/diy/ - Do It Yourself

>>1758363
>120V AC
*laughs in transmission losses*

I have to solder pins to my Raspberry and wires to some other boards. Can anyone give a run down on how to ESD proof myself? I know that there's a bracelet i'm supposed to wear to take care of the charge i might build up but i have no radiator to hook it up to, and honestly I have no idea whether a water pipe will do.
Any way to just wing it with basic precautions like discharing myself before beginning and not wearing synthetic clothes or will the few meters between the pipe and the workstation fuck me up?
Also how do i check if my hobby soldering station is ESD-safe?

>>1758390
Honestly I've never bothered with bracelets, ground planes etc. and never fried anything. I think as long as you're not wearing double wool socks over a carpet and a synthetic vest you should be ok. Just touch something metal before, like you said.

>>1758392
is a large metal mass enough or does it have to be grounded?

As someone without any formal training in electronics, I thought this might be a good place to ask:
I want to design a PCB that has one wall adapter input (Since I'm in europe, that's 220V/AC) and can drive devices that need the following voltages:
5V DC
8.5-36 V DC
9-30V DC

IIrc, common voltage levels are 5V and 12V DC, so I'd like to supply those. To get the proper voltage I thought I would use transformers. If I do so does it make most sense to use two transformers in parallel, both with their respective windings to get me proper voltage levels? I'd make them DC by using a full bridge rectifier afterwards and a capacitor for smoothing.
Alternatively I could use the full bridge rectifier first and use a switching voltage regulator afterwards.
Anything else (stepping down voltage with a linear regulator or something) seems wasteful to me if I am honest.

Basically how do I get a proper power supply for my circuit, assuming I want to design one and not use a pc power supply unit or something.

>>1758395
Much better if it is grounded, a large mass should be enough however. I always wear a bracelet for soldering on the RPi, but I didn't in the past and nothing ever happened. Don't worry too much about it.

>>1758398
I think in theory you could use a single transformer with two taps on the secondary side, one at the full 12V and the other at 5V. Dunno how practical a design that is though.
Is your project specifically building this supply? If not, maybe just buying adapters (think phone charger) might be an easier, better, faster, and cheaper solution.

>>1758399
Many thanks, can i do the same trick for the solder, touching something with the tip once it's hot or does the charge build up again shortly?

>>1758405
meant soldering iron

>>1758406
Do it once before starting, after that, don't bother. If your iron is so shitty that it builds up charge while working, tapping it against a grounded object isn't going to save you. If not, discharging once before starting is enough.
All in all, I've had some pretty extreme stuff happen to my raspberry Pis and so far, no issues. They are surprisingly resilient.

>>1758403
>I think in theory you could use a single transformer with two taps on the secondary side, one at the full 12V and the other at 5V.

A great idea, but I'm not really going to get a custom made transformer. Maybe there's good ones out there. I might have to go with buying an adapter, I'm sure there's some that can supply 5V and 12V DC at once, not like those are rare Voltages.

>>1758398

if your current requirements are low, you can get away with using 3 nearly identical circuits, all powered from the same 30Vac transformer. the LM317 is noob-friendly. sure it's not efficient, but AC power is substantially free, so who cares. to get the 9V minimum, you'd add a fixed resistor in series with the pot. you can figure the value by trial-and-error or using the formula in the LM317 wikipedia page.

>does it make most sense to use two transformers in parallel

never! you only connect transformers in series. you could, for example, wire a 24V and 6V to get 30Vac, which when rectified and filtered, gives you about 40Vdc entering the LM317, and 38V exiting it.

>>1758193
It's the lamp end. The battery pack has the same connectors, but there's no length to them so I can't reach them into the case to plug in the other wire.
>>1758207
I don't have easy access to a soldering iron and even if I did, the wire is too far in for me to effectively do that. There's about enough space for me to reach in with a couple of fingers to attach a wire, but not to do any kind of work with a soldering iron.

>>1758412
>I'm sure there's some that can supply 5V and 12V DC at once

There are, and i own one

>>1758423
is that a capacitor? Holy that's a big boy

may project is finally finished, yay!
god i hope it won't explode when i plug it into the mains

How do I make a PCB? Is there a good YouTube video on it

>>1758434
>>1758437
i made this one on a desktop mill, much better than nasty fecl3

>>1758438
link desktop mill pls

does anyone know why a 13 line program uses 13 kilobytes of memory? Do the arduino libraries really add that much overhead?

I am tired of flopping around with wick like a pleb. Fucking useless piece of shit.
So i want to buy either a high end solder sucker or low end desoldering iron (they both cost the same)
which one do you think is best? I imagine the desoldering iron even a cheap one should be leagues above any plunger sucker right?

>>1758450
>Do the arduino libraries really add that much overhead?

all libraries do that shit. adding 1+1 and doing ''cout'' with the result creates a 250K binary in C++. that's like half of the entire original Macintosh OS

>>1758437
>Is there a good YouTube video on it

unfortunately there are thousands of such videos, and too many different techniques to count. find the one done by the 12yo girl, that's the best one.

>>1758495
>I imagine the desoldering iron even a cheap one should be leagues above any plunger sucker right?

if, by a desoldering iron you mean a soldering iron with a desoldering attachment, that is shit. if you mean a dedicated desoldering station, that's way better.

>even a cheap one

wrong. cheap ones will
- not melt solder
- get tip clogged
- get solder reservoir clogged
- be too weak

>>1758495
>I imagine the desoldering iron even a cheap one should be leagues above any plunger sucker right?

if, by a desoldering iron you mean a soldering iron with a desoldering attachment, that is shit. if you mean a dedicated desoldering station, that's way better.

>even a cheap one

wrong. cheap ones will
- not melt solder
- get tip clogged
- get solder reservoir clogged
- be too weak
- have a pump that weakens and dies after a few months.

>>1758501
>f you mean a dedicated desoldering station, that's way better.
when i say cheap one i mean cheap one as in this shit
https://www.aliexpress.com/item/32951350667.html

Have you tried using them tho? I just can't imagine they are worst than cold suckers

>>1758503
>Have you tried using them tho?

nope, never tried anything under $300 or smaller than a 4 boxes of kleenex. have zero confidence in a $20 machine.

>>1758363
that pic should be reversed. so that L is common ground and N goes straight to the LM7805.

>>1758390
I just put my feet on the radiator
top comfy especially during the winter months

Redpill me on stm32 nucleo boards /ohm/.
What are they good for? Are they best-in-class? Should I ditch lolduinos and espee32s? Should I be posting this in /mcg/, the deadest general on this board?
So many questions, so little time...

https://www.st.com/en/evaluation-tools/stm32-nucleo-boards.html

>>1758543
learn to pick the right tool for the job
i use attiny13 for like 90% of my projects

>>1758542
No radiators in my house, got underfloor heating

>>1758543
they cost less then arduinos, have already a lot of examples available so you can get a hang of it.
If you're just starting with stm32 I'd suggest getting some discovery board, it does cost a bit more but you get onboard electronics/sensors with proper examples.
It really depends what you need them for, sometimes bluepill is just enough and costs next to nothing, getting a bigger nucleo board is a must only if you know exactly that you will need most of it's functionalities. Read about the differences between boards and see what you need for your projects.

>>1758495
Use the right wick. Cheap wick sucks. Try some Chemwick. Same with desoldering tools. Cheap ones aren't that great. $2800 weller station is pretty good but even that craps out sometimes. And out comes the Chemwick.

What's a simple way to go about digital-to-analog conversion without resorting to a DAC IC?

I'm taking an input audio signal into an stm32, then low/band/high passing it, and outputting the respective bands on the output pins. It's output as a PWM signal though, not as an actual analog signal.

I know that Class D amplifiers do something very similar with a low pass filter. Could I just rig together an active op-amp RC lowpass filter and get decent results? I just need it to drive an LED display, not for listening.

>>1758398
Do the outputs need to be able to sink current as well as sourcing it? Do they need to be isolated from one another? How much current do you want to draw as a maximum?
A multitap transformer is a pretty good way to go, and 12/5V transformers shouldn't be uncommon if you browse about on alibay. Might need to buy some big caps though. Using a linear regulator to get variable power is one idea, but you can buy an LM2596 buck converter module for $1 that can handle 3A and will spit out anything from 36V to 2V IIRC, and has spots for current and voltage adjustment. Just have to replace the trimpots with potentiometers, and run it off a ~36V SMPS. Then there's self-inclosed ac-dc SMPS modules for a few dollars each that can handle an amp or three, which you could shove in a case together.

>>1758434
I'm 80% sure that's FR-2, FR-4 doesn't look nearly that brown. Ya might have got chinked.
Running a craft-knife through an off-cut to see if you hear the glass fibres should be a test enough. Or scratch a bit off and burn the resin off to see if there's any fibres left behind.

>>1758437
I use BigClive's method with the inkjet transparency and dry film photoresist.
Directly printed to my printer from KiCAD, the best circuit design software.

This is a Wheatstone Bridge. Say something nice about it.

>>1758716
ur not even a real bridge faggg

Question for bored hobby-fags, any recommended kits? I'm going to build a LED cube for my nephew, but aside from that I've got nothing. I understand electronics enough to work out a schematic, but I can't ever find something to do.

>>1758720
I soldered together a guitar pedal kit the other day. That was fun. https://buildyourownclone.com/

>>1758721
Reminds me of a USB foot pedal idea I had for an fps, best I had was a sewing pedal at the time, witch inspired me, never followed through. How do guitar pedals feel?

When submitting pick and place positional data to a board house how do they know what orientation the part starts at? For example, if a part is rotated 45 degrees do they match up pin 1 somehow or is there a standard 0 degree orientation?

>>1758720
Elektrosluch is pretty simple but offers some fun, so long as you don't go deaf when the microwave turns on. How old is the lad?

>>1758735
What would you use it for? I'd quite like to have a rudder bar or something for tank steering myself. I know joysticks use potentiometers in a sort of gyroscope housing, not sure what on-off pedals use but probably just a heavy duty button and an enclosure built to actuate it. I'm assuming continuously variable pedals use potentiometers in some sort of housing.

Does anyone know any joystick ICs that I can just solder to a potentiometer and go? One that handles the zero-point properly (self calibrates on boot) would be nice too, I've had issues with that before on a cheap joystick.

>>1758739
I think it's more that there's a standard "pin 1" for every footprint under the sun, and that pin numbering information is contained within the gerber file or whatever format it is.

>>1758741
It's not in the Gerber. At lease not on the copper layer. You can indicate pin 1 on the silkscreen but a human would have to look at it and decide. Which may be the case, I just assumed it was automated. Positional data gives an X, Y, and rotation. It's easy enough to see which way the part must go on but most parts could be installed 180 degrees without issue.

>>1758619
turns out it is, fr2 is used for single player boards because it is cheaper and the main advantage of fr4 is that you can make vias and multilayer pcbs much easier
So if the pcb is single layer it doesn't really matter which flavor of fr it is

>>1758748
For example, this TQFP in the positional data is rotated 45 degrees. Where is pin 1?
In the packaging section of the data sheet Pin 1 is top left. Which means the output from the design software uses counterclockwise rotation. Is this standard?

>>1758752
But FR-2 in my experience delaminates easier and is more prone to water damage. Not to mention its better mechanical strength, which naturally matters more for a single-sided board as there isn't a piece of copper on both sides for support.
It's pretty easy to find single-sided FR-4 on alibay, finding 2-sided FR-4 is the hard part. Still bought like two of those 10x20cm 2-sided bits though.

>>1758755
But is that vector file the only information the pick-n-place uses to place components? That would require some sort of advanced OCRing to figure out which footprint is which, as opposed to a simple svg file of part IDs, coordinates, and angles.

>>1758756
>But is that vector file the only information the pick-n-place uses to place components?

No there is also the positional data which is a separate file. So each component will have an X, Y and Rotation for it.
The board house does say in their FAQ that positive rotation value is counterclockwise. Might not be the same everywhere but I'm okay in this instance at least.
The last assumption I'm making is the starting position (0 degrees) must be what is shown in the datasheet. Otherwise the assemblers would not know.

>>1758759
Like so.

>>1758761
Well that could easily be tested by going through datasheets of different manufacturers and see if they all have the components in the same direction, or have a default orientation specified, and also have the same pin layouts (which can get a little confusing when ICs have multiple pins commoned together, or a missing pin like some HV ICs and optos. I imagine a general rule is "pin 1 is closest to the top left" and the rest can be figured out by the pin numberings, if the pin numberings are truly universal

>>1758756
that is only problem in wearable and outdoors electronics, if you have your project at home it won't be an issue and if i want to use the thing outside i will cover it in molten silicone

>>1758763
Delamination when soldering is a very real problem in any situation. That's why I stay away from FR-2 perfboard, if ever I want to desolder something it just pulls the pad right away with it twice as easily as if it were FR-4 perfboard. Which is still easier than desoldering from 2-sided perfboard with plated through-holes. I just hate perfboard in general though, so ymmv. I also like to use PCBs as semi-structural components, so the strength helps, but I can see how this isn't an issue for most people.

>>1758767
One huge advantage tho is that since fr2 is basically just paper and resin, you won't get instant silicosis when you mill it and the dust is far less abrasive to the machine.
And as i saw from my prototype fr2 can be milled down to 0.2mm traces without any issues, so i would actually say it's better for me than using fr4, and much cheaper too. 10 pieces of 5x10cm single sided fr2 boards cost only $3. There is no way they sell fr4 of that same quantity for cheaper than that

>>1758781
>you won't get instant silicosis when you mill it and the dust is far less abrasive to the machine
Oh yeah didn't think about that. Not sure about the resin dust though. I usually do that shit in a place with good ventilation but that isn't really an option with a dinky little indoor router/mill like yours. And for price and cheap prototyping there really is no downside. Though I'd still keep a couple FR-4 sheets lying around for those occasional permanent quick-fixes you mentioned.

It's working!
But now my hunger for diy electronics projects grows and i need MORE

to charge a two cell lipo properly do i have to charge each cell separately or is it safe to charge them in parallel without reducing their lifespan?
Should be fine right? Since if they are always used in paralel they can never become unbalanced

>>1758900
Iirc Li-Ion packs have a small PCB that keeps them balanced when you recharge them. No idea about LiPo.

kek at 80s slav engineering
can you guess what this monster thing does?
no?
it's literally just a clock that shows time on a 7 segment display that was on soyuz
what the fug slavs, one small ic could probably do the job

Hello lads! I'd like to build an antenna for my baofeng and was wondering if anyone here has plans or instructions on how to do that? I've never done anything with radios/antennae before but I'm looking for a fun weekend project so I can learn more about how everything functions.

>>1758973
What are they using for the timing?

>>1758981
You mean like a large stationary antenna, or a portable one? If it's portable, just buy it. If it's stationary, you might still be better off buying one, because a baomeme doesn't operate at frequencies low enough to need a terribly large antenna.
Still, try asking on /ham/ or /hrg/ or whatever it's called. The amateur radio general. There's a bunch of interesting antenna topologies to experiment with anyhow, like a Yagi-Uda, Magloop, ground-plane (the one with 4 reflector things sticking out below the antenna)

>>1758981
you might be better off over in >>>/diy/ham/

>>1758973
those ICs look aesthetic as fuck

Whats the point of /mcg/? Was it made when the Arduino and RPi became suddenly popular?

Reposting:
>What are some resources about SBCs? I'm lost on what to get for specific budgets and applications.

>>1758716
Let's say I have a Wheatstone bridge. I don't know any of its resistor values initially so I take an ohmmeter and measure each resistor in the bridge. The measured resistance of each leg is not the true value of that measured resistor but actually the measured resistor in parallel with the other three resistors in the bridge.

Is there any way to derive the true resistance value of each resistor in the bridge after having measured the loaded resistance of each resistor in the bridge?

This is a real Wheatstone bridge, not an ideal one so none of the legs have exactly identical resistances. They vary by a few 10s of ohms. In the actual application the difference in bridge resistance is nulled in out in software by an ASIC.

>>1759261
It's the containment board for people who think that flashing an LED requires a microcontroller

>>1759261
/mcg/ popped up within the last half year. Some college student was shitting up the thread with easily google-able questions and people told him to google it. Eventually a thread was started about easily researchable questions and it's persisted for some reason ever since. That's not to say that it's bad in and of itself, there's quite a few of those threads here, but general topic of discussion is limited and a lot of the questions would get quicker responses on /g/ or simply require looking up documentation/datasheets. So it's kinda in limbo.

>>1758434
>>1758841
nice

>>1759263
Just take your measurements and write a python code to wander through a phase-space of the four resistances looking for the point of least error. Actually it might only need to be in 3 dimensions since you could normalise them with respect to one of the four. The hardest part would be constructing the algorithms that represent the effective resistances.

Applying a voltage to it and measuring the voltages at each node, then measuring the total current might give you easier results to work with, possibly even ones that are easy enough to solve analytically.

>>1758363
I need help with some light bars.
12v, 7w, 36 LEDs
I have 2 of them, and I need to power them with 24V power source

Easy just going to wire them in series and it should work fine from a power standpoint.

My question is, they are both too long.
If I cut the strips in half, is that going to affect the voltages?

>>1759282
>If I cut the strips in half, is that going to affect the voltages?
No. Every 3 LEDs is a place you're meant to be able to cut through the LEDs, as each set of three of these LEDs is in series. Then each group of 3 is in parallel with every other group of 3, so removing one or two or half will just remove some series LEDs (and resistors), having no impact on the voltage across the rest of the LEDs. But the total current will go down. If you want to use both strips worth of LEDs, solder wires to the ends of the other half too and wire them up. For a 24V source, it wouldn't matter if you soldered the two strips in series, or the two half strips from a single strip in series. If using all four half strips, you could solder either to both sides of your strips, or just one side. Plenty of options. If you just want two half-strips worth of LED lighting, you only needed to buy one in the first place.

>>1759279
I am able to measure all the node voltages. The only thing I cannot do is physically break the bridge since it's not made from discrete resistors but rather laser trimmed thin film resistors similar to pic related. If I break it there's no way to restore it.

>>1759288
Yeah I'd measure it like this:
If I'm correct, both voltage measurements will only be dependant only two resistors, assuming the PSU impedance is significantly small. You'd be able to find the ratio of both sets of two resistors just like that. Then flip the bridge 90° and measure it the other way and calculate those ratios. Now you've got the ratios of both orientations, you should be able to calculate the ratios between all four resistors. Then use the current measurements to backtrack what the total value of each resistor is.

>>1759287
Alright that makes sense. Thanks

>>1759291
Can't put an ammeter in circuit. Thin film network. Breaking the connection to ground to insert an ammeter would involve breaking a wire bond that cannot be easily reattached.

>>1759295
It isn't breaking a fucking connection unless it's already somehow connected across the fucking voltage source with your thin fucking film network

can someone post the nixie tube 4chan /diy/ banner that an anon made? i just saw the noose one that got picked in the end and it made me depressed

Just a friendly reminder that if you are from EU and are stocking your parts for cheap monies from chinks, make sure to save up your money and order massive amounts of anything you will need in the span of this year (and in the time to be delivered this year) because on the 1st of 2021 it's game over for chink shops in eu

>>1759361
>on the 1st of 2021 it's game over for chink shops in eu
How so? A new tax? Or a flat-out embargo? What about ordering from more professional sellers like LCSC? Does this apply to industry?

>>1759364
the 22 eur limit is going bye bye and EVERYTHING will be taxes and on top of that you will have to pay handling fees for processing the tax and on top of that the delivery will take several months because the customs will no have to go through 10 times more packages (since 90% of them were under 22yuros and were just left through without any taxing)

>>1759366
>charge more taxes
>have to spend even more to get them inspected
sasuga eu

>>1759172
>What are they using for the timing?
Crystal right there in the pic.
But there was an option for external input.

>>1758973
>>1759373
I swear there has to be some sort of discrete digital formatting circuitry going in there, unless the ICs are all a bunch of NANDs or NORs making T-FFs for clock division. Might as well use discrete transistors at that point like those novelty clocks, though the TO-92 presumably wasn't around back then so that would just mean more metal cans.
The left-most board has some large components that make me think it might be a power supply or maybe an analog circuit of some kind.
And is that 2-board stack their method of 2-sided PCBs? Why not just put copper on both sides of the fibreboard?

YES! My baby is finally here

>>1759273
What about a containment thread for people who think that flashing an LED requires an IC? :v)

>>1759387
As opposed to using 10 different components to do it the analog way instead of using a single micro you smug cunt? Fuck you

>>1758973
the board on the left is a prehistoric switching power supply used originally by cavemen
those big round white things are optocouplers, but not actually because they weren't invented yet so they are actually isolation transformers used as octocouplers

>>1759391
Dude, bistable oscillators are literally the easiest thing ever, give desynced pulses, and don't take up half a breadboard

>>1759375
>The left-most board has some large components
It's a power supply. There's a video about this thing on YouTube
https://youtu.be/GmetgNPL-HQ

>>1759282
>I need to power them with 24V power source

what the other dude forgot to mention is that if you put them in series, each one must have the same number of LEDs so that the voltage divides by 1/2 exactly. otherwise the one with more LEDs will be overloaded.

>>1759282
>I need to power them with 24V power source

what the other dude forgot to mention is that if you put them in series, each one must have the same number of LEDs so that the voltage divides by 1/2 exactly. otherwise the one with fewer LEDs will get more than 12V.

How do i re-cap my amplifier without killing myself /ohm/?

Anyone know any retailers in Canada that sell 300v/500v electrolytic caps so I don't get fucked by customs?

>>1759515

active tech has several walk-in stores, and they all handle mail orders AFAIK. and each city will have diff stores, so check yellow pages.

OK /ohm/, help me ID / find a part. My customer needs Q33 in pic related. Clearly it's a MOSFET (G D and S leads labeled on board) and manufactured by International Rectifier. The part number should be at the top, which I read as either 093055 or D93055. The 747D is the date / production code (week 47 of either 1997 or 2007) at production facility D. I can't find one iota of info on the part number 093055 or related suspect variations. Can /ohm/?

>>1759559
Nevermind /ohm/, I found it eventually, it's a house numbered part for a Miller welder. Shame the customer couldn't just tell me what it came out of.

>>1759275
>quicker responses on /g/
Quicker, maybe. Useful? Unlikely. /g/ is a fucking dumpster fire of shitposts.

what's the point of a segmented display like this?

>>1759595
AD-24 alphanumeric display. 24 segment. Designed by the Asyck company in Japan. Google is your friend.

>>1759606
bruh I know what it is, that's where I found it.

I mean what kind of alphabet won't look like total ass on this

>>1759609
not that bad

>>1758720
I've seen you post on /biz/ before, can you tell me where you buy those bulbs that go around the LED number? Just curious anon. I'm not trying to steal your side hustle.

>>1758496
>adding 1+1 and doing ''cout'' with the result creates a 250K binary in C++
>that's like half of the entire original Macintosh OS
Look how far we've come anons. Sometimes it still surprises me.

>>1759622
You mean the nixie tubes? I had to travel back in time to buy mine.

>>1759387
The gain of a comparator makes it much more consistent than NPNs, but I'm not opposed to transistor-based oscillators. Especially sine wave oscillators. But if I'm already using a comparator or two, I'll use one for my oscillator.

>>1759613
That's actually really good, better than those 14/16-segment displays. Probably better than those whacko EL displays too, but those are cool for their own reasons.

>>1759622
Those aren't LED numbers, those are nixie tubes. They cost like $20 each, often more, making them a premium item for electronic projects. They're also a pain to drive, requiring high voltage transistors/drivers and some sort of HV power supply. But very rewarding when you put them in a project thanks to their retro/soviet aesthetic. Unless you put fucking blue LEDs underneath them.
You can buy them from ebay, but tubes-store.com is a source I'm somewhat fond of for the russian ones. And remember, the upside-down 2 as a 5 is a good thing.

>>1759609
Admittedly the T is a little weird, but thinking about it, I can see your point. When you go as far as 24 segments you might as well go full dot matrix. 5x7 characters would only be 35 segments.

>>1759661
I actually really like stylized segmented displays but I don't think this exact configuration makes sense. I've seen other eclectic alphanumeric displays that can do far more with fewer segments, and I'm wondering if there's more out there I haven't found yet.

https://www.youtube.com/watch?v=D2cyZoCat_U

>>1759661
>>1759676

Specifically what I'm mulling over is a way to create custom segmented displays that can display futhark runes in a similar stylized way. I could do led matrix in the end but I'm trying for a specific visual flair.

>>1759659
What can you tell me about wacky looking segment displays and EL displays? I wanna know

>>1759679
This is full of awesome. If you look at the Asyck company page about the AD-24, they offer custom displays.

>>1759682
https://tubes-store.com/product_info.php?products_id=660
Can't find a datasheet, but IIRC there's an old looking site somewhere with a bunch of different EL segmented displays being reviewed. The name IEL-0-IV refers to a smaller size than the IEL-0-VI, so I imagine all their names follow that scheme. Pic related and the link is the IEL-0-IV, but some VIs show up when googling. Hunt about and see what you can find. In a pinch you can make your own EL displays, Applied Science did a more complicated one; the apollo guidance computer's display.

Also see:
>http://www.tube-tester.com/sites/nixie/nixie-tubes.htm
A great resource listing every nixie I've ever seen with comprehensive reviews of them all.
>https://www.gstube.com/index.php
A good resource for datasheets of all sorts of vacuum tubes. May or may not come translated.

>>1759710
bruh this is WILD

>>1759297
The circuit has two "grounds." The LO side of the bridge is not connected to the ground pin on the device. The LO side of the bridge is supplied by buffered voltage reference and there is a difference between that "ground" and the actual circuit ground.

I'm not stupid you know. There's a reason I can't just put an ammeter in the circuit, now you understand?

>>1759710
Please stop linking people to these sites. The more people that are aware the faster stock depletes and the faster prices rise which hurts you and I. Since most people here will probably never complete the project they buy the tubes for it they end up wasted and we pay more for no reason.

>>1759751
At this point most of the coveted tubes are sold-out anyhow. You know, the nice big tall ones. All the tubes are getting poached by clock makers anyhow, who are bound to be better at googling than you or I assuming that a significant fraction are from russia/ukraine. Plus they have nearly the same selection as eBay, which isn't exactly an obscure site. And the prices are sufficiently high already that I can't see people buying them just to leave in a parts bin for a decade, they're the sort of thing you specifically buy for a project, so let them have them. Not to mention it increases the market for new stock to be manufactured.

I'll likely never buy nixie tubes unless I have enough disposable income and a sufficiently cool project which uses the Ω, °C, µ, k, etc. symbols, but also one that doesn't suffer from being large and delicate. Like the IN-1, 7, 7A, and 7B, or the IN-14, 19A, 19B, 19V. In particular the IN-19A has a bunch of SI prefixes while the 19B has a bunch of units that would be nice on a DMM and/or LCR meter.

More pressingly is my desire to make diy EL displays, as they're more energy efficient and can get some nice diffuse light out of them. More obscure, more fun.

business idea: make new nixie tubes for the basedman

>>1759751
>>1759809
bruh don't worry im poor and im more interested in EL displays anyway.

I'm a far way away from messing with those anyway, I'm still knee deep in Ben Eater tutorials. Just made my first register.

I don't know if this or csg is where I should ask : when an ebay seller lists a led flood light in 10w, 30w, 200w, 300w are they talking about total current draw or "equivalent light output to a incandescent bulb of this rating" ?

>>1759848
>current draw
I mean power draw

>>1759845
>Ben Eater
My man. His error correction videos (checksums and such) are a good watch, and his subsequent explanation of the cyclic redundancy check is mindblowing. Makes me want to do more math papers. Also good work actually using resistors on your LEDs.

>>1759848
>>1759849
It really depends. If it's meant to be a drop-in replacement for a normal light-bulb (i.e. edison screw or bayonet or possibly automotive) then it's a bit more likely to be an "incandescent equivalent" than otherwise, but even then they should have writing alluding to it being an equivalent. If it's something like a bare 12V LED COB then "50W" means 50W electrical power. I suspect this is the case for floodlights too, but I'd check the lumens just to be sure. Normal LEDs have something like 150lm/W (check wikipedia page for luminous efficacy), so compare that value. Watch out for obviously exaggerated luminous flux values too.
https://en.wikipedia.org/wiki/Luminous_efficacy

>>1759871
It wasn't without sacrifice that I learned I needed to do that. I've spent days fighting my shoddy work through trial and excruciating error looking for missed connections, weird feedback, voltage fluctuations, IC problems...

I had to order more LEDs and resistors to keep working after the number I've destroyed, and I'm finding that not every IC I have has been working properly, so I had to build smaller test circuits to make sure the IC function is intact and weed out the bad parts.

I had a very weird problem where my clock wasn't pulsing correctly and lights were coming on in places there should be no power, and it wasn't until I dismantled most of it did the problem vanish and I never figured out why.

But ever since I started being extremely thorough with my powers and grounds and ICs and LED resistors, all the weird spurious problems seemed to vanish entirely.

attached: a rubbish part I was really proud of until it fell apart

>>1759871
Looking at this
https://www.ebay.ca/itm/283688298148

150lm/W means 50W is 7500 lm which is close to 5x 100W incandescent bulbs!

Looking around the web, I'm seeing 84lm/W (https://i.pinimg.com/originals/c8/8d/c2/c88dc202761797bd84466e571333ad9a.jpg)) and 80 lm/W (https://www.ovoenergy.com/binaries/content/gallery/ovowebsitessuite/images/guides/how-many-lumens-do-you-need-120v.png)) but those are for a19 bulbs

>>1759880
That listing doesn't say how many lumens it is, but since the 50W one just has a single chip in there (like a common 50W LED COB) I can say with 90% certainty that it's 50W in real electrical power.

>>1759882
I think those might be old. 20W for a modern LED light-bulb, even a 1600lm one, is pretty high. Check out your local hardware store or supermarket's catalogue and see what they have for sale.

>>1759884
>the 50W one just has a single chip in there
Good catch. I hadn't thought of that.

I have a Luminus branded LED bulb here. It claims 14.5W = 1520 lumens (equiv 100W) I know from experience this bulb provides less light then the 100W incandescents I have. As an estimate, I'd give it equiv to 80-90W.

That said, I have no light meter to know for sure, just subjective observation.

>>1758900
if in parallel, charge in parallel, and always keep in parallel until death do them part

>>1759515
based amp. Future Electronics also seems pretty well stocked

yeah i just started learning electronics just for the hell of it. resistors, and diagrams and all that

>>1759921
>always keep in parallel until death do them part
Isn't it just fine to get oddly charged cells and temporarily connect them via resistors (100Ω?) to balance them?

>new helping hands come with soldering iron holder
>soldering station handle goes straight through it without stopping
great

>>1759929
looks good on the ole resume too, kek.

this mfer mgiver holy shit. starts throwing money at me, kek

>>1758759
So IPC-7351 states how footprints should be created, as in the 0deg rotation and what not. But hardly anyone follows this...

When you generate your pick & place file, in one of the columns you will notice component rotation, this is relative to how the footprint was created. But again no one follows the rules (IPC-7351) so most of the time this rotation information is almost irrelevant

So, how does the PCB assembly house handle rotations? They rely on the silkscreen/overlay layer, and if you don't show pin1/polarity markers then they will bug you for a PDF that does or ask for the design files

>>1759880
For what kind of light source to light source distances do these 7500 lm apply?
What I mean is that if the LEDs were 6 meters apart you probably wouldn't get the same brightness?
I don't know much about optical stuff

>>1759939
It's luminous flux, not luminous flux density. Look up photometry units.

>>1759939
>>1759940

what is a good substance to spray pcbs with as poor mans conformal coating?
Acrylic spray paint can?

I'm trying to do battery powered lighting under a pool table, but I don't know much about this kind of thing. Using an ESP8266 w/ a simple FastLED library design.

I've got two (2) 25800mAh 5v power banks. Would this be the way to hook them up to the length of addressable LEDs? (WS2812B 5v).

Thank you for your help.

>>1759996
I think your design will work, but I have a question. You are installing lighting under a very heavy object that presumably doesn't move. You are using rechargeable power banks. Why? You will have to charge these things, requiring plugging it into a charger connected to a wall outlet. Why not just power the project with a 5V power supply connected to the AC power to begin with?

>>1759930
>temporarily connect them via resistors (100Ω?) to balance them?
yes, you should do that before assembling any rechargeable pack. even 10Ω should be fine
>just fine
possible but not just fine. Li+ batteries are wear components. leaving aside the uneven charge/discharge load borne by each and the reduced pack total energy capacity, the pack's service lifetime will be that of its most worn cell. Li+ failure modes are unpredictable to put it mildly. you're much better repaid by spending time to improve the power efficiency of your design or procuring a properly sized battery

>>1759959
could work if your application doesn't flex much or at all, but consider RTV silicone potting too
of course clean your board thoroughly so you don't go putting polish on turds/fingerprints

>>1759996
Why not use a smaller MCU, or do you need wifi to control the LEDs? Also yeah that's a waste of power banks, at the very least I'd just use one.

>>1760118
>the pack's service lifetime will be that of its most worn cell
Still doesn't matter as much for parallel packs compared to series packs. Unless it shorts on you, then it really fucking matters. Would having one of those protection circuits on each cell help with that?

I'm gonna dump some pics from my work camera.

This is a really failed solder joint on a power resistor.

>>1760180
This is a DIY ISP interface on a Tesla centre console computer that melted during soldering.

>>1760183
This is the capacitor that shorts out in a 2004 Ford Focus, causing either the throttle position sensor or engine temperature sensor to show peaked values, I don't remember which.

>>1760188
This is the thermal paste application in a $5000+ audiophool amplifier.

It exploded because of this.

>>1760188 (You)
>>1760190
Whoops.

>>1760191
Customer tried to replace their own Ipad charge port, picture taken after cleaning up all the burnt solder and flux they left there.

>>1760194
I don't know if it's the 1359 or the SI9120 that makes this happen, but it happens a LOT when they get together.

>>1760190
>>1760191
Better have got a refund for that shit

>>1760196
Balanced line driver added to the stock stereo of a 2015 Nissan nav head unit.

>>1760202
The manufacturer supplied a replacement board for free, but the customer paid for the repair work.

It took weeks to get the board and it was assembled wrong. It was awful. Don't buy audiophile trash, you're paying for nothing.

>>1760203
Low-pass filter and mono mix injection on the rear channels of a Hertz 4-channel class D audio amp. Tapping inputs in the top image, injecting output in the bottom.

>>1760207
This is the computer of a 1999 CAT forestry machine. Note the awful bodges and conductive glue everywhere. I spent about 2 days scrubbing glue and repairing corrosion damage in it, but it did get repaired.

>>1760211
"How should we insulate the hard drive?"
"Masking tape."

Also note the definitely intentional and not afterthought tier power resistor bolted to the floppy drive.

>>1760212
And I'll leave you with a Gamecube RGB SCART adapter, made with a new old stock fancy ass SCART cable. It got rid of all the awful crosstalk in the original unshielded China cable.

>>1760194
lame. I wanted the before picture

>>1760178
thta much is true. individual protectors on each cell could protect against that one failure mode, ofc. but they introduce complications. e.g. it may be hard to reset one if only one protector trips due to uneven current sharing or something like that

>>1758973
>display that was on soyuz
>what the fug slavs, one small ic could probably do the job

The key word is "probably".I don't think USSR had a rad hardened space approved real time clock. They surely did have such IC's for generic alarm clocks and such but for the space?

They did overbuilt and overengineer that thing to get the job more than done. Lot's of gold plating and other rare metals I believe. There were no compromises. lol

>>1760227
but was it Y2k compliant?

>>1760227
How many cosmonauts died in the soviet space program? How many soviet space craft have exploded?

>>1759515
I buy a lot of my caps from JustRadios which is out of Canada. The rest I try to buy in bulk. It would be nice to get direct replacements for that amp. It improves the ease of install and looks professional should anyone ever open the amp.

To discharge, just test lead ground to the other side of a preamp plate resistor. 100k draws down 450v slowly and safely enough.

So I've come upon some LiPo batteries but I can't think of any small electronics uses for them since I'm not into RC stuff. What would /ohm/ use a bunch of lipo batteries ranging from 200 to 2200 mAh for?

>>1760315
i was thinking just to eat the customs/shipping charge and ordering directly from mesa

i honestly don't even know if it's the caps, the input transformer hums hard when i turn it on, with standby engaged and no input

the hum reaches the speaker with standby off and is increased by volume

>>1760319
Filters are a good look. Sometimes I'll scope the HT with a 10x probe too see if it's quiet enough. Even my cheapo pocket scope handles this which blew my mind. If filters were bad enough, you'd think they'd take a fuse with them. Nothing from the preamp should draw enough to hum the power transformer, but the hum is in there since you can throttle it with gain. Master volume would be in or near the power amp.

There are other suspects but you're probably on the right track.

>>1760316
>I have junk, what do?
sell them on ebay to someone with a clue

I have a laser printer and wanted to use it to make pcbs, but I also have a 3d printer and realized I could laser etch it. Im also about to build a cnc machine for milling so I could mill pcbs.
Which of these gives the best result with the least effort. I know I could order from china but they arent shipping now because of the virus and nows a good chance to learn.

>>1760327
best result can be gotten from photoplotting, least effort is milling. photoimaging will be slightly to much faster than photoplotting, with finer trace/space capabilities vs. milling
the nice thing about the photo processes is that you can get liquid photoimageable solder mask which you will truly appreciate for SMD jobs

>>1760327
Low-effort, fast, medium/high-quality method is rattle-can black + laser to make your photoresist. There's an anon here who was posting that a month or two ago.
For maximum quality, I think a laser-printer transparency is slightly better than the toner transfer method, but I'm unsure.
An anon or two mill their boards and get pretty good results: >>1758434.

Rate my setup, and then tell me how to make these helping hands hold up their own weight without putting 15000 tons of clamping force on the wing-nuts

>>1760342
>helping hands

Panavise or Panavise Junior

>>1759263
4 equations, 4 unknowns. Seems possible.

>>1759748
post circuit diagram

>>1760325
do you think a power conditioner would help?

>>1760342
6/10 needs more desk space and a Panavise

hey /ohm/. im quite retarded when it comes to electronics, but i got myself into a situation where i need to make a gadget that would detect the rotation direction of a shaft.
from mechanical side its as easy as pressing a wheel mounted on the gadget to the shaft to transfer its rotation. shaft speed should be ~1000rpm.
at first i was thinking about making a pseudo encoder with two hall effect sensors 90 degrees apart. the problem is, i really would like to avoid using any uC or batteries. i figured i could use a small motor (size 130) as generator and flash either red or green led depending on direction of rotation.
pic related is what i came up with. the battery in the center is supposed to represent the motor. i also added the two zener diodes to clamp the voltage below 4.7V, but i dont think it works the way i expect it to, ie if i raise supply voltage to 10V, the voltage is still 10V going to led branch.
do i even need the whole clamping branch? i plan on having a 3:1 gearing (the motor will spin ~3000rpm).

>>1760360
I'm in a 1.8mx3.5m bedroom, I'm doing the best I can. Also I'll look into that panavice thing.

>>1760361
If you go for the twin encoders, it would be fairly easy to use transistors or logic ICs to make a quadrature decoder. Be it with on-off hall-sensors, or analog hall-sensors or inductors with a comparator (or TL431 or similar), but an on-off sensor would be easiest.

A single up/down counter IC might work as your decoder.

>>1760364
>If you go for the twin encoders, it would be fairly easy to use transistors or logic ICs to make a quadrature decoder.
i dont really care about the position of the shaft or its rotation speed, only the direction of rotation. also, as ive mentioned previously, i would like to avoid using any batteries, thus my idea of using a small motor as a generator just to power one of the leds.
i think i managed to make the zener diode work as expected in pic related.
now the question is, are there any other problems im not seeing?

So this converts pwm into analog voltage, but if you connect into some high impedance input like an opamp then you can ever only icrease the analog voltage since the cap has no way to discharge, so how does that work?

>>1760374
Depends on whether the PWM output is a push-pull (totem pole) output, or just an open-collector. If it is a push-pull output, then it will push current into the capacitor from Vcc when high and pull current back out of it to ground when low. If it's a collector with pullup resistor you'll notice a difference in speed between rising edges and falling edges due to the impedance of that pullup, so you'd need to select your filter's impedance to be significantly greater than the pullup.
Naturally, you'll need to ensure the corner frequency is significantly below the PWM frequency. Having a multiple order filter here will give you a sharper cutoff and hence let you get higher frequencies showing up in your output for a given PWM frequency. If it's a power device, to avoid wasting power in the resistor, you'd use an LC low-pass filter instead of an RC filter, which is exactly what a class-D amplifier does.
Usually you'd want a push-pull output (or half-bridge?) as your PWM output, feeding into your filter, which itself leads to a buffering amplifier such that the potentially variable load impedance doesn't interfere with the frequency response of the filter. Class-D amplifiers naturally can't do this, so they use a feedback loop from after the filter back into the PWM circuit in order to correct for possible nonlinearities.

>>1760373
Yeah you wouldn't exactly be storing the positional data, just getting an almost-live reading of the direction. Say you used a 74LS194 bidirectional shift-register (http://pdf.datasheetcatalog.com/datasheets/150/375496_DS.pdf).). You'd put 5V at the shift-left input (D_SL) and 0V at the shift-right input (D_SR). Then the clock input (CP) would be sent to one of the on-off hall-effect sensors, and then the mode-selecting inputs (S0, S1) would be tied to the other sensor (though one of them needs to be inverted). Having S0=0V and S1=5V means on the next clock pulse it should shift left, and if they're inverted then it will shift right instead. Then you take one of the middle parallel outputs (or combine multiple outputs via diode logic) and use that as your direction output. When rotating one way, the rising edge will always happen when it's configured to shift left, and the continuous left shifts will fill the register with 5V outputs, when it spins the other way each rising edge will always happen when it's configured to shift right, and the continuous shifts will fill it with 0V outputs, from the D_SL and D_SR inputs respectively.
So with one 16-pin IC and a transistor-resistor inverter and maybe some diodes you've got a direction-sensor / quadrature decoder. Should take up less room than a motor and will have less losses. Could even be waterproof if your sensors could operate from the other side of some sort of seal.

>>1760378
Wait I'm retarded, a single D-flip-flop will work. Same principle, but faster reacting and with less bloat. A single 74LS74 would work just fine.

>>1760379
oh or you could go for one of those neat miniature SMD single chip buddies, they're kinda cool, but not that common outside of specialist bulk supply sites like digikey. picrel is a NC7SZ175

>>1760378
>>1760379
>>1760380
but i would still need batteries to supply the 5V to hall sensors and flipflop. i would like to avoid that if possible. also, a small motor like pic related (size 130) would definitely fit in the envelope i have and would take up way less space than a 9v battery and i dont really care about any losses, as long as it can power one of the leds.
do you think the circuit in >>1760373 work as intended? also im not quite sure which half would function better, the top or bottom. afaiu, the top one should have a brighter led while still clamping the voltage going to led to 4.7v, right?

>>1760380
>not that common
only all over aliexpress if you don't mind catching a little cold
1G79 and 1G80 would also be good choices if only one output is required

>>1760361
many of those diodes could be eliminated

>>1760393
very neat, thanks. i was wondering if those diodes are necessary, since led are basically diodes themselves.

>>1760384
I guess it's better to use a motor if you don't have a source of power nearby, but what's the damn shaft connected to that there's no power anywhere? What's powering the shaft? Is there room for a solar cell or something? Or even use the quadrature decoder coupled with a little rectified brushless generator on the shaft?

>>1760393
>all over aliexpress
Oh, they are? Neat. I should buy some T-FFs

>>1760398
>but what's the damn shaft connected to that there's no power anywhere?
its a handheld gadget that has to be easily transportable (ie i can just put in my pocket and go elsewhere). i definitely could use a 9v battery, but as ive said previously, i would love to avoid that.
>Or even use the quadrature decoder coupled with a little rectified brushless generator on the shaft?
whats the point in using a generator to power a quadrature decoder, if i can just use the generator to power a simple circuit like in >>1760393?

>>1760402
>a handheld gadget that has to be easily transportable
So what powers the spinning shaft if not batteries? Or is the device external to the spinning shaft and you're coupling the shaft to the magnets/motor with a wheel?
>whats the point in using a generator to power a quadrature decoder, if i can just use the generator
brushes wearing out at 1krpm, probably not a big issue if it's not on all the time.

>>1760397
they are but they don't like a lot of reverse voltage or current. conveniently, in the antiparallel connection, the reverse voltage they'll see is limited by the other LED's forward voltage. indeed, the two-pin red/green bi-color LED is just that configuration in a single package
aslo conveniently, the zener itself is also a diode with a normal forward characteristic

>>1760398
then you probably want the 1G80. TI's Little Logic handbook contains a competitor cross table which can give you other part numbers you can try searching on chinkbay

>>1760404
>is the device external to the spinning shaft and you're coupling the shaft to the magnets/motor with a wheel?
exactly. i have a device thats completely separate and i need to measure the direction of its rotation only once in a while.
>>1760405
cool, i didnt even think about it that way, thats why i added all those diodes in the first place. thanks again for simplifying my circuit.

>>1760278
>How many cosmonauts died in the soviet space program? How many soviet space craft have exploded?
No idea.

I read somewhere (wikipedia? not sure) that Russia switched to digital flight computers/controllers in like 2005 or so. Mind boggling. IIRC there were using mag amps or stuff like that.

>>1760421
vacuum tubes are much more resistant against EMPs and cosmic radiation, I imagine the same applies for discrete logic (or µm transistor sizes in general). IIRC the russians were using vacuum tubes in their aircraft at some point well after the americans had swapped to silicon because it was actually lighter when including all the shielding needed by the american jets.

I built something, will answer questions if anyone is interested https://www.youtube.com/watch?v=XunVbPj_gwE

>build strange op-amp circuit
>effectively open-loop feedback, V+ ≠ V-
>keeps clinging to one rail or the other
>output is unpredictable and useless
I don't know what I expected, but at least it gave me some inspiration.
So you know clipping circuits, which trim the upper and lower parts of a waveform, how about a circuit that does the opposite? Squeeze the middle of the waveform down to ground and only the peaks show. Picrelated. I think an 2-diode exponential amplifier with an attenuation stage would approximate this.

>>1760755

FFS, just throwing the lights in a random pile does nothing to illustrate the function of the circuit, and just makes it confusing to watch.

>>1760760

it's not like you invented anything new. every amplifier circuit will do that if you turn down the supply voltage.

>>1760760
if you are arsed about precision, you could use two half-wave rectifiers and sum the results
if not, just use an LM324 at a low voltage kek

>>1760424
>IIRC the russians were using vacuum tubes in their aircraft at some point well after the americans had swapped to silicon because it was actually lighter when including all the shielding needed by the american jets.

Plausible.

BTW check this out https://en.wikipedia.org/wiki/Yantar-4K2M

reconnaissance satellite that uses FILM. And thats in 2015 LMAO

>>1760768
>amplifier circuit will do that if you turn down the supply voltage
>>1760771
>an LM324 at a low voltage
Wait really? I figured it would do clipping, but I guess it depends on the output stage. So the effective open-loop gain isn't sufficient at low supply voltages to keep the two inputs approximately equal to one another when near 0-crossing. Probably due to not surpassing the collective minimum Vbe in the output stage, or a current-source getting stuck with too high an impedance to push.
What sort of voltages are we talking, ±2V?

Now I want a PSU with variable split-rails that can go anywhere from ±24V to ±0.5V. Since I'd like to drive speakers with it, it would probably need to handle up to 3A continuous.
I'm imagining a shitty circuit with diode drops slapped between the V+ rail to the output voltage in order to facilitate feedback to a buck converter and to the linear dropper at the same time, so the switcher voltage is always ~1V above what the linreg is set to. And then somehow reflected to give the same out of the -ve rail.

>>1760792
>better image quality
I bet that's going to change now that CCDs are getting better and better. Actually it still should have been at that point back in 2012, covering three of the missions, but perhaps they can be blamed on legacy hardware.

If I use an emitter follower as an output stage of a PSU, what's the collector-emitter ripple rejection like? What about a darlington emitter follower? How do these compare to a low-dropout linear regulator (operating with a low-dropout)? I'm only considering using transistors because of the low-dropout.

>>1760795
oh, you want a controllable effect. that's different. the LM324 is famous for its poor crossover distortion. any under-biased class B amp will do it if you don't mind fiddling around a bit

>>1760792
ever hear the old saying that the only secure computer is the one that is never plugged in or turned on? likewise, don't broadcast anything you don't want heard. while you're lmaoing, they're collecting and you don't know what they've got

>>1760821
enjoy this dude's thesis http://rincon-mora.gatech.edu/publicat/books/thesis/ldo_book.pdf

this uses two ordinary npn mosfets to switch AC, how the fucked does it work?

>>1760855
First of all, NPN refers to the structure within a BJT, which have N, P, N doped silicon within them in that order. A MOSFET has no such junctions, rather it relies on the field effect to create an ohmic region. Instead they're called N-type, NMOS, N-channel, etc.

The answer to your question is that the MOSFETs have body-diodes as can be seen in the diagram, right under the words Q1 and Q2. These body diodes conduct the other half of the waveform.

However, the gate-source voltage still needs to reach a certain level to turn the transistors on, and that source is in the middle.Consider 2 situations, one being the positive peak, and one being the negative peak, assuming the upper wire is live and the lower wire is neutral. In the negative peak, Q2 will be the switching element while Q1 acts as a diode, to get the gate-source voltage high enough the gate just needs to be brought above the negative peak, and to turn it off it needs to be brought below the negative peak. In the positive peak, Q1 will be the switching element, and so for the gate-source voltage to reach the threshold, it needs to be brought above that peak voltage relative to neutral, and to turn it off it needs to be brought below the peak. In other words, to have it off in the negative side and on in the positive side requires a fairly extreme bootstrapping circuit. If it's an optically charge-coupled device then this doesn't matter, but if it's tied to a circuit that's referenced to neutral you'd have a hard time switching it. Simulate the circuit if you want.
So I'd instead use one NMOS and one PMOS on the neutral rail, possibly in parallel with series didoes.

>>1760858
Wait so if you connect ground to the drain and + to the source on an n channel fet, it will always conduct no matter the gate voltage? Just like an ordinary diode?

>>1760863
holy shit it does, i just tried it with a multimeter and it shows 0.5V voltage drop between source and drain, this changes everything!

>put electrolytic capacitor into circuit
>it starts getting really hot
>even though it wasn't in reverse
>turns it it was only 16V and was across 24V
Oof, never had that problem before, I should be more careful.

>>1760864
Yep, but for some circuits it's unwanted. The diode is a natural consequence of the semiconductor geometry that forms the field-effect transistor. You can see it in the diagrams of how a MOSFET is constructed that include the source-body connection, which is the connection which makes a capacitor between the gate and source in order to activate the field-effect. Without it you'd have something closer to a JFET, which can never be turned as completely off as a MOSFET (in terms of off-impedance) and operates in the depletion mode. The JFET is constructed without a body connection at all, instead the gate is on both sides of the drain-source connection, so it acts a fair-bit differently. But this makes the JFET useful as a non-polarised electrically variable resistor, which is very useful for high-impedance analog inputs and odd circuits like AGCs.

>>1760821
Ok, preliminary testing done. Basically, there's no difference between a linreg's PSRR and an emitter follower's, at least at these frequencies and to what my scope can see. Pic related is input to the regulator on top of the output of the regulator, in AC-coupling. Note the different Y axis scales for input and output.
Getting the rippling power supply was a bit of a pain though, my signal gen can do DC offsets but it has a high output impedance, so I buffered it with an op-amp and used that as my rippling power supply. I was comparing the output of the buffer amp and the output of the regulator for the images, the input and output of the buffer were basically identical. In both cases I was using a 2kΩ resistor as a load.

I assume a darlington will be no worse than a single emitter follower, though I'd only use one where I need the current output to be too high for the base-driving circuit to handle otherwise.

>>1760837
>, they're collecting and you don't know what they've got
They got nothing bud
the film capsules got lost
That's it
End of story

>>1760875
I may try out MOSFET source-followers next, but considering I plan on adding an op-amp to include the emitter/source follower into a negative feedback loop, it probably isn't necessary. This is for a split-rail adjustable power supply, by the way.
But some testing in spice tells me that I'll be able to get right down close to 0V with this sort of power supply, and it's a pretty elegant analog circuit. Or it is if you ignore the fact that it should have adjustable current limiting but has nothing of the sort yet. No clue how much more complicated that will make it. I also have no clue how the voltage reference will work, at the moment my diagram just has the adjustment (digital) potentiometer between V+ and GND. I'll post the circuit later.

The un-elegant part is that it has adjustable two buck converters, one for each rail as part of the power-saving output, some sort of rail-splitting/mirroring switching converter that might also be a booster too, AND the AC-DC power supply the whole thing is plugged into in the first place.
Then again I've got a multitap transformer lying about somewhere that could work as the first two stages.

Is there any meaningful difference between a CPLD and an FPGA?

I have this seemingly simple logic table but I have been trying to wrap my head around a possible solution for a while now. It's almost a XOR gate but the third expression requires another gate that would probably need 3 inputs bnut I cant figure out.
Can anyone help me please?

>>1760863
yep, ain't depletion mode great?

>>1760918
FPGAs have a different form of routing more suited to parallel data handling and processing. FPGAs usually have internal RAM. CPLDs are easier to find in true-5V and 5V-tolerant I/O types. ultra-low-density FPGAs blur the lines a bit

>>1761002
invert IC, AND the result with COM

>>1761006
oh god I'm dumb
thanks

Just bought this for my "stock car stereo with bluetooth" project.
It will receive Bluetooth and output 3.5mm so it should work for my needs.
But now I need to find a way to add a display for the metadata and media controls.
I bought a separate arduino for that.
I've looked into bluetooth modules for this purpose and found some forum posts talking about the modules requiring AVRCP 1.3 to be able to display the meta data.
I tried searching for modules that support that but all I could find were a few bare chips I would need to solder to a board myself rather than an arduino plug in module.

Am I missing something?
Is AVRCP actually not required?
Am I biting off way more than I can chew? Probably.

>>1761012
You should learn how to do actual boolean algebra, it's not very hard and you can use it to simplify boolean expressions. Learning karnaugh maps would also be really useful, especially for larger truth tables.

>>1761121
I smell a ground loop approaching, you might have to use an isolated power supply to power it.

For your question, an ESP32 might be able to work, else an STM32 with the right peripherals is relatively likely to do this, plus I think some versions had a decent built-in DAC. I'm assuming you can control the details of a bluetooth handshake in code on either of those two.

I have a Walkman that lacks a tape type selector. I found this: (http://www.learningelectronics.net/circuits/playback-amplifier-for-cassette-deck.html)) about building circuit for cassette playback, and it seems the only difference between the two EQ settings is one resistor.
Pic related is the preamp section of my Walkman. The 5.5k and 20k resistors are inside the IC, so the values are only measurements. The resistor in red is supposed to be there according to the datasheet's block diagram (https://0x0.st/izmH.gif)) , but I can't measure a value, so I'm really not sure.

I get that the the whole thing is a low pass filter that does the playback EQ, but I don't understand why there is a voltage divider (the resistors inside the IC) and what the extra resistor (the 10k one that isn't in the circuit for the self built amplifier or in the datasheets example circuit) is doing.
I think if someone explained to me how this circuit works, I could figure out what resistors I need to solder to the board (between pins 3 and 4, and 21 and 22 respectively) to change the EQ to the desired 70µs.

>>1761173
It's all to make a particular frequency-response profile of the filter/preamp. Build it in spice and generate a Bode plot of it if you want. Also note that trying to measure resistance in-circuit isn't always accurate and may have distinctly skewed values.

>>1761166
What the best resource for learning how to do proper karnaugh tables?

>>1761187
I learnt through a university paper on digital logic, and we had a textbook called "digital logic fundamentals" or something similar. I never used the textbook though. I think it's a simple enough concept that the wikipedia article should be enough. Just put grey-code inputs on the X and Y axis, then transcribe your truth table. Find the SoP/PoS expressions right from that.

Also note that logisim is software that (among its many features) can turn a truth table into a boolean expression, and can even make karnaugh maps for you. So even if you want to learn how to do k-maps anyhow, logisim is a great tool for troubleshooting or working on larger projects or whatever. If you can get it to run on your computer, that is. I've had trouble with java versions or w/e a few years ago and haven't bothered reinstalling it since.

>>1761193
just keep your old JVM around, is all good

>>1761121
>his default search engine is aliexpress
AVRCP is a Bluetooth profile, not a board-side protocol

>>1761173
if you're talking about what I think you're talking about, that's not a voltage divider, it's two resistors to the virtual ground, probably to load the head. values might not be critical and might be mentioned elsewhere in the datasheet

>project has 4 instrumentation amplifiers and 4 standalone op-amps
this is getting bigger than I anticipated
suck my lm324

If I'm measuring the voltage across a current shunt, is there a reason to use an instrumentation amplifier instead of a simpler differential amplifier? I'm wondering if the low-impedance of the shunt will mean I won't have an issue since the differential amplifier's impedances will be much higher. But since there's a very likely situation where the current through the shunt is 0A and the output impedance of the source-follower would be really high, I can't say whether the differential amp would read properly.

How do I cheaply and easily light LEDs? I have a few things I want to light up in the vein of pic related or similar which don't need any sort of controls, just on/off. Since they're scattered around or small objects, I don't want to (or, rather, shouldn't) run USB or wall power to them since it seems like massive overkill for 2-5 LEDs. I've considered coin cells but heard they're not particularly long-lasting, and I'd like this to be more of a set-and-forget than something I have to consciously take care of. AA/AAA is probably possible but is a bit bulkier to add sometimes. Any suggestions before I do something particularly dumb?

>>1761291
>simpler differential amplifier?
It has been done before.

>>1761299

>>1761299
>2-5 LEDs
A capacitive dropper off mains is always an option, a few hundred nF should do you. Pic related, use the upper one for minimal costs (with equal numbers of LEDs in each antiparallel column), or the lower one if you can't stand the flicker of the above one. You'll want to stay below the maximum current of the LEDs, for which I don't think there's a proper equation so I just simulate mine in LTspice until my chosen capacitance gives me the correct current. But since almost all the voltage is being dropped across the capacitor, 270nF or thereabouts might cut it just fine, assuming the LEDs are 20mA each and that you're on 240V mains. X_C = j/(2πf*C), where j is the imaginary unit, f is the mains frequency, and X_C is the reactance of the capacitor.

>>1761322
I don't understand the circuit, isn't the 2nd half just an amplifier? Why not just do amplification with A1? And it still doesn't address what will happen in the event that the current is zero. I think I'll go for a proper instrumentation amplifier after all.

>>1761335
It was used for high side current sensing. If the voltage of the upper rail is higher than twice the max. common mode voltage (Vc) of A1, you can adapt the circuit with (R2/R4=R3/R5) >1 rather than 1 and use A2 to recover the lost gain.

Any reason why I should use pic related over...

>>1761488
this?
current is roughly 100mA

>>1761173
I had a closer look at the circuit from learningelectronics.net and came to the conclusion that what matters is the relationship between the capacitor and series resistor. I also did some maths that was probably wrong and came to the conclusion that shorting the 10k resistor should bring me close enough, and in fact the tape sounds fine when I do that.
Now I only need to buy a little switch and do some soldering.

I guess the next step would be to add Dolby NR decoding, but I don't think that's realistic. I can't seem find the needed IC anywhere.

>>1761220
>AVRCP is a Bluetooth profile, not a board-side protocol
Im new to all this.
Apparently not all mudules support avrcp
How can i tell if a bluetooth module will support it?
Where should i buy from?

Finally finished this mockup of a switch reader circuit. It consists of triple CD4017 counters, whose outputs are diode-OR'd through switch contacts into a single data wire. This way the microcontroller (an STM8 device) can read the position of a total of 24 switch contacts through a single data line.
The micro still needs a status wire to know when the counters have reached the count 0. Or a reset control line to ensure the counters are reset back to 0. Currently the STM8 just clocks the 4017s, reads the data line, and drives a bi-colour LED (green=switch closed, red=switch open).

>>1761504
Approximate schematic:

>>1761506
what are the diodes for?

Im looking for a 74hXX IC that have 8 (more is fine but its unlikely) AND gate in a package where one of every input of every gate is common. Essentially like a buffer with an output enable, except it doesn't go high impedance when disabled but outputs all lows.

>>1761522
diodes allow electron flow in only one direction

>>1759263
Open Vout
Apply power across the bridge
Measure current out of each side
Measure voltage drop across each resistor
R = V/I

>>1761299
>batteries
>set and forget
pick only one
classic USB is so easy to use though, you can get break out boards and just hook up your ground and VBUS wires and truly forget about it. it's not like USB-PD or anything

>>1761488
>>1761489
"this" can use a MOSFET with lower Vds(max). whether you would prefer to do that is up to you and the rest of your design

>>1761501
if you really don't want to bake your own board, you'll need to write code. you could start with an ESP32 bluetooth speaker project like https://github.com/MhageGH/esp32_BT_Speaker and add code to listen for ESP_AVRC_CT_METADATA_RSP_EVT events, then do what you like with the data you receive. read the developer documentation for the ESP32 IoT development framework for details
there are full-featured audio development boards for the ESP32 with headphone DAC/amp and onboard mics and all that shit, or you can save a few bucks by wiring the ESP32 of your choice to a headphone DAC/amp module of your choice
if they don't say they are AVRCP 1.3 capable, they probably aren't. single-chip standard devices that support AVRCP metadata don't really fit into the economics of the consumer electronics business. if a designer wants a value added feature like that, they can save a chip by using a bluetooth MCU and the manufacturer's existing devkit

>>1761571
that doesn't sound like a generic function. are you sure that an octal driver with pulldowns won't work? anyway good luck https://en.wikipedia.org/wiki/List_of_7400_series_integrated_circuits

>>1761504
doesn't seem like you'd be buying anything vs. a chain of HC595s

>>1761645
>octal driver with pulldowns

Didnt think of that lol. Thanks

>>1759385
Dope, what are you doing for the rest of the build?

Sorry if this is a stupid idea as it's been a long time since I've messed with electronics and I'm drunk right now.

I have a device I'm looking to modify in a way that increases power consumption, but I'm concerned about the battery life I can expect after modifying it. Given that I know how much current the modification draws at a specific voltage, wouldn't I be able to power the device with an appropriately sized resistor between the battery and the device to simulate the additional draw of the modification I wish to make to see if the battery life would still be satisfactory?

>>1761668

very very obviously, yes.

>>1761695
So if the modification I'm doing draws an additional 70 mA at 3 V, then I can just throw a 43 ohm resister between the battery and the device and get a mostly accurate representation of what I could expect battery life to be? Would I run into any issues due to the device stepping up the voltage from what the batteries provide?

I really need some advice.

What you see here is a garage safety sensor (receiver) that doesn't work. I can't seem to figure out why. I can power it with a 5V (from my arudino), but the LED doesn't even light up even though there is a voltage drop across the LED. However, other than that, I think the circuit is kinda working. the IR receiver voltage changes when I try to use my TV remote against it. I originally thought the caps might have died, so I replaced it a while ago, but still nothing.

>>1761771
Did you test the diodes?

>>1761812
The diode should be working. I do see a voltage drop across it.

>>1761506
What kind of software do you use to make such neat schematics?

>>1761746
Or you could just calculate the battery life from the total current and the capacity of the battery in amp-hours. Time = capacity / current.

>>1761827
It probably isn't what he's using, but KiCAD has a reasonable schematic creator. Best piece of FOSS I've ever used, hands down.

>>1761827
looks like KiCAD. maybe EAGLE

>>1761746
a switching converter is a constant-power load. it will draw more current as the available input voltage drops, in order to maintain a regulated output to its load. the resistor does not approximate this effect, quite to the contrary. you can treat it as a very rough approximation (factor of 2-3)
for best accuracy, if you don't have the actual device available, but do have a chinky dinky step-up converter module laying around, use that and put the resistor on its load side, such that the converter draws the known measured current at the known given voltage. but you're probably better off just calculating it against the capacity of the battery

>>1761771
>I can power it with muh arduino
they're part of a current loop. they're not meant to be "powered" like some chinkshit module. try a 24V supply with a 1k-2k resistor. also try reading before you just go jamming voltage into things

>>1758390
>and honestly I have no idea whether a water pipe will do.
as long as it's a cold water pipe you're good in the hood. unless... nah never happens.

I'm half finished, r8 me

>>1761327
>110mAh
I might be stupid but that doesn't sound like it would even last a few hours with 20mA LEDs.

>>1761329
Might give this a go, will see

>>1761645
I might resign to using USBs, then. Maybe I'll chain them to only use one cable; I had just hoped to avoid taking up any extra wall sockets or USB ports, since both are at a bit of a premium with my current setup.

>>1761869
If you are considering USBs, buy some 2-pin micro-USB sockets. They can be soldered onto basically any circuit and allow any old micro USB cable to power the thing. Plus they're a lot easier to solder than the 5-pin ones.

>>1761868
Also suggest what to do with these buttons on the rotary encoders I have

>>1761870
I know it's on its way out, but any reason why specifically micro- and not mini-USB? Mini always struck me as more robust, and a little bit easier to visually identify which way is up in low-light, so I was considering those.

>>1761839
>>1761851
Battery life would be satisfactory based on what I've calculated it to be, but I've been lead to believe that I can expect less so I wanted to be able to test it myself. I could probably attach the resistor after the power supply rather than at the batteries.

>>1761872
It's just more common, at least in my house and where I live. It was more prevalent as a cellphone charging standard, which might factor into that. It also has a more positive locking/detent mechanism, not that I've ever had any small connectors slide out on me, perhaps with the exception of well-worn 3.5mm audio jacks. But since cheap arduinos and my DVD drive and that one logic analyser I have came new with mini USB, perhaps there's not much to it. I'd agree that mini is a more robust connector. I haven't checked whether you can buy the mini-version of those handy 2-pin sockets however.

>>1761868
>using pulse-driven digipots
sad
>non-RRIO opamps for high-side current sensing
lel
>using a full inst amp for current sensing
>32 position digipots
much precise
>suggest what to do with these buttons
turbo boost

>>1761872
>why micro
because you probably don't even have to leave your chair to find a small, discreet junk charger to power them

Would this power up latch circuit work?

Once the button is pressed, the MCU waits a seconds and then brings the LATCH line up, thus keeping the system running. When it the time to power down comes, it just brings the LATCH down.

>>1761891
>sad
It will just be a bit slow, that's all. If I can figure out how to make pushing the button in quadruple each turn pulse then that would be great, but I've looked into it once before and it looks like a right pain. But now that I think about it, I'd only need the oscillator on the [Increment] pin, meaning I don't have to worry about shitting out higher frequency quadrature. Maybe if I can jack into the existing oscillator with only a few extra parts I might give it a shot. But even as is, it still gives me excellent precision for a mostly analog solution.
>lel
Oh yeah. I don't intend on the PSU output voltage getting too close to the upper input rail, it's mainly designed to give me nice low supplies (should be able to go down to 0V) but it's still somewhat of a concern.
>much precise
Actually they're 100-position digipots. With a ±15V range, that gives me 0.15V minimum resolution. I'll have a dinky current/voltage panel meter on here, btw. Probably an automotive one, because the voltage matches. I'll consider using different op-amps if I happen to desire more than ±16V (which I think is the LM324's limit).
>turbo boost
Like, an output current/voltage boost? Sounds a bit dodgy (i.e. dodgier than the circuit is already). If you meant a console-hack style turbo button, then yes I've considered that.

Any ideas on how to get LEDs that tell me when it's limiting current or limiting voltage, without putting diff-amps across the diodes?

>>1761897
1. be careful what you initialize before you have pulled the enable high
2. enable voltage detection/brown-out reset
3. check the leakage current through the reg

>>1761900
>turbo boost
I meant nothing by this, just shitposting. thought about posting a VTEC meme then nah
channel on/off, maybe
>how to get LEDs that tell me when it's limiting current or limiting voltage
Pic related

>>1761854
unless?

>>1761917
>just shitposting
What, should I make the buttons cut-spark so I send flames out the exhaust?
>Pic related
Oh yeah, I don't need the negative feedback of a diff-amp, a comparator will work just fine. But now I have an extra comparator...

If you hadn't noticed, what I posted was the top half of a split-rail supply, with the bottom-half being mostly just a mirror of the top. But it leaves me with a small amount of freedom as to how much autonomy each side has, e.g. whether the top and bottom will limit to the same voltage in an overcurrent situation, and whether I have a seperate meter for the negative side. I'm thinking no on both counts, but I'm not entirely sure.
>channel on/off
Well there's only one channel, and I'll just use a DPDT switch to turn everything off anyhow. No point in throwing a JK-flipflop in there, I'll have enough ICs. Or maybe a switch to pull the feedback loop to ground, if for whatever reason I want to keep the switchers running and the FET warming.
I think I'll do some testing on a console-style rotary encoder turbo button, both in sims and on a breadboard. I'm suddenly

You didn't happen to see any issue with my feedback loops, did you? Aside from possible polarity issues, I'll sort those out on a sim or partial prototype. Not sure about the comparator's output though, whether I should buffer it with BJTs in order to pull the switchers' FETs' gates properly to the rails and back.

>>1761921
Maybe he means that pic of the pipe next to the hot water cylinder with so much current flowing through it that it's glowing? The one people post in threads alongside the Russian LED.

>>1761928
>What, should I make the buttons cut-spark so I send flames out the exhaust?
this but not all that ironically
>I'm suddenly
I see that
>issue with my feedback loops, did you?
yes, I see no voltage feedback

>>1761934
>I see that
Shit, I was going to say I'm suddenly wanting to use tin-can op-amps (TO-99?) for everything. The voltage feedback should be there, between the left-most encoder+digipot, beeding to FBLin+, which feeds the op-amp near the NMOS FET's feedback loop just after the switcher.
I'm rather proud of my use of a diode drop in order to provide feedback to the switcher through FBSW+. Vosc+ is a triangle-wave, btw.

>>1761928
>whether I should buffer it with BJTs
you'll want to buffer it with something fast and strong. you'll burn a lot of power in turning off the FET if you don't
>nesting one feedback loop part-way inside another
what happens when you place a dead short on the output?

>>1761941
>you'll burn a lot of power in turning off the FET if you don't
Well I probably will anyhow, but I'm not sure what kind of frequency I'll be running at.
>dead short
The voltage across the current-shunt will skyrocket, causing FBI+ to hit the + rail, which will cause FBLin+ to be pulled below 0V, which will turn off the source-follower FET. When, because of this, FBSW+ decreases to only 0.6V or so, which is lower than SWoutbuffer+, the differential amplifier on the left will output a higher voltage, causing the comparator to output a lower duty-cycle, and hence turning the transistor on for less time. Or is it more time. I get confused with PFETs, but it's just a swap of FBSW+ and SWoutbuffer+ if it is wrong. Don't criticise my node naming.

Work for a small company.
The company is looking to assemble small PCB's in house.

What is the best option: to get a chinese T962A, or to buy a toaster oven and a reflow controller like X-Toaster?

Where can I get a bulk amount of Germanium Diodes, european shipping?

>>1762010
Ebay from russian dude selling D9E's

I would like to build an SDR transmitter using a DVB-t usb dongle.
the idea would be to extract the current frequency/clock from the dongle and use it in my AM transmitter.

the dongle uses rafael micro r820t chip and i can't find datashheet

>>1762064
You would be better off trying to use an Si5351 or RPiTX or FL2K than you would repurposing an RTL-SDR.

>>1762097
>>1762097
well shit, thats awsome i didn't know those existed.
thanks anon

>>1761917
Thanks anon, i tried it out, it worked.
t. >>1761897

>>1762064
>>1762133
Consider a standalone AM transmitter so it doesn't swamp out your SDR by being right next to it
Or an SDR with more bits and RX/TX modes (I think they're all half duplex though)

>>1762133
All good, what are you trying to do?

What's a good low power wireless transmission system? I want to be able to remotely trigger arbitrary IR remote code blast, and I want the IR remote device to be able to run for months on a coin cell battery (so higher power-draw protocols WiFi is not suitable).

I also want to be able to communicate to this device from the basement equipment rack, so BLE might be too short-ranged for this. Any recommendation for communication module or integrated microcontroller board?

>>1761984
the toaster oven itself is a huge variable

>>1761949
ok, that's fair. now what happens when you lift the short?
what I'm seeing in this circuit is way tf too much gain and too little compensation to be very stable. load it into a good SPICE and see if you can get it to ring
btw if you were really trying to make this period-accurate you'd be using more plain old BJTs

>>1762216
>months on a coin cell
there are very few or no radios that consume only a handful of µA in receive mode. instead of indulging childish managerial fantasy, maybe you should start reading real datasheets for real radios

>>1762216
Well it sounds perfect for LoRa. But RF connection to an IR remote feels a bit stupid, why not just solder the RF receiver directly in place of the IR receiver on the TV/other device? Be easier since you don't need to modulate the IR (since those receivers contain their own demodulators) plus you could just jack off the power rails inside the TV. It wouldn't be terribly destructive either, since the power rails should be going directly into 2 of the IR receiver's pins. I'd just cut the receiver out though, desoldering 3-pin components can get messy. Buy a couple replacements just in case you need to revert, or even add a switch on the RF board to swap back to an IR receiver.

>>1762225
>what happens when you lift the short
The voltage across the shunt drops to 0V, FBI+ drops to 0V, which is lower than the digipot's set current, meaning the voltage at FBLin+ is no longer being pulled down through the right-diode but rather the left-diode instead, returning to voltage-mode feedback.
>way tf too much gain
You mean in my direct control loops (the current-feedback op-amp after the right-most digipot, and the voltage-feedback op-amp next to the source-follower that takes FBLin+), or the two differential amps?
I decided that a gain of 100 was reasonable for the switcher's feedback, since the output of the diff-amp will need to ride all the way up and down the triangle-wave, which will go from 0V to +5V. Hence being at maximum duty-cycle would be indicative of FBSW+ and SWoutbiffer+ being 0.05V different. I could probably change this to 0.1V (gain of 50) but we'll see. At maximum current (5A) the source-follower will drop potentially upwards of 0.35V, so we're cutting it close to the ~0.5V drop granted by the diode to FBSW+. If I'm correct, too high of a gain for the variable duty-cycle diff-amp will cause the switching converter to act poorly and possibly oscillate, but I thought a gain of 100 should be fine.
The inst-amp for the current feedback also has a gain of 100, reaching nicely up to the 5V rail, so I don't see any issue with that.
The two direct control loops should both be within closed feedback loops; the high gain shouldn't be detrimental since there shouldn't be any delay in them, but I guess a simulation will tell.
>too little compensation
Not sure what you're referring to, what does compensation mean in this instance?
>period accurate
Not my intention at all, I just like analog feedback loops and the idealness of op-amps and comparators. Though I also have a thing for obscure packages, like CDIPs and TO-99s. I'd use military dual flatpacks if I could get any, it would make PCB designing kinda fun.

This anon again >>1761121

Installed the bluetooth module into the car today and it works! but as >>1761166 predicted, there is quite the ground loop goin on.

I've watched a few videos and they suggest a DC/DC converter
I searched around for a 12v > 12v isolated dc/dc converter and I have found quite a selection
some with 3 pins, some with 4, some with 5
and a whole range of wattage and amperage

Here is one I'm considering: https://www.mouser.ca/ProductDetail/MEAN-WELL/SPU01M-12?qs=sGAEpiMZZMvGsmoEFRKS8Koqt8Pjkl39hTBx1N%2F4CBIf9fy9utQErw%3D%3D

I know I need a 4 pin one so I can just pass the power and ground wires through it.
But how can I tell what amperage converter to get?
There isn't mention of Amps on the BT recievers listing (https://www.amazon.ca/gp/product/B07KXG3NG4/ref=ppx_yo_dt_b_asin_title_o00_s01?ie=UTF8&psc=1))


If it wasnt already obvious, I'm a little new to this stuff...

when was the last time you needed to solve a differential equation in your line of work, /ohm/

>>1762225
How about weeks on AA batteries then?

>>1762231
It's an electric heater. I don't feel comfortable messing with power electronics.

>>1762298
>If it wasnt already obvious, I'm a little new to this stuff...
You're doing pretty well for someone new to this. You're right that that sort of 4-terminal isolated DC/DC converter is the kind of thing to go for in this case. The other option is magnetic or capacitive isolation on the audio signal itself, but these might well degrade audio quality. What you're looking for isn't an amps rating on the BT receiver's listing, but rather a part number of the main chip(s). You then look for a datasheet of that, and search in there for a "quiescent current", "standby current", "maximum input current", etc. I imagine the output current won't be anything terribly high, though I don't know whether or not it will be significant compared to the current required for BT operation. I'd guess maybe 500mA at the most, but since it's only a receiver it could be below 10mA.
Unfortunately, I can't seem to read the part number on this image, and the listing doesn't say it either. Hunt around on forums or other places. If you already have one, read it yourself.

Ensure that your audio wires are properly shielded multi-core wires, not just some loosely bundled seperate wires.

>>1762301
>I don't feel comfortable messing with power electronics
If it's got a low-voltage side then there's nothing to worry about so long as you keep your electronics well away from the higher voltages. Is it a complicated enough machine that you can't just use a smart wall socket on it?

>>1761984
heyo, I was in exactly that same position!
I chose to go for a Controleo3 build. I bought a kit on ebay for ~$200 excluding the toaster. Built it in a couple of evenings.
I have absolutely no regrets. I get beautiful results and it catapulted me into the wonderful world of SMT design.
A guy I work with has used a T692 at a hackerspace and he's said that the Controleo3 toaster is much better in terms of even heating of the board and components.
If I were building it myself, I'd forego the kit and use salvaged parts to save cost. Since the company paid, I did what was quickest.
But either way, I do thoroughly recommend the Controleo3.
I don't think there's much better way to get not only bang for your buck, but bang for your space. I don't know about X-Toaster, but I think that toaster mods in general are the only way to get good reflow in something that won't take up too much space on your workbench.

>>1762313
internals if you're interested. The controleo build independently controls the two elements, and adds a third "boost" element.

Can linear regulators cause multi-megahertz noise if badly set up?

>>1762301
most of the 2.4GHz radios I'm seeing consume ~10-15mA with the receiver active. weeks on AAs is a bit more likely but you still have much to gain through creative power management, like turning the receiver on during defined time slots every few seconds. the nRF24 series is pretty easy to program and use

>>1762298
figure a micro USB port wants 2.5W
if you're just running line out to an existing head unit's aux in, you shouldn't need 12V on the BT receiver side. are you sure you want 12V-12V and not a 12V-5V converter?

>>1762313
noice, but
>that hobby servo to open the door
makes me cringe every time

>>1762318
absolutely

>>1762321
>are you sure you want 12V-12V and not a 12V-5V converter
Not him, but it has a dedicated power slot for <35V, not just the USB input. It might just go into a 5V linear regulator though, so looking at the circuit would tell for sure. But A: 12V-12V isolated converters are more common than 12V-5V converters, and B: having it go through a linear regulator would mean less power supply noise.

>>1762310
Here's a pic of mine. Ill try to find that datasheet. Thanks!

>>1762321
>are you sure you want 12V-12V and not a 12V-5V converter?
:^)
I didn't stop to think that the output doesnt need to be 12v since its only powering the BT reciever which is 5-35V.
That simplifies things!
Thanks for pointing that out!

>>1762334
>power supply noise
like noise in the signal? or noise from the converter itself
either way thats a good reason to get the 12V-12V


>Grampa used to run a TV repair shop by himself
>died before I went to college for electrical engineering
>tfw will never be able to ask him for help or talk tech with him
it hurts anons...
At least I inherited all his equipment like the magnifying ring light I took this pic with and an old ass oscilloscope.
Hopefully I can carry on his legacy. Maybe even open a business under the same name as his TV repair shop.

>>1762318
>>1762321
>absolutely
Wait, it doesn't actually use linear regulators, on the ±12V rails, only on the +5V rail. The ±12V rails use MJE181 and MJE171 transistors in some arrangement. The service manual doesn't come with a circuit diagram, but it has the next best thing: an assembly diagram! Turns out it was originally a kit meant to be assembled by the user! Which explains the subpar soldering. The original model was made in 1975, but mine has some slight revisions (an LED instead of a neon as the mains indicator, different fuse setup, etc.). So now I just need to trace it and find the likey source of the noise.

>>1762342
>AS19AP1Q757
I'm still coming up blank even with that part name, that's a pain. Maybe there are documents on the Drok website.
Good luck with electronics from here on, I never had any family into ham radio or electronics, but thankfully I've been able to take home some old university devices that they weren't using anymore. Sadly, their bin of Fluke DMMs weren't up for grabs.

>>1762347
I couldnt find anything in my searches either. The Drok site was shit; i couldnt even find the board I bought.
Is there any way to meter out the amperage?

>school tech equipment
Holy shit i would love to take home some of the shit they have at my college.
Each lab station has to at least be over $20k worth of equipment.

>>1762352
>any way to meter out the amperage
Yes, plug it in and put a small (1Ω?) resistor in series, measure the voltage across it. It shouldn't use more than 0.5A so the resistor will be safe. Even better, plug it into a USB tester (got mine for like $4 off aliexpress and it does V, A, mAh, W, etc.). Looking at the circuit, the DC connector looks like it goes directly (through a diode so you don't fry your USB) to the SE8533, a <40VDC to 3.3V linear regulator. So getting a 12V-12V converter isn't necessary and a 12V-5V converter will work fine too with likely no more noise and only half the wasted power. 5V one might be cheaper too.

Is the 1A DC-DC converter particularly more expensive than the lower power ones? Because I'd just get that to be safe.

>>1762347
>1975
if they're using the old emitter-output discrete regulators, under-capacity capacitors or a particularly ill-behaved load could cause that sort of thing

>>1762342
>>1762310
the AS19 is one of Jie Li's famous bluetooth chinkchips. still safe to assume you can run it on 5V esp if you're just using line out levels and not trying to drive an actual speaker directly
you could measure the current with any DMM with an amperes range, or you could just assume 3W, have 15-20% headroom over the typical maximum of a micro USB port, and not worry about it
make sure that the isolator you're using is good on an automotive power supply. the + supply in a car is notoriously hostile. look up "load dump" for example

>>1762356
>>1762357
so a 1 Amp, 3 Watt 12v-5v converter would be ideal?
>make sure that the isolator you're using is good on an automotive power supply
how do i check for this?
would the listing outright say "car power supply safe" or something?

>>1762359
here is an option
https://www.mouser.ca/ProductDetail/RECOM-Power/RKZ3-1205S?qs=sGAEpiMZZMvGsmoEFRKS8Koqt8Pjkl39mupWO7WlPWVQw%252BfybXiphg%3D%3D

>>1762357
>old emitter-output discrete regulators
Maybe, but there's an 7805 in the same circuit. No clue why they've chosen emitter-followers here, but they have. The noise is measured at no load. A cap on the output fixes it almost entirely, but I'd like to get to the base of the matter. Pic related is the circuit I found on some shill site that made me make an account, I don't understand where the noise could be coming from, I'm in the process of probing around. But the weird part is that the noise on the 5V rail is at 10.15MHz (at the moment) while the noise on the -12V rail is at 3.6MHz, and the +12V rail has 18.8MHz on it. And just now the noise cut out entirely. It's hard to say whether those values were all different as the frequency changes between measurements.

The circuit path of R18 looks like it's trying to somehow provide feedback to the -ve rail because the +ve rail is controlled with R6, but I haven't a clue how. Perhaps R17 and R18 make a sort of current balance, such that very little current flows into Q7's base and making it a rail-inverter of sorts.

Some probing tells me that I'm getting the noise on the zener diode ZD1, and that it's the same frequency as the +5V rail, oddly enough.

>>1762359
for best chance of success, look for something specifically marketed for automotive usage. for example https://www.aliexpress.com/item/32845197212.html

>>1762366
>why
because they're relatively easy to work with when high precision isn't required
checking all the elcaps and resistors for correct value would be appropriate, especially C3, C5, R2, R3, C4, R4, and since it seems to be happening on all supplies, check C1, C2, C6 too. a few cercaps in strategic places would be helpful. usually I just bung 1µF on the output and hope for the best
>but I haven't a clue how
it's a tracking regulator. R18 and R17 sum to about one diode drop below ground when the supply is at equilibrium

>>1762368
https://www.amazon.ca/Converter-Regulator-Waterproof-Converters-Smartphone/dp/B07H7X37T6/ref=asc_df_B07H7X37T6/?tag=googleshopc0c-20&linkCode=df0&hvadid=292997009026&hvpos=1o1&hvnetw=g&hvrand=16493333825573640673&hvpone=&hvptwo=&hvqmt=&hvdev=c&hvdvcmdl=&hvlocint=&hvlocphy=9001597&hvtargid=pla-600746528101&psc=1
oooooo this looks like it would work
and its usb which is easier to use compared to the wire pincher on the board

>>1762368
At one point putting a cercap on the output would increase the noise, but now it just decreases it from 2Vpk-pk to 1.5V.

>connect probe ground to itself
>scope reads mostly clean (<500µV)
>touch grounded probe to circuit ground
>8mV pk-pk noise at 86mV
>DMM reads 0VAC and continuity between scope ground and circuit ground
>noise persists with circuit turned off
what the fuuuuck

>>1762373
Knowing how to remove cancer from links is a valuable skill you should pick up.

https://www.amazon.ca/dp/B07H7X37T6

>>1762373
You sure it's isolated? If so it looks like a pretty good choice.

>>1762374
>it's 88MHz
Am I picking up a radio station over here?
>still senses signal when the grounded probe is touched to the scope's own ground lug
This isn't making any sense. I'm turning my scope off to see if it still happens when I turn it on again soon.

>>1762376
fuckin rip
the search continues

>>1762377
An audio isolation transformer might be cheaper

>>1762373
>>1762376
>>1762377
>>1762378
hey this fits the bill
https://www.amazon.ca/dp/B00N0UNJEQ
1.5A is alright right? I know most USB ports are around there
I guess if something outputs power via usb, its shouldnt blow up anything that accepts usb power correct?

>>1762378
this, and also subject to much less hostility from the rest of the car

>>1762380
hmm now how bad would it be to put a usb splitter or hub on that port and giving my car a charging port?
Is there anything to worry about when charging a phone in your car like this?
Like really bad car battery drain or overheating this little board?


>>1762378
>>1762382
The tutorial I've been partially following that recommended the DC/DC converter tried the audio isolator first and it didnt solve the ground loop.
And they also seem to be more expensive from my quick search.

>>1762380
A review of this one seems to be exactly what I'm doing.
I think I'm going to buy it.

is it just me, or would this work?

>>1762389
it'd work, as long as the turn-on pulse due to grounds finding their own respective levels doesn't blow anything out

>>1761984
>reflow controller
is that a joke? make your own from an arduino and an ssr in like 5 minutes you absolute peasant

>>1762401
add protection diodes to +12V and 0V then

>>1762384
car battery is 40ah, so you can charge your phone from it about 13 times

>>1762406
Even more if it's turned on!

>>1762380
>$17
haha what the fuck
just buy a new car instead
https://www.aliexpress.com/item/33052193101.html

>>1762410
>Estimated Delivery: 30-50 days
The amazon one will get here on thursday, so $15 for 10x faster shipping is alright by me

>>1762411
fucking burgers man
no wonder you are always so poor and your economy is in the shitter
>hey let me just pay 10 times the price of the item for shipping because i want it NOW!
Based on your car pics i can tell you definitely don't make a lot of money at your frycook job, so stop being so impatient and and wait a few weeks instead of wasting money on instant gratification
i bully because i care <3

>mfw chinky rotary encoder isn't bouncing at all
well that makes things easier
the pushbutton barely bounces either, though wouldn't matter even if it did

>>1762417
>can build turbo rotary encoder from one leg of the rotary encoder put through a high-pass + schmitt-trigger, the button with a pullup, and an oscillator
>it only takes 6 diodes, 4 resistors, and 1 transistor
thanks, that one anon who brought up diode logic years ago and prompted me to look it up on wikipedia, this will work just fine at 50kHz, so long as my 1N4148s actually arrive

>>1762384
>hmm now how bad would it be to put a usb splitter or hub on that port and giving my car a charging port?
actually pretty gud... but you may find it faster for a device to charge from its own converter that already does charger ID/power delivery notifications and has the USB connector on board

>>1761522
The circuit would not work without the diodes. CD4017 has outputs that can both source and sink current. If they were connected together without diodes, one of the outputs would always try (and fail) to force current into all the other outputs.

How do you go from ugly blocky pwm or inverter output to beautiful smooth sine wave?

>>1759651
>>1759651
>>1759651
>>1759651
>>1759651
>>1759651

>>1760373
can someone tell me what software is this please?

thanks