[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip / qa] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Settings Mobile Home
/diy/ - Do It Yourself

4chan Pass users can bypass this verification. [Learn More] [Login]
  • Please read the Rules and FAQ before posting.
  • There are 77 posters in this thread.

08/21/20New boards added: /vrpg/, /vmg/, /vst/ and /vm/
05/04/17New trial board added: /bant/ - International/Random
10/04/16New board for 4chan Pass users: /vip/ - Very Important Posts
[Hide] [Show All]

File: 8742.jpg (384 KB, 1740x1344)
384 KB
384 KB JPG
Thread killed by big oil:>>2404526

>I'm new to electronics. Where to get started?
It is an art/science of applying principles to requirements.
Find problem, learn principles, design and verify solution, build, test, post results, repeat.

>Incredibly comprehensive list of electronics resources:
Additional resources below:

>Project ideas:

>Don't ask, roll:

>Archive of Popular Electronics magazines (1954-2003):
>Microchip Tips and Tricks PDF:
>Li+/LiPo batteries required reading:


>Principles (by increasing skill level):
Mims III, Getting Started in Electronics
Geier, How to Diagnose & Fix Everything Electronic
Kybett & Boysen, All New Electronics Self-Teaching Guide
Scherz & Monk, Practical Electronics for Inventors (arguably has minor issues with mains grounding)
Horowitz and Hill, The Art of Electronics

>Recommended Design/verification tools:
KiCAD 6+
Logisim Evolution

>Recommended Components/equipment:
eBay/AliExpress sellers, for component assortments/sample kits (caveat emptor)
Local independent electronics distributors

>More related YouTube channels:

>microcontroller specific problems?
>I have junk, what do?
Shitcan it
>consumer product support or PC building?
>household/premises wiring?
More rules-driven than engineering, try /qtddtot/ or sparky general first
>antigravity and/or overunity?
Go away
File: blacklight_social.jpg (206 KB, 1200x628)
206 KB
206 KB JPG
You know what would be fantastic, if equipment manufacturers would print part numbers on components with UV reactive ink. Then you'd be able to see them.
in what software/by what means are datasheets and application notes typeset/written
is it latex?
File: fume extractor.jpg (84 KB, 960x1280)
84 KB
just finished my soldering fume extractor. i need to make some shitty filters for it, but honestly, i'm probably just gonna point this toward an open window and be done with it.

total cost was like $8, maybe a bit less. first real /ohm/ project though, it's a total hack job but it's a job done. now i can work on other stuff without worrying so much about rosin/lead fumes.
update: if i turn it around it's a great compact desk fan too. works surprisingly well for a 12v fan running on 9v
what are some appliances i should look for that have a decent amount of salvageable parts?

basically, i want to build up my collection of electronic shit very non-specifically and cheaply. ordering piecemeal on amazon will kill me in the long run, even with prime, and i figure ripping parts out of existing goodwill/junk appliances makes more sense. i'd love a ferrite core or two, more resistors, switches/buttons and anything having to do with reusable ICs.
File: 555 kit.jpg (92 KB, 1421x553)
92 KB
>anything having to do with reusable ICs.

this is not 1980.
90% of all the silicon is gonna be custom parts.
all the SMD caps are gonna be unlabeled, and the resistors unusable on breadboard.
your best bet is kits of resistors/caps/transistors/op-amps/regulators/logic from Ali.
or arduino starter kits with sensors and shit.
If they used white ink it would be visible either way. But they don't use ink in most cases, laser engraving is cheaper.

Maybe for some more neutral datasheets, but I imagine the custom templates used for particular manufacturers (especially TI's ones) aren't optimally done in a typesetting program. I know I've seen datasheet pdfs with "Microsoft Word" in the title/metadata, so I wouldn't be surprised to see them using that sort of wysiwyg with custom templates.

Old analog stereos and amplifiers. Lots of big capacitors, power transistors, that sort of thing. The older ones will also have plenty of THT components in general, but if you're at the point of salvaging passives you'd better just buy a pack off ebay for $2. Also power supplies often have big caps, inductors, transformers, that sort of thing.

>ordering piecemeal
Buy variety packs. Or order on LCSC. If you're making a ~$100 order the shipping is worth it, and you can get a shit-ton of parts from LCSC for $100. Other sites may have cheaper shipping, or free shipping thresholds, but the generic parts on LCSC like passives and jellybeans are generally quite a bit cheaper.
>you can get a shit-ton of parts from LCSC for $100
You can buy the entire company for about $tree fiddy.
I have a thing (OCD) about taking motors and speakers out of junk electronics. Optics too. HV stuff is fun. Anything that's heavy-duty rated.
Also, proprietary plug & socket combos, just because.
Most cheap UPSs here in brazil have some cheap ass tht mcu for control (usually atmega 328). I also like collecting inductors and transformers, because they are the most expensive things to ship and find. Also connectors, buttons etc. Resistors and capacitors I buy from ali, because you will spend more time organizing your salvage than using them if you get resistors and caps. (only get the high power ones).

Hard drivers, disk readers and the like usually have pretty powerful and small magnets, and some brushless motors have hall effect sensors you can reuse.
trying to understand how a 555 voltage doubler works. i got it working in ltspice, roughly doubling my input voltage. is there a way to fine tune this without changing the input voltage or input zener? i want the output to be 20V rather than 22V.
File: Capture.png (76 KB, 1266x1400)
76 KB
Change R2 to 2k.
that didn't decrease the PUMP_OUT voltage, it just changed the duty cycle.
File: 1638149954645.jpg (64 KB, 900x900)
64 KB
>this is not 1980.
i know
No. A voltage doubler gives you double the input voltage, minus diode drops. The only way you can decrease it is to use smaller capacitors, which will result in a lower average output voltage under load by virtue of having more ripple. The output voltage (and ripple) will be more load-dependent; it will still provide the original 2x-minus-drops at no load.

If you need a converter with more control over the ratio, you need to use magnetics (transformers or inductors). Boost, SEPIC, flyback, bridge, etc topology.
how feasible would it be to make a car stereo starting from an empty double ISO 7736 rack and putting a raspberry pi in it?
Just needs to play music and I'll design the UI and everything, needs to cost less than 30 euros (taking out the raspi which I already have) since there are analogue devices on aliexpress for that price.
I guess making the analog controls and fitting them into place would be a pain in the ass since I want it to look something like this.
>design the UI
just install android? I guess most of the work would be in the knobs, buttons and screen.
yeah, about that, I think I also have some 7 inch displays I could fit in but that doesn't help much, knobs are gonna be the difficult part I guess.
Any good books to get started on RF and antennas?
>i want the output to be 20V rather than 22V.

3 series diodes on the output will lower it by 2V.
which is kinda obvious to everyone but you and to those who responded to you.
File: tiiny android box T95X.jpg (381 KB, 1280x720)
381 KB
381 KB JPG
>putting a raspberry pi in it?

i'd use a tiny android box like the T95x, plus an HDMI display, and control everything using the remote control that comes with it.
this requires almost zero work. add a 12V to 5V car charger and you're done in 30 minutes.
if getting the IR from the remote into the box is too much work, replace with wireless mouse.
Stuff an amp in there with 3.5mm input and Bluetooth/USB/SD, or an old tablet/phone for the interface.
File: 1564655972507.jpg (64 KB, 600x818)
64 KB
>3 series diodes
That's the niggiest way to do it.
File: emoji.gif (1.47 MB, 382x308)
1.47 MB
1.47 MB GIF
File: proxy-image.png (10 KB, 600x324)
10 KB
>trying to understand how a 555 voltage doubler works
It's a "charge-pump". That's your key-word. Also "diode clamp" may help you..
>It's a "charge-pump"

not with the PWM signal connecting right to ground it isn't.
lets blame it on the intern.

PWM signal is shorted to ground.
the line directly joining C23 and C24 must be removed for this to work.
File: 1655677604916.gif (21 KB, 3200x1600)
21 KB
what? are you talking about this?
File: 1539293453097.gif (992 KB, 500x315)
992 KB
992 KB GIF
in a resistor voltage divider, is there any guideline to picking how many ohms to use? i keep hearing that it will work with any value of resistor, i just don't get how that works.
File: yes.gif (24 KB, 3200x1600)
24 KB
>are you talking about this?

precisely that.
>is there any guideline to picking how many ohms to use?

If the resistance is too low it might draw more current than you would prefer. If the resistance is too high then the resistance of whatever is connected might affect it significantly.
File: resistor divider.jpg (25 KB, 480x300)
25 KB
>i just don't get how that works.

pretty obvious when you look at the formula.
it's the ratio that matters, not the actual values.
100K and 220K gives same answer as 10 ohms and 22 ohms.

>is there any guideline

of course. you want the load resistance to NOT affect the calculation.
by making the dividing resistors 10x smaller than the load, you limit the error to 10%
lower resistance = more current
10+100 will give you the same voltage as 100+1k, 1k+10k, 10k+100k, etc.

Modern Antenna Design - Thomas A. Milligan
Electromagnetic Waves and Antennas - Sophocles J. Orfinidis

When you complete those two, come back and type "NIGGER" in the chat for a quiz. lmao
>ratio that matters

so basically what i'm getting from this is
>voltage = ratio of resistor values
>current = current in -resistance of resistor
and the resistance of the resistor has no effect on voltage aside from the ratio of the two resistors in sequence.
thanks sir, see you in two years I guess
fuck, i wish there were more people on here who knew about machining (not /emt/). would have helped me greatly in college years back. thanks for the explanation
File: 1635959458015.jpg (188 KB, 900x895)
188 KB
188 KB JPG
You can cut it down to a year if you read while you poop. Shitreading is more useful than shitposting.
No problem, Anon. Nobody knows everything and you gotta start somewhere.
I poop really fast, can't even read the news
File: 1562506440916.jpg (62 KB, 807x812)
62 KB
I went to grade school with a kid who took lightning shits. He was in and out in less than 2 minutes. I think he kept a stick of butter jammed in his ass at all times. Fuckin weird.
>Pretty sure those are all really common ranges for ceramic caps. If you've got the capacitance formula C = perm*area/distance then you can start designing caps pretty easily. Keep in mind the formula for dielectric strength too. Oiled paper or mylar with aluminium foil is likely your best bet for starting out, but it's going to be pretty tough to get up to 1µF. Custom ceramic dielectrics aren't going to be easy, and supercaps are kinda easy but aren't suited to audio uses.
Thanks. I think I'll start with plastified magazine paper and alufoil. Maybe I'll also use kraft paper since that was used in manufacture of paper caps many decades back.
What would the formula be if the caps are rolls of paper-foil-paper-foil ?
Draw the circuit diagram with the load attached. The load is in parallel with one of the resistors, which will affect the ratio. Typically the load isn't a constant, fixed resistance, but will draw varying current. If you want the output voltage to be stable, the current through the resistor has to be significantly higher than that through the load so that variations in load current don't have much effect upon the total current.

Needless to say, you don't use a voltage divider to generate a supply voltage. You use it to generate a reference voltage for something which draws minimal current (e.g. an amplifier input).
File: orp.png (54 KB, 434x368)
54 KB
Anyone have an idea how ORP Redox meters work? I basically want to measure and display the ORP value of ozonated water. With my very basic knowledge about the whole procedure, I assume you just measure the voltage between the two lines (BNC connector) of the probe which is then stuck into the solution? As I understand, the probe basically acts as a battery?

It can't be that easy can it?
Area = total area of both pieces of foil (with negligable amount of loss due to geometry)
distance = thickness of paper
permittivity = whatever paper has (often expressed as relative perm multiplied by perm of free space) look it up on a wikipedia data page

I'm assuming they use ion selective glass encapsulated electrodes. Hydrogen ion selective membranes are used for pH measurement, and I imagine ORP measurement is done with electron selective membranes. The actual measurement is piss easy, just voltage measurement as you say, the tough and expensive part is in manufacturing the electrodes and keeping them from drifting. Also you may need temperature compensation. If that's already being done inside the probe there may be a voltage or current bias you need to apply, or there might even be full digital circuitry in there running through a BNC with phantom power. I doubt it though, hook it up to an oscope and dunk it in some lemon juice or whatever. Chances are it is just a voltage.
File: Capture.png (43 KB, 656x293)
43 KB
it is that easy
>the tough and expensive part is in manufacturing the electrodes
high end probes exist but they are rare and usually dont come with BNC. temperature compensation etc is usually done by the electronic readout unit; but then you dont have a BNC plug, but something like pic rel. is common
>Also you may need temperature compensation
I have a table for the measurements to ozone concentration and the required temperature is 25°C, so I'll probably need to compensate.
>it is that easy
That's good to hear. The controllers for ORP start at around 100 Euros and they're pretty limited, so hooking it up with a microcontroller would be great.
Well it's good to know that my bumblefucking guess managed to be more or less correct. Do you work with that sort of thing professionally?
So how do I calculate I B ?
File: IMG_20220620_083133.jpg (1.22 MB, 5056x3792)
1.22 MB
1.22 MB JPG
File: pic related.png (8 KB, 400x400)
8 KB
Take the circuit and simplify the base-emitter junction to just be a diode. Then you've got a kinda funky voltage divider with a parallel resistance. Solve for the voltage and then the current in one branch.

Not sure what kind of voltage drop you assume for diodes, 0.5V might be ok for a small signal amplifier but I'd usually lean towards 0.6V if not 0.7V. Better yet, see if there's a Vbe characteristic I/V curve you can cross-reference on the datasheet.
Wait, but the voltage across RE will also include the collector current's voltage drop.
Guess you should treat R_E like it's beta+1 times smaller, because the voltage drop actually across it will be proportional to the current through it, which will be beta+1 times higher since it's operating in the linear region.
Is it just me, or could LC filters be used without much difficulty PSU filtering for audio circuits? Not a choke + cap, but a real LC filter. I know they're a bad idea for digital circuits, but something like an op-amp circuit basically has a fixed current draw, so the inductor should keep ripple down without side-effects. I'd still need a zener / TVS to handle spikes on startup or shutdown, or when plugging in a load, but I say it's worth a try.

I'll also test chokes, assuming I have any that are good at audio frequencies. I'll have to figure out how to measure the difference between chokes and inductors with my scope since I have a few of both lying about. Just ringing the coil with a cap using a square wave and eyeballing how much it rings looks good to me, but I'm open to other ideas. No my half-bridge inverter for switching circuits like this is not currently working, unfortunately.
>LC filters be used without much difficulty PSU filtering for audio circuits?

it's overkill.
you can get Brita-approved levels of filtration with a resistor and cap, like with R1/C2
Thank you for taking your time writing that down, but I think I still don't quite understand the schematic.
I also found this site:
So taking their schematic, VBB is the voltage drop through R1 (I don't know R1 yet) and VBE is the Base Emitter voltage drop? Now and what is beta? Shit it's been too long since going to school...
>a resistor and cap
For removing power supply ripple? The voltage drop across the resistor and its resultant heat is usually a massive turnoff. I can see it working for low-power circuits though, which that circuit fulfils mostly fine.
I was getting really confused until I realised there were two R1s. Freaky.
>mic preamp that doesn't use a JFET or op-amp
Is that not an effective input impedance of 680*hFE? Or like 60k? Or am I crazy? I'd rather see a JFET there.

Anyhow, I was going to use them for SMPS filtering. What you're talking about seems to be more the "chasing 100s of microvolts" level rather than "chasing 100s of millivolts" level. I definitely wouldn't want the kind of noise I've been looking at recently to be capacitively coupled to my output like is going on with C6.

Judging by the original worksheet, I assume you're supposed to choose the collector current and base your amplifier design around that. From that you derive the load resistor. Then you choose an emitter resistor based off what you want Ve to be, which determines Re. Only then do you calculate what base current you need, and hence figure out what kind of biasing resistors R1 and R2 are required to do so.
You set R1 and R2 such that the voltage at them and Re is a divider that sets the needed DC level, which still has a degree of freedom that you get rid of by assigning a required base current. I think. Use Desmos or excel as an interactive calculator for figuring it out.

Beta is the hFE of the transistor, the current gain.
If I wanted to make a bode plot creator for audio frequencies, how difficult would it be? It would ideally create a log-log plot to view live on my oscope in X-Y mode. I already have a function generator that can make a sine wave with an input sweep voltage, which I could set to be a linear or logarithmic ramp or whatever. Obviously I'd need a rectifier, to convert the output sinusoidal signal into its amplitude, but I think I'd also need to divide it by the input amplitude for normalisation. Which sounds like a pain, but I'd be making it logarithmic anyhow so I think that's just a difference amplifier in log-world. Assuming diode-based log amps are accurate enough for that. My output amplitude might actually be constant enough across audio frequencies, so long as I buffer it.

The rectifier itself would need to be good at 20-20kHz frequencies, which doesn't really apply to diode+capacitor+leakage resistor circuits, whether they have an op-amp ideal diode or not. I think I may need a sample+hold-based ideal diode circuit, where I change the rate at which I reset the stored voltage as a function of the input frequency. Like, I'd short it once per 2 cycles, do that twice out of phase, and get the greatest of the two. Anyone have ideas? I could just ditch the rectification and show a blurry mass on the scope, but I'd rather not.
File: IMG_20220620_122119.jpg (1.24 MB, 5056x3792)
1.24 MB
1.24 MB JPG
Right, I already have the load resistor, R L based off the max I C value of 100mA.
Why not just use "bode plot creator" built in into every PC? Your sound card will have better performance and more accuracy compared to some thrown together analog solution with exotic log amps and precision rectifiers.
But V BB and R B is the load resistor, right?
File: ce amp.gif (124 KB, 929x979)
124 KB
124 KB GIF
>based off the max I C value of 100mA.

that's an odd design choice.
do you know that a transistor passing its max rated current is gonna be too hot to touch?
in the real world, you'd choose the collector resistor based on the desired gain, which is approx Rl/Re.

there's a standard step-by-step procedure to designing a common emitter amp, which i dont think you've learned.
or else you wouldnt be asking about base current.
coz base current is not part of the design procedure: it's whatever it happens to be after you've satisfied your other design criteria.
If there is FFT how come there is not FLT for the Laplace transform?
Is there any reason you don't want to use a microcontroller and do all of that in software?
How would you make the X-axis a semilog ? You'd probably need to take 10^x for every frequency or something like that.
There are some algorithms that could fit your description of FLT, but in practice I have no idea where such algorithm might be useful, because most of the times you're interested in steady state operation, so decay terms have no value there.
So LT is overrated?
no nigga. Fourier transform is a subset of it. You usually do not need to calculate the LT aside when you are designing control systems.
File: 1558219168383.jpg (93 KB, 960x898)
93 KB
Joe Theismann didn't calculate the LT and look what happened.
>Fourier transform is a subset of it.
Or rather the LT is a generalization of the FT by adding an exp decay so the integral would converge for certain "bad functions". So FT is useful

>You usually do not need to calculate the LT
and LT is overrated.
>not posting the vid

you had one job, anon.
He became Joe Calvesmann after that day.
Can I please get a QRD on offline converters? I'm considering making a tiny ATX power supply for a 100W living room PC and I'd want to do it without a big transformer. My understanding is that you basically hook up mains to an offline converter with a few discrete diodes and you can get whatever lower DC voltage you want. But a quick search on digikey shows a bunch of offline converters with supply voltages in the 10-40V range.
PicoPSU with a power brick?
Is there no CPU power connector? If there is then this looks perfect
Oh, this one https://www.mini-box.com/picoPSU-160-XT has a 4 pin. That might be the solution
Make sure you get the wide input version if you intend to use a power brick over 12V, because the regular version is a direct pass-through (no voltage regulation).
You're probably right about that being the best method for accuracy, but I do have a nice CRT hybrid scope so being able to see instant results is at least somewhat of a nice idea.

You mean DSP the whole thing? Yeah not a bad idea. I've got the audio-quality DACs and ADCs for such a project lying around, and some ESP32s that should have more than enough grunt and the DMA to pull off such a task, but ESP32s are a fucking mess so I've been procrastinating.
>How would you make the X-axis a semilog
The frequency sweep of my function generator can either be running on an internal sawtooth ramp, or an external arbitrary ramp. I'd use an external logarithmic ramp, maybe using a CD4046 locked to a DDS'd square wave of the desired frequency, though maybe with some calibration a logarithmic ramp using diode log amps would be sufficiently precise. I'm fine with a tolerance of 5%.
modern high end scopes have a bode plot function. (most of them have sig gens builtin). it is very easy to do it in software and a really big pain in the butt to do it in hardware.
newfag here, what does "Shitcan it" mean?
File: Bode Plot.png (36 KB, 859x738)
36 KB
I don't know how you envision you complete set-up since you already have a signal generator, but I'd build a complete device, complete with a cheap DDS signal generator, an MCU and an LCD, and a 4-port connector for your circuit under test. The input is connected to the signal generator and the output is connected to the MCU ADC. That way everything is synchronized, so you set the sweep parameters via some buttons and LCD, and push the start button. The rest is done in software. The circuit response is measured for every sweep freq, and the bode plot is displayed on the LCD.
I thought of something like this a while ago.
toss it
File: 20220620_210422.jpg (1.7 MB, 2961x2155)
1.7 MB
1.7 MB JPG
I'm looking for help in repairing a Switching powersupply from the 80s. This is for an Osborne 1 portable computer. The issue I am having is that it's outputting perfectly 0 volts on all rails, it has 120v , the fuses are good, and frankly I am still new to repair and am still trying to understand how these supplies work, and don't seem to get why it gets exactly 0 on every rail.


Here is the service manual if you would like to help. I really would appreciate any tips.
I also can't find what kind of power transistor this is, it doesn't have any print on it, and the manual does not say.
File: 5V and 12V supply.jpg (117 KB, 1319x517)
117 KB
117 KB JPG
>Osborne 1 portable computer.

yummy. still have the 5-inch monitor and motherboard from mine.

>appreciate any tips

no point trying to revive this geriatric patient, coz if you do some other organ is gonna give out soon after.
instead of repairing, try replacing.
new 5V 12V dual supplies are easy to find, then add a 12V charger for the -12V line.
My CRT scope is not modern, and I doubt it was high-end either.

Yeah LCD with a UART output to a computer is probably easiest. Maybe use the scope as a display output. Would likely ditch my function generator from the equation too, since I doubt it can get 3 orders of magnitude in a single sweep anyhow. But I could easily get that from a DAC and appropriate filtration.

I think I’ll just build that ESP32 DSP board and have this be one of its programs. I might even just make it a nodeMCU shield, that way I can stack on an LCD shield or half-bridge drivers or anything else I desire. Less hardware design for me.
i made a new amplifier called the uncommon source amplifier. it works on a new principle called constructive feedback
Check the electrolytic capacitors. One or more might be shorted.
does it use biodes?
>does it use biodes?

it's powered by dad jokes and bad puns.
File: EQUIVALENENT CIRCUITS.jpg (823 KB, 3640x3216)
823 KB
823 KB JPG
I have some generator questions.
I have an answer: right hand rule. learn it, love it, be it. and stop making silly images that pretty much have no point.
How can I get that -12 v. What is a charger?
Oh, or do you mean like, a charger, a separate unit, not like some sort of circuit that can switch the polarity.

Dumb question, can you use any amperage PSU, will the unit use the amperage it needs or is that a good way to damage something.

Neither. You have a fundamental misunderstanding of how generators work, it seems.

At their core, all dynamos make use of the fact that a conductor in a changing magnetic field produces a voltage. If the conductor forms a complete circuit, it will, therefore, produce current. It can only do this as long as the strength of the magnetic field that the conductor resides in is changing.

By extension, this means that you can never produce DC from this effect. In order to do so, the magnetic field would need to be capable of increasing in flux density (strength) infinitely. Obviously, this cannot happen. If you want to keep producing current, you must start reducing/reverse the magnetic field that the conductor resides in. As a consequence, this will also change both the magnitude of the induced voltage and its polarity.

In other words: Generators that make use of a dynamo can only produce AC. If you want DC, there has to be some sort of rectification going on. This can be something electromechanical (like a commutator and brushes), semiconductor based (like diodes), or something really esoteric (like a mercury-arc valve). Regardless of what you use, the exact nature and phasing of the AC waveform you feed the rectifier is going to be very rarely, if ever, of any consequence. The arrangement of the magnets doesn't matter, because all that does is change the phase of the AC waveform (which, again, we don't care about) relative to the clock position of the dynamo's rotor.

The only major exception to this is in brushed motors/generators. By design, the brushes are positioned to change their connection to the dynamo's coils at specific clock positions of the rotor, which reverses the polarity of the applied/produced voltage of the coil. But this only means that reversing the direction of rotation also reverses the polarity of the voltage a the terminals. If that voltage is positive or negative depends on the commutator design, not the magnets themselves.
does left-to-right passing a magnet perpendicular past a parallel wire induce it?

As was pointed out: Right-hand rule. Look it up.

As to your specific question: Still means nothing, because the the direction of magnetization isn't known.
>How can I get that -12 v. What is a charger?

like a phone charger.
except it's 12 volts instead of 5.
you only need it to be able to supply 100mA, according to the specs in the repair manual (page 10, figure 4)

>can you use any amperage PSU,

once again, check page 10, figure 4
+12V needs to be 2A or more.
+5V needs to be 2.5A or more.

the one i suggested (after a 30-second cursory search) meets those requirements.
feel free to make a more thorough search.
you might even find one that incorporates a -12V line.
File: 2022-06-21_16-53.png (58 KB, 350x356)
58 KB
I bougth this motor for a CNC im doing, i want to control the speed but its not easy to me due to the motor being a 20 volts DC and i dont have any experience with voltages higher that 5v, How do i achive this /ohm/??
The brick my laptop uses outputs 20 vdc. Try looking for one of those.
>How do i achive this

search aliexpress, amazon or ebay for ''DC motor speed controller''.
there's 1 billion models to choose from.
File: IMAG1723.jpg (766 KB, 2688x1520)
766 KB
766 KB JPG
With some difficulty I managed to construct this capacitor from alufoil and paper soaked in baby mineral oil I had lying around. It measures around 11 nF, although if I press on it hard it measures upwards of 50 nF.
I need some way to keep all of the material strips aligned as I roll them up together, a way to roll them up very tightly, and a more robust method of attaching the leads; I used transparent sticky tape.
Not bad, I guess.
Is there a higher power density topology for a mains-12V converter than
>step down transformer
I feel like
should give it a run for its money but I haven't actually seen anything to back that up
>a more robust method of attaching the leads

How to Solder copper wire to aluminum foil.
https://www.youtube.com/watch?v=_mYkM9lHMho [Embed]
havent tried it myself, coz i used up all the Crisco lubing my boyhole.
I want a diy solution :)
Did you know that most power supplies already do that? They rectify the 60Hz mains, filter it, then switch it at hundreds of kHz through a flyback transformer, since the resulting output caps and transformer are far smaller than an old iron transformer. You may still have some old iron transformer wall-warts lying about, but in general they’ll all be switched-mode. You can tell the old-style ones from their weight. Only big audio equipment and test equipment will use mains-frequency transformers these days, since they’re not as electrically noisy.

How much current do you need? That will dictate how beefy the resulting power supply will need to be. Does the power supply need to be isolated from mains? If so, you’ll need a transformer-based topology, like a flyback or forward. You’ll first need to choose a switching IC, Texas Instruments make some decent ones, so do Maxim. The off-line ICs are going to be easier to design with, but if you want a particular feature from a standard DC switching controller you should be able to get it working with a bootstrap power supply like a zener and an extra winding on your transformer. Switching regulators are ICs with the switching transistor built into them, but chances are you’ll want external transistors for motor driving.
I’d practice your game by designing some DC-to-DC converters before jumping headlong into isolated AC-to-DC.

Just buy a professionally made PSU and put the wires into the screw terminals written on the box.
File: wakandan technology.png (2.22 MB, 1797x971)
2.22 MB
2.22 MB PNG
what's this thing in s6e3 of better call saul? huell gives a valet driver's key to some guy to dupe, he uses a drill to get the physical key right and an oscilliscope presumably to get the fob's frequency. after that, huell hands this thing to saul and tells him pressing 4 will unlock and 5 will lock the car the valet driver was driving.

what is this thing and can it be built to be reprogrammable?
It's a movie prop.
If you had a software defined radio transceiver you could probably do something similar with the fob

Tv show*
Is there a way to test electrolytic caps when they’re still in circuit? Won’t anything else in parallel fuck with a multimeter?
>Won’t anything else in parallel fuck with a multimeter?

you can only test ESR if you have a meter that's design for in-circuit testing.

>what's this thing

actual devices exist to create either a replay attack, or a man-in-the-middle attack.
pls contact your local DSN (dark skinned neighbor) to acquire one.
lol nobody in this thread knows but the people making the show sure did
it's their version of that hackable childrens toy. I guess if they put the real one in people would think it was too crazy!
The chinks might even have devices that just listen and repeat what they pick up, like what you can get with NFC.

There’s basically no way to test them, aside from the extra smidge of information you could get with 4-wire sensing. Arguably you could put a tiny AC current through the cap and measure the voltage at different points to characterise the cap, but it would be a nightmare.
File: DALL-E mini.png (2.68 MB, 1528x1328)
2.68 MB
2.68 MB PNG
it appears the world isn't yet ready for N-chan and P-chan
File: 20220622_104325.jpg (231 KB, 914x1218)
231 KB
231 KB JPG
I dont know what is wrong, i am trying to solder this wire and the solder just goes through the hole and bubble up like this onder side. Why it does not stick the board?

Pic is other side of the board, i was soldering from under
Solder won't stick on surfaces that are oxidized. If adding flux doesn't help then you need to remove oxidized surface with an abrasive (sandpaper or scrape with knife). Make sure that wire alloy is actually solderable, because some wire alloys won't wet properly.
Are those annular rings covered with white silkscreen or does it just appear that way in the pic?
scratch the hole with the rough side of a sponge
>Most cheap UPSs here in brazil have some cheap ass tht mcu for control (usually atmega 328).
That's pretty fancy. Here in India we usually have PIC16 in UPS type things
>>2412232 >>2412305
>does left-to-right passing a magnet(s) perpendicular past a parallel wire induce it?
Yes, if followed the same way by another magnet(s) with poles flipped, then you'll have full AC.
Rotary of flipped pole magnet(s) perpendicular passing a parallel wire.
use Halbach array.
It is just metal, has a slight yellow tint but looks silver on camera
Thanks, i think i was not using enough flux. Although i think the board/my new wires sucks a bit as well

I bought a solder wire with lead, expecting it to work better but somehow it is worse. It crumbles and it is weird
File: diycb.jpg (551 KB, 2048x1866)
551 KB
551 KB JPG
ive been making prototype boards at home and now its time to take a prototype into a real industrial application. i want to take my device to market but im too jewish to get pcbs made until i get the build right, because the design keeps changing and certain parts are quite expensive. the problem im having is these fucking things are so fragile. any serious movement and some solder connection comes loose. anyone have any tips to try to protect a ghetto board like this against movement and vibration damage without spending money?. im gluing the connectors with a glue gun but it doesn't seem to help and then makes it impossible to repair
Change the pin header to a low-profile surface-mount connector. Lay the capacitor flat and glue it to the board or change it to surface-mount.
Or just learn how to solder.
>low-profile surface-mount connector.
or a card-edge connector
and use quality solder, flux, and a real soldering iron
What's the most common SMPS controller IC these days?
Is TL494 still the industry standard?
thanks breh
>Change the pin header to a low-profile surface-mount connector
actually these dupont connectors are 90% of the problem, i will spend some shekels on better connectors, thanks

>Or just learn how to solder.
getting better but everyone has to start somewhere
File: diycbc.jpg (36 KB, 527x445)
36 KB
something like this ? how do you actually secure it to the board - glue?

Is this the programming header? You could just use pogo pins in a jig pressed against test point contacts on the board. Zero extra parts.
File: file.png (1.77 MB, 1264x824)
1.77 MB
1.77 MB PNG
my first project is going well, soldering is halfway done.

This is basically the controller circuit for arduino. I am using a low power PIC that mostly sleeps and it turns on the arduino when needed. Arduino will signal back when it is no longer busy and pic will turn it off.
File: output - Copy.webm (2.93 MB, 800x450)
2.93 MB
2.93 MB WEBM
and this is what arduino does. it is a photo frame. it will be updated roughly once a day and also using the buttons if wanted
>Yes, if followed the same way by another magnet(s) with poles flipped, then you'll have full AC.
Thats why my pic showed both N-S and S-N magnets in series.
>Rotary of flipped pole magnet(s) passing a parallel wire.
Continuous 360 rotation perpendicular passing parallel wires.
>use Halbach array.
Obvious choice.
Thanks, anon.
Can someone enlighten me on why decent humidity sensors seem to be hard to come by?

Somehow i still can't wrap my head around the fact that dehumidifyers and such work with a 2 Dollar sensor while the free market is not able to produce reliable hygrostats for home automation. I read through reviews and they confirmed my suspicion: its all unreliable crap. Last one i heard whas my buddies coffee storage completely molded because the ventilator didnt run for a week straight because the hygrostat just said "error".

I ditched all type of arduino fuckery years ago, and switched to Siemens Logo. It was a good decision because i am using the same modules for years reliably for all my projects.

Now i am pondering if i should get an 80 bucks combined humidity and temp sensor that gives out a 4-20ma signal and hook it up to a Siemens Logo.
As long as the magnets are parallel to the wires and passes them perpendicular, you'll be fine.
File: lm386n.jpg (93 KB, 800x800)
93 KB
A fast retard question here
The lm386n can be a voltage comparator?
File: photomasked soldermask.jpg (1.77 MB, 2451x2725)
1.77 MB
1.77 MB JPG
>It crumbles and it is weird
That's a sign that you don't have enough flux. Chances are your iron is too hot, causing all the flux to burn off before the joint is properly made.

>tin plating
>but no solder mask
up your game man
>those bodges
Do a thorough check of your etch resist before etching, and go over any dodgy parts with nail polish (or UV-cure solder mask)

>card-edge connector
Don't those need hard-gold or nickel plating? The latter can definitely be done at home though. But more importantly, where do you get the card edge sockets that aren't the standard PC-size? The PC-size card edge sockets are way too large for what I want to use them for.

Whatever suits the purpose most. Usually general-purpose controllers aren't that well suited to certain types of conversion, like forward or current-mode-boost or what have you.

Pogo programmers are the best, yes. Even make a peg-based programming clamp like I did.

IIRC the Bosch ones might be ok. The /3DPG/ guys experiment with hygrometers a lot for filament storage, ask them too.

Maybe, but not very well. They've got internal negative feedback, so they'd be awful at small differential voltages. Might be able to leave some pins floating, or assemble a long-tailed-pair out of ~4 BJTs, but at that point I'm wondering why you bought surplus audio power amps but not any op-amps.
>Maybe, but not very well.
>but at that point I'm wondering why you bought surplus audio power amps but not any op-amps.
I think about buying 10 op-amps, in México it is a robbery to buy one lm386 for 56 pesos (2.79 dollars)
any best option?
>it is a robbery to buy one lm386 for 56 pesos (2.79 dollars)
Bro they're 9 pesos at Steren:
Don't tell me you don't have Steren in your rancho.
nice, I used to run similar setup as poor mans desk fan.

and not only was Carmen very pleased, you get 0.9 pesos off on your next compra.
wtf?, a, they sell the lm386, i need the lm386n for a mono op amp chink, but i will try buy it in that store, thanks anon precioso
File: USB_3_Pins_static.jpg (82 KB, 546x468)
82 KB
So, I just replaced smashed USB 3.0 socket in my laptop, but couldn't get broken GND Drain pin out of the hole, because of the ground plate acting as heatsink. It was connected to mounting holes of the shield so I bent the pin in new socket upwards and soldered it to shield. Have I fucked it up? It appears to work with USB 2.0 devices, but I have nothing to test the 3.0 standard.
The electrical connections are fine, but the port will be less mechanically stable.
File: ZD1880.jpg (971 KB, 3251x1273)
971 KB
971 KB JPG
Just got a 7-segment VFD that I plan on using for a benchtop power supply. For real cheap.
It has 11 digits so I can have both voltage and current on the same display, with a crude 7-seg V and A after each number. Reading about the displays has a bit of ambiguity about whether I should be pulling and pushing the grid and/or segments, so I'm wondering if anyone has more experience than I. I plan on hooking the thing up to a power supply to test what the segments do at different voltages, but for now I'd still like to be able to make some plans as to what drivers I need to use.

This part itself comes with the recommendation to use it with the UDN6118 display driver IC, but it's just an array of darlington pull-up transistors, so if I need to pull the grid and/or segment lower than the filament I can't use something like this. Or maybe I should set the filament voltage to be 5-10V higher than my GND rail? Pic related was in the "datasheet" that came with the display, which I think is alluding to something like that but it's strange to see it being set as a voltage with reference to Vcc not GND, backwards zener notwithstanding.

Also whether the filament needs AC or can be run off DC or a square wave would be good to know, since I don't plan on making this PSU out of an AC transformer.
File: sinc4.png (47 KB, 1062x609)
47 KB
I am writing my own matlab, and I started with FT of a square wave. When the range of the square wave is 0 to 1, I get the correct sinc function. But once I shift it down to -0.5 to 0.5 so the average is zero, the spectrum function looks like this. It is almost a sinc, but something happened with the central frequency. Shouldn't sin(x)/x be 1? Why am I am getting zero? Is it because the average is zero value (DC) is zero? So it is not really the proper sinc function? I can then run an inverse FT and get the correct square wave back, so this spectrum shape seems correct. But I've never seen that shape of a sinc before for a square pulse. It is always the standard sin(x)/x with the peak at zero. Why am I getting zero at zero?
got a question, what is a good way to get very cheap wire that is bundled together.

im thinking hdmi cable, but for practical purposes a few flat Ethernet cables may be better. I don't have any spare cables so I need to buy something, anyone know the best cheapest source?
If you use lookup tables for FT and use linearity of FT you can write
>rect(...) - k --FT--> sinc(...) - k delta(...)
which means that if you add any non 0 offset that offset will produce dirac delta at 0 which explains the dip at 0. This also makes sense because if you remove DC offset you also remove component in spectrum at f = 0.
HDMI cables are usually not particularly cheap for what you're getting. Compare that to a DB25 serial cable, though they're not as cheap as they could be since nobody makes them anymore. Cat5e is going to be about as cheap as anything, plus you can buy it without the ends on it for crimping your own, unlike HDMI. Or just generic ribbon cable.

Do you need it to be shielded? Any reason you need parallel lines instead of some sort of multiplexing?
rect(x) is integrable. rect(x)+k for k≠0 isn't.

If you're talking about a DFT, then F(0) will be the DC component. The DFT of a square wave approximates a sinc function, but isn't one.
ribbon may be the cheapest way to go

ill ask this as well while I'm still in a 'parting out' what to get phase for the project.

I want to have a connector in between 2 sections of cable, it doesn't need to be water tight, but it does need to be secure enough to not undo itself by its weight alone.

and I would like to be able to attach something to a circuit board that would be a female receptacle and i plug a male one into it, I was thinking dupont connectors and super gluing them, but im not sure.

I also need a circuit board/like material, let me explain, I need to attach some mechanical switches at exact spots so I cant use the boards with holes, and it would honestly be FAR more of hassle than its worth to to etch my own, id rather just attach wires and go from there, but I cant really find anything like this that either has no copper on it, or is not pre holed. any ideas?
>I was thinking dupont connectors and super gluing them
Connectors like JSTs and most Molexs have locking tabs, so do a lot of other connectors.

>circuit board/like material
Why not sheet acrylic/polycarbonate?
Honestly I'd just get some copperclad FR4, drill the required holes, and just go over where you need any pads to be with nail polish. 0.5-1 hour of work to draw the traces and stuff, then dunk it in etchant for a while. Nail polish is way more resilient than toner transfer or photomask, so it doesn't matter if you overcook it. Or just overcook a slice deliberately to use the FR4 without any copper.
my main issue with that is I have no etchant, as this is likely going to be the only time I use a bare pcb like this, I have no need to get any for future use, and if lets say I lets say do end up making bigger runs, I would go with a professional third party, the prices I see are cheap enough to be worth it, but a bit more expensive than I want for an absolute prototype.

you have any thoughts on takeing a dremmel with a diamond burr and cutting away the copper or would this just be a hassle?
>takeing a dremmel with a diamond burr and cutting away the copper

this guy did it that way: https://www.youtube.com/watch?v=iqEwLXAkuJI
i think it's easier and faster to cut a v-shaped groove with a steel ruler and x-acto knife, or box cutter.
Hey frens, quick question
if i want a tec module to quickly warm up
can i just say pump 10 amps into it for a few seconds and then reverse the current to quickly change temperatures?
basically my goal is to switch between ambient room temperature +20c and ambient r.t. -20c as fast as possible without damaging the module.
It's one of those small 4cm ones.
Would rapid switching damage it?
File: square wave.png (108 KB, 690x608)
108 KB
108 KB PNG
Apparently there is a difference between rect() which is an even function with the range of 0 to a, and the square wave function which is an odd function with the range of -a/2 to a/2.
Both seem integrable. The integral of rect() is sinc. The integral of square wave is apparently sinc() - f(0)? Based on an example in this textbook, it is the same series as sinc() except for the missing sin(0)/0. The difference between rect and square threw me off, since I thought of them being the same. And I've never seen a graphical representation of the spectrum for the square wave. It would be the shape that I got: sinc with the central frequency missing. Maybe the square wave is not important enough from the signals and systems theory point of view?
rect(x) = 1 if |t|<1/2
rect(x) = 1/2 if |t|=1/2
rect(x) = 0 if |t|>1/2
Note that this is 0 everywhere except for the finite interval [-1/2,1/2]. Which is why you can integrate it over (-∞,∞). But ∫rect(x)+k = ∫rect(x)+∫k, and the integral of a non-zero constant over (-∞,∞) doesn't converge (i.e. is undefined).

Also: pic related is about Fourier series, not the Fourier transform. And you haven't clarified whether you're talking about the (continuous) Fourier transform or the discrete Fourier transform (DFT), which is closer to a series than to the transform.
I built another capacitor this time with paper from a different magazine, which seems to have less plastic in it. I rolled the four layers very tightly, though I think they could be tighter. The leads are rather loosely "crimped" with the foil instead of taped to it, and they still flop around. I soaked the whole thing thoroughly in the same baby oil I used before. Measures around 30 nF which I think is a massive improvement. Something that concerns me is the fact that resistance doesn't climb to infinity. What could that mean? High leakage?
File: IMAG1724.jpg (1.18 MB, 2688x1520)
1.18 MB
1.18 MB JPG
>rect(x) = 1/2 if |t|=1/2
I don't really understand this part but it probably means that the value doesn't jump from 0 to 1 instantly so it "equals" 1/2 on its way from 0 to 1. I don't know how to achieve that programmatically if i were to generate rect point by point since the next point after t/2 would be t/2 + some dx, so it is not a vertical jump. But it doesn't matter in our case. I get what you are saying about the limits of integration and convergence. I set the limits to some interval, so it is the fourier series so I don't worry about the convergence.

>Also: pic related is about Fourier series, not the Fourier transform. And you haven't clarified whether you're talking about the (continuous) Fourier transform or the discrete Fourier transform

Well I do numeric integration to calculate the fourier coefficients. And I set the limits of integration to some interval, say -pi to pi. So I guess it is the "discrete" series not continuous series, since it is still summation not integration, with a certain dx. I tried simpon's 1/3 and 3/8 and 5 to 10 point gaussian quadrature, and gauss gives weird results while the simpson's works well. And no, it is not DTF. It is brute force numeric integration:
is there any sort of chip that you can communicate with somehow over serial or i2c or some shit and use it to trigger high or low on its pins?

basically i have 30 fets i need to turn on and off but my micro only has 10 gpio pins so i need to add a slave chip and communicate with it over say serial so that way, i get more gpios to trigger the fets. (10 on master chip 10 on slave and one pin for coms where master tells slave which pins to trigger)
So i was wondering if there is some standardized chip for this? since that situation surely happens all the time
>S key stops working
>open keyboard, solder joint is cracked
>fix solder joint
incredible to think that normies would have to buy a whole new keyboard if this happened to them.

>communicate with somehow over serial or i2c or some shit and use it to trigger high or low on its pins?
a shift register is the correct answer, and most can even provide effective level shifting if you don't have logic-level fets and you don't need to drive them very hard. if you insist on a specific protocol then use another microcontroller. there are dedicated i2c io expansion chips but imo it's bad form to use things like that.
>High leakage?
yes. maybe the ink's a little conductive?
>situation surely happens all the time

sure, and it's always solved by applying a string of shift registers like the 74hc495
I was thinking about that. I also saw here: http://physics.bu.edu/~duffy/semester2/c08_dielectric_constant.html that the dielectric strength of paper is quite low. I calculated the strength of my paper to be around 1.6 V for a thickness of 0.1 mm. That's really poor; way lower than the. I don't know how much that increases once saturated with mineral/paraffin oil.
I meant to say way lower than the arduino cap meter voltage and the digital multimeter voltage
Calculated "K" is aproximately 20, which seems like a lot for paper soaked in oil if the tables : https://www.engineeringtoolbox.com/relative-permittivity-d_1660.html are to be believed. Maybe my paper is a lot thinner?
>Maybe my paper is a lot thinner?
you WILL buy a micrometer
>i want to take my device to market but im too jewish to get pcbs made until i get the build right
how jewish are we talking? pcbway will do $5 + $20 for shipping (2-4 days.) 3 iterations will cost you $100.
ive fallen into the trap before of "i have to get this right". what ends up happening, is that I never "get it right", theres always something im not sure of or happy of.
my advice: design a board, and _just make it_. it can have imperfections, even really bad ones. but once you have a real prototype (that you can actually show and send to people) in your hands, it will make it obvious what actually matters and what doesn't.
Thank you based anon!

I will check them out
If a bluetooth receiver is part of a pcb can I just solder an antenna (sma) connector onto that to boost the range outside of the enclosure I'm putting it into?
You drill holes in the board and solder it into place. You might need to get a staked via kit and press them in.
Is there a footprint on the board for a ipex / u.fl connector?
I lost power while working from home. No internet and all my work was lost.
Wife was hosting a meeting and was instantly booted.
I want to prevent that from happening, but pic related seems like a half assed solution.
I'd need 4.
personally, I stole my UPS from the office. how long do you want to have backup power? if more than a few minutes, why? also you only need enough power for the modem, router, monitors, and towers.
Nah not really, it's just a cheap Chinese amplifier board this is it and the bt receiver, just want to be able to have it run properly in a metal speaker box
Or just stack a bunch of sheets and measure with normal callipers.

Yes, but any modification to the existing circuit may change its impedance, the significance of which depends on BT version. 95% chance you're not going to be able to improve the existing antenna design without good understanding of microwave circuit design or use of a VNA. An SDR might be good enough to measure efficacy.
But if it's moving an antenna outside a metal case it's closer to 50%, to the point I'd consider it worth a shot. Just watch out for dB loss through coax.

>all my work was lost
Do you just never save?
/g/ might be better to ask for specific UPS models. And also how most efficiently to configure them.

I swear there should be some sort of ATX UPS that bypasses the whole inverter+AC-to-DC inefficiency thing, and also tells the computer itself to initiate a full RAM save to disc. Can't be too tough to figure out, and that way you'd only need like 30 seconds of battery. Maybe longer if you've also gotta dump your VRAM if you're rendering or whatever. Automated quicksave scripts for every open program might be better for user-friendliness, but honestly getting that to work with proprietary software could be a nightmare.
You need to disable the PCB antenna by moving a capacitor to the new antenna feed line. Because the board isn't fitted for a proper connector there will be loss in gain and increased noise. As the other anon said you have to mitigate impedance mismatch to optimize performance.
>Or just stack a bunch of sheets and measure with normal callipers.
but then how will you characterize the variation of thickness from sheet to sheet? your capacitors will be amateurish at best.
>variation of thickness from sheet to sheet
If it's within 10% who cares? You're averaging across the thickness variations within each sheet too, is that an issue? Honestly I think mylar would be a much better material than magazine paper. Maybe really thin oil-soaked tissue paper.
it's just a shitpost anon. micrometers are a vanity instrument like SMUs or function generators.
The true samurai gathers paper wasp nests for building capacitors.
File: IMG_20220623_215318__01.jpg (1.57 MB, 2098x3551)
1.57 MB
1.57 MB JPG
If it happens again, I run the risk of driving somewhere else and working - or getting caught and losing the privilege.
I made this battery box - inspired by drunkenly watching YouTube ice fishermen put ones together.
>12Ah lithium
>Flood lights
>Night lights
>Ham radio
More or less a portable generator that has served me in many ways.
Yesterday I plugged the modem into here and had enough battery on the laptop to pretend to work.
That cost me about $150, the same as a UPS as it's more versatile. When I'm shoveling snow at night, that illuminates the whole yard. A UPS won't.
But I have reservations to the longevity of a lithium serving as a built in capacitor for perhaps years without fail. I thought those were the ones you discharge fully, then charge.
>Do you just never save?
Kek. It's far from the 1990s where I lost complete school essays, but a damn battery and $100 of forethought would have saved me.
I probably lost 30 minutes of work but 30 hours of access/productivity.
can this be fixed just with a soldering iron?
Yeah. The middle joint is cold but all 3 joints should be reflowed with 60/40 or 63/37 and good flux. Clean the board before and after soldering, followed by a continuity/resistance check.
cheers, i'll give it a go
>it's just a shitpost anon
I'm getting real bad at noticing those recently

>30 hours of access/productivity
Ah so you're wanting something to keep the computer alive for a long time. Might be cheaper to just work from a laptop, or even something with an ARM or RISC-V (lol) chip.
If you're going to install 1kWh of battery anyhow, why not just go for a solar installation?
Made a dumb mistake. I was using meters instead of millimeters. Strength is approx. 1.15 kV. for my ~0.072 mm paper.
I will use different paper next time. I have some brownish packaging paper lying around.
we expect to see a whitepaper on this when you're done
File: electro-coil.jpg (93 KB, 768x860)
93 KB
>Getting Started in Electronics
Quick Question.
Rolling axial-wound coil for inductor or electromagnet or transformer, etc.
[whatever application]
COIL-WINDING: start 1st winding-layer on left & wind to right.
Upon reaching right end of 1st layer winding:

Do I wind 2nd layer right to left?
Or return wire to left, then wind 2nd layer left to right again?
>wind clockwise from top to bottom
>return by winding clockwise from bottom to top
is this what you mean? the current just needs to be going in the same rotational direction. you can wind it clockwise or counterclockwise in whatever braindead jumble you like, as long as it's consistent.
>the same rotational direction.
1st layer winding clockwise left to right...
2nd layer carrying on clockwise right to left..
"continue as is" clockwise.

Like how a fishing line is reeled/unreeled. As in, the wire doesn't make any sudden angular turns. An ideal coil has no upward/downward direction to its winding, it just goes round and around in the same space in the same direction. So yes technically the chirality vector flips each time you get to an end and start on another layer, but the actual angular direction vector remains constant.
File: OPTIMIZE-RATIO.jpg (547 KB, 2384x1832)
547 KB
547 KB JPG
>Getting Started in Electronics
Quick Question.
core diameter
coil height/diameter
winding current -vs- turns

Is there a general rule-of-thumb height/diameter/volume ratio?
>core diameter
Smaller core diameter means stronger magnetic field inside that core, but that is limited by the saturation of the core. Long solenoids also have stronger internal magnetic fields, and hence higher inductance. But long solenoids are generally less convenient for arranging your magnetic fields how you desire.
>winding current -vs- turns
Not sure if you're talking about this, but there's a trade-off between wire gauge and number of turns. The result is the volume of windings and the maximum amount of amp-turns won't change, just the voltage required to get those amp-turns. More turns = higher series resistance = more voltage required, usually this is optimised for the battery or power supply to give the maximum desired (instantaneous) amp-turns. Amp-turns (continuous) are thermally limited, so for a given winding geometry won't change significantly whether you're making 20 turns or 200, unless you're getting cooling fluid between the wires. But you can optimise your geometry to be better thermally, this usually means getting a higher surface area to your coil (i.e. significantly under- or over-square).
Amp-turns are what your magnetic field strength is proportional to (for a given core material) so you generally want to maximise them.

Also consider square/rectangular wire to get a higher fill-factor.
i see, that seems like a great solution, i can even chain those fuckers and only use 3 pins on my micro to do it
i just hope it's fast enough, i need to change the states every 10ms
File: file.png (241 KB, 981x452)
241 KB
241 KB PNG
so i ordered a 50pack of these.
i better fucking not get chinked like with the attinys i bought
Yeah you can spit data at them at a few MHz no sweat. There are slightly more convenient I2C multiplexer ICs, but the old 595 is pretty decent.

7400s from alibay are usually pretty safe. Especially the common ones like that.
>There are slightly more convenient I2C multiplexer ICs
like what? it doesn't have to by i2c or serial, as long as whatever method it uses is simple to do. also it has to be available on ali, which usually means older shit. can't buy any new fancy ships if chinks don't stock them or they have to ship from like us with $100 shipping
File: file.png (34 KB, 665x342)
34 KB
so if i understand it correctly, i generate some clock PWM, and when the clock is HIGH is set the data pin to high or low and this writes that value into the register.
and then when i have written all of the data i send a single high latch pulse on the 3rd pin which causes the IO pins on the chip to actually change states
thats actually really simple
>and when the clock is HIGH
ok so it's rising edge, this means i need to have the data line in the state i want when the clock line goes high
wait, is this the same things that LED strips use?
okay, just to confirm i'm reading this correctly
the chip can work no problem 2v - 6v, it uses 80ua to power itself, it can output at most 20ma on one pin and at most all pins combined can output 70ma
can someone confirm this? reading datasheets is my kryptonite
File: CUSTOM-FIT-COIL.jpg (1.31 MB, 2240x3288)
1.31 MB
1.31 MB JPG
>/ohm/ - Electronics General: Decappinated Edition
>Getting Started in Electronics
Electronics Self-Teaching Guide
>>2414163 >>2414165
>2 criticals
>core saturation
Is there a spreadsheet or free online coil-design webpage that "beginners" can use?
ONLY needing novice/user to enter simplest of inputs?
>all cooling thru 1 value to offset heat-gain
-room temperature background
*user vary this for easy grasp of extra cooling needed)
-core material/permeability
-core diameter
-core density
-coil density
-coil diameter
-coil height
-wire type/shape
-wire gauge/size
-wire Ohms per length

then "it" does all the calculations?
-total surface-area/volume/weight
-total turns
-total winding-layers
-total winding-turns per layer
-total wire-length
-coil temperature
[just the basics]

Something that allows us "/ohm/" beginners to alter values for "what-if" comparisons?
Just good enough for hobbyists to get started.
you had ONE job
File: sensor.jpg (1.24 MB, 1920x1440)
1.24 MB
1.24 MB JPG
Found this going through my bag of sensors. Anyone know what it's for? Chip has the words LM393 and DN828
>like what
Don't know any off the top of my head, just remember seeing them around the arduino community once or twice.

>i need to have the data line in the state i want when the clock line goes high
Doesn't really matter, just write the serial output pin separately before writing the clock pin. If you write them both at the same time you'll end up in limbo, otherwise you should be fine. You may want to make the three pin writes a bit further apart by putting other operations between, in the event that a single clock cycle is too short a time for the shift register.

Also arguably you could use SPI or USART hardware for writing serial to a 74595, but don't quote me on that. No clue how you'd do the latch pin, might be easiest on your MCU to just have an external clock divider IC running off the first clock, but that feels like putting the cart before the horse.

Mine says only 4µA quiescent current, ±35mA for each output pin, and ±75mA maximum supply current. Well it's the MC74HC595A, so chances are it's just differences between manufacturers.
Real question is, if you have +20mA from four of the output pins, and -20mA from the other four, does that exceed the IC's limitations?

Also if you're looking for versions that can drive more current (say, for multiplexing a shit-ton of LEDs) you're shit-outta-luck. Gotta go with transistor arrays or even discrete transistors. I think I found 2-pack transistors made the most sense for my 16x16x16 RGB LED cube design.

There are calculators that do simpler stuff, like calculating inductance and calculating magnetic field strength, but nothing that I've seen that do the whole lot. I'd just steal the calculations from them and make a spreadsheet collation of equations.
Also why are you showing pictures of solenoids with three seperate windings on them? I'd also ask why you're writing like a schizo, but I don't think I'd get a coherent answer.
File: proxy-image.png (37 KB, 625x337)
37 KB
LM393 is a jellybean comparator. I'd guess it to be a simple threshold comparator, maybe with a bit of positive feedback. I've seen ones like that being used that are explicitly designed for use with an LDR for measuring light intensity, which are notable for having a pullup resistor on the input. The DO will be the digital output of the comparator, while AO may either be the analog input voltage, or the analog voltage at the trimpot. There's also caps on there to filter out ripple on the input.

Trace the circuit if you're curious. And read up on what a comparator is if you're unfamiliar with them. Op-amps too.
File: Finger-of-GOD.jpg (14 KB, 240x240)
14 KB
>Encouraging /DIY/ team-build inventory for electronics General OP-post and have a lol...............
Jeez, then everyone for themself
& fuck you all.
Ah, thanks. I think these are used in conjunction with a couple of soil-moisture sensors I have.
what is the best way to read a bunch of termistors? i don't want to waste analog pin for each one.
maybe some sort of digital termistor that can be chained? is there something like that that is used alot by diy cucks?
>fuck you all
Every general has at least one cock-free pedant hackaday contributor. Long-winded, too. They are totally intolerant to non-faggotry so if you want to get in their good graces, chop your dick off and go read books about being cockless to toddlers.
to be more precise i have some peltier modules and i need to maintain their surface at a certain temperature, like say 45C (+/-1c)
File: THIS thrad is OPs.jpg (40 KB, 600x434)
40 KB
Encouraging /DIY/ team-build inventory for electronics General OP-post links to and have a lol...............
Now noobs get /DIY/ tools like other board /generals/ have.
No more fuckabout on own, ....then need that thread-ringleader sun-nigger's dance.
>YES, everyone for themself
>& fuck you for trying to furnish general.
Well it's not like I haven't been responding with what I assume is somewhat useful information, but the extremely odd layout of those posts just rubs me the wrong way. Maybe I'm wrong and it's just ESL.

Am I the only one who finds that way of talking uncanny? If the method of speech is just a reference to something obscure I'll happily eat my words, but until then I'm pacing between the bullshit detector and the schizo alert. Not to mention I'm 90% sure he's the same guy who was talking about designing a super efficient motor/generator while trying to wind a stator coil sharing an axis of rotational of symmetry with the permanent magnet rotor, and honestly I'm kinda losing patience with that kinda samefagging. The questions like >>2414210 are also rather detailed to the point where asking on an austronesian stingray wrestling forum gets kinda ridiculous. I can see why /g/ and /sci/ have rules against homework posting. Don't get me wrong, I find magnetics interesting, and I bet this topic would be more engaging if not for the sensation that I'm talking to a brick wall while trying to explain E&M concepts.

Bro if you're reading this, there's at least 4 free magnetics sims, watch some tutorials and give one a shot. It will be more productive than this game of chinese whispers, trying to convey applied vector math through text and 2D images alone, with something akin to a language barrier impeding progress from both sides.

I mean, I didn't want to say it
>Am I the only one who finds that way of talking uncanny?
No. He's a Tongan canoe builder who's trying to build a flashlight battery for night fishing/surfing. The problem is that Tongan public schools only teach you how to cook pork wrapped in plantain leaves buried under the sand, and bone-breaking techniques for efficient raping.
None of this changes what I said about eunuchs/hackaday contributors.
File: you're the star, anon.jpg (135 KB, 780x587)
135 KB
135 KB JPG
Don't worry, anon.
This /general/ is (You)rs.

"we coooooom for thee, alone"
File: tegaki.png (10 KB, 400x400)
10 KB
>soil-moisture sensors
Watch out for those, some of them leech metal ions into the surrounding soil via galvanic corrosion. There's probably a hackaday article on it. Or two.

You can multiplex them with a bunch of MOSFETs pretty easily. The simplest method is a 1D matrix with an array of N FETs controlled by N digital outputs (could just be those outputs acting in open-drain/source mode) going to one analogue input with a single resistor. Then you could also have M thermistors coming from each MOSFET for a total of NxM thermistors and M resistors, feeding M analogue inputs. You may be able to get even less analogue pins required if you use diodes like a keyboard matrix, but maybe not.
If you do this kind of analogue multiplexing, I'd buffer each ADC input with an op-amp circuit.

Digital thermal sensors are also an option, but unless they're designed to be serially addressable like a neopixel you'll need a lot of CS pins, which aren't as easy to multiplex as an analogue signal.

Also how are you driving all of those peltiers? Because that sounds like a more difficult task. Arguably you could measure the peltiers themselves for their temperature, though I'm not sure how effective it would be if the reverse sides weren't all at exactly the same temp ±1C.
use toroid-coil, slowie!!
Go easy on the fermented pineapple juice, Atamai.
Using toroid makes that project work.
fuck off with your irrelevance.
File: 1602030529893.gif (773 KB, 333x358)
773 KB
773 KB GIF
What are you so upset about? lmao
keep shinin' anon
we relish thy spectacle
File: 1559467451821.png (160 KB, 512x512)
160 KB
160 KB PNG
and a HONK HONK to you.
> And I set the limits of integration to some interval, say -pi to pi. So I guess it is the "discrete" series not continuous series
The Fourier transform is the integral over (-∞,∞). The function being transformed must be integrable over that range. If the function has a periodic component, its Fourier transform will include a Dirac delta at that frequency.

The Fourier series only applies to periodic functions, and is a sum of sinusoids whose frequencies are integer multiples of the frequency 1/T, where T is the period. The coefficients are obtained by integrating over one period.

The DFT is essentially the Fourier series applied to a sequence of discrete values, with summation replacing integration. For any given size N, the DFT can be expressed as a matrix where a[m,n]=ω^(mn) where ω=-2πj/N.

The Fourier series for a square wave with duty cycle k has coefficients:

c[n] = (sin(2πkn) + i(cos(2πkn)-1)) / 2πn

IOW, the real part is a sinc function; the imaginary part isn't. The amplitude is

|c[n]| = √2·√(cos(2πkn)-1) / 2πn

However: for a real function, the series has conjugate symmetry: c[-n]=c[n]*. The amplitude of the real sinusoid obtained by adding the complex sinusoids for harmonics n and -n is

|c[n]+c[-n]| = |c[n]+c[n]*| = 2 |Re(c[n]))| = |sin(2πkn) / πn|

which is a sinc function. Adding or subtracting a constant (e.g. removing the DC component) will just shift this up or down. Except, you have to consider that you're shifting it before taking the absolute magnitude.

If you plot the c[n] and c[-n] coefficients separately, each will contain an equal-but-opposite imaginary part which cancels when summing to form a real signal but which will affect the magnitude which you're plotting. If you remove the DC offset so that the real part of c[0] is zero, the amplitudes of the separate c[n] and c[-n] components near n=0 will be dominated by the imaginary parts.
File: enjoyin' da ride.gif (1.82 MB, 236x173)
1.82 MB
1.82 MB GIF
luvin' the view
File: electro n sheit.jpg (174 KB, 1920x811)
174 KB
174 KB JPG
I am but a mere man, Sifu.
I merely scribble.
i noticed the peltiers generate voltage when turn them off, so like 1.5V at about 50C that is something i could measure .
the voltages should correspond to certain temperatrues right?
since the votlage generate is linear to the temp difference between the two sides.
Actually doesn't have to be precise to 1C, i can tolerate precision up to +-5C at a maximum
do you think most phone charger wall warts are designed to survive a short?

no unless they have an apple logo.
if i draw apple logo on mine, will they become short protected?
Thanks anon. Why is complex DFT summed from 0 tp n-1 not from -N to +N? I thought the whole idea of the complex FT or complex Fourier series was to combine a0 and a=cos and b=sin to bake all three in c but the interval has to be -inf to +inf or -T/2 to T/2. Is that because in case of DFT the signal is sampled in "real" time, i.e. there is no negative time? Then how does the math work?

This is what confuses me as well. So for the FT, as a result of summing from -inf to inf we get positive and negative frequencies. But we can show geometrically or via the Euler's that the imaginary parts cancel out and the real parts are doubled. So we get cos(w).
But why cosine? What are we supposed to do with this cosine? Did we end up with a cosine transform for an even function? But what if we are transforming an odd function.
File: gfd.png (23 KB, 840x789)
23 KB
>bought a Goldmine Elec Surprise box
>it's filled with a bunch of garbage and some neat shit I'll never use
I thought I was immune to loot boxes transisters but they got me too in the end
File: IMG_20220624_132903.jpg (1.55 MB, 2688x1512)
1.55 MB
1.55 MB JPG
it did have this cool looking thing in it though
> some of them leech metal ions into the surrounding soil via galvanic corrosion.
Lol. This warning came about 4 years too late. Found this out the hard way.

Luckily this can be minimized by only turning on the sensor when you actually need to take a reading. When turned on, the sensors cause a small amount of hydrolysis which greatly increases the oxidization and corrosion. Turning on the sensor for a brief moment and then turning it off when you take the reading makes it last far longer.
Damn I think I am confusing direct and inverse transforms, and also both time and frequency domains can go from 0 to N or from -N to N. So for example when we calculate f(x)e^-jw, both x and w could be negative or positive. This shit is so elusive.
> f(x)e^-jw
that should be f(x)e^-jwx
I need to make some geometrically precise and intricate coils on a PCB. Does kicad allows that?
>what you are doing
rogowski coil
File: 41003657-1-1.jpg (503 KB, 1600x1200)
503 KB
503 KB JPG
I have been soldering a prototype board, I am getting some high resistance in between pins that were supposed to be unconnected. Did I fuck it up while soldering? The resistance varies, like 5k to 100k

on a fresh board I don't get such resistance
Could be a board defect, but it's 100x more likely you fucked up somewhere. I'd just start with a new board.
It is probably me, I am noticing same issues with other boards I soldered as well. While fresh boards/unsoldered areas seems to be fine

Is that because I am using too much heat? I suck at soldering so I tend to heat things a lot and my soldering iron is shitty and has no temperature control
File: IMAG1725.jpg (1.09 MB, 2688x1520)
1.09 MB
1.09 MB JPG
Another paper capacitor. This one has tanish packaging paper as the dielectric soaked in the same oil as before. Measured ~11.3nF before oil and ~13.5nF after oil. Leakage is still high, but leak resistance did increase by more than twice after oil. I read that paper capacitors are very leaky and no good for AC coupling (and also decoupling, I deduce), so that's just something I'll have to live with. How does capacitor leakage affect AC signals?
>Is that because I am using too much heat?

i've seen PCBs become conductive when they turn black from overheated components.
presumably, if you use the cheapest chinese boards, this could happen more readily.
File: 20220530_163041.jpg (79 KB, 741x1131)
79 KB
/diy/ please help, I'm retarded and can't fix my headphones. I've looked at multiple tutorials so what the fuck am I doing wrong? I just need to solder 4 wires to these contacts. I strip the enamel, I -see- the copper, and I think I'm getting proper contact. I've checked the instructions to ensure I'm putting the correct wire to the correct contact. I even got my headphones to work perfectly for a fleeting moment, but 90% of the time all I get is a faint distant sound of whatever it is I'm trying to play. I accidentally broke the audio jack a few months ago, there is no issue with the wires themselves.

Headphones: Sennheiser HD 630VB

File: 20220530_162943.jpg (117 KB, 1350x851)
117 KB
117 KB JPG
File: 20220530_163209.jpg (126 KB, 681x689)
126 KB
126 KB JPG
I believe this shitty chinese solder might be to blame, but this is my first electronic repair.
You deserve nothing but failure.

oh, you're soldering to a plug.
i'd say your chances of success are zero in that confined space.
what you need is a plug with wires already attached. like a cut-off pair of old headphones.
then you splice your cable to those wires, and wrap it up in electrical tape.
it'll look like shit, but it'll work.
It's what came with my iron. Is it that bad?

That seems a lot more practical than fucking with these plugs. Thank, anon.
>Is it that bad?
It's really just handfuls of random waste products that chink slaves sweep up and throw into a solder pot. China's national motto is "it kinda looks like the real thing". Enjoy your nervous system disorders.
What's the use for the RLC circuit that every university insists on teaching?

>Ressonance so it becomes a purely resistive circuit

Well, just use a resistive circuit then.
>Well, just use a resistive circuit then.

you're way off, Padawan.
resistance is boring.
resonance is the essence of magic in the world of electronics.
File: 1624138301087.jpg (87 KB, 640x640)
87 KB
Fuck, I shouldn't have bitten into the roll like a pack of Bubble Tape.
File: 1640258812507.gif (2.61 MB, 498x278)
2.61 MB
2.61 MB GIF
RIP, Anon. We'll use your misfortune as a lesson for the others. Your slow, painful death shall not be in vain.
Resonance circuits make guitars work. Therefore, resonance circuits are cool and you should learn about them.
> Why is complex DFT summed from 0 tp n-1 not from -N to +N?
It doesn't make any difference. The input is one period of a periodic waveform so x[-k] and x[N-k] are aliases. The output can be -(N/2-1) to N/2 or 0 to N-1. e^((2π-x)i) = e^2πi.e^ix = e^ix, so again y[-k] and y[N-k] are aliases. In terms of programming 0 to N-1 is how arrays are normally indexed, but -(N/2-1) to N/2 makes for a better visual representation (central peak with tails vs a bathtub).

> But we can show geometrically or via the Euler's that the imaginary parts cancel out and the real parts are doubled. So we get cos(w). But why cosine?

That's only true if both coefficients are real and equal ce^jx+ce^-jx = 2c.cos(x). This implies that the component has zero phase, i.e. is a cosine wave.

For a real signal, c[-n]=c[n]* so the component with frequency ωn is:

c[-n]e^-jωnt + c[n]e^jωnt
c[n]*e^-jωnt + c[n]e^jωnt
= (a[n]-b[n]j)e^-jωnt + (a[n]+b[n]j)e^jωnt
= (a[n]-b[n]j)(cos(-ωnt)+j.sin(-ωnt)) + (a[n]+b[n]j)(cos(ωnt)+j.sin(ωnt))
= (a[n]-b[n]j)(cos(ωnt)-j.sin(ωnt)) + (a[n]+b[n]j)(cos(ωnt)+j.sin(ωnt))
= (a[n]cos(ωnt)-a[n]j.sin(ωnt)-b[n]j.cos(ωnt)+b[n]j.j.sin(ωnt)) + (a[n]cos(ωnt)+a[n]j.sin(ωnt)+b[n]j.cos(ωnt)+b[n]j.j.sin(ωnt))
= (a[n]cos(ωnt)-b[n]sin(ωnt)-a[n]j.sin(ωnt)-b[n]j.cos(ωnt)) + (a[n]cos(ωnt)-b[n]sin(ωnt)+a[n]j.sin(ωnt)+b[n]j.cos(ωnt))
= (a[n]cos(ωnt)-b[n]sin(ωnt)) - j(a[n]sin(ωnt))+b[n]cos(ωnt)) + (a[n]cos(ωnt)-b[n]sin(ωnt)) + j(a[n]sin(ωnt))+b[n]cos(ωnt))
= (a[n]cos(ωnt)-b[n]sin(ωnt)) + (a[n]cos(ωnt)-b[n]sin(ωnt)) - j(a[n]sin(ωnt))+b[n]cos(ωnt)) + j(a[n]sin(ωnt))+b[n]cos(ωnt))
= 2(a[n]cos(ωnt)-b[n]sin(ωnt))
>like 1.5V at about 50C
Yeah but the voltage is a function of the temperature difference between the hot and cold sides. So you need to take into account the rear temperature in order to use that information. Also your ADC circuit will need some sort of protection circuit to not blow up when you switch on the peltier.

t. put a 4700µF filter cap across my apple USB charger and killed it just from the inrush current

Also ensure that the tin plating isn't getting chipped, since if you get exposed copper it's going to wear away. I'd also really want to ensure that the plating is lead-free.
To be honest I think the capacitive sensors (with a properly durable insulating cover) are likely less of a hassle.

Measuring in-circuit is unreliable at best. If it's just wires you've fucked up, but if you've added transistors, diodes, or ICs then it could well just be current leaking through those.

Wonder if packing tape is worth a shot? Space blankets?
>How does capacitor leakage affect AC signals?
It lets a DC current alongside them. Which for AC-coupling, means you're no longer AC-coupling. For DC-coupling, it means you're losing a bit of that DC, which may or may not be an issue depending on other impedances.

Set iron to hotter temperature so you can solder quicker. Add extra flux to workpiece, and have a blob of fresh solder on the iron so you don't have to waste time feeding solder in. Get in and out fast. You may need to do this two or three times to get the wire properly tinned, taking breaks between each in order to let the wire cool down and not melt the PVC insulation.
Also check the copper bundles for nylon strands in the middle.

>like a cut-off pair of old headphones.
Some of either have enamel wire inside with nylon strands, I think aux cables are a bit more likely to be easier to solder.

Good flux is more important than good solder, but they're both important.
>Good flux is more important than good solder
What if it's flux-core solder? How about that, Mr. Fung?
Just add better flux to the shitty flux.
>Yeah but the voltage is a function of the temperature difference between the hot and cold sides.
ye i know, i will have one thermistor to get the abient temperature to know the base to start from, but other than that it should be very reliable no? as in i will known that for example if peltier is producing 1V the temperature difference between its two sides is exactly 50C and so on. I should def be able to hit a 2c precision with that
File: ey b0ss.png (9 KB, 400x400)
9 KB
My fear is more that an array of thermopiles would each have different enough temperatures on their other side, since there's enough reason to drive them individually in the first place. If you could assume the background temperature to be equal for all of them, then you could just measure the front temperature only once too and assume it will be the same for all. If any thermopile has to work particularly harder or less-hard then that means it's going to pump a different amount of heat from the others, hence changing the temperature on its back side. Only if the back side is all really well thermally connected together would that kind of assumption stand. Which could be the case, I think the required condition is that the thermopile-to-thermopile thermal resistance would have to be significantly less than the thermopile-to-ambient thermal resistance.

Actually now that I think about it it's not that difficult to multiplex, pic related. Much worse if you need to H-bridge them though.
File: tegaki.png (11 KB, 400x400)
11 KB
oh, i wouldn't measure just one, i would measure the voltage for each peltier separately since i need each of them to maintain different temperature (+1 thermistor to get ambient room temperature)
what i'm planing to do is a simple mosfet for each peltier to turn them on or off, and to measure temperature i will quickly turn them off take a reading and turn them back on. since the voltage i will use to power the peltiers is 5V and it seems like the peltier will never generate more than 2V since max temp i need is 50C, i can just hook it up to analog pin directly

I drew a p fet, but i think i would use nfet on the ground of the TEC instead tho. Just a dead simple circuit.
I hope tec doesn¨t do some nasty inductive shit where it spikes voltage or something
>oh, i wouldn't measure just one
NO that's not what I meant. I meant you're still sampling a differential voltage between the front and back of the thermopiles. If the temperature on the other side of the thermopile isn't the same across all the thermopiles (which it won't be if you're driving them separately) then you won't be getting accurate temperature measurements with just one thermistor on the back of the whole array. You'd need an array of thermistors to measure the back-side temperature of all the thermocouples, worst case scenario.

>If you could assume the background temperature to be equal for all of them, then you could just measure the front temperature only once too and assume it will be the same for all
>If any thermopile has to work particularly harder or less-hard then that means it's going to pump a different amount of heat from the others, hence changing the temperature on its back side
Whether or not this temperature difference is significant or not depends on your thermal situation, as I outline with my comparison of thermal resistances.
File: file.png (171 KB, 640x640)
171 KB
171 KB PNG
I want to connect boards to other boards or non attached components. I tried soldering cables directly but it does not look like it will be reliable so I want to use sockets like pic related.

What should I search for/buy? I am looking at digikey but I am very confused, there are so many options and when I find one it requires some kind of specialized tool for the wires.

Ideally I want sockets that are solderable to a pcb board with regular 2.54 mm spacing and plugs that are attachable (or solderable) to cables I already have without a specialized tool

This is all low power btw
File: file.png (1.95 MB, 1500x1413)
1.95 MB
1.95 MB PNG
I think I will buy one of those
File: USBasp.jpg (75 KB, 800x800)
75 KB
Right-angle 0.1" pin and socket headers may do what you want (the moulded ones not the crimped ones), but they're not ideal. I prefer SMD edge-mounting normal vertical headers as it's a bit more low-profile and the forces are more inline with the boards themselves. Especially if you use double-row headers spanning either side of a 2-layer board, that works really well, pic related.
There may some form of card-edge connector you can use also, but I'm not familiar with any that are designed to be edge-mounted.
JST connectors are commonly used in this application. The pins are a pain to crimp properly without an expensive crimp tool, so maybe buy bulk JST cable assemblies to avoid frustration and added cost.
I bought a new desk fan, when I switch between the three speed modes I hear my speakers popping and sometimes my screen turns off. Wtf is going on
Suddenly switching high current load will produce lots of EM radiation that will interfere with sensitive electronics. Your fan doesn't have adequate suppression.
A magnetic field from fan's coil. Place your fan away from PC or buy a USB 5V fan.
I've posted about this before but still haven't figured out my ass from a hole in the ground. One thing holding me and my projects back is I have no idea how to design custom portable power solutions (battery packs with arbitrary arrangements of cells).

Let's assume I'm working with standard samsung 25R 18650 2.5 Ah batteries. Let's say I need 20 volts, 2 amps max (not continuous) output.
Obviously there's a curve of solutions I could implement.
For example, I could have 5 in series and do some light boosting to 20 volts, but this has current (and energy density) drawbacks. I could do 6 and buck down to 20 volts, also with the same power density issue (but possibly slightly more efficient). There's also the stupid idea of 1 battery boosted up to 20 volts which would have an absolutely hamstrung current response and just be stupid period.

My question is, IN GENERAL, would it be better for me to
>go for less batteries and boost it up to the required voltage
>get as close to, but not exceed, the necessary voltage and just boost the final amount
>slightly exceed the voltage and buck it down to what I need
or something else in-between? (I want to use this for multiple laptops so in the end I'll probably have a variable 20 +-2 volts target output voltage)
i understand that BNC connector ground on oscilloscope front ends are tied together and then tied to mains ground. is it possible to use multiple non-galvanically isolated high voltage differential probes at once (something like a micsig dp10007 probe) to e.g. probe high side and low side MOSFETs at the same time, or to probe anything in general without paying attention? or would this blow shit shit up?

if this can be done safely, then what is the point of products like tektronix A6909? is this just an additional level of safety?
File: IMG_20220625_223736.jpg (653 KB, 1248x1664)
653 KB
653 KB JPG
To get 20V you need 2 batteries of 10V.
>or to probe anything in general without paying attention? or would this blow shit shit up?
Yes, you will only blow shit up if you exceed voltage ratings.
>then what is the point of products like tektronix A6909?
Same thing, but some better specs. It's basically an isolation amplifier which allows you to use existing probe in a same way you would use active differential probe.
i finally ordered the shit i need to make a portable power supply the right way. instead of LM317 wasteful niggatry, i got a buck/boost converter to run off a 9v. also, some 10k and 1k potentiometers to to replace the trim pot without losing all the accuracy. gonna be wild to see if i can actually make something pseudo-useful
As the batteries discharge, what does the voltage curve look like?
At what voltage should you stop drawing power from the battery?
wouldn't that be covered by the BMS?
I was more thinking it might make a difference in how many cells you went with if using a buck converter.
Oh, I see what you mean. Honestly questions like that are why I'm here, because there's woefully little information I can find on the proper design of these things. It's all proprietary and locked down tight, or you gotta synthesize it all yourself from basic principles of power electronics (that I never learned since I specialized in DSP and didn't take any power electronics classes)
Yeah I get some of that... but when negative frequencies in the spectrum are explained, the usual explanation is that if Cn and C-n are added together, the imaginary parts get cancel and we get (a-ib) + (a+ib) = 2a. Maybe I am confused and we actually need to add Cn.e^-jwx and C-n.e^jwx? Because Cn is just a phasor, right?
>that calculation
>= 2(a[n]cos(ωnt)-b[n]sin(ωnt))
I tried the same in the complex form, it seems a lot easier. and I got

Cne.^-jx + C-n.e^jx
= (a-jb)/2.e^-jx + (a+jb)/2.e^jx
= a(e^jx + e^-jx)/2 + ib (e^jx-e^-jx)/2
= a(e^jx + e^-jx)/2 - b (e^jx-e^-jx)/2j
= a.cos(x) - b.sin(x)

I think I know why. You have
> = (a[n]-b[n]j)e^-jωnt + (a[n]+b[n]j)e^jωnt
but doesnt c = (a - ib)/2 not just a -ib?
and still something isn't right.
I think the proper answer should be
a.cos(x) + b.sin(x)
a.cos(x) - b.sin(x)
And damn this is wrong
>Cne.^-jx + C-n.e^jx
It should be
Cne.^jx + C-n.e^-jx
then the final result
a.cos(x) + b.sin(x)
If you have the space go for 6 in series for the added capacity. Having to step up the voltage from 18.5V or stepping it down from 22.2V really doesn't make a huge difference, you will get 90-95% efficiency on the conversion either way.
Really depends on the charachteristics of the battery you want to use, for example if you use Li-Ion or lead-acid batteries the discharge curve is pretty flat from like 90%-to-20% so you connect as many as you need to exceed minimally the voltage limit then you slap a buck after it and you are good to go. Alkaline and NiMH batteries have a steeper discharge curve so if you connect the minimal number together to exceed the voltage level you may find yourself under this given voltage as they discharge. In this case I either add one more battery and step it down for the added capacity or if I don't have the space for it, I use a buck-boost converter, so when the battery pack's voltage gets under the limit it starts boosting it.
tldr;: Is there an graphic that explains the components on an Arduino PCB?
I'm not referring to the pins but more to all the bits that protect the ATMega and make it run reliable.

Long story:
I am pretty much a noob in microelectronics, so far i used raspberry pi 2, arduino and siemens logo.
I want to understand how to implement a CPU like the ATMega in a PCB in order to safely and reliably work with both sensor and card readers and relays.

From practical experiments i learned that:
-The magnetic field of the relay coils fucks with the circuit and can create all kinds of damage. There is solutions for that though on a component and circuit level, but i feel like some healthy spacing plays a role too and should be considered when planning a casing for a device (does it?)

-Multiple libraries can be conflicting in C++ , the theoretical 128 adresses you can put in an I2C network seem utopic regarding the stuff i witnessed when i tried to apply RTCs, Datalogger and a bunch of sensors. From my current understanding my solution would be to run several ATMegas, have them communicate with each other to prevent having all libraries on one chip. This would also help locate error sources if malfunction happens.

Feel free to point me in a direction to improve on
this is a great tip. Especially as it makes assembly a little more intuitive. Great if you want to sell your device as a "kit" in order to avoid all the legal stuff related to marketing an electrical device.
File: meh.png (36 KB, 737x600)
36 KB
I need advice on a matter. This should be simple, but I'm too retarded to piece this together. Anyways, I have a microcontroller outputting a sound effect that lasts for 2 seconds. I attached the output of the uC to a NPN for amplification (yes, I am aware this heats up the BJT, but I don't care due to how little this thing is going to run). But my main problem is that whenever the sound turns on, I hear this "thud" or "pop" sound before the sounds gets played and after the sound finishes (the sound itself plays just fine). I think it has something to do with influx and sudden switching, but I'm not sure what I can do to dampen or eliminate it.
The issue is that you haven't taken any care of biasing. That arrangement works fine for switching but not for amplifying a signal. You're after something like this: https://en.wikipedia.org/wiki/Common_emitter#/media/File:Complete_common_emitter_amplifier.png

What are the specifics of your microcontroller output?
File: ozzy eats a bat.jpg (10 KB, 480x360)
10 KB
>what I can do to dampen or eliminate it.

one easy thing to try is just to put a cap across the speaker; like 10 - 220uF
this should dampen the thuds but also make the sound much less grating.
listening to square waves will turn your brain to mush.
just ask Ozzy Osborne.
>fine for switching
My mistake, your load should be on the collector (higher voltage) side anyway
DC offset
>listening to square waves will turn your brain to mush.
Mushwave is a genre of electronic music consisting of monophonic square wave tones played through an array of piezo earpieces driven into the ultrasonic frequency range.
File: 1615847010768.png (1.22 MB, 1125x1394)
1.22 MB
1.22 MB PNG
need help coming up with a circuit to power a low power electric motor both ways. I am assuming that to send either power or ground to either pin would make it work how I need. What transistor would work best for this task? npn, pnp, hopefully a 2n906 or 2n2904 could suffice?
look up "mosfet h-bridge motor controller"
File: arduino nano schematic.png (379 KB, 2228x1176)
379 KB
379 KB PNG
Clip a ferrite bead on the cable. Maybe even solder in some suppression caps/chokes inside it. A single suppression cap at the input of the switch would be my attempt.

If you have a dumb load (motors, heaters, etc.) then 90% of the time it's better to just set your battery voltage to approximately the voltage you want for that load so you don't need any boosting or bucking. This is what e-bikes and such will do, the maximum power output will drop as the battery drains, but it's a lot less complicated to just not have a converter in there, especially for high-power loads. Especially especially for loads that have some form of feedback (speed-controller, thermostat, etc.).

In the event that your load ABSOLUTELY NEEDS a fixed voltage, yeah there's a spectrum of solutions, which I'd pick by function of what's easier to get working. 1S is always simplest for balancing reasons so you may want to go for that, but it does mean thicker wires and making high input-current boost converters is a bit of a pain, so I'd only do it for low-power cases. Bucking is always going to be a bit more efficient, so if you have a low-voltage high-current load (that again, NEEDS a fixed voltage) I'd go for a higher series count with buck conversion. Avoiding buck-boost is probably a good idea, but SEPIC really isn't that bad, and may possibly be more efficient if more bulky. Charging circuitry is also something to take into account, it may be easier to buck-charge from a 12V or 19V power brick than trying to step-up to 6S.
The above cases also apply if your load needs a switching converter anyhow, like a constant-current LED string.

Yes, it's called a schematic. They're easy to find for any dev-board out there, and from there you can reference the MCU's datasheet if you need to. LED on SCK pin, bypass caps on VCC, crystal across XTAL pins, reset button on reset line with pullup resistor (it's active low) with a high-pass cap to the UART transceiver's DTR, etc.
>The magnetic field of the relay coils fucks with the circuit and can create all kinds of damage
Sure you didn't forget the freewheel diode? The coil's magnetic field really shouldn't be causing issues beyond maybe 5mm distance, and even that's just noise to the ADC.
>Multiple libraries can be conflicting in C++
You're often better off just ignoring the user-written libraries since they're often written by retards. Sticking to wire.h or whatever and just writing to the dedicated TWI hardware's registers should always work fine, plus it gives you a better picture of what's actually happening under the surface.

How much current? That's going to determine whether you can get away with complimentary BJTs, complimentary logic-level MOSFETs, complimentary power MOSFETs, or N-channel power MOSFETs with drivers. There are ICs out there that have built-in H-bridges (like the common L298) but they're at tier-1 of that list so they're not particularly capable, especially if you want them to be compact and work without a massive heat-sink. More modern dedicated motor drivers with integrated switches do exist, which are an option if you're not at risk of getting fakes via alibay. Actually half-bridge drivers from alibay are also often fakes, as I found with my IR2184s. Maybe they have legit EG2131s.
let's say someone wanted to make an "EMP gun" for funsies, is there any definitive way of doing it?

because thinking through the hypotheticals in my head, it feels like it would just be blasting a god-almighty current through a highly directional antenna like a yagi-uda antenna and hoping you don't microwave your balls in the process
>a god-almighty current
god-almighty current pulse, that is
Heyo, having some trouble finding a specific component, and I figure someone here might be able to help.
I'd like to mod a fight stick to have a retractable USB cord, that behaves similarly to a retractable extension cord. This seems like the sort of thing you should be able to buy anywhere, but everything I've found so far comes with a weird catch. They only work if you pull from both sides simultaneously. Since one side will be in the stick, that won't really be possible.
Does anyone know where I can find one that'll work for my use case?
File: abscom 1-3-4.png (37 KB, 748x937)
37 KB
ABS module for a car that never came with it. Any advice before I start buying parts?
Conventional non-nuclear EMPs are made by putting a lot of current through a coil, then detonating an explosive around that coil to inject even more energy into the magnetic field, which collapses. Or something like that. Doing that in a way that can be repeated would not be easy at all, the kinds of energy required would just be impractical to be stored as anything other than explosives in terms of size and speed.

If you just want to shit out enough interference to damage electronics, a rotary spark-gap tesla coil might do you well. Could use it for a bit before the FCC bashes down your door.

>Any advice
Lay it out in KiCAD or otherwise sort your schematic out with labels so you can actually see what you're doing. And ensure that your chosen comparators are push-pull or give them pull-ups. Also why use BJTs in that strange arrangement?
>local electronics merchants don't differentiate between chokes and inductors
That would explain why the boost converter inductor I wound on one of their toroids doesn't work very well.

Also does anyone know if wrapping a second winding around an inductor and shorting it with a resistor is a useful thing? I get the feeling it should act somewhat like a choke, since it should pass DC well, but I've searched and I can't find anything on this online. I was thinking it could be a good way of suppressing ringing in a low-pass filter, and a few simulations suggest that it might be the case. That said, I'll keep experimenting with snubbers and other simple ways of reducing Q without impacting the DC characteristics too much.
There was an article about an EMP on hackaday.

The commenters were mad at the author for not cautioning that pace makers existed and could be affected.

i need to use 10 i2c devices on one bus but they all have same i2c address. do you think that if i put fet on the sda line for each device and then only turn on the fet for the one i want to talk to it will work?
Yeah if they don't have a CS line you'll need to do that. But often the "alternate address enable" pin can be used as a CS line. Trying to block the SDA line may not be trivial, SDA would be simpler but even then both are bidirectional if the thing does clock-stretching.
>Sure you didn't forget the freewheel diode?
Arent tose already art of those 5v relay cards that get sold for arduino?
do any of the cheaper (<$150) LCR/ESR meters have input protection in case i connect a charged capacitor by mistake? i wanted the de-5000 until i learned it has no protection.
Thanks. Not entirely sure I understand the question, but my uC is the ardiuno nano (atmega 328p), so its output from the uC is 0-5V and max current is ~50mA, but I can't tell you anything on the waveform other than the frequency I am trying to recreate was originally sampled at 11kHz.

Anyways, let me try some math as well as putting the load (speaker) to the collector's side of the transistor and see what happens. I might also try a 9V VCC as well
I'll give these options a try as well if the above doesn't work. Thanks!
Oh yeah they probably are. So long as it's not damaged. Show a wiring diagram just in case it's wrong, but those things are pretty easy to wire up so should be fine so long as you're not sourcing Vcc or GND from a GPIO.
File: file.png (49 KB, 553x834)
49 KB
There has to be a program that helps me designing circuits. Like where do I put my components, where to put wires and solder them etc.

Right now I am doing pic related in google sheet
File: 595.png (258 KB, 1318x746)
258 KB
258 KB PNG
KiCAD + trial and error. The cad itself does the schematic to footprint conversion, from there it usually only takes 2-3 attempts to get a sensible layout. Maybe a few more times fine-tuning the pinouts if you've got micro pins or op-amps to swap about.

Pic related, I got a bunch of 74HC595s hooked up to 7-seg displays without too much difficulty. I also got all their pins on PORTA, making it easier to directly address them.
I laid it out in SPICE, and labeled just about everything. FL/RR/B/L, those point to wheels on the car: front left, rear right, back, left, etc. And the transistors, they're just like that for ease of access.
File: 1578088534278.jpg (195 KB, 1000x750)
195 KB
195 KB JPG
>google sheet
google sheeit
thanks friend, looking good so far
>looking good so far
My display driver is doomed to never be completed though, because it takes too many MCU pins, is too large, has other flaws in its circuit. Which is why I bought a 4-digit LCD panel meter instead. And also a raw 4-digit LCD and 11-digit VFD for future projects. I’m never going to use these shitty THT 7-segment displays or THT 595s because of how large they are. Yet for some reason I haven’t yet bought any SMT 595s, even though I bought SMT 4046s without ever using that kind of IC before.

This is like the third time I’ve made a doomed schematic with 595s and/or 7-segs. The only time I actually made a 7-seg PCB was that FM receiver with an RFA5807M and a single digit that had to cycle through all three digits in an FM frequency to be useful.
absolutely for general use, just buck down. this also means you can use the same arrangement of batts to more efficiently boost higher, should you need to in the future
File: 4511 decoder.jpg (63 KB, 848x299)
63 KB
>doomed schematic with 595s and/or 7-segs

wrong tool for the job.
using a 4511 and multiplexing means you only need 8 control lines to light 4 digits.
if i want to control a small motor speed, can i simply power it through a mosfet and then feed square wave into the gate of the mosfet?

google ''pwm motor control schematic''
3 million results.
Whats "the best" RPM pickup transducer for my next home made motor speed control circuit?
File: CC LED DRIVER.png (24 KB, 350x310)
24 KB
>BCD to 7-seg latch
>can source current
Oh hey that looks really useful, especially since my displays are common-cathode. For some reason I never thought about using 7-seg decoders for multiplexing before, and now I feel like a right retard. I should probably rid myself of any desire to use latches instead of relying on persistence of vision. It also has a bunch of circuits showing how it would be used to drive VFDs and LCDs, both of which I'm going to attempt.
>BJT for current sourcing and MOSFETs for logic
Feels kinda odd to me, idk.
>no hexadecimal decoding
Why do they always do this? How has nobody ever made a hexadecimal-to-7-segment decoder IC?

And I guess that R8/C1 in your schematic keeps it blank on startup? Neat trick, though I'd consider doing that with LT instead just to check the display works properly.

Also considering that you could use a pair of 2-to-4 decoders to run both the decimal-points and the cathodes, so a CD4555 (or 556?) wouldn't be remiss if you care about saving pins and/or want multiple ranges. They'd be able to drive enough current, I suspect.

I'd use a hall-sensor and a tiny magnet JB-welded in place, though it doesn't really count as a transducer. Looking for anything more specific?
File: RPM sensors.png (146 KB, 306x789)
146 KB
146 KB PNG
>"the best" RPM pickup transducer

- if you can fit a disc on the shaft, then a light interrupter would work great. it'd be just like a mouse.
- second best is to paint the shaft and use a reflective sensor.
- third best is just paint and a photo-transistor. assumes shaft will always be lit.

is reCAPTCHA now opining about social issues, or lauding me for sucking dick?
is it coz i said ''shaft'' 3 times?
Dipshit here, can I run multiple Li-Ions (18650) off of a single TP4056-Protect in paralell without any major issues for a while? I don't have a load balancer or something fancy on my hands but I need to power a small arduino project (RS458 sniffer) off of some batteries.

>TL;DR: 2 Batteries, 1 TP4056?
TP4056 is designed for use with a single lithium cell. It's possible to charge 2 cells in parallel if their voltage is equal beforehand, but the time to full charge is doubled (500mA per cell instead of 1A for one cell).
Why don't cars have retractable piss funnels built underneath the seats so you can keep a reserve tank of coolant for emergencies?

it's gets very odorous with time.
best bet is to drink it down again: keep it in a permanent cycle.
The international space station must smell like an antifa drum circle. Dirty butts everywhere you look. Stink pits. Cheese taint. Sneeze breeze. A little pee in the hole in your car wouldn't be bad as long as it was pumped into a sealed reservoir under the hood. Then you could boil it off and piss it out of the exhaust. This would increase fuel efficiency by 42%.
Thanks Anon, that's all I needed to know. I don't need to charge it, I just need to sniff a serial line for a few hours at my workplace.
File: urine into water.gif (19 KB, 1028x173)
19 KB
>The international space station must smell like an antifa drum circle.

they actually recycle urine into water (as i suggested) and feces into hamburger.
this one again. solution is simple: most if not all i2c devices have a so called AD0 pin. that allows each device to have two different addresses. you can leverage that by using a demux, shift reg, or discrete pins to "slave select" which device youre using. ie set all ad0 pins on the devices to 0. when youre about to do something with one, set its ad0 pin to 1. that way you only care about one i2c address, and know which one is active, because you told it to be
File: 1579709993450.jpg (81 KB, 800x600)
81 KB
>successfully docked
>crew safely inside
>change clothes and have some dehydrated space coffee made from Tang
>Hey, guys. What's for dinner?
Your turdburger, Captain.
i've seen some people probe a 3 phase motor driver (usually brushless dc motors) by tying all 3 phase together with a resistor to form a virtual ground, grounding the probes to that virtual ground, and then probing the phases with 3 probes. i understand the resistors will limit current but how does this not blow up the oscilloscope? isn't the virtual ground floating at a point above oscilloscope ground i.e. you're shorting the motor controller to mains ground, through the scope front end?
File: qg7RG.png (183 KB, 1612x1008)
183 KB
183 KB PNG
If you have a balanced load, which is true for a motor there won't be any current flowing from virtual ground node to scope ground. You then have 2 cases with
>floating power supply: it doesn't matter where you connect ground, because system is isolated
>grounded power supply: it already has defined 0V as mains ground, same as scope, which means virtual ground will also be at 0V, from theoretical view you shouldn't need resistors, but because of signal integrity it's best to keep ground loops small, so making star connection could be helpful
thanks, that makes sense. i'm still learning oscilloscope safety and i have watched the eevblog video. in general, is using a decent quality DMM to measure the following a foolproof way to avoid blowing shit up?

DMM red probe: my oscilloscope ground clip, that is not yet connected to the device under test.
DMM black probe: something on my device under test that i want to connect my oscilloscope ground to.

if i see any voltage between the two, it's a no go? if there is no voltage then it's safe, notwithstanding the maximum voltage input limit of my scope and assuming i don't do something stupid like connect two probe ground clips to two different ground potentials?
If you're going to use sufficiently high resistor values >10k then you won't blow up anything, even if you mess up connections. Only real way to blow up scope is to connect ground to high power power source - either directly to mains or a powerful PSU. Ground of scope is connected diectly to earth of power plug, so only thing that can blow up are copper traces on PCB and probe itself.
i mean in general, not specific to motor controllers powered by an isolated source. is using a dmm a foolproof way of avoiding blowing up a scope probe? or do i need to go a step further and use some sort of current limiting resistor in series to account for various things (capacitive coupling i guess) that might give a false indication that the floating reference is grounded?
>isn't the virtual ground floating at a point above oscilloscope ground
Not if the motor + ESC is floating with respect to the scope.

Diodes from VSS?

>is using a dmm a foolproof way of avoiding blowing up a scope probe
Yes? Only thing that could blow up a scope probe is either overvoltage or shorting something through ground. The former can be solved by using a high-voltage DMM, while the latter can be solved by wrapping kapton tape around your scope's ground lug. Or running it on an isolation transformer like a good boy. Also not clipping ground clips to different circuit nodes, but that should be obvious.
I posted a while ago, but I'm still struggling. This is the powersupply to an Osborne one computer. It outputs exactly 0 volts on all rails and I don't know what would cause that. I keep checking parts one by one and they test fine. The power transistor gets pretty hot, but I was able to switch it manually by hand out of circuit. Both black caps, C6 and C7 are fine. All the diodes are fine, a random sort of Resistors are fine.

I'm just shooting in the dark, does anyone have any idea on an educated guess as to what would be the problem.

Here is the service manual.

File: 20220627_201355.jpg (2.92 MB, 4032x3024)
2.92 MB
2.92 MB JPG
Here's a photo
Nevermind I forgot to test these vertical diodes on heat sinks, they seem to pass voltage both ways, all of them. Maybe a common component they are hooked to is shorted?
If the something runs on +5 v and 2.5 Amps, can you put a powersupply rated for +5 volts and 11 amps?
Yes. Fuse the PSU output at the very least just in case of a short.
I'm almost entirely certain this anon is a trained AI whom thinks he's a boy from Brazil.
>hall-sensor and a tiny magnet
yes, by far the least effort.
>JB-welded in place
crikey, I think I'll neatly embed the magnet into a machined hole(s) on the rotor.
Maybe 2 or 4 magnets around the rotor so its still balanced under high RPM?
>super efficient motor/generator
It now uses a toroid coil, slowie.
Again, Thanks for that solution, anon.
It halved the length.
I’m almost certain this is samefagging.
I'm absolutely certain you're a tard.
Is there something like multi channel mosfet? I need to control a bunch of devices, each around 200ma a would rather not use like 20 stand alone fets
File: bregh.png (34 KB, 316x414)
34 KB
It's trivial to fake a (You), but that's neither here nor there.
Two (You)s and another related post coming in quick succession after 3 and a half days is just kinda strange. The fact that I was called "slowie" days after posting is icing on the cake.

And something about "super efficient motor/generator" being responded to with "It now uses a toroid coil", as if that somehow has anything to do with whether or not it's a motor or generator or whatever, is another oddity.

A further note on the actual topic, I'm pretty sure toroids are designed to keep the magnetic field INSIDE the windings, the magnetic field outside the windings is minimal. So for something to rotate you'd want it inside the toroidal core (see: tokamak, synchrotron), at which point trying to get its motion outside the toroid again would be tough.
Plus you say "coil" instead of "coils", as if a single coil will impart motion in any efficient method at all. Though that lack of plurality may not be intentional.

MOSFET arrays are likely what you're after. There's a few on Digi-key, though it may just be easier to go for darlington arrays as I think they're more common, like the ULN200_ series. Dual MOSFET packages may be acceptable if you can't find full-on arrays with common-sources, and a bunch of AO3400 packages is pretty small either way and still easily handles 200mA.

Multi-source FETs exist within integrated circuits and I don't understand them.
File: Schematic.png (2.23 MB, 1810x1304)
2.23 MB
2.23 MB PNG
SCR1 and Z1 are a crowbar circuit that shorts the +12V rail to 0V when the 5V rail goes too high. Doing this could damage anything along SCR1's path of the circuit, like D7,8,13, maybe other branches from T2, a potential thermal fuse in T2 (cont-check the windings), and DB1. Q2 and D4 and that path may also have seen more strain for a moment. Not to mention F1 and R1, but I guess you've already checked those. The chokes should all be fine, they're really not the kind of parts to die from a split-second of overcurrent. Neither is the thermal fuse or SCR itself, if the crowbar was doing its job.

The lack of regulation on the ±12V rails is somewhat unsettling, but if they're not going into anything critical I guess there's no issue with it.

Which diodes are they on the schematic, D6,7,8? An output short could have fried the diodes, but I'd expect the fuse to pop before the diodes. The same bit about Q2 and D4 applies, maybe. Q1 might also be in the main switching current path, not sure.
They may just be passing current through the transformer windings, so try to desolder them and make further measurements.
File: yorks.png (229 KB, 402x254)
229 KB
229 KB PNG
Oh and your X caps look like they've seen better days and that Y cap (guessing it's C3 or 4) looks exploded. No clue what C5 and C24 are trying to do, but they look ok from a distance.
Is it tripping a breaker or RCD?
Its the same series equivalent of the 2 wires AC-OUT.
Length of looped wire in series with lots of series N-S poles half the time and lots of series S-N poles for the other half.
I'm not showing you the end arrangement, anon.

take meds
>the windings, the magnetic field outside the windings is minimal
Those windings are not part of the motor.
Is there something like a Quinled but with constant current feature? Want to smartify an aquarium LED light.
if i have several motors and each has its own switch is there any way to use just one flyback diode for all of them so i don't have to use one per motor?
>use just one flyback diode

absolutely not.
absurd to even ask.
unless they're all in parallel.
No it just sits there, outputing 0 volts. I'm not sure what I'm doing or how this is supposed to work.
If you don't know what you are doing, you shouldn't be trying to fix a power supply of any type, let alone a switch-mode.
I don't get it, how do you learn to do this stuff if you're not allowed to try to fix it. I can replace parts, but how does anyone know how to narrow it down to what part(s).
Do you own a multimeter (pronounced mul-TIM-it-ter)?
Before you start, try to find a schematic. Look over both sides of the board with a magnifying lens and bright light. Reflow any cold solder joints you see. Test components near any brown/discolored areas on the board to make sure they're still within spec.
Not him but the schematic is right here: >>2417356.
Can someone answer this here >>2417223
>Maybe 2 or 4 magnets around the rotor so its still balanced under high RPM?
Should I do this for high RPM shafts?
>Should I do this

yes, the more magnets the better.
it increases resolution, and also the odds of one coming loose and embedding itself in someone's face.
which is always good for a laugh
File: 1535841685670.jpg (36 KB, 657x527)
36 KB
>embedding itself in someone's face
Happened to me. The guys in the shop nicknamed me Compass Face because if I float on my back in a swimming pool my head points north.
>>2417882 >>2417888
>the more magnets the better.
I only want the high RPM shaft balanced, anon.
Maximum of 3 (120deg) apart would balance.
Getting a machinist to balance it would be difficult due to its geometry.
Nothing rattling around inside can exit the casing.
Has nobody ever designed a high RPM rotor?
>I only want the high RPM shaft balanced, anon.

i just remembered working with motors like 20 years ago, and they used ring magnets to measure speed.
that certainly cures the balancing problem.
oddly enough, these were synchronous 3-phase motors, so the speed was varied using a clutch.

soft magnets similar to this.
>ring magnets to measure speed.
[-] one diametric ring magnet
[-] continuous arc-magnets full 360 around shaft.
your post has nulled my previous thinking;
one magnet can be used, or many....
and be balanced.
Thanks anon, this solves my problem.
This is spot on.
Unshielded bearings fear the magnetic o-ring
thanks for the heads up

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.