brightness of all different colours at once, or just one or two, and swap between which ones are active? Because if the latter is fine, you can use the existing driver board, and a few hand-assembled high-power LED arrays that you just toggle with switches. I'd do this personally, one for a colour of your choice, and the other always for white. With PWM for both. If you require the former however, then you'll need either a bunch of constant-current driver boards, or will have to settle for the compromise of using resistors.
As I've said before, if you use resistors, then just using a MOSFET like pic-related will work fine. No need for a bunch of constant-current drivers.
> Instead I need to work with my 12V and cram as many LEDs as the Vf allows, to reduce the bulk of the dropping resistors
Yes, 3x 3.8V UVs should fit (doesn't look like there's any dropout voltage from the CC driver, if you want to use it), try 4x 2.7V blues, etc.
>step-down the red/IR channel to 6V
Shouldn't you just have more red and IR LEDs in series instead?

>>2007784
>>2007820
to add to this, not just part cost also assembly cost.
A single driver IC might cost $0.8 while a bunch of fets and resistors total would cost $0.50 but now you need to pick and place more parts which blows up the assembly cost higher than the savings from the parts.

>>2008912
That's a bit above my level but thanks for sharing. I'm gonna use the 555 timer circuit with the working simulation: >>2008676

It does what I need, with imperceptible frequency fluctuation at the high end of the PWM input range (about 15 to 17 kHz). All the PWM board needs now (besides = resistors) is a set of preconfigured 12V to 3V step-down converters where I can attach like 3 pins.

>Another method of getting this performance is via a broadband light source (halogen) and some filters.
Most oldschool scopes use it. The bulbs only last 100 hours and the spectrum is diffuse: they fail to efficiently excite flurophores. My green bandpass filter only has a 30nm wavelength so I need bright blue light to get a good image. But I love the warm light of halogen and have space NOS Osram Bellaphot bulbs.

>variable thin-film interference filter
Real filters are 25mm disks of dichroic glass. I have a couple films and plastics, polarizers and IR blockers. Also a good rotating polarizer from a film SLR. The real scientific imaging filters are 25mm glass disks that I hold in front of the dSLR mirror with circles of craft foam.

>$1200 thing you mentioned
Thorlabs sells LEDs encased in lenses (basically lasers). You still need to drive it. I'm trying to make an equally powerful compromise for about $600, turn on the blue/cyan light and you can excite any common green fluorescent protein.

>Shouldn't you just have more red and IR LEDs in series instead?
No. The problem is that blue and cold white (blue) light are high-voltage low-current. Red and IR light are low-voltage high-current. UV is high-voltage high-current. Basically I need to step down for red/IR, to avoid big hot resistors on the LED board.

Pic related is probably the final version of the LED board. It's gonna be crowded in there (50mm diameter board) with the LEDs and a PLED open fault chip underneath. Most of the LEDs are tiny microchips like 1--2mm of the kind you see on star boards.

Is there anything I can do when designing the PCB to make hand soldering (with an iron) some harder packages like QFP/QFN easier?

>>2008939
>the 555 timer circuit with the working simulation
Note that it can't go any less than ~2% duty-cycle. And if you want multiple channels, you'll need more ICs than my solution. It's not any more complicated than the internal diagram of a 555 either.

>Pic related is probably the final version of the LED board
First of all, you're going to need current limiting for every single channel. Can't just say "oh 12V for 4 3V LEDs", you'll thermally runaway and destroy them.
Secondly, for TVSs/zeners, you don't really need them in parallel to the LEDs at all. Put just one set after the power supply, if any at all. You're not going to get voltage spikes unless there's something to cause them. An isolated power supply won't give you a voltage spike at all, maybe with the exception of a particularly violent lightning strike, but I still doubt it.
Are you using the constant-current driver board or not? Are you going to have a bunch of different PWM channels, or switches between a few PWM channels, or what?

>Red and IR light are low-voltage high-current
If the IR and red LEDs will always be on at the same time, and require the same amount of current, you could put them in series. There's also the option of selecting LEDs with more similar characteristics. Sounds like you've got a real hodgepodge of different types.

>>2008956
Add longer solder pads, ones that stick out from under the IC by 2mm or so. Should be possible to drag-solder like that.

Will I fuck my computer if I use a buck converter to step 12V DC up to the 19V it takes at the barrel plug? Do I just need to throw some capacitors across the output? And if so, what value caps should I use for 19VDC at a max of 156W?

>>2008984
Anything I can do about the underpad?

Which would you pick?

>>2009018
It's just used for battery charging, the computer has internal buck converters for powering logic-level stuff, I wouldn't worry about needing too much capacitance. Perhaps 1mF should be more than sufficient.

>>2009047
Ohh, it's got an underpad? Best thought is to add some vias to a pad on the other side, spread some solder paste underneath the QFN, and heat it up from below.

>>2009018
>Will I fuck my computer if I use a buck converter to step 12V DC up to the 19V it takes at the barrel plug?
I assume you have a laptop you want to power by stepping up 12v to 19v? That would be a boost converter, not buck. Or a buck-boost topology. Buck by itself drops the voltage.
>Do I just need to throw some capacitors across the output? And if so, what value caps should I use for 19VDC at a max of 156W?
You don't use large capacitors on the output of an smps unless the datasheet specifically calls for it. Too much capacitance can actually hurt the smps when it tries to start up. You should look for a boost converter to output at least 9 amps at 19v (9 amps is rounded up to the whole ampere). Internal capacitors will handle the smoothing, and again, any external caps will be recommended by the datasheet.
Also keep in mind that at 12v, the input current will be 13 amps to make 156 watts, plus whatever else is needed due to efficiency losses. 80% efficiency would come out to a bit over 16 amps.

>>2008173
>tried to brush it off gently
Assuming it was corrosion from a leaky cap or battery, corrosion will not have those types of strokes left after attempting to clean it.
I've dealt with a lot of old school electronics, that is certainly from a factory worker.

looking to convert an old atx psu into a benchtop psu
it works but makes the high pitched noise just barely audible, even at no load
how do i determine which component is causing it and fix it
just touch all the caps/xformers and see if it stops?

nvm, found it
i need to connect something to the high voltage transistors' heatsink
either a small cap or a large resistor to gnd fix it, but the strange thing is the heatsink seems to be electrically isolated from everything, theres 0v ac and dc between it and anything else i touch

>>2008984
>PWM
I'll try out your suggestion in a simulator. There would be 4 such circuits in total FYI.

>current limiting for every single channel
What should I put there? Got a couple calcs open playing with different voltages and wire lengths (normal: 3 ft 22 AWG). With or without cable voltage drop, most bulb values say "0 ohm resistor needed." Isn't that the same as hookup wire?

https://www.gtsparkplugs.com/Dropping_Resistor_Calc.html
https://www.calculator.net/voltage-drop-calculator.html

>TVSs/zeners, you don't really need them in parallel to the LEDs at all. Put just one set after the power supply, if any at all
Okay thanks. My symbol may be wrong but I was looking to put PLED "shunts" I think they're called (pic related). Basically an open circuit protector that doesn't fuck a whole array if 1 bulb fails. Seems easy enough to solder them to the underside of the bulbs, cuz I was gonna add leads to the tiny microchip format bulbs anyway.

>Put just one set after the power supply
I'm looking at these PolyZen things that are encased zener diodes, 1 after each driver board and then a small PTC fuse too.

>Are you using the constant-current driver board or not?
Yes I already have the boards.

>Are you going to have a bunch of different PWM channels
There would be 4 PWMs controlled by 4 pots.

>Sounds like you've got a real hodgepodge of different types.
All the LEDs are different, it depends on the properties of each wavelength. I found the highest-power ones possible. Each wavelength has its own properties:
>660nm: 3A, 2.3Vf
>740nm: 0.7A, 1.9Vf
>850nm: 1.5A, 1.5Vf

>>2009128
>What should I put there?
That calc is shit, nvm. Resistors come in values < 1 ohm, duh.

>>2007710
It's the ghosts, anon

>>2009123
Sounds like an inductor, or "coil whine" as some call it.
Does the noise change at all under load?
You may just be out of luck, if its a cheap psu, its probably a bad design.

>>2009127
>i need to connect something to the high voltage transistors' heatsink
>either a small cap or a large resistor to gnd fix it, but the strange thing is the heatsink seems to be electrically isolated from everything,
Yeah, there's a reason its isolated from everything, don't go grounding it out with anything.

>>2007827
you can do this with TP4056, what you need is to connect a schottky diode from TP4056 IN+ to LOAD+ and TP4056 OUT+ to LOAD+, this will make LOAD pull power directly from INPUT if there is one, and pull power from battery when there isn't.

>>2009172
how come it makes such a difference in the screeching if it's isolated?
its enough to touch one multimeter probe to the heatsink, the other left floating to stop the noise

This headphone amp specifies max input voltage as pic related, which means it should be ~10Vpp with VDD 5V right?
Also how would you design an attenuator/preamp for a headphone amp? I feel like the DAC output is already pretty high and there should be negative gain.

>Phono
>PreAmp
>Amp

What are the differences?
Signals are just Signals, right? Do You really need multiple stages to give a sum of sine waves more amplitude? I'd have thought one Tube or Audio OpAmp should be enough, right?
(Or multiple tubes when you filter between lows and highs, but I hope you get where I'm coming from)

>>2009212
What happens if you touch it with something insulated, like plastic or rubber?
If it stops making noise when you touch it with a probe on a multimeter, but the other probe is floating, it honestly sounds like a vibration issue. Check all the screws/bolts connecting things to the heatsink, and the mounting points where the heatsink is mounted to the PCB.

>>2009212
What happens if you pee on it?

Hey is there anything obscenely wrong/stupid with the way i'm taking the signals out from this BGA? It's an eMMc so I'm gonna have to impedance and trace length match. I have a (1+2*15+16) 4 layer setup. 2 and 15 on power and gnd plane, respectively.

by the way when choosing between long lines and less vias (like pic) or shorter lines and more vias, what's the pros/cons?

Cheap ali iron update.
I got impatient while waiting for KSGER to arrive, so I decided to dig up my older, shittier iron and try to fix the one I have right now. I replaced the 1k resistor with 47k one to reduce the heat. In a meanwhile I also bought brass tip cleaner to wipe the tip from the black smudge.
It fucking worked, I soldered some bare pads to see if it won't overheat after a while, and it didn't. I keep the trimpot at the lowest value, the iron is still hot enough to melt solder, but it doesn't overheat.
Now I can finally transfer some of my projects from breadboards to perfboards while waiting for the soldering station.

ok guys, give me your best zener diode jokes. GO!

>>2009415
It's no joke, but you can prevent rape if you stick a zener up your pooper. Just make sure it's oriented in the proper direction or it'll actually help you get raped.

>>2009421
No no. Stick it up that way and you'll be able to bring iit the easiest in to the police yourself.

Be careful of breakdown potentials though, that gets messy.

Newfag first time working on a PCB here. Trying to replace the USB-B port on a keyboard. Did I fuck up this pad? Can it be fixed?

>>2009435
Yeah, do you have a multimeter? Scrape the solder mask off the trace next to the burnt pad and cover it in flux. Get some solder on your iron and tin the trace. Then use a component leg or short piece of solid wire that you have pre-tinned to bridge the trace and the pin. Check for continuity and test.

can you reprogram bios chips while they're still in the motherboard or do i need to unsolder
it's an msi p67s c43 b3

>>2009470
You can program them on the motherboard.

>>2009460
>Then use a component leg or short piece of solid wire that you have pre-tinned to bridge the trace and the pin.
Like this, correct?

>Yeah, do you have a multimeter?
I don't. I know it's essential.

>>2009506
>Like this, correct?
Exactly like that.
>I don't. I know it's essential.
Just get a cheap one with basic features.

>>2009128
>Got a couple calcs open playing with different voltages and wire lengths (normal: 3 ft 22 AWG). With or without cable voltage drop, most bulb values say "0 ohm resistor needed."
The voltage drop of an LED changes, it isn't constant. It decreases with increasing temperature. If the LED heats up as it's lit, it will decrease its effective resistance, and draw more current, which will heat it up more, and decrease the resistance, etc. It's a positive feedback loop; thermal runaway. Without some form of current limitation, you'll likely encounter thermal runaway, which will destroy all the LEDs in the string.

If you've got constant-current regulation (like that driver board), then great, you don't need a resistor. But if you've just got a constant-voltage power supply, you're going to need a resistor of ~10% or more of the total voltage drop. 12V white LED strips could use 4*3V LEDs, but they use only 3 of them and a resistor to drop the other 25% of that. 10% is just a guess, I'd go for 15% to be safe.

>1 after each driver board
Hold up, you don't need resistors if you're using a driver board for each series string, because the driver boards are current limiting.

LET ME SAY AGAIN:
EITHER RESISTORS OR CONSTANT-CURRENT DRIVER BOARDS
BOTH IS WHOLLY REDUNDANT
CHOOSE ONE

>I'm looking at these PolyZen things
There's basically nothing that can cause a voltage spike at the output of a buck converter if there isn't a spike at its input. Don't bother.

>an open circuit protector that doesn't fuck a whole array if 1 bulb fails
Absolutely unnecessary. LED lights across the world function just fine for thousands of hours without anything of the sort. They're not incandescents. Check their datasheet, they're designed only for 3V LEDs, and they shunt to 1.2V when active. So if you're using current-limiting resistors instead of a CC driver, the other LEDs will end up getting overdriven. They're only meant to be used with a CC driver.

>>2009144
a .1Ω resistor won't work

>>2009401
Also are there any good resources on impedance matching short of going through a transmission lines book? I use Eagle so it doesn't have a tool for it but I wouldn't be able to use it anyways considering I don't know the concept behind it (I am 4 years deep on a EE degree but I don't have a hard E&M course yet)

how hard is it to make your own PCIE expansion card

>>2009536
Some youtuber called "FesZ Electronics" has a simple video or two on the topic. He uses LTSpice to make transmission lines and shows the effects of proper and improper matching of both lines and terminations/loads/sources. Not sure about anything more complicated, but it was a good intro for me.

>>2009560
I'm also rather interested in this. I'd first take a look at simple things like diy sound cards and maybe some ham radio cards. I think the pinout is reasonably transparent, but actually interpreting whatever protocol that's meant to go through it might require some reasonably fast logic/mcu.

>>2009230
>Phono
Usually a very high-gain amplifier, or sends voltage through the source to get sound. Imagine a guitar, it has non-powered pickups, which are basically inductors with the strings over them. A Phono amp sends voltage through the pickups so that a signal can be produced.
>PreAmp
A smaller amplifier, used to bring a low voltage source up to a higher voltage with correct impedance for a high-end amplifier. A preamp isn't designed to drive a speaker directly. It takes a signal and conditions it to go to the larger amp.
>Amp
More commonly a "Power Amp" this is what most people think of with "amplifier". It drives speakers with lots of power.

Most commercial units have a preamp built into them so its a preamp and amplifier in one. Or, a lot of modern devices can properly drive an amplifier that isn't very high power.

>>2009412
>In a meanwhile I also bought brass tip cleaner to wipe the tip
Did you never have one? You should always have a brass sponge or regular sponge (dampened with water) to regularly wipe the tip of your iron while you are soldering. Flux will build up and burn up on the tip, so you need to keep it clean. Also, before shutting the iron off, coat the tip with a blob of solder and let it cool/harder on the tip to seal it while not in use.

>>2009596
>Did you never have one? You should always have a brass sponge or regular sponge (dampened with water) to regularly wipe the tip of your iron while you are soldering.
I used a dampened sponge, but the black coat would not go away, It was oxidized too much because of heavy overheating.
>Flux will build up and burn up on the tip, so you need to keep it clean. Also, before shutting the iron off, coat the tip with a blob of solder and let it cool/harder on the tip to seal it while not in use.
Thanks, I did coat the tip before unplugging the iron.
Before fixing the iron I couldn't even solder a single joint after 1-2 minutes, since the temperature must have gotten so high that the flux in the solder melted before reaching the tip and the tip oxidized extremely quickly.

>>2009506
>Like this, correct?
>>2009513
>Exactly like that.

wrong. it's better to run a wire between the two endpoints of the trace. that way, there's no possibility of
- you cutting yourself with the x-acto knife
- you destroying the trace even more
- you not removing enough solder mask to make a good joint
- you causing a short circuit coz of cramped quarters

>>2009230
Also a phono preamp corrects for changes made when a signal is engraved in a record master, look up RIAA correction for a better explanation

>>2009653
Do you think he could do it without destroying the other pad?

>>2009653
>x-acto knife
Tiny slotted screwdrivers work pretty well.

>>2009661
>Do you think he could do it without destroying the other pad?

yeah, coz a noob will have an easier time soldering to a pre-tinned solder pad, than an untinned trace.

>>2009595
>Imagine a guitar, it has non-powered pickups, which are basically inductors with the strings over them
>A Phono amp sends voltage through the pickups so that a signal can be produced
Isn't it normal to have pickups made from wire coiled around a magnet, which don't require any phantom power?

>>2007321
metal oxide > carbon film because of stability but i dont think you can find many metal oxide potentiometers. even 2 wats are carbon most of the time

>>2009560
I forget where, there was some old blog where some guy DIY'd his own soundcard. I think it was PCIE.
Coulda sworn I bookmarked it somewhere...

i want to upload a binary to my stm32 via an stlinkv2 dongle
>says CKS32F103C8T6
it's fake isn't it
should i dispute? apparently the cks fakes work fine

>>2009706
It's one of the more popular clones. They started using those instead of stm32s in those T12 stations.

>>2009712
oh really

>>2009706
If you ordered it from a reputable source it would make sense to dispute it. If you ordered it from ebay or aliexpress, you got what you paid for.

>>2009516
>If you've got constant-current regulation (like that driver board), then great, you don't need a resistor.
Okay. I'll remove the dropping resistors each each LED string.

>So if you're using current-limiting resistors instead of a CC driver, the other LEDs will end up getting overdriven. They're only meant to be used with a CC driver.
I'll keep the shunts because they make 12V and 6V versions and I'm using constant current drivers.

That's it for the LED board I think. Onto the main board.

>PolyZens
It's hard to know what's a meme, and these are apparently a meme. Gone from design.

>PTC resettable fuses
Are these a meme? My AC power input is good and it has a fast blow fuse in it.

I guess the easiest question is, what in pic related can I afford to lose, while still having a safe circuit for ~30 really bright LEDs? Consider the PolyZens gone, also the dropping resistors from the remote LED wafer.

>>2009716
It was from ali.

>>2006757
unironically this

I need to crimp two fucking sides of one fucking fuse.
Why do I need to buy a whole fucking set?

>>2009827

good point!
also, why do i have to buy a whole box if all i need is one tooth-pick. life's just so so unfair.

>>2009836
It's true, though. We could all just hold individual stores of common materials like this for people who literally only need one.

Will pic related work?

>>2009719
>each LED string
By "each LED string" I meant each series string. So if you've got a 365nm series string next to a 395nm series string, you'll want a CC driver for both. Properly adjusted to 1A and 1.4A, of course. If you're using resistors instead, then you'll naturally want one resistor for both of those too. Technically there's some compromise to be made, by setting the CC driver to 2.4A and balancing the two strings to one another with resistors. But you'll need to drop 1-2V across each resistor in order to get the balancing done right (or use some sort of constant-current diode/regulator), which will still mean a fair bit of power.

To be honest, considering you're currently planning on having 4 of those $16 CC driver boards, you've got the room to put them somewhere away from the LED arrays. So I'd instead just use resistors for the lot, and have the resistors somewhere externally. So long as you're not running on a battery, the ~75% efficiency isn't going to matter. Throw an LM2596 on the lower voltage ones if you want.
Actually LM2596 modules are cheaper and smaller than what you're currently planning to use, and you just have to replace the inbuilt trimmers with external pots to control their current.

>>2009862
What is the BMS supposed to be doing? Is there even a battery in your circuit?

>>2009866
it's going to be connected to the batteries, I am wondering if it will charge the battery and powering the load at the same time

>>2009868
The CN3722 IC is designed to give a proper lithium ion charging curve. It will provide the programmed charging current, to the system, regardless of how much current the load is pulling. Say you've programmed it with a resistor to be at 1A, the battery would normally be being charged with 1A without a load, but if an 800mA load kicks on (measured at the input of the XL4015), you'll only get 200mA going to the battery. So it will work, just at a lessened capacity.

Since the charge current varies with the MPPT's solar reception, the only way to get ideal UPS charging is by somehow modifying the current feedback signal on the fly as a load is pulled (possible, but not terribly easy) or buying an MPPT controller specifically with that feature.
Going more into detail of the first option, you could remove the connection from the current shunt resistor to the CN3722, add an identical current shunt resistor at the XL4015's input, subtract the two values with an instrumentation amplifier circuit, and feed the result back into the CN3722. There might be other ways of wiring it to do this passively, like connecting the XL4015's input before the shunt resistor, or even directly to the solar panel, but I'm not sure if the battery would power it when there's no solar if you did that.

>>2009876
>So it will work, just at a lessened capacity.
ok, that's good enough for me. I was asking this because I am currently using TP5100 and it doesn't work if there is a load connected to it. Thanks a lot anon, I can't believe 4chan is more helpful than fucking stackexchange.

>>2009877
Any charger should work like that if you just put the load in parallel with the battery. Though in either case it's possible that the charger objects to the effective load resistance jumping about and cuts off.

Are PLCs really needed for all applications where robustness is needed? Or are corporatives just insecure whiners?

>>2009878
yea I thought the TP5100 setup would work, and I heard TP4056 doesn't work too if there is a load connected while charging.

>>2009827
use pliers

Oscillating guitar effect anon back!

>>2005308
>>2004711

My schematic was wack and I had no idea what I was doing.
This time I have it looking decent enough for some troubleshooting.
Volume seems to be way too high and when the knobs are turned all the way up the second stage op-amp oscillates in a very high freq.
What components could I swap out to reduce the volume?
C11 looks like a good candidate but I don't want to change to tone of the output by swapping components without consdering how that works with all the other components.

>>2009939
This schematic MAY have mistakes, it's fresh from last night and decifering veroboard layouts is a little convoluted https://tagboardeffects.blogspot.com/2014/01/ibanez-ts808-with-bass-control.html

move over you hobos
>>2009959
>>2009959
>>2009959
>>2009959

>>2009879
No, just use a Raspberry, it will work just fine.