/diy/ - Do It Yourself

reposting it from /mcg/ because no one answered :

Is there an easy way to know if your VHDL code can run on a specific FPGA ? I'm pretty new to this stuff and i'd like to run everything on a small and inexpensive chip, but i'm not sure the 1200 logic cells of a XO2-1200 is enough...

>>2645470
Nobody ever answers in /mcg/, is that thread just full of bots or what? I also stick to /ohm/ even for my mc questions, /ohm/ always answers those well.

Anyway, as far as I know you have little choice but to place-and-route and hope for the best, although vendors often have eda software that let you synthesize and map the design.

>>2645487
how many people do you recon even answer questions in this thread on a regular basis? probably one or two?
i mean obviously the person who creates a new thread is obligated to answer all questions, but any other takers?

>>2645567
>i mean obviously the person who creates a new thread is obligated to answer all questions,

what kind of fucktarded rule is that. It's a general moron.

>>2645042
>screws on the outside are all fucked and impossible to get off
any suggestions that don't involve it getting destroyed?

>>2645487
>Nobody ever answers in /mcg/,

Bullshit. Most of the simple questions get answered, as well as many of the advanced questions. The issue for you is that not many people on this board are qualified to answer stuff like "Is there an easy way to know if your VHDL code can run on a specific FPGA", and when someone comes by who is capable then they will probably try to help.

So stop crying and be patient, or maybe go find the answer yourself.

>>2645573
Post pics.

>>2645573
wd40 or similar shit, leave act for some hours then try again, but post pics first

>>2645470
This is a vague question.
Obviously when you synthesize your vhdl into gates for a device it will tell you.
Once you get closer to that stage you’ll know.
Don’t put in a purchase order for 5 million xo2-1200’s before getting to this stage. Maybe just buy one (preferably a dev board) and try the end-to-end process first.

>>2645578
Only in /ohm/. There are no replies regardless of question on /mcg/. Simple, complex, qualitative, quantitative, shitpost, effortpost, doesn't matter. Hence why it almost seems like it's not populated by actual people.

>>2645573
>any suggestions

if they're recessed, then drill off the heads.
if not, then grind a slot with a dremel, then unscrew with flat screwdriver.

>>2645690
You're the worst type of newfag. Did you bring your paranoia from /pol/? Maybe that garbo board is actually dominated by bots.

Is this white power a sign of a failed cap? It’s the only one on the whole board that looks like this
I have a family vhs camcorder that I’m working on for a hobby. Basically when you put a tape in, the transport feeds the tape to prep for play/rec but the spindles do not turn at all. So after 5 seconds the whole unit shuts off. The read head still spins at full speed though, so I’m trying to determine why the tape does not actually move. Belt still looks okay and gripping

>>2645732
That looks like a fuse. There's a tiny stick of dynamite at the end of it so be careful.

>>2645732
Hold up, I rubbed the powder off after I took the picture and it’s a actually an t2.5a 125vdc onboard fuse, so I don’t think that will help me after all since the unit powers on and works
>>2645734
Yeah disregard

>>2645639
Why are you responding to something you have no clue about ?

That's not a vague question at all, modern FPGAs don't synthetize anything into gates. Everything works with LUTs and the compiler can and will optimize your processes so you effectively have no direct clue on how many logic cells you need just by naively looking at the code.
Also, it's obviously at a prototype stage for my own stuff which is why i don't want to order a devboard that might be completely useless to me if it turns out i can't run the software i already wrote.

I have found the answer anyway, i just need to look at the RTL detail level of a simulation and i'll see how many registers, lut and what not is used.

>>2645735
>I don’t think that will help me after all since the unit powers on and works

come on.
the fact that it's a fuse doesnt mean it's the only fuse.
the white powder is worrying.
especially as it looks like it leaked from inside.

>>2645763
>the white powder is worrying.
>especially as it looks like it leaked from inside.
Don't worry. The Sekret Cervix will have the investigation wrapped up by this time tomorrow, two weeks from now.

>>2645708
Have you tried taking your meds, schizo?

>>2645736
Of course we didn’t have a clue about what you’re doing, you didn’t tell anyone, nor post any details about your code.

LUTs are the same thing as gates, and, in fact, implemented in terms of gates.
Your response is akin to a silly statement like “FPGAs aren’t semiconductors, they use transistors”

>>2645780
He posted all the details required to answer his question. You clearly have no clue what a FPGA even is so why do you bother posting?

>>2645781
LBGTQFPGA+2FA?
What's a FPGA?

>>2645763
Its located right next to the AC power and the fact that the buttons motor and viewscreen work but not the record or microphone makes me curious. I could replace it but why would a fuse leak? It either works or it doesn’t.

>>2645732
Press play and probe for voltage at the motor contacts.

>>2645732
Try manually twisting the spindles with your fingers to break them loose.

>>2645800
The motor will spin to ff and rw, but that is when the transport is pulled back ie the tape is not around the heads. When I hit play I manually turn the motor from the backside but the tape is getting stuck somewhere

>>2645802
Spindles themselves spin freely. The last time this was fully working was 2019 and I had no issues with it then. Worse case I just sell it and rebuy the same model and hope it too doesn’t have the same problems

>>2645781
Why do you refer to yourself as “he” and didn’t bother to google what the G is for in FPGA?

>>2645819
Take your meds, schizo

>>2645812
https://www.repairfaq.org/REPAIR/F_vcrfaq2.html

>>2645825

>>2645830

>>2645780
what the fuck are you talking about ? i'm the guy who asked the question btw

if you want to make an 2 input AND gate your FPGA uses a LUT which is implemented as a 4-1 multiplexer with a register on one side and the output on the other. your inputs go into the address and you end up with your usual AND gate truth table. There's no actual gate, why would there ???

>>2645822
don't bother with him lol

>>2645470
You will have to run at least synthesis to know your cell utilization.
Lattice diamond should be free for those small devices. Yosys and nextpnr should have support for those parts as well if you want open-source tools. https://github.com/YosysHQ/oss-cad-suite-build/
This should have everything pre built and ready to use.

>>2645812
>>2645825
Apparently it wasn’t circuit related but for some reason a take up gear gets stuck halfway in the turn so I have to force it with my finger another 1/8 turn to get it fully in position. After that it works like a charm
I’m guessing it sticky grease from 35 years of being around, I put some white lithium grease there to help it out but no luck. Good news is that it works now and I can record to tape
Bad news is that I cannot remove the tape until it’s done recording or I take it out early, but it resets the gear position so I’d have to open the entire side panel again just to nudge the gear.

>>2645567
I mostly OP the thread now, this thread too. /mcg/ is sorta a containment thread to keep programming discussions from taking up space here. There’s a fair bit of overlap between the topics but mostly that defaults to being in /ohm/. It is a lot quieter than here, but it’s never in danger of not hitting its bump limit even without unsubstantial posts. I’m not some guru. I do my best but I’ve never used an FPGA in my life. I also use it as a way to shock arduino kiddies into giving up or learning how to actually program in C instead of copy-pasting function calls to pajeet code.

>>2645732
Did someone say white power?

>>2645870
>there’s no actual gate
There is what effectively works as a gate to the point the device is called a “field programmable gate array”, you’re all varying degrees of autistic. Senseless semantic argument solved, thank me never.

>>2645940
I agree with you about the grease. Did you clean off the old stuff or just added more? I'd encourage you.to clean it off and apply just a little dab then cycle it by hand 10 or 20 times to spread it evenly.

How do I repair this 50s tube radio?
Thing doesn't make a sound. Voltage on rectifier seems fine.

>>2646026
Chances are some caps are dead, though since it looks like it's already gone through a replacement run or two they could all be fine. Here's some other things to check:
>that all filaments are getting warm
>that no heavy components have caused solder joints to crack
>that no components have discoloured from heat
>that potentiometers aren't full of gunk
If none of those are your problem, chances are you'll need to use an oscilloscope of some kind to find your problem. Assuming it's a superhet you'll need to find the IF oscillator and check it's oscillating at the right frequency, then find the mixer and check that it's properly mixing, then check the amplifier and diode detector. Following the signal as it makes its way through the circuit.

>>2646032
Thanks anon. I'll checl that in the coming days.
I also have an analog scope and could check for frequencies. Do you perhaps know of an online guide or book that covers how to do this systemetically more in depth? Perhaps specifically about tube radios.
Will "how to fix anything electronic" be of any help? Its recommended in the OP.

>>2646042
>how to fix anything electronic
It's worth a shot, check it out on libgen. The basic principles of repair apply anywhere, but tube circuits and radios can be more particular. A full reverse engineering of problematic areas of circuitry will be required for a proper qualitative understanding of the problem, hopefully it doesn't come to that. Never reverse engineered a tube circuit or a point-to-point circuit before, I'm guessing you'd write down a netlist for those tag-strip nodes.

Okay I'm clearly getting filtered by soldering iron maintenance. My tip has got so oxidised that it can't even melt solder. Scrubbing it with a sponge or even steel wool doesn't do anything. I've dipped it in flux and that powdery tip cleaner shit and it hasn't done anything. Is my tip fucked? What can I do to stop this from happening again?

>>2646059
Doesn't look good. You can try more active (acid) fluxes to try and attack that black layer, or maybe a molten salt. Molten citric acid might do the job well if a liquid solution doesn't, and it only melts at 156°C (decomposes at 175°C). If you have good temperature control you could just jam the iron in some citric acid powder without worrying about decomposition.

I've heard you can file back the layer to bare copper, then electroplate nickel and/or iron atop it to get a nicely solderable coating back again. But I've only heard of such a process in passing, you'd have to research if it's effective or sufficiently easy.

Back on the topic of temperature control though, do you have any? That's the kind of oxidation I'd expect from a tip drifting up past 400C for hours on end, as opposed to a tip that can hold a nice and steady 320C. Or a tip exposed to tarry organic compounds, or a tip that's been scraped too vigorously, or a tip that gets left completely free of solder while hot, or just a really shitty tip. These are the points you can try to improve on.
Thermal shock from a wet sponge causing the plating to flake off is, I think, an unsubstantiated claim, I'd guess that a wet sponge would be better for your tip than vigorous scraping on copper or brass. Especially since those metal wools may likely be work-hardened. Not that I haven't got decent tip life (3 years and counting) from my chinky T12 tips scraping on chinky copper wool.

>>2645960
>There is what effectively works as a gate to the point the device is called a “field programmable gate array”
That logic works until you take into account the fact that a single LUT can effectively describe many many MANY gates. This isn't as simple as you think it is, and the only one arguing semantics is you. I asked how to translate from "VHDL" to "logic cell" and you answered "just count the gates". That's like asking "what's the output capacitance of the circuit" and being told "just count the capacitors". Absolutely retarded is what it is.
A LUT/mux unit can describe any kind of combinational logic unit, but it can also describe sequential unit, it's much more powerful as a concept than a single gate. the engineers that built the compilers obviously know this, which is why you can have a gigantic 1000+ gate system reduced to effectively 3 components.

There is nothing inherently bad about being wrong but for the love of god at least don't act like a dick while doing it.

>>2646069
>a single LUT can effectively describe many many MANY gates
I would suggest that "There's no actual gate" is still closer to false than true. Between the transistor layer of abstraction and the LUT layer of abstraction, there's going to be gates. That said, "LUTs are the same thing as gates" is definitely pretty false.
>you answered "just count the gates"
No I'm not that guy, I turned up late with broad generalisations after half-reading the replies.
>you can have a gigantic 1000+ gate system reduced to effectively 3 components
That's pretty cool.

How the fuck does the interconnect on a FPGA work? How can you connect N components to N other components?

>>2646072
>How can you connect N components to N other components?

one easy way is for everything to be interconnected during manufacture, then un-needed connections blown with high current.

>>2646074
>blown
FPGAs are typically reprogrammable.

>>2645377
are fiber optic isolated oscilloscope probes expensive as fuck because they use exotic components? or is it another case of low volume production and the goy tax.

>>2646084
>and the goy tax.

What's stopping you from undercutting them and getting rich you fucking /pol/tard.

>>2646072
It is arranged in a grid with switchboxes at every intersection to route signals.
LUTs also have some extra pickup and output switchboxes to connect to the interconnect.
That's roughly how it works, however every manufacturer might have a slightly different implementation.
Cologne chip documented their routing architecture in the datasheet:
https://colognechip.com/docs/ds1001-gatemate1-datasheet-latest.pdf

I'm not sure if this is the correct general but I I couldn't find one more related.

I got a laptop recently and I'm dissatisfied with the cooling performance, it has a ryzen 6800U and the entire PC seems to draw at least 45W under load (it has an oled panel as well which are quite hungry). I tore it down and found that it has a shitty little 1 x 12 x 210 mm heat pipe which probably explains why it throttles when playing something as old as FNV. I measured the dimensions of the inside of the laptop and figured I might try upgrading the heat pipe to a 2.5 x11 x 250 mm one but when looking through the data sheets the cooling capacity of the different sizes don't seem to make any sense to me. Eg. how/why does the 'cooling power' drop as the size of the heat pipe increases?
Have any of you tried something like this and if so did it work, what should I watch for/avoid? Are the rated values their cooling capability without active assistance like that of a fan?

>>2646086
it's okay to say "i don't know."

>>2646064
I'm using a cheapo iron that doesn't have temperature controls, and supposedly has a temperature of 420C.

Electroplating sounds kinda overkill for this, and I don't even know if my tip actually has copper underneath.

I managed to tin a tiny bit of the tip with solder but I had difficulty with getting it on. It would just form into a ball on the tip instead of spreading out and coating it. I flux I used didn't help so I guess I could get something stronger.

>>2645736
FPGAs don't synthesise anything.

Synthesis/compilers do.
And even FPGA toolchains synthesise into generic gates first, before mapping to cells and then implementation maps a netlist onto target hardware.

Any non-shit tool should be able to give area estimates either in microns squared or gates.

>>2645792
Female-Pretending Gay Asshumper

I am going to add an internal 50 ohm resistor to my scopes BNC input inside the case and add a switch to the front panel to make my very own 50 ohm input! One thing I am not sure about is the sockets and the circuitry are hidden inside some kind of a metal box that looks like an EMF shield. Am I going to disturb anything if I open it up?

Also, it seems like adding a resistor is a trivial task and a logical thing to do instead of messing with potentially expensive and ugly feedthrough terminators. Why bother if you can just add a simple resistor and a flip switch? But I couldn't find any discussion about it. Nobody had that idea in a community of DIY modders? Weird.

>>2646069
> single LUT can effectively describe many many MANY gates
So what? And yeah, before it’s programmed. After it’s programmed—which is exactly what we’re talking about here—it’s exactly a gate. You can mathematically prove their equivalence.
You’re confusing implementation and the concept of “gate”
How many gates are on an unetched ASIC wafer? is a nonsense question.
Here’s another way to look at it. Take the 74LS21. How do you know it doesn’t use a LUT internally? You don’t. How do you it doesn’t use multiple NAND or NOR gates internally? You don’t. But you can prove all conceivable gates can be implemented just NAND or NOR alone. Just like it could be implemented as a pre-programmed LUT-like structure.

>>2646253
>add an internal 50 ohm resistor to my scope

sounds pointless and dangerous unless you're specializing in transmission lines.
image scoping a power amp, and inadvertently adding a 50-ohm resistor across something.
not only will you fry the 50-ohm, but potentially some expensive power transistors as well.

>>2646273
A LUT represents many gates when programmed.

The actual implementation of a LUT itself consists of many many gates, so if it was only emulating a single gate it would be so inefficient as to make FPGAs totally worthless.
Majority of LUTs, when programmed, model 10s or 100s of gates.

A LUT is a specific type of gate.

Attention FPGA niggers and faggots:
Take your stupid argument to /mcg/ or better yet /lgbt/ you summerfaggot trannies.

brainlet here, i got this little magnetic pogo connector thing, when i solder to pins on the back the pins shift since the plastic housing is melting a little bit. is there a proper way to connect these pins or is it a get good situation?

>>2646288
> A LUT represents many gates when programmed
Absolutely false. If you have LUT4s (which is what OP has) a 5-input gate needs more than one LUT4 to represent it. This (opposite) situation is extremely common in FPGAs when dealing with busses for example.
>>2646290
This guy need to be a teacher

The useless semantic argument boils down to the fact that there are some situations in n input (where n>2) gates that don’t have traditional names as given to 2-input gates (not, and, or, xor, nand, nor, xnor, etc.) but they’re still gates and the IEC representation definitely wouldn’t have much problem with that.

Back in the day, we would do something similar to FPGAs, except with EPROMs (eg 2704) to replace some of the discrete logic gates with lookups where speed wasn’t that critical which is where all this shit originates from.

>>2646309
Tin the contacts and the wire separately before you make the joint, and use flux.

>>2646319
ended up mounting it to a little piece of stripboard. i am too spastic for this
thanks for the advice anyway i did get further than before.

>>2646315
Modern FPGAs are majority LUT5 or greater.
Modern synthesis tools predominately produce netlists utilising 2 input physical gates.

There is a reason that most gate counts are based on: total cell area / nand2x1 area.
NAND2X1 is the best approximation for average gate size across the overwhelming majority of chips. Because most gates are 2 or 3 inputs

>>2646309
>>2646348
Buy some low melt solder.

>>2646309
>plastic housing is melting

you can get these or make your own by squishing an alligator clip in a vice.
you place them as close as possible to the plastic.
another option is to dunk the whole thing in alcohol except for end of pin sticking up.

>>2646253
Do you know why 50 ohm terminations are used, when they are used, and how to diagnose whether they should be used?

>>2646353
> majority LUT5 or greater
Is there a device that has native LUT5s?
The reason native LUT6 exists is because a MUX4:1 can be implemented using a single native LUT6… considering the whole multiples of 2 thing in computing, like 8-bit busses and whatnot.

I have a laptop with vertical lines of pixels getting stuck as R/G/B on the LCD. it changes if I flex the LCD and I've taken it apart and determined that it's probably not a faulty connection but broken traces in the ribbon cable. The ribbon cable connects to the LCD near the top of the display and curves over a PCB at the top of the display. When I played around with it it seemed like flexing the cable where it curves would make the lines on the screen disappear or get worse.

I can't source a ribbon cable for a reasonable price at the moment. I'm thinking of putting aluminum foil on sensitive parts and then gently heating the ribbon cable in the area where I think the broken traces are.

Will a heat gun be able to reflow broken traces in a ribbon cable or is the heat not going to be enough?

>>2646433
https://www.amazon.com/s?k=foam+insulation+neoprene+sheets+with+adhesive
Use something to wedge between the ribbon and the case.

>>2646435
lol, I'd actually had strips of cardboard in there. They helped, but didn't solve it, and I thought maybe I could totally fix it with reflowing. I do have some neoprene at home already which is my fallback plan, but if I can fix it with a heat gun that'd be ideal.

>>2646441
You can buy T-tips that come with a silicone pad, but it's easy to ruin it even if you're careful. A heat gun is way overkill.

>>2646443
I'd rather have 3 lines across my screen than a broken cable. I'll try the foam and probably call it a day then.

>>2646443
forgot pic
>>2646444
Try the wedge trick.

>>2646309
Too late but I’d consider crimping a butt connector on its end.

>>2646433
I assume you mean a modern flex cable, not an old ribbon? The traces are copper, you’re not going to melt them without incinerating the polyimide plastic. Repairing flex cables is doable, but it’s really fiddly work, especially so when the traces are really close together. Basically you scrape off the insulation in the broken area and solder a bodge wire across the break. But in your case that would put more stress right on the bend, so you’d be better off backing off and soldering the wires further from the bend. Or even soldering them on the PCB itself.

You may be able to cut out a few millimetres of bad flex cable, strip the ends, overlap them a little, and solder the two sides together, if you have enough spare length. You’d then need to do a solid job of covering the joint up to prevent flexing of the joint, and add decent strain relief after the joint to prevent tight bends.

Honestly I’d look for junk laptops to steal the ribbon cable out of.

>>2646446
>I assume you mean a modern flex cable, not an old ribbon?
I suppose so, it's not like an IDE hard drive cable, it's the thin flexible transparent substrate type less than 1mm thick.

unfortunately, I don't have the cable length nor the expertise to repair it in such a way. it's not a big deal, it's just an old laptop that I wanted to get back in like new shape for sentimental reasons. it's a thinkpad T40. I could get a whole new panel on eBay but for the price it's not really worth it.

>>2646042
If you're okay with youtube videos, there's always Mr. Carlson's Lab, he has pretty extensive videos of repairing and restoring tube equipment
https://www.youtube.com/@MrCarlsonsLab

FPC vs FFC
https://focuslcds.com/journals/fpc-vs-ffc-vs-elastomer-strips/

>>2646447
If it's a FFC connector you can run magnet wire parallel to the whole length of the broken traces and bodge it at both ends, then cover it with Kapton tape.

>>2646447
Check the contacts on the flex cable and make sure they aren't de-laminating.

>>2646253
>>2646255
please do not mod you scope's inputs. You can use a simple BNC tee and a 50 ohm term like this. This also lets you switch out terms to 75 if that matters to you. If you want to make your own 50 ohm terms you should put two 100 ohm resistors in parallel
>>2646285
Some clock generation/distribution eval boards come connectorized and matched to 50 ohms. It's really for signal analyzers but if you want to measure rise times you have to use a scope

>>2646059
Try using plumbing flux or tip refreshener to tin the tip. If not, try using wire wool or brass scouring pad to clean it off (while it is cold) and then use plumbing flux or tip refreshener thingy.
In worst case, try using 2000 grit sandpaper, which is not a problem as tip is fucked anyway, and you can't fuck it up more in case you just sand off all coating (which would expose copper, which gets eaten away with high temperature, but tip is fucked anyway so whatever)
>What can I do to stop this from happening again?
Also, stop using water sponge, it leaves limescale on the iron which is just as hard to remove.
I run my tip at 450C because you know, gotta solder fast, and no issues what so ever. Typical china tip, nothing fancy.
Leave bit of solder on tip after you're done.
>>2646124
>I'm using a cheapo iron that doesn't have temperature controls, and supposedly has a temperature of 420C.
Is this a blue mains iron that uses 900M tips? Those do overheat way past 450C.
>Electroplating sounds kinda overkill for this, and I don't even know if my tip actually has copper underneath.
There is copper on all tips.

>>2646394
Yes, they help with reflections and ringing at high frequencies, typically when the cable length is > 1/10th of the wavelength and so it should be treated as a transmission line that needs to be terminated. Standard 10x probes are relatively high BW thanks to the lossy coax. But I specifically need it to test a "transmission line" probe which is just a 50 ohm coax and a resistor at the tip. So it needs to be terminated. Once terminated, it can provide a fairly flat frequency response up to 1GHz! The only disadvantage is low impedance.

>>2646499
>please do not mod you scope's inputs.
Yeah I took a closer look, it would be hard to solder a resistor and then wire some kind of a switch all the way to the panel. I could use a signal relay close to the frontend circuitry, but I am afraid I am going to introduce some parasitic inductance. The circuitry is very neat and carefully hidden inside the metal box to avoid any kind of interference. It would be bad if I mess it up.

>If you want to make your own 50 ohm terms
I looked into that but it seems hard mechanically. The resistor part is easy but I don't know how to connect the male and female parts together. I've seen some examples where a female BNC socket is screwed onto the male connector but the ones I have look very different and there is no threading on the female part. I probably need to shop around and find the right parts.

>>2646511
A threaded BNC is a TNC

>>2646400
Happens quite often in asic design that buses are not a power of 2, or even number of bits.
You need odd input LUTs as well as even input LUTs to best represent all possibilities. Its more efficient than using a higher input LUT and having one pin tied off (e.g. LUT6 with one input grounded, instead of a LUT5)

>>2646595
Also plenty of logic that you specify individually gets amalgamated or shared into a bigger LUT by synthesis algorithms.

What you think might give you some LUT3/4s in series could end up synthesising as LUT2/5s in series instead. Total num LUTs required is lower. Bigger LUTs are always more power efficient than multiple smaller LUTs.

>>2646400
>>2646595
> native LUT5 exist?
Not really. So Lattice, for example, can give you what I’ll call a LUT4½ but they call it a LUT5. It’s just that they hack one of the native LUT4s with an extra MUX, and the 5th input is the output of the other LUT4.

When counting LUTs, thats where they get weird LUT/gate equivalences, the device has a theoretical LUT4¼ since only half the LUTs can be hacked that way into limited LUT5s.

Another reason why this isn’t a “real” LUT5 is it incurs another layer of propagation delay. An ASIC CPU can do the entire instruction decode in 2 layers of gates and gate-like things without names.

>>2646059
always coat it in fresh solder before you leave

>>2646618
LUT4s are 2 LUT3s and a 2to1 mux
LUT3s are 2 LUT2s and a 2to1 mux

By your logic there are no "native" LUTs at all

An LUT is actually an 8051 core with a JSON parser running to tell it to map which inputs to which outputs. The bytestream is actually just JSON text
An FPGA is just a microscopic breadboard with many very small 8051 microcontrollers running in parallel

>>2646668
A LUT is a transexual transister that kills itself 41% of the time.
>>2646641
Native LUTs crossed the Bering Strait in 1492. It's why their penises are so tiny.

>>2646507
Yeah had the same thought. Might sand or file it since its essentially dead anyway.

>Is this a blue mains iron that uses 900M tips?
lel yes how did you know

>>2646627
When I was putting solder on it when I finished (before it got this bad), it just balled up. Was it already over at this point?

Why the fuck I can't pick some solder on my soldering iron? It just refuses to stick to it, even though I'm doing exactly what youtube tutorials do, I set the temperature like they say, I clean the tip, I use flux. What the fuck is this mystery?

>>2646674
I don't know. I could do it when I was a grade school kid, youtube and internet didn't even exist in my world.

>>2646675
I did it when I was a kid as well, I had one of those huge old soldering iron and it just worked like intended. But now it just does not work, how chinese managed to fuck it up?

>>2646677
>how chinese managed to fuck it up?
They're busy trying to prove that glorious China numba wan. In the meantime, buy Hakko or Weller. If you can afford Metcal, JBC, or Pace, even better.

>>2646674
>Why the fuck

90% of soldering is the tip: good tip = easy job.
the other 15% is the solder: must be brand name, have lead, and a flux core.

>>2646674
if the iron is hot enough to melt the solder and the solder just rolls of then the tip is probably tarnished

>>2646674
The tip is an iron core surrounded by copper and plated with nickel chrome. When the plating wears off from too much heat (and using the tip like a chisel), the solder won't stick anymore and the core dissolves.gns8rj

>>2646687
> the tip is probably tarnished
What does it mean and how to untarnish it back?

>>2646691
copper core surrounded by iron

>>2646693
>how to untarnish it back?
Sand it smooth and plate it with nickel chrome. lmao

>>2646693
>how to untarnish it

only sure-fire way is to order a reputable one from Amazon.

>>2646693
impossible, you'll need to replace it. how long have you used it? i've had cheap tips slef-destruct on day 2. you can look at guides to maintain them a quality tip is goes a long way but they all tarnish eventually

>>2646699
>reputable
>from Amazon

>>2646700
> how long have you used it? i've had cheap tips slef-destruct on day 2
Basically second time, yeah. Last time it kinda did the job but it was shitty. Today it just does not work anymore lel.

Okay, can you suggest what iron tips should I buy next?

hi guys! can you help me identify the type of this motor? it's from an electric weed whacker. i believe it's a universal motor but would appreciate help. line voltage goes to the stator, and rotor is brushed with each brush side having a coil in series with it. pic is the motor. i'm wondering is it possible to use it with a battery (18V nominal) if it is a universal motor...

>>2646706
i use weller which is crap but it's all they carry at the local hardware store. i'm not too meticulous about selecting tips cuz like i said, the tip industry is a giant racket. try to turn the power off when you're not using it, don't use flux with acid in it, don't clean it too often or use brass wool for cleaning and you'll get as much life out of it as possible

>>2646706
Any temperature controlled iron that uses T12 or T18 tips will be better than what you've got now.

>>2646708
>is it possible to use it with a battery

sure, but not an 18V one.
you'll need 5 or 6 in series if you wanna whack any weeds.

>>2646710
>i use weller which is crap but it's all they carry at the local hardware store.

there's your problem.
the red-handled wellers from hardware stores have useless tips.
the tips used by the blue-handled weller irons are top notch.
you cant substitute them, tho, unless you lathe/grind the shaft down.

>>2646722
I used a blue Weller that was in use for a total of like 25 years with zero maintenance. Got rid of it because touching the metal felt like touching electricity.

Rossman is wrong 99.9% of the time but he was right about soldering irons
If you can afford it get one with with heaters integrated into the tips
i.e. hakko 951

>>2646722
I've had ws51 for less than 7 years and only used it occasionally and the pencil is falling apart. The crappy plastic threading got stripped and the tip holder is popping out all the time. Great tips though. But overall build quality is crap. And a replacement pencil costs almost as much as the whole thing. Fuck that shit.

>>2646704
Not him, but it is shocking for me that Amazon in addition to stocking highly reputable brands also allows random chinese sellers to sell their crap just like they would on Ebay or Ali.

Has anybody used these PCI/PCIe translators and had them work? Not specifically electronics, I'm making a PCIe bracket from a blank, but here's hoping it'll come out ok.

>>2646735
That's bad, but it's worse for scamazon to repackage used items and sell them as new with a markup (but FREE SHIPPING with PRIME lmao). It's their shitty judeo-chinktian business model.

>>2646738
Amazon bakes the shipping cost into the item price

>>2646732
>Rossman is wrong 99.9% of the time
He was also right about microscopes. I love my Amscope, binocular version of his tri.

>>2646739
Of course. It's a monopoly that was effectively used to kill off independent, local businesses. Digital Walmart.

>>2646740
I like LX/Luxo but they're very similar. Mantis someday

>>2646742
Disgusting.
No more Circuit City.
No more BBB.
No more Fry's. I could actually walk in and spend 30 minutes digging through transformers and cement resistors.
But apparently there's Fry's Food. I wonder if that's the same brand.

Business idea: an opamp buffer as an active probe. What do you do if you need to probe a high impedance source with a low impedance probe? Right, you buffer it. But instead of buffering on the board, why not build a buffer into the probe? Why do they use some weird two-gate FETs in the standard active probes? There are plenty of cheap high BW low capacitance opamps.

>>2646671
molten solder is a liquid, so it's going to be affected by gravity. let it cool for a few seconds and rotate the tip as you apply.
if it's actually repelling like water and oil though, you bought trash and should get a refund.

I had an idea of making a chess board that lights up the square you're on depending what side you are. like black pieces light up the square red and white pieces light up the square blue. I originally thought about using like RFID readers under each square, but that would be 64 readers. what would you guys consider a better way of doing this? I also though of having the individual pieces light up on the board by having contacts on each square and an LED in each piece but I decided I want the actual squares to light up.

>>2646745
i'm not 100% clear what you're asking but there are definitely some diy active probes that use op-amps

>>2646764
Put a small magnet in the base of each piece and reed switches under each square.

>>2646778
would that allow the option of 2 different colors depending which side of the magnet is facing down? I am unfamiliar with reed switches

>>2646782
2 reed switches per square, each sensitised to opposite pole.
Then youll have a digital circuit to decode what colour should be. (None, blue, red)

>>2646787
damn that sounds so much easier than any of the ideas I had originally had.

>>2646782
>which side of the magnet is facing down?

reed switches dont care about polarity.

>>2646788
Your original idea would work too, it's just the long way around and more expensive.

>>2646788
Actually reed switches are not pole sensitive. Use a hall effect sensor for that.
1 per colour

>>2646789
>reed switches dont care about polarity.
>>2646795
>Actually reed switches are not pole sensitive.

is this fucking real?

>>2646818

if you google around, you'll see 1000's of variations of this circuit.
that makes it real, despite the hard-to-believe R1/R2 resistor values.

>>2646836
yeah ive seen them with mosfests, but this one is super simple. whats the catch? it says 10A thats insane.

>>2646818
Does it make mustard gas?

>>2646840
*mosfets and transformers
and btw can i shock a belligerent melanated with this?

>>2646840
>10A thats insane

garden-variety triacs can handle 6-12A easily, with heatsinks.

>>2646818
at least it's kinda isolated. Is this for driving a MOT?

>>2646641
No, I mean uncombined. So if you look at the data sheet for xilinx vertex 6 they mention it. Same reason an 8-input NAND gate is not actually implemented as a bunch of 2-input nand gates.
>>2646669
> transistor
That is closer to the truth, this guy gets it.

>>2646745
Don’t forget to put the op amp in a socket and provide a tube of replacements

>>2646868
Bro LUTs are just SRAMs and MUXes cascaded

>>2646942
Dopes anyone here even know what LUT stands for? kek

>>2646971
Maybe. You got a dollar?

>>2646985
I'll bite. Lookup Table. Basically the logic table defined by the programmer. Pretty much implemented in SRAM, it can't be using anything else.

>>2646988
>Pretty much implemented in SRAM, it can't be using anything else.
Why not? If it is a static const table it goes to FLASH. If you mean the speed is a concern then yes, you could make sure it is allocated in SRAM.

>>2647027
What is fanout of FLASH

Answer: not enough

If I short out a motor's windings and put a constant torque load on it, will measuring the current give me the correct current/torque ratio for using as a motor? I'd wind a string around the rotor's outside with a weight hanging off it, presumably if I pick the weight right it will reach a steady-state (constant speed) pretty quickly, at which point the forces are balanced and the measurement can be done. I'd put 1Ωor 10Ω or whatever resistors in between the windings just to make it easier to measure. It's a 3-phase outrunner BLDC, and I'll be using an oscilloscope to measure the hopefully sinusoidal peak current.

>>2647033
the hell is a "fanout"?

>>2647038
https://en.wikipedia.org/wiki/Fan-out

>>2647044
mindboggling
what does this have to do with a humble LUT being placed in FLASH vs RAM?

>>2647065
fan-in

>>2646699
Is this the real reason? I've been buying tips for two bucks off ebay and treating them as disposable. Should I just bite the bullet and get a quality $20 tip? Will it end my problems?

>>2646988
> can’t be anything else
Lol, even roms, proms, eproms are used all the time to replace gate logic. Usually where performance doesn’t matter. A FPGA is just something purpose-designed to implement gates, and can be reconfigured… most of them use sram now. Doing it old school: https://www.youtube.com/watch?v=BA12Z7gQ4P0

I used to work at a company that made large scale raid arrays, we used to use asics (very expensive) then we switched to fpgas (mid) for about 10 years when they became fast enough. Now we use off-the-shelf x86 as they are now fast enough to get the job done. It’s mainly high-bandwith xor operations.

>>2647073
>bite the bullet and get a quality $20 tip?

a single good tip is $5 - $12 (weller or hakko)
for $30 you can get a kit with maybe 10 diff sizes.
if you can fit it in your iron, which is far from certain, it'll work fine for 5 - 10 years.
obviously, you want to avoid aliexpress. amazon is iffy. digikey, newark, mouser, future, etc are reliable.

>>2647035
>bolt motor to door
>wrap string around with 100g weight on it
>let weight go and film motor
>oscilloscope shows 1.16V peak to peak
>through 3.9Ω resistors that's 0.105Arms per phase
>diameter is 71mm so torque is 0.0348Nm
>torque per amp per phase is hence 0.331 Nm/A
>torque per amp is 0.110 Nm/A

>oscilloscope also shows 20.53Hz on its roughly sinusoidal wave
>motor turns every ~20.5 frames at 30fps
>1.463 rotations per second
>ratio between electrical frequency and rotation frequency is hence 14
>makes sense since there's 14 permanent magnet pole pairs

Cool I did science today. Only thing concerning me is the motor is supposedly 170Kv, which works out to 0.056 V/(rad/s), which is only half the torque/current ratio. I wonder, would this have worked better if I'd measured the current on the output of a 6-diode rectifier instead of putting diodes across all three phases? Didn't really want to try that considering there's like no voltage to saturate a diode in the first place.

Also I'm realising that using this as a ground motor at 12S is almost certainly overkill. I need like 27.5Nm maximum, which equates to 250A total. Which is doable even at 6S I think, though my calculations suggest this current is directly inversely proportional to motor winding resistance even at high speeds, which doesn't sound right.

>>2646671
>lel yes how did you know
I know somebody with that exact shit.
You can't do anything really, as iron has no temp control and even at 20W it will overheat to god knows what temperature.
Because you know, Hakko 936 doesn't consume much when it is maintaining temperature, maybe 10W or so.
Proper iron is either thermally balanced (your typical hardware store iron) or has PTC heater (Ersa mains irons are like this I think, but not limited to ersa obv), or thermally controlled (936 soldering stations, T12, etc)

What size/type are these squashed SMD caps? Never seen a form factor like this. Why are they wide but short?

>>2647216
Maybe Johanson S49/1825

>>2647038
Flash memory cannot have every individual bit wired out to other logic as an FPGA needs. SRAM does, this is why they use SRAM

>>2647216
I think they do that for lower inductance. IIRC there was an EEVblog episode or something about them, but I can't find it. Maybe it was about wide power resistors instead.

>>2647299
Good point. I think the main reason for the sram is speed, and the sram is fabbed right on the fpga die. They could fab, say, nand flash that way if they wanted to, but the process might not be compatible with flash. Flash tends to be accessed more serially because of the incredible density and fpgas need a lot direct parallel acess. Some fpgas have 4-port sram elements.

>>2647324
Basic schematic for a LUT2 is 2bit flip flops and a 2 to 1 mux

Actual SRAMs might be used at scale for optimisation but a lot of FPGA reconfigurable portion is still flip flop + mux

Why oscilloscopes don't have built-in microwave ovens? It would make sense, since they already have built-in displays.

Retarded manufacturers spend their R&D time and money on this shit instead of putting an HDMI port and letting me connect my 27in overhead display that will cost 3 times less than anything the R&D team can come up with on their scopes. Or even better, remove the graphical part all together, which will make the scopes like twice cheaper and easier to manufacture, put a fucking data port like Ethernet on the device and let the community make open-source tools to work with the data.

Doing a diy mouse, can't seem to find the optical sensor that's on the bottom of mice. What are they called?

>>2647351
They stopped selling those for whatever reason. Earlier you could buy them off Mouser for example.

>>2647350
High frequency probes & scopes need good quality high frequency display capabilities.
Also making a need for an external monitor just makes scope more difficult to use, not less difficult.

>>2647361
Yeah, just build it the microwave oven inside so it's easier to operate during the day. Fucking retard.

>>2646710
Are you supposed to flux the tip? The solder that came with "my first soldiering kit" doesn't have rosin core.

>>2647361
>high frequency display capabilities
HDMI 2.0 (std. 2013) can output 240FPS at 1080p without even chroma-subsampling. If 720p @120FPS is sufficient, then HDMI has managed that since 2005. Displayport can handily manage >500FPS with no compression.
These are off-the-shelf consumer-grade standards.

huh

>>2647361
>scopes need good quality high frequency display capabilities.

Why? The human eye can't see any more than 24fps, anyway.

>>2647364
>Are you supposed to flux the tip?

absolutely not; that'll just burn it all away.
you apply flux to the parts you wanna join.
when it's heated, it turns acidic and burns off oxidation present on the surfaces you're joining.
the flux+oxidation then flies off and goes into your lungs where it remains safely stored.

>>2647370
>The human eye can't see any more than 24fps

who still uses human eyes in 2023?

>>2647351
its just a monochrome cmos capable of detecting IR

>>2647427
>capable of detecting IR

yep, unless somebody stuck an IR-blocking filter in front of it.

>>2647323
>>2647241
>1825
! That's it.
>Low Inductance Chip Capacitor (LICC®) sometimes referred to as Reverse Geometry Capacitor (RGC) has its terminations on the longer side of its rectangular shape.

Thanks, anon.

New challenge. What are these cylindric resistor symbols? Coax cable? Why 25 ohm then. And I am not sure about the setup, two runs of coax on the output side?

>>2647460

>>2647460
yeah, they're coax.
>laser driver
I don't see any laser diodes

>>2647497
It is a test circuit. The two 25ohm outputs are for the diodes. And the 50ohm ones are connected to the scope. That explains two runs of coax cables. Here's what the real circuit looks like.

>>2647460
>>2647507

the 25ohms you're talking about are explained through transmissions lines.

Look up "Characteristic Impedance", basically it's what the impedance of an infinite cable would be OR the impedance of the load needed for the signal to not "bounce" back.

if you used a 50ohm cable you would get echoes.

>>2647519
)))echoes(((

>>2647519
Yeah but I don't understand where the characteristic impedance of 25 ohm is coming from. Why not 50 ohm.

>>2647582
Ah sounds like the laser diode's input resistance is 25 ohm so the 25 ohm coax is needed to drive it.

Bought a bunch of micro controllers, switches, LEDs, and other shit. This is my first time trying to build an electronic
Am I going to need shit like resistors and diodes to make it function right or will I be fine just soldering it right to the board? The switches all say they are made for Arduino, but the board is not arduino

>>2647582
Basically if you look up SWR you'll see that a matching impedance is needed to maximize power delivery through a transmission line, among other things. I don't know the specs of your laser diode but it seems to me that the input impedance of your load is 25Ohm (at least, it is at the intended working frequency.)
This also means that you won't get signal reflections (same reason why you need a matching impedance for the oscilloscope)

>tfw failed electromagnetism course twice and was only saved thanks to the transmission line class

>>2647610
>trying to build an electronic
what.

Resistors are more or less needed for every application imaginable. So if you're making your own shit, yes you do need them. They're not expensive at all though. I'd suggest buying at least 10 resistors of each decade of E6/E12, from a few ohms to 1 or 2 MOhms.

I don't understand what you mean you talk about soldering to the board, do you mean soldering directly to the MCU board ?? you can easily get protoboards adapters for MCU boards (usually in arduino format), you'll be able to solder on that directly then "plug" your MCU and get a working prototype immediately.

How tf am I supposed to calculate an inductor for Bright Power buck converter? I've found a nice buck converter IC for LEDs (BP2863) which has everything built in and I just have to add couple capacitors, current shunt resistor and coil
Well, first of all, datasheets I find are all in Chinese... But ok, no big deal, right, as formulas are same for their other IC (BP2836, which doesn't have flyback diode and rectifier). I still don't quite understand what tf I need.
Obviously I need to respect on/off times at all conditions in 220V outlet, but still, how do I pick a frequency of operation for this shit?

But yeah, buck regulator is waaay cheaper than capacitive dropper, and smaller too, so I can layout PCB on one side, and use other for heatsinking, lowering LED temperature down.

>>2647633
An electronic device*
It's a vertical mouse with 20 thumb buttons and 3 clickers and some joysticks
>Resistors are more or less needed for every application imaginable
Any resistor? How do I know which to use
>do you mean soldering directly to the MCU board
Yes.

>>2647625
Yeah thanks anon, I know what TL is and how to match it, I was confused by the 25 ohm characteristic impedance which is an odd number to me since I am more used to 50 ohm termination. I don't know anything about laser diodes but I looked it up and yes, this particular type has a 25 ohm input impedance. Perhaps all of them do, no idea.

>>2647610
Stop. Put down that microcontroller, you're clearly have no idea what you're doing.

Now. Yes, you will need resistors. They are everywhere. Come in all sorts of flavours, but I suggest getting a kit from China that has 600 or so 1/4W or 1/2W through hole resistors. They are of shit quality, but you're not flying to space either. You might also want to get a breadboard and battery holder with wires, and DT830 multimeter.
After you get them, figure out what the fuck is Ohm's law, and Kirchhoff laws. Make a voltage divider. Then figure out Thevenin equivalent. Then figure out behavior of diodes, and try to figure out why do you need a resistor for LEDs. After all this, get a pack of bipolar BJT transistors (they are cheap too) , and figure out how do they work too.
Once you've figured that all out, build a multivibrator circuit. Got blinking LEDs? Understand why do they blink? Good. Now you can pick your arduino whatever and make LED blink using computer.

You can figure that shit out in, idk, maybe a week of dicking around.

How do you guys check if analog ASICs off Aliexpress work before you install them? I have this chip from Texas Instruments that keeps failing in projectors. Do you measure ground relative to VCC? I would imagine semiconductor parts failing as a short circuit. Are there other tests for this type of chip?

>>2647669
Qfp socket in a known-good test projector.

>>2647663
> They are of shit quality, but you're not flying to space either.
The’re OK if you’re making controls for extreme-depth submarines though.

i need to measure airflow within a certain room and I am trying to find the weak spots (in various conditions). Would it be ridiculous to use microphones to measure the air flow? or should i go more for a more normal approach?

I am building a DIY standing desk using some linear actuators and would like some input.

The desk itself is pretty much done (I'm using it right now), but I haven't done anything with the actuators yet. I have 4 of these PA-14P actuators that have potentiometers that I need to control and keep in sync: https://cdn.shopify.com/s/files/1/0061/7735/7891/files/Linear_Actuator_with_Potentiometer_PA-14P.pdf

My idea is to control each pair of actuators with a L298N H-Bridge, and the potentiometers to a separate ADS1015 12-bit ADC. I would be able to control the H-Bridge with simple PWM, and use the differential mode on the ADC to get accurate readings of each actuator.

Finally, for safety I'll use a 4-channel SSR between the H-Bridge and actuators. I'll hook this up directly to a switch to ensure that it won't get power or move unless I'm actively holding down a button. Everything else will be controlled by a spare RP2040 microcontroller that I have.

Does this sound like a solid plan? This is pretty much my first project using strong mechanical parts.

>>2647669
>DMD1000
I have a projector from 2005 with that chip in it.

>>2647661
Alright well sorry for bursting your bubble but if you have no clue about how electronics work in general your project will at best take a really long time.
The ATmega 328p you bought can hardly run USB HID devices without major optimization, maybe if you can import V-USB to them you'll be able to but you'll get USB 1 at best.

your switches aren't made "for arduino" they're just regular switches. you'll need pull up and pull down resistors. same for your LEDs, you want to limit the current through them or you'll fry them. Your sensor module is just that, a sensor module. do you know how the data is sent to the MCU ? what protocols are used ? You need to multiplex your inputs because you only have so much pins on your board but you have around (at least) 23 pins for the buttons + at least 2 pins if you're somehow able to use a muxed ADC that talks in I2C and you're lucky enough to get a sensor module that speaks I2C.

long story short this is a pain to deal with even if you know what you're doing, and you don't even know how to get a correct resistor value through ohm's law.

>>2647683
one of the flaw in your design that i can see directly is that it implies that each potentiometer is equal and linear. you'll need to normalize them otherwise if you're aiming for a single value each potentiometer will be slightly off. this can be done easily enough with the microcontroller, but you'll still have some error.

>>2647699
Yea, I can definitely confirm each actuator has different min/max values, but it would be quite easy to save those values in eMMC and just linearly interpolate between them, right? Is that how normal desk-controllers work?

can't listen to w2aew for too long. he is one of the worst lip smackers on youtube. makes loud mouth sounds ever few seconds.

>>2647663
>>2647694
Oh I see.

>>2647717
git gud summerfag
frfr no cap ong

>>2647655
You want to go for as high a frequency as you reasonably can, since that means a smaller inductor and hence a cheaper BOM. But I wouldn't go too close to the maximum frequency specified in the datasheet, go a tad lower. In this case it tells you the minimum off time of 3µs, so I'd aim for 4-6µs. 10-20µs might be more efficient in terms of switching losses if the cost for the larger inductor isn't prohibitive, but it might be less efficient in terms of ohmic losses.

>>2647661
1kΩ for LEDs, 10kΩ for pullups at the bare minimum. 100kΩ isn't a bad idea also. 100 and 1M are used less often but handy to have. Also get some 1N4148 diodes for your key matrix for proper N-key rollover. Some bypass caps isn't a terrible idea (100nF ceramic).

Also as the other anon says, the ATmega328P can't do USB HID. Get an ATmega32U4 board like an old Teensy or Arduino Pro Micro (or other USB capable AVR like the 16U2), else get an ESP32S2/S3, RP2040 board like the Pi Pico, or an STM32 board like a bluepill/blackpill (beware fakes). If you go for an AVR, I'd also get yourself a USBasp so you don't have to fuck about with bootloaders.

>>2647678
Not ridiculous at all, if you characterise the signal strength as a function of wind speed AND direction first. An easier method might be measuring the speed of evaporation of a liquid by its temperature, the simplest form would just be a bunch of thermistors coated in water if you can ensure that the amount of water on each one starts off equal. Using thermistors does require that temperature is uniform though.

>>2647683
Why not just gate the power input to the L298N, instead of the outputs?

>>2647726
>You want to go for as high a frequency as you reasonably can, since that means a smaller inductor and hence a cheaper BOM. But I wouldn't go too close to the maximum frequency specified in the datasheet, go a tad lower. In this case it tells you the minimum off time of 3µs, so I'd aim for 4-6µs. 10-20µs might be more efficient in terms of switching losses if the cost for the larger inductor isn't prohibitive, but it might be less efficient in terms of ohmic losses.
So I guess I'd have to calculate stuff with lowest socket voltage and highest socket voltage to make sure it is within range and call it a day, right?

>>2647744
Yeah.

Can I use a cfl backlight driver from a 50" screen to drive a bunch of CRT flybacks simultaneously?

>>2647763
How would a CCFL backlight driver drive CRT flyback transformers? The former is designed for moderately high voltage DC output, while the latter needs to be switched at relatively high frequencies to produce high voltage. Or do you mean to replace the flyback transformer? If so, you'd have to compare the voltages of the two but I'm pretty sure CRTs want a significantly higher voltage than CCFLs.

>>2647775
https://www.youtube.com/watch?v=11YHeqdaHco

>>2647460
>This device is optimized to drive a differential TOSA with a 25Ω flex circuit. The unique design of the output stage
enables DC-coupling to unmatched TOSAs with laser diode impedances ranging from 5Ω to 10Ω.
>TOSA: transmitter optical sub assembly
They did it that way on purpose to ka3 your life easier when actually using it but it makes life crappy when taking measurements

>>2647460
>This device is optimized to drive a differential TOSA with a 25Ω flex circuit. The unique design of the output stage enables DC-coupling to unmatched TOSAs with laser diode impedances ranging from 5Ω to 10Ω.
>TOSA: transmitter optical sub assembly
They did it that way on purpose to make your life easier when actually using it but it makes life crappy when taking measurements

why are fans so expensive, like why are some small pc fans 10$+ when i can also get a box fan for around 20? i am genuinely confused by this

Is this to say the lockout needs to be configured for low voltages when there -is- a bootstrap diode, or when there -isn't- one?

>>2647447
piss easy to take out. done it on numerous cameras. also bare cmos doesnt have anything on it.

>>2647427
These days they're a lot more than that. Not only are they a CMOS sensor, but a high-frame-rate low-pixel-count CMOS sensor with image processing hardware and USB HID built into one package. While you could pass this processing off to another IC or two, it would make it a lot more expensive.

>>2647455
>LICC
mmm

>>2647892
I assume there's an internal bootstrap diode coming from an internal 3-5V LDO. A diode, internal or external, is basically a requirement for operating a high-side N-channel MOSFET. If you don't have such a diode, then you won't be able to operate with a Vin - Vout dropout of less than the gate threshold voltage (3V in this case). Even with a bootstrap diode, if the high-side FET remains on for too long without going low, the capacitor will drain below 3V due to the IC's quiescent current and it will turn the FET off. But that condition wouldn't happen during low/no load conditions, rather the high-side FET will only be on for a prolonged period of time if there's a really high load or if the output voltage setpoint is set to equal (or be greater than) the input voltage. I'd look for an internal diagram in the datasheet to explain it better.

Anyhow, assuming it's backwards logic is somehow correct, if the thing is going under low-load conditions it shouldn't matter whether there is or isn't an external diode. If there isn't an internal diode and you haven't added an external one, then that voltage lockout should be configured regardless of duty-cycle.

>>2647892
>I assume there's an internal bootstrap diode coming from an internal 3-5V LDO
The one under the pMOS, yes? I think I get it; The external diode keeps the nMOS' driving opamp Vs pulled high enough to switch the nMOS in the event that the internal regulators don't.

>>2647907
Meant to reply to >>2647905

I want to replicate this DIY solar e-reader cover:
https://www.mobileread.com/forums/showthread.php?t=273119

But apparently the guy wired the MicroUSB cable directly to the panels, which is a no-no (despite most solar-to-USB panels on Aliexpress being wired the same way).

So I'd probably have to add a solar controller/regulator - but is that even enough for safety? This other guy added an over-discharge protector, a short circuit protector, a boost converter, an MPPT, a flux capacitor and fuck knows what else:
https://www.youtube.com/watch?v=kEttqWJrdww

Can enough of whatever necessary components be flattened enough to fit into a cover flap? Their overall point would be to not screw up whatever device I'm charging, while not being grossly inefficient.

>>2647686
This chip wears out over time and needs to be replaced. You can only buy it from Aliexpress (not TI) therefore every second chip is faulty. I wonder if there’s an easy way to tell.

>>2647907
The "5V source + MOSFET + diode + op-amp" circuit is presumably a low dropout 5V regulator, though note that the polarity is backwards since it's a P-channel transistor that turns on with a low voltage. Likely a drawing error, maybe there's a revision. Using an external diode bypasses this regulator with the positive rail's voltage, but since the bootstrap circuit can't handle more than ~5V it should only be done on sufficiently low voltage rails. There isn't really a situation where the internal diode + regulator circuit becomes unable to work, only if the current drain on the bootstrap circuit is particularly high (e.g. high duty-cycles) might you want an external diode for thermal reasons. Also schottky diodes may have a voltage drop improvement over the internal one.

>>2647911
All you really need is a voltage regulator. MPPT is nice to have, I'd go for it if you're really pushing for high power output. All that battery management shit is only useful if you need an external battery that is being charged by the solar panel, but in your case you'll just plug it directly into the Kobo.

As for flatness, using a thin (0.5-1mm) PCB with thin parts on it should be able to make your circuit compact enough. I'd cram it in the gap between the PV panels for protection if nothing else. You may be able to find an existing switching regulator that has a small enough footprint, but I'd probably make one. You also have the option of using an MCU instead of a dedicated IC for a switching controller, and by adding extra components to measure solar voltage and current, you'll be able to write an MPPT algorithm if you need to. Just ensure whatever circuit you choose doesn't boot-loop and get into thermal issues with low light exposure.

>>2647919
Thanks for the reply. Designing circuitboards is sadly currently way above my pay grade, but would wiring something like this between the cables be enough?
https://www.aliexpress.com/item/1005005223784703.html
Though I'm not sure why the guy used a pair of panels or whether it was necessary, I was going to order a single larger one instead.

>>2647905
yes it is a more specific type of sensor. and i would still say its a sensor + controller combo more often than not. as it would make sense that brands would want their own firmware on the device.

>>2647922
That sort of might should work fine, but remember it can only increase voltage, not decrease it. So ensure that the maximum voltage of the panel in full sunlight is no more than 5V. The panel size he used is a pretty common one, likely cheaper to get two of those than going for a larger and less common panel. It can also only handle a certain minimum input voltage before it stops boosting, so check that you're not losing much power when that's the case.

That design in particular doesn't say a maximum output current, you may need to check the datasheet for that. Looks like it might be too small, considering you'll likely be aiming for ~5W of solar and that size of converter is probably rated for less than 1A. And chinky boards never have brilliant thermal performance either, best to assume they're not getting the maximum power output stated on the datasheet. I'd go for a 1.5A or 2A "capable" chinky converter if you're aiming for 1A. Though thermally bonding a 1A converter to an aluminium sheet that acts as backing of your case is a viable option. Bonus points if it's on an aluminium-backed PCB or otherwise has no solder joints or parts on one side.

You can design a PCB without much knowledge by following appnotes and reference designs, but I'd probably recommend against it unless you're up for a steep learning curve and probably an additional revision or two. Finding an existing circuit on EasyEDA's community collections might be ok though.

>>2647923
>i would still say its a sensor + controller combo more often than not
Last time I checked that was the case, for example look into the Pixart PMW3360, which the Ploopy mouse uses.

>>2647928
>ensure that the maximum voltage of the panel in full sunlight is no more than 5V
I'd test the panels anyway, but I presume a panel labeled as "5V 1.9W" shouldn't output more than 5V?

I found a less shitty listing with more data on the output (though google/deepl were powerless to decipher the non-english pages) - the 5V setting is 1A, as far as I can tell. Meanwhile, my Kindle's max is 500mA. That's 2.5W max? I am still clueless on this topic, but that seems ok, even if chinkshit is less efficient than advertised. If it's not ok, I will prepare my anus for PCB designing.

Although... I found this step-down converter which might work better (or just differently):
https://www.aliexpress.com/item/1005005084434791.html
Should I go with the step-up one or this step-down one for this? As long as the panel doesn't go above 5V, I presume it's the former.

Thanks.

>>2647939
Or what about using this thing? I'm not sure if ChatGPT is bullshitting me as usual.

>>2647939
>I presume a panel labeled as "5V 1.9W" shouldn't output more than 5V?
Well maybe, but what we call 12V panels are actually panels that output more than 12V (~18V at peak power, ~22V open-circuit) because that makes it easier to charge a 12V lead-acid, so I couldn't say for sure. Look for the "open-circuit voltage" specification, or better yet a voltage / current curve (pic related).
>the 5V setting is 1A, as far as I can tell. Meanwhile, my Kindle's max is 500mA
If it can do 1A and your e-reader only wants 0.5A, then no problem. Though if your solar panel can only handle 1.9W and your kindle wants 2.5W, you may find that the kindle pulls the solar panel well out of its most efficient range. See how in pic related, if the load tries to draw more than ~3.1A, there's basically no voltage and hence P=V*I=0. This is what MPPT circuits solve, basically the result of using one will be the output voltage of the charge controller is such that the kindle draws a lower current.

>>2647942
>>2647947
No that's a step-down linear regulator (less efficient than a switching regulator) that outputs a constant current, then a constant voltage of 4.2V in order to charge a lithium ion cell. Equivalent to part of the battery circuitry in that youtube video, but not something you need if you're just sending 5V into the bottom of the e-reader.
Pretty sure none of those have male micro USBs on them, they're all female mini-B, micro-B, or C. They're designed to step 5V down to a voltage for charging a lithium ion cell, not to step a voltage up to 5V.

Also the protection chips (DW01) on some of those boards are ass, they only cut out at like 2.4V, which is well into damaging a lithium ion cell.

>>2647942
No idea about this, but AI is all about bulshitting. I once asked bard some nonsensical question and it gave me a detailed answer in a very convincing and even patronizing tone, quoting part of my bullshit question. It's talking to a mentally ill person.

sup boys, half retard here. I need to run a 24Vdc/20A PSU to a bunch of LED strips. This is fine to stick at the end of a 110Vac/13A rated extension cable yes?

>>2647982
Yes. The PSU is outputting a maximum of 480W (500W).

>>2647982
Is it a AC-DC PSU?

>>2647988
thanks

>>2647989
yes

I'm about to start designing the PCB for my senior design project. Practical implementation and PCB design is one of the things I'm worst at. I know some of the beginner pitfalls and I can always google for more.

Instead, I would like to ask you guys what the more subtle errors that can crop up are. Things that aren't a quick google away; the little things you only pick up from years of PCB design, that can really elevate a project.
(note: the highest frequency on the board will be roughly 600kHz, so nothing too bad)

>>2648026
Add footprints and traces for alternate configurations/parts to make your board more versatile, and also for times when you can't find specific ICs or MCUs.
Add test points for easy probing, and headers for debugging.
Be aware of loop inductance, stray capacitance, power planes and signal planes.
Keep the board thermally balanced and shielded if necessary.
Make sure your documentation is complete and accurate.

>>2647726
> Why not just gate the power input to the L298N, instead of the outputs?
The L298N has a secondary function of turning my 12v into 5v, which I can use to power the ADC and microcontroller.

>>2648029
single-ended vs differential pair transmission lines
trace width, routing angles, propagation delay
via stitching, ground plane, star ground
EMC testing

Where can I order PCBs? What quality can I expect from them?

JLCPCB wont ship to my shithole... Well, they will but for $30 which is a bit too much. You can order via aliexpress with aliexpress shipping, but dunno.
PCBWAY they do ship to my shithole. Overall quality is decent, but I've noticed that silkscreen sometimes has skid marks or misalignment.
What about elecrow? Can't get any reviews on jewtube, but their factory in ads looked kinda not that professional.
local services are moronic, as they don't even have approximate pricing and im not feeling like calling them. Plus, they have only green soldermask.

>>2648026
See: >>2642616

>>2648032
Just use an external 7805. It will handle more power than the dinky SOT-223 5V regulator on the breakout board.

>>2648261
But the L298N module uses a 7805 as the voltage regulator though. Where did you see it using an SOT-223?

>>2648026
Use all your free copper and free solder/tinning.
The copper traces should look like you just outlined them with a thin PCB line. Generally.

>>2648121
If you’re picky about the quality and the brands of clothes the factory workers are using, consider DIY. There is this one autist, Mr. Carlson, who is also very picky and has some nice vids on making PCBs just the way he likes them.

I just use bare archer perfboard without any pads myself.

Who is Electronics General here?

>>2648264
My breakout board has a SOT-223-packaged 7805 on it. Maybe yours has a TO-220-packaged 7805 instead, idk.

>>2648296
>single sided without solder mask
Not that anon, but getting a design to work single-sided is tough. Getting a design to work single-sided without running traces beneath passives is near impossible. Solder mask greatly helps prevent shorts between an SMD passive and the trace beneath it. While you can just hand-paint it, it can be time consuming enough that automation is worth considering.

>>2648312
I'm closer to Electronics Staff Sergeant.

If it still ain't broke fix it again

>>2648296
>If you’re picky about the quality and the brands of clothes the factory workers are using, consider DIY.
It is not about the quality of clothes, but overall organization of process.
PCBWAY and JLCPCB looked really clean and well organized.
Electrow looks a bit more ghetto... Like idk, soldermask silkscreen machine is covered in soldermask paint entirely, and everything is kinda dirty lol. But then again, it is factory equipment, not medical equipment, nor IKEA showroom, it will get filthy eventually.

But then again, even shittiest pajeet factory would beat DIY pain in the ass when toner doesn't transfer for shit, or you over-etch. Not to mention the luxury of double-sided boards that is really hard to achieve at home. Plus price of FR4 is comparable to ordering entire PCB so why bother. Yeah, I know phenolic shit is cheaper, but you know, i want my tracks to not lift when i solder.

What's the difference between this type of biasing and just a resistor from V+ to ground? Why also apply voltage?

>>2648394
To adjust the DC offset.

>>2648400
Hmm. This circuit is to correct for Voff, Ibias and such. R4 would introduce a DC offset since there will be some voltage drop on it due to Ibias. Why is that not sufficient?

>>2648321
> design to work single-sided is tough
Please consider using jumpers

>>2648400
Here is what I mean

>>2646693
Means you need to become Elden Lord.

You know of any "latching transistors IC"?
I believe I want about a dozen of them in series for a challange of enable each one in the correct order. Feels generic enough to exist but I'm not fluent with the /ohm/ terminology to have found it myself.

>>2648439
...latching transistors engaging the inputs of an AND gate IC seems correct. I'll lookup some keypad code lock kit meanwhile.

>>2648439
There is DB3 diac that is sorta of a latching transistor, but not what you want

>>2648439
Why not just use an SR latch IC? Or make NAND/NOR latches out of logic gates? Comparators can also make decent latches.

>>2648473
>SR latch
Yes, thanks that seems like the thing! Will try to simulate in the weekend, counting on having to "sum them up" with an AND IC as well but perhaps not.

>>2648493
>simulate
I recommend Logisim evolution for this.

>>2648321
>Solder mask greatly helps prevent shorts between an SMD passive and the trace beneath it. While you can just hand-paint it, it can be time consuming enough that automation is worth considering.
It is not a big problem. Most SMD resistors have conductive film only on the top side of the package, bottom is just bare ceramic.
It is a bigger issue when you surface mount a through-hole resistors, as those usually have thin glaze layer on top of conducting film and that one can withstand 250V typically.

But DIY PCB is a pain because success rate is 50%, since nothing you're using is meant for what you're doing. Toner isn't meant to be transferred, and if you're doing photoresist method, toner isn't meant to block light, etc.
>>2648406
I wonder, why didn't they invent SMD earlier? Through-hole seems so wasteful and weird to automate.

>>2648504
> why didn't they invent SMD earlier
Before PCBs everything was point-to-point on solder lugs, so all components were that way. Thru hole just used available components in a mass producable way.
And axial capacitors are getting rarer as PCB technology progresses, although audiophiles can still get their yellow polyester caps because they really believe they hear a difference.

>>2648494
> simulate
Get a few Virtex-7 2000T devboards and design your latches in verilog and you can try them out for real in an actual hardware implementation.

Simulation is very unsatisfying.

>>2648504
> Toner isn't meant to be transferred, toner isn't meant to block light
I’m lucky enough to have an old lexmark laser which has world-class toner deposition. It’s great for doing both. So, it definitely varies… not sure if they make a printer for this purpose or not, but they should. Also might be able to hack a laser to enthicken the toner layer.
Mine’s the next best thing to a cricut thickness.

Anyone looked into building thermal camera? 3..4 hundred is quite a bit to pay.

I built my own soldering microscope for a total of perhaps 100 bucks with HQ picam. Very nice result compared to the 800+ i wouldve paid for the same-ish quality. 30 fps, accessible via browser, mounted on a gas piston articulating monitor arm.

Now im wondering if i can do the same with thermal camera?

>>2648504
>success rate is 50%
>toner
>photoresist
Laser ablation of etch-resist is basically 99+% success for copper, with the right software (Lightburn + custom inkscape script). I'm currently in the process of fine-tuning the lasering of the mask, though I've put it on hold until my rigidifying upgrades to my laser CNC machine (a modified ender 3) are done. Being able to get more than 2000mm/min would be nice. If you have a half decent XY laser then you can probably hit 10000mm/min without trouble, 5W (optical power) diodes are good enough, 0.5W isn't. If you go for an Erbium or Neodymium laser with a galvo, you can go far faster and ablate resist far more precisely. Lightburn for a galvo is more expensive though, though the machine cost will typically dwarf the software cost.

If you have good photoresist then look into using a resin 3D printer with UVTools. I found the LCD contrast wasn't good enough to prevent the developer taking away some of the exposed resist, but it was a cheap 2nd hand Photon Mono with a backlight running at 1/3 brightness. A setup using two LCDs, or using a DLP, would almost certainly work well. An downward projection setup might be even better, since you could use liquid solder mask with ease.

Hello frens
Is building a wind turbine too big a task as a first project?

>>2648534
BIRD KILLER

>>2648534
>too big a task as a first project?

if you're 12 or over, you can do it.

>>2648538
thanks
>>2648536
I'll put up a 'no birds' sign so they know to stay away

>>2648534
Generally yea, though it depends on what you have experience with. Building a wind turbine has a few facets to it, namely mechanical, electromagnetic, and electrical.
The mechanical side is pretty straightforward, though not lacking in engineering rigour. Other people on YouTube seem to go for the vertical cylindrical turbine design for simplicity, which I’d likely do also. You can either do lift-based or drag-based, which will mainly depend on how complex and consistent of a curved airfoil you’re able to manufacture. A CNC foam cutter and fibreglass would probably make some neat airfoils.
Electromagnetic will require a little reading on what sort of generator hardware to use, the easiest being a car alternator or a brushed universal motor. Car alternators have the advantage of having built-in voltage regulation. Going brushless (induction or synchronous) is also an option, but here are more complicated. It’s also an option to construct and wind your own generator, but I’d recommend against this u less you have a good reason.
Electrical will probably be trivial. Either you use an off the shelf MPPT wind turbine charge controller, regulator, and/or inverter, or you just feed an inverter or DC-DC battery charger directly, assuming you’ve done the math be ensure the generator voltage doesn’t get too low or high. Then there’s things like a BMS, dynamic load switching (e.g. heat up your water cylinder while the wind is blowing) grid-tied-inverter, that can also be bought off the shelf. Again, making your own is technically an option, but these are very complicated devices that you’d want half a decade of electronics experience before even thinking about building one. If the DC output of a brushed universal motor is high enough, you could actually feed it directly into SMPS devices like phone chargers without even needing an inverter.

TL;DR watch a YouTube video of people making them and decide for yourself whether it’s something you could do.

>>2648548
>>2648548
thanks I wasn't expecting such a detailed reply
I'm going to save the text of your post for the future, and you're right probably checking out youtube is a good idea, I've already seen a design on there I like (made out of a fisher& paykal washing machine)
thanks

>>2648539
>'no birds' sign
lmao

What's your favorite rarely used stock symbol/model in LTSpice? I just discovered a time controlled switch. Not sure if I'll ever use it though since I can achieve the same result with voltage source functions: pulse etc.

>>2648512
Still, I don't understand why did they drill holes to install components.
>>2648533
Laser is expensive.
If we're talking about specialized equipment, I'd rather buy a special UV flatbed printer.
But yeah, Id rather pay zheng to suffer with over/under etch :D

Are there any developments in low-voltage ebeam lithography yet? Or in nanoscratch.

>>2648654
I heard that MIT has made progress with advanced ligmography techniques.

>>2648600
>Laser is expensive.
If you already have a cartesian CNC machine like a 3D printer it's a $50 upgrade.

>a special UV flatbed printer
What technology do those use?

>>2648804
>What technology do those use?
UV cure inkjet I think.

I need to know how I can de-nigger myself with electronics. Is there a book or youtube video you can link that guides me step-by-step? What kind of sewing iron will I need? I live in Zanzibar so no Amazon.

>>2648121
just do point to point or manhattan style

>>2648828
The first step of the process is to stop relying on hand holding and dive in to the subject directly on your own. That includes the ability to find your own resources and do your own research. Only resort to asking others when you have exhausted every possible resource, otherwise you won't gain an ability to solve problems on your own. There are some possible beginner resources in the OP, but those are just suggestions, because everyone learns differently and some people prefer one book to the other and some would rather listen to lectures. You can also direct your questions to ChatGPT as it will be knowledgeable enough as a basics "teacher".

>>2648854
The problem he is experiencing is precisely that (((chatgpt))) isn't even capable of teaching nigger-tier anything, least of all things like electronics (as opposed to, say, js-based webtardation).
The books in the OP are a way better way to start, but they don't really span a large skillrange, it's really just hyper-beginner (really only good for kids) and late-college-tier.
The suggestion to "just do it" is a good one, but the problem is that you can't "do" what you can't even name, and electronics is particularly jargony, the only thing more jargony than electronics is probably biology.

Overall the following suggestions work best:
- Start with the beginner books in the OP, skipping anything you already know and stopping when it becomes too much.
- Get a simple project list (there are tons of them) and practice those.
- Look up how others tackle those same exercises, this will teach you about the existence and use of all kinds of components you have never heard about before.
- 6 months from now you should be ready to start the non-beginner books.
And the rest is actually researchable from there because you now have the background and jargon sufficiently down to understand a wiki page about something, and to relate it to what you're actually doing.

>>2648846
Can't really mount SMD LEDs manhattan style, not enough cooling.
>>2648828
>I need to know how I can de-nigger myself with electronics.
Uh, I don't think you can get strong bleaching effect with electricity, chemical way is much better way to go .
>Is there a book or youtube video you can link that guides me step-by-step?
Yeah, but idk. >>2647663 check this out.
>What kind of sewing iron will I need?
None. Use solderless breadboard. Well, maybe 30W-40W mains iron that is sold in any country for rather cheap.
Maybe soldering station which is more expensive (depends), but I'd not really care about soldering until you know the theory, and soldering is just a dumb skill, like welding or driving a car. You just need experience.
>I live in Zanzibar so no Amazon.
Aliexpress brother.

>>2648871
> it's really just hyper-beginner (really only good for kids) and late-college-tier.
Recommendations are open. I think it would be better to have a selection of more focused works, like one on SMPS design, one on instrumentation, one on analogue audio design, etc.
>the only thing more jargony than electronics is probably biology
I think optics is up there.

>>2648890
> Can't really mount SMD LEDs manhattan style, not enough cooling
You also can’t make a nest of thermal vias unless you’re autistic enough to make a 2-sided board and spend ages drilling holes and riveting copper wire in them. I’ve done it. I’d rather make a electroless silver plating setup.

>>2648902
>Recommendations are open. I think it would be better to have a selection of more focused works, like one on SMPS design, one on instrumentation, one on analogue audio design, etc

just one + one + one?
how about 1075 books on every electronics topic under the sun?
21gig of automated, cybernated, electroplated goodness.
magnet:?xt=urn:btih:KRK6F3CR2YLLMTHF2KXVDFXVDIEOWOLW

>>2648902
>I think it would be better to have a selection of more focused works, like one on SMPS design, one on instrumentation, one on analogue audio design, etc.
That would definitely be an interesting direction.
Another issue with the collection in the OP is that it's a bit of a messy mix and match for the purpose of learning. For example How to Diagnose(...) is worthless for learning purposes, All New(...) is a great resource as a way to get up to speed with whatever it is you forgot from not touching electronics for 15 years, or to generate exam questions for your students, but is a very poor resource for learning, and Practical Electronics(...) is a really good "solutions inventory" (it really fits well with the theme it's going for), but again won't help with principles at all (you already need to know your way around electronics and then it tells you "oh by the way you can use this to achieve that result or this result"). Thus, if you want to put a "curriculum" in the OP, maybe taking references from good organized programs (because they will expose the right progression, though one should always vet their choices of books...) might be a valid strategy.

Potential picks for beginner books include:
Newnes Electrical Engineer's Handbook
Electrical and Electronic Principles and Technology (J. Bird)
Electrical and Electronic Principles and Technology (J. Bird)

But I'm not too sure what to put for intermediate levels. As for advanced, TAoE and PEfI should be as much literature as anyone should need, I think. The rest is purpose-specific.

>>2648922
Anyone can type "electronics" in libgen.is (to give a caricatural example). The problem is to know which are worth a read for the right level. The vast majority of books on any topic are complete rubbish, and even if you find 10 gems, they're worthless if you can't understand them because you're too new to this.

The best electronics book I've found so far is the arrl handbook.

>>2648854
In my experience, some of the stuff ChatGPT says is imprecise/incorrect, but sounds legit. Stuff that you will pick up if you have at least a basic understanding of the topic, but will completely go by you if you are a total noob.

I want to make a switch for my garage lights that turns on when I get close with my car but only if it's dark outside. I have put pic related together but I have no idea if it will work. Can someone please check this over and tell me in what ways I fucked up before I go out and spend money on components (read: shipping from mouser).
I was thinking maybe have transceiver only blast ultrasound when LDR says it's dark?

>ebike charger
>fan off when no charging leading to excess heat and possibly a house fire in the long run
which of these trimmers would be the most likely candidate to control the fan, anon kun?

>>2649035
The 2 pole plug is the fan

tl;dr; how to go from designing / routing 2 layer PCBs to 4 layer PCBs ? Are there any rules on what to do ? I do Vcc/Sig | GND for 2 layer, for a 4 layer PCB, do i do Vcc/Sig| GND | GND | Vcc/sig ? Do you predecide that you will route 50% parts on the top layer and 50% on the bottom ? kthx

>>2649037
4 layer boards are failures to do all the routing on 2 layers, which were, in turn, failures to route a single sided board. Thats why you see the majority of boards with, like, 95% of the components on one side.

Nobody wakes up one morning and says “I’m going to design a 4 layer board today because I’m in a 4 layer mood”

Usually you’ll hand a rev 3 board change to bhupinder that requires the addion of a few components, maybe it’s a package change from the mfg because of stock issues or whatever, and they can’t fucking handle it so they just order up another layer and call it inflation. It’s cheaper than calling me back to fucking do it.

Same way software is maintained nowadays.

>>2649037
Just watch Robert Feranec's videos on youtube. After that you can also watch talks by Eric Bogatin and Rick Hartley. The main thing you need to learn is how EM energy spreads around. You'll then be able to layout anything, even high speed USB, switchers and everything else.

>>2649033
So, I think you need to breadboard this before doing anything. Some quick thoughts:
• you probably want some sensitivity adjustment trimpots for the tranceiver and the LDR.
• you don’t appear to need a dual rail op-amp
• voltage is usually a bit higher, like ±15v
• the negative input should be more negative than VCC. Like GND.
• might want to redraw the gain to not look like a trick question on an EE quiz
• it’s highly unlikely that the op amp can drive that relay coil without a transistor driver.
• i suspect ultrasound is pulsed with some duty cycle. What’s wrong with a PIR module?

>>2645377
What's the fastest approach to control two stepper motors, endswitches and a few LEDs over USB with a PC, with a Python program? It's not a time-critical operation like a router or 3D-printer, just a demo setup.
I don't mind soldering and doing stuff on my own but I'm in a hurry so please help a stressed out anon out

>>2649047
Ok, so I understood some of this.
Trimpots make a lot of sense and I'll put them in.
For opamp I just put in what I found on the internet. Is LM321 a suitable replacement?
So I should put in a different ACDC converter that outputs 15V like an IRM-02-15?
>the negative input should be more negative than VCC. Like GND.
On the opamp? Should I just put it to GND or -15V from the ACDC then?
1/2

>>2649047
>>2649063
>might want to redraw the gain to not look like a trick question on an EE quiz
That's what I have no idea about.
Transistor driver? Is this different to a gate driver?
Doesn't say anything about pulsed duty cycle in the data sheet of the
>CUSA
>TR801
>065
>2000
>TH68
>What’s wrong with a PIR module?
I was thinking that ultrasonic would be more directional and would eliminate roaming cats from triggering it.

Stupid fucking chan won't let me post
2/2

>>2648531
Getting a good sensor is the hard part. Sure you can get a cheap-ass sensor, but it'll be a very blurry slideshow.
The smartphone dongles are basically just a sensor module in a case, so IDK if you will have much luck beating the price.
There aren't many suppliers and US-market sensors are usually limited to 9 FPS by law so you don't really have many options available.

>>2648531
There was a recent paper about a low-cost method for producing thermal lenses.

>>2649035
Reverse engineer it yourself, for which you'll have to hope those ICs aren't no-name chinky chips. Look for a voltage divider being created and fed into an op-amp or comparator, where the other end leads to some sort of temperature sensor. This is likely a thermistor, but it may be a diode. If it has no temperature sensor and is hence open-loop, maybe it's running off the output current instead, so check if the current sense amplifier is going to anything else. There might also be an MCU PWMing the fan, doubt they have an analogue PWM circuit. In this case, just hard-wire the fan on permanently or with a switch. Or with your own temperature sense circuit.

>>2649040
Some people use 4 layer boards for signal integrity purposes, or to fit massive power planes. And some people put all components on one side to make soldering with a reflow plate easy.

>>2649058
For a quick and dirty solution, it's going to be feeding an arduino nano or uno serial data using the python serial library. The arduino itself is telling a pair of L298Ns what to pulse. Even easier if you can get a pair of stepper driver shields for the arduino uno, that way you barely need to wire anything, though you may well find that they both try to use the same pins and so won't work independently. For end-stop switches, just wire some microswitches or whatever with one lead going to an unused pin on the arduino, and another lead going to ground. Tell the arduino to enable an internal pullup on that pin. As for LEDs, you can either go the meme route with WS2812s (individual or strip), or just use individual LEDs with a ~1kΩ resistor on each LED. Normal (non addressable) RGB or white LED strip lights need to be controlled with a transistor on each negative pin, so it's less straightforward. You'll likely want PSUs with both 12V and 5V rails, the 5V regulator on the arduino *might* be fine.

>>2649077
Thanks for the input! I'll get a adafruit motor shield. meme LEDs sound interesting, but I'll stick to one or two LEDs for now, that's sufficient.
>>2649071
Could you hack a cheap sensor with a peltier cooler to get nicer pictures?

>>2649105
>Could you hack a cheap sensor with a peltier cooler to get nicer pictures?
Maybe, don't know about that. Unfortunately that wouldn't change the fact that cheap sensors usually have really low resolutions/framerates

>>2649033
>a switch for my garage lights that turns on when I get close with my car but only if it's dark outside.

so you're expending money and effort to replicate the exact functionality of one of these gadgets.
which cost me $3 at the thrift store, or $20 new.

>>2649063
>>2649065
> lm321?
I haven’t used a 741 in over 20 years, that’s a 70’s era component. LM358 is used everywhere nowadays, LM321s are also hugely popular.
> ac/dc, voltages, gain
App notes are your friend, here. Some of the data sheets will have a dozen or so sample circuits you can adapt to your needs.
You might want to use multiple op amps, the photosensor will probably use one op amp by itself acting as an independently adjustable-sensitivity comparator.
> transistor driver
This is one of the most common circuits on the internet due to things like the arduino being barely to get enough current together to drive a LED.
You’ll need one with a flyback/snubber diode.
> duty cycle
I don’t know much about ultrasonic transducers, but I assume they just send out short bursts every half second or so. If you just blast it, your cat will hate you.
PIR modules are pretty directional as far as light is directional. Depends on the plastic lens and how you mount it (e.g. if you mount it in a depression or a short tube and point the opening to the target it will be more directional)

i have a government surplus auction place near me and they always have a ton of lab equipment and shit for sale. i've gone and looked around and they usually have scales like pic related, but there's shit wrong with them like the screen has failed or the glass is broken or whatever. i don't know all that much about electronics but i have considered just buying a couple of the same brand and seeing if i could frankenstein them together. do you think that'd be possible?

>>2649235
If you can figure out how to calibrate them properly then it might be worth it for precision scientific instruments. Otherwise it's a waste of money and effort.

>>2649239
i'm a chem E undergrad and i do chemistry shit at home. my gf paints and i decided to get into synthesizing dyes and pigments during covid. nothing serious, just a hobby, but if i could get an actual precision balance for pocket change that would be really badass.

>>2649243
Something to get you started:
https://www.nist.gov/pml/owm/laboratory-metrology/calibration-procedures

>>2649257
so i get that, my question is more like, if anyone has any idea of how daunting the electronics would be. like, i bought a hot plate before for like, 10$, and i took the bottom off and it was like one broken wire that had to be soldered. but a hot plate is a braindead thing, just potentiometers and heating elements, but a scale is going to have microcontrollers and shit in there.
i guess, worst case i could just buy one and take it apart and find out myself.

>>2649268
This is just a basic rundown of the components:
https://hammelscale.com/scale-parts/

This gives you a better idea of the circuitry involved:
https://www.analog.com/en/applications/markets/instrumentation-and-measurement-pavilion-home/weigh-scales.html

>>2649271
interesting, thanks friend.
might be a fun project and would be really cool if i could get an super precise scale working for under 100$.
worst case i could probably part it out on ebay or something and recoup it if i really wanted to.

>>2649280
>part it out on ebay or something and recoup it
If it's something that's in decent shape and complete there's no reason you couldn't. Just don't get sucked into the trap of buying broken stuff in the hopes of fixing it because it's a toss-up most of the time. Slut-tronics, because there's no telling how many retard hands have been inside before you get to take a turn.

>>2649308
well the thing is, an analytical balance is like a 1000$ piece of equipment. i'd never buy one, but if i could fix one up for a couple hundred bucks, that isn't so bad.
discovering government surplus auctions was one of the coolest things i've done recently. they'll sell off like ultra low temperature freezers for 20$, busted up crown victorias for 800$. sure most of the shit is in fuckin awful condition. if a university can fix up a mass spectrometer, they'd sooner do that than toss it.
but i mean, sometimes you get shit like this,
https://www.allsurplus.com/asset/22138/782
a fuckin cnc mill, how bad could it be?

>>2649330
The vast majority of scientific equipment is overpriced by a factor going as high as 10000x, and you can build the same shit for pennies on the dollar and achieve 90%-110% of what the stuff providers will try to scam you with can do. This goes for everything from SEM and NGS kits ($500-$1000 to make it yourself) to AFM and STM ($50) to brightfield microscopes (as low as $1, no really) to (likely... :^)) analytical balances. In truth what you're really buying with them is the "brand proof", which makes publishing scientific results easier because "everyone else has used hardware from that brand" so they all implicitly trusts it without requiring you to write a 100 page dissertation published bit by bit over a few years before they start accepting your results may be valid from using that instrument.
If you aren't in a scientific institution and pressured by publish-or-perish, then you don't have to suffer those scams. Just build your own shit. There are plans and CADs for most anything in a usual chem or bio lab.

>>2649330
Even if the circuitry is fucked, so long as the load cells are intact you should be able to whip up a PCB with MCU, good ADCs, good op-amps and matched resistors for like $30. It’s the mechanical side that would be expensive to replace, and also more likely to be dead than the main circuitry. That or the user interface components, like buttons and rotary encoders, and the display. Those will be cheap. The circuitry might be damaged if some chemical got spilled and went inside I guess.

>>2649035
If it's building up heat on idle, it's probably still transforming DC
Ork solution: splice a pair of leads off the fan's supply and jack them into into the output so the fan is always on.

Electrocuted the fuck out of my arm today even though I was using a grounding wire on my wrist. Is this something I should see a doctor about or can I treat it at home? Is the part opposite of my elbow where it bends
https://imgur.com/a/wTRweLH

>>2645377
Hi, got a friend who runs a small personal server in town where archives and stuff are hosted for free, a recent power failure made him change his backup power system's (2500w inverter I believe it was) batteries after realizing these only lasted for about an hour or so (should've been 4 hours supposedly) but while doing so another power outage occurred and some files got fucked up or something since the whole system shut down.
Left me wondering if there was a way to connect a secondary backup power system with a relay or something by reusing some of the old batteries which he's never thrown away; like something to keep the system functioning while somebody performs maintenance on the main battery series, keeping the secondary system charged while sensing when the primary system is disconnected.
Is anything like this doable? Only became aware of these problems while helping him reconnect the new battery system since the delivery guys from the company he got the batteries from just left the things lying around (They were supposed to hook them up again.)

>>2645377
This is the problem with boards and threads on chan, it's all hyper focused autistic shit I don't care about. Even the stuff I do care about is smarmy faggot know it alls that are pants on head retarded.

It's full of people who know nothing. Just like this thread. It's DIY but what projects have any of you faggots actually completed and then posted?

>>2649745
Sorry, I know nothing.

>>2649745
i think the only time i come to post is when i have a question about something i can't solve by googling, or where i want an opinion from people who will actually respond to me in a timely manner and who probably have no reason to bullshit me.

the only time i contribute is while waiting on answers to my questions. i don't understand people who lurk these threads full time.

>>2649752
>i don't understand people who lurk these threads full time.
This is the digital globohomo information superinterstate classroom of the 21st century, sponsored by tiktok and Bud Light.

>>2649757
>tiktok and Bud Light.
aliexpress and amazon

>>2649781

Hey guys, I want to repair an old oscilloscope and I found what I believe is broken. Anyone have any experience de-potting these hybrid ICs? Can I just soak it in acetone, or should I be more surgical in my approach? In case if you are wondering, I do have the schematic for this part, but the transistor types have been left out, so I want to see if I can identify them and maybe do a component level repair.

>>2649837
It looks like a capacitor network / array.

I'm designing an absolutely retarded guitar pickup, should it have a resonant frequency lower, higher, or at the frequency of the string it's attached to?

>>2649745
Ok and ? If you actually were in these threads you would know what sort of cool stuff people make all the time.
>>2649752
>the only time i contribute is while waiting on answers to my questions. i don't understand people who lurk these threads full time.
Then I am sorry for you, and I am happy people here are not like you.
Having a passion for electronics, microcontrollers is cool and there arent many forums where people who like it can come together.

>>2647350
It's your fault grandpa
Why use this ancient crap when you can get yourself a Saleae

>>2645567
I try to answer questions in my free time but I stopped after constantly being harassed by know-it-all cunts who actually know very little.

>>2650340
When did that happen? How do you get harassed in /ohm/?

>>2650348
Who made this gif? I need to kill them.

>>2650359
Lemme guess- you followed the gif tutorial?

>>2647717
Don't get sad, you can watch youtube videos about simple projects and just copy it exactly. Then listen to how it works and then dig a bit deeper into the math/physics. That way you don't get overwhelmed with 2 weeks of textbook material and 0 funtime. as the other guy said, a basic electronic kit should have a few leds, resistors, capacitors and transistors, preferably BJTs (get npn, and pnp if you feel like it in that order) since they are significantly cheaper than through hole mosfets. Can't stress enough, your components must be through hole, surface mount components are pretty much useless for electronic projects unless you build your own pcbs. buy yourself a bread board as well + a few jumper wires. a typical arduino kit will have all of those in it for 30$ or even less. These days I'd recommend getting the esp32 instead of an arduino since the esp32 is just superior in every way and dirt cheap. It has wifi/bluetooth + adc/dac and much more so no need to fuck around with shields.

>>2648121
check out allpcb, i think they give you free pcb sample under a specific size too.

>>2648026
As a beginner try to stick to through-hole components as it makes it easier to solder. Make sure that your high frequency lines are properly matched to prevent reflections and your traces are spaced apart to prevent cross talk. For signals that require length matching, make sure to do so to prevent racing conditions. Differential pairs should also be wired using the differential pair tool on PCB designs. Even low frequency traces can have a lot of cross talk if you have a lot of amps flowing through them, so keep that in mind. If you have a multi-layered board, try to have at least 1 ground plane and if you can a power plane as well. Avoid using too many vias, use them only when necessary. If you can't have power plane or ground planes, try to use copper pours for those. Watch out for voltage drops across your traces, this can be a big issue. Use an analog and digital ground as explained here
>https://www.arrow.com/en/research-and-events/articles/principles-of-grounding-for-mixed-signal-designs..
Avoid cuts in your return path, especially for hf signals as it creates a lot of EMI. if you are using components that have antennas (such as an ESP32 with pcb antennas), be careful about where you place them, you should avoid having a ground pour next to it as it can have a negative effect on it. If you don't have time to do multiple prototypes add contingencies to your board in the event that one section doesn't work such as solder jumps and pin connectors to redirect some traces to other places. Add test points everywhere you think might be required. Last thing I can think of is be very mindful of the 3D footprint of your components. Sometimes just because the footprint on the PCB seem to be okay, it does not mean that you have clearence in 3D. For example, say you're using a terminal block connector with wires connected to the side parallel to the PCB and place a relay right next to it. If you're not careful, the relay might block them.

>>2648439
Not sure exactly what you're trying to accomplish here, but if you need a dozen of latches in series, what about using a shift register? If you post more info about what you're trying to accomplish there might be a better way.

>>2649042
+1 for robert feranec

anybody recognizes this burnt part? 5 pin semiconductor, pretty near power input, if not immediately connected to it

>>2650388
what is the board? 5 pin sot-23 could be an Op Amp. check if the component is connected in a way that might be op-amp related

>>2650388
>>2650393
otherwise could also be a voltage regulator.

>>2650393
its a wah predal, its def some sort of regulator/LDO but dont know which one exactly

>>2650388
Is there an inductor+diode just north of there? Could be a step-up (boost) IC.

>>2650395
sot23-5 are standard packages. no way to tell the exact type without a number unfortunately. if you know the in/out voltage you could replace it with another type of regulator and see if it works. if you can tell me the numbers on the component i can try to find it for you.

>>2650396
no inductor afaik
>>2650397
it ends on "06", everything else is burnt. pedal takes 9v. i think ill write dunlop, ty anon

>>2650399
do you have the exact model of the pedal? I can try to look the same item online, maybe they have a pcb picture without the burnt component

>>2650401
crybaby sc95 "slash"
did same search and best i got was a reddit pic but still cant see shit

>>2650403
any chance of posting a picture of the entire PCB, i found some images that look similar, need to know where the chip is located exactly for comparison

>>2650405
reddit thread "Any mod ideas for a dunlop crybaby gcb95F??", right corner down, just under DC jack

>>2650388
It's a regulator to drop the 9V battery to 5V or 3.3V.

>>2650409
ending 06 might be to 6V perhaps

>>2650411
78L06?

>>2650413
pretty sure its LDO

>>2650411
it has 2 resistors connected to one of the pins and on the other side the 2 pins are shorted, so i think it's some kind of adjustable LDO.

>>2650403
sorry, looked around and couldn't find it. It looks like they all have different chip numbers. i managed to enhance a few of the pictures and could make some numbers and they don't end in 06. It's definitely a voltage regulator and older revisions used a simple zener diode to regulate the voltage. my advice, find out the output of that chip and replace it with a simple zener if you cannot fix it under warranty. Just watch out, the chip burned for a reason, you might have a short somewhere on your board, find it before replacing the chip.

>>2650420
ye
>>2650430
thanks anon. the pedal was burnt by some faggot with 18v. i really wanted to replace with the exact part but since its there only to keep battery consumption i say fuck it and will just bridge it lol :^)

>>2650434
i think i saw some guy make a video of himself fixing the exact same model on youtube, perhaps you could ask him to snap a picture of the board for you if you're not in a rush.

WE HAVE A NEW THREAD ALREADY:
>>2649781
>>2649781
>>2649781
>>2649781
>>2649781

>>2650434
I'd look for LDOs in the same footprint on Octopart or wherever. With any luck you can find one with the same pinout (reverse engineering required), and see what reference voltages they use to see if that tells you what the output voltage would be (assuming they're adjustable). I'd definitely check the voltage rating of every component on the board to see the probable voltage it's being stepped down to. If that fails, maybe there's some biasing cues to take from discrete transistor circuitry. If that too fails, then just try it on different voltages and see what sounds right. An LM317 or equivalent using bodge wires should work fine for testing.

>>2650634
>>2650634
>>2650634
>>2650634
>>2650634

YOU MADE THE THREAD TOO EARLY YOU COCK SUCKING CHINESE NIGGER MONGOLOID SHRIMP DICK

Will this thing suck the gate charge out of a MOSFET when the output is low?

>>2650656
I'm not the thread maker, I'd normally have made it much later until it got sniped at like page 5. But there's literally zero reason to post on a half-abandoned thread.

>>2650824
Yes, the output of an op-amp is push-pull. Most op-amps will pull the gate of most N-ch MOSFETs below their threshold voltage, though logic-level FETs can have quite low threshold voltages, and non RRIO op-amps can have somewhat high minimum output voltages. Comparators are typically better for that.