/diy/ - Do-It-Yourself

>>2032342
OLD THREAD

this thread brought to you by the 7814 SMD 4mm Square Sealed Rotary Switch from Mr. Bournes. with 2ohm series resistance.

>>2037826
comfy

>>2037826
Oh we doing this again? I'm down. Gives a good opportunity to see kinds of parts you've never thought about before.

As for this particular part, I like it. It's interesting to see SMT versions of large or mechanically stressed components. It's also an easy one to turn the THT version into the SMT version just by clipping and bending the leads, makes it look like a PLCC.
Maybe I should make my DIPs SMT the PLCC way, instead of the QFP way. Will save more room, which is important because I'm not going to buy SOIC versions of all my DIPs.

>>2037814
i want to get into gps stuff, mostly for /o/tism
first project would be a simple 0-100 acceleration meter
then a data logger to combine position with various engine parameters
and finally if all goes well some super precise rtk system to accurately measure on track position and analyze different cornering lines

it looks like the rtk stuff is still fairly expensive (200€ per module and you need two)
so i want to start with something simpler and obviously cheaper (~50€)

do you think the ublox neo-m9n is a good start?
any other options with similar or better performance?

for a specific devboard i was looking at the mikroelectronika gnss 7
it is pretty cheap (40€ from mouser shipped with tax) and offers everything i need (uart output and external antenna connector)

I ordered a broken sony RX100 for 50e, it has a "turn off and on again" message, i fixed a way cheaper sony camera made around the same time that had the same message, do you think it is doable to do it again?

>>2037897
>doable to do it again?

absolutely. but it'd be more productive if you typed ''sony RX100 turn off and on again'' into google than seeking betting odds on a pedo website.

>>2037826
based and traditionpilled

>>2037796
>not very smol, not very fast
can be done with a fairly basic lab
a good-enough multimeter, Aneng AN8008 or AN8009 are aight
a good-enough soldering station will help much with the LED cube. get a 936D type soldering pencil clone (the smol meme irons will not be as comfy as they look) and a tip assortment. it's cheap, you can develop technique with this and it will serve you well. upgrade to the pocket iron meme later if you want to
desk lamp. magnifying lens is a big bonus you'll one day find yourself glad to have
a fan, to keep solder fumes out of your nose
solder, the 0.8mm stuff is cheap but prefer 0.6mm or smaller. leaded solder is easier for beginners to work with. also get a flux pen or some paste flux
needle nose pliers, tweezers
a work holder. the little $5 examination stands with two alligator clips and a magnifier arm are ideal for attaching wires to things, not so ideal for building boards. buy two, cut the magnifier off of one
wire cutters, wire strippers. the bladed kind with the diamond notch in them are enough, the automatic kind with the jaws separate from the cutter are really nice
insulated hookup wire. 24awg solid for the breadboard, 24awg stranded for everything else, maybe 20awg or larger depending on the fan you want to control
that should get you started on building at least one of those things into a permanent form

>>2037812
>Make that ~30 now that I've jumped to SMD
based
>You wouldn't happen to have generic logic-series specs, would you?
there's a TI 74 series logic databook on archive.org, dated but useful. I don't have a full complement of 74 series data but I do have a decent lot of the smol logic data to hand. generally, if I look up a datasheet, there's about 50/50 chance I save it
my part number search looks like ls mfrs/*/*4017*
>Might make it less latency or something idk
not especially, processing introduces latency. on the host, that latency can be synchronized with other sources

>just found out I've been using my bench oscilloscope incorrectly for years

>>2037994

according to your GF, that's not the only tool you dunno know how to use.

Woot, my inverter is famous now- and still running. New inductor on order.

>>2037994
Is there a wrong way to use a scope? What have you been doing or not doing?

Someone tell me if i'm losing my mind, i'm doing this simple circuit over and over because its driving me nuts:

5v source, 1 LED, 1 Resistor.
LED is 2v/20mA
If I want 8mA thru it I do (5 - 2)/.008 = 375Ohm for the Resistor.

I put these values in Falstad and I get a current of 9.085mA thru the LED.

I put these values in EveryCircuit and I get a current of 8.25mA thru the LED.

Am I missing something or is this just the implementations of the simulators?

>>2038030
each simulator modelled its default red LED on a different real-world example. the LEDs you get in the real world may be like either or different from both

>>2038030
In reality, an LED doesn't have a fixed voltage drop, but rather a curve. The 2V specified by the datasheet is going to be at a specific forward current. At 0mA, the potential is going to be 0V, obviously. If it's not one of those chinky datasheet stubs with a single table, there should be a current/voltage graph, showing something like an exponential curve. There should also be thermal data, showing how the forward voltage can change with temperature.

If you want to calculate it more accurately by hand, look into describing the diode with the Shockley diode equation. But half the things I ever want to do with that equation just end in an analytically unsolvable equation, so don't get your hopes up. The sims should be accurate enough.

>>2038068
A guess-and-check method is effective here.
Assume the voltage drop, then test it with a real LED (or accurate sim, I guess) and see what the actual current you get is.
Now you know the voltage drop at a current close to the value you want, use that voltage drop to calculate your series resistor, and you will get very close.

>>2038071
Yeah but IRL you just choose whatever standard resistor value is closest. I generally don't use <5% resistors for an LED. If I wanted precision, I'd use some sort of linear CC circuit.

>>2038074
oh sure, but anon on the edge of sanity up there is getting value from this exercise by learning useful things about diodes AND simulators and when (and how much) to trust either one

Anyone got recommendations for a cheap-but-cheerful PCB mill?

>>2038068
>In reality, an LED doesn't have a fixed voltage drop, but rather a curve.

it's actually more opaque than that, because the curve is for a ''typical'' LED. actual LED characteristics (of the exact SAME part number) will vary according to a bell curve.
so, the poor sap is looking for certainty in an uncertain world.

>>2038108
got to measure current and voltage of the led and vary the supply to ensure that power remains constant
it's the only way to be sure

>inb4 thermal power doesn't scale linearly with total power

>>2038074
>>2038082
>t.often uses 1% resistors for LED ballast because I lacked the ability to care about 0.09 vs. 0.08 cents per resistor on my last LCSC order

>>2038112
>being that sure
>for an LED

>>2038138
>being that sure
>for an LED
it was a joke

>>2038024
I've been using single ended probes like they're differential probes, connecting different points to the ground clip of each probe. I'm about to graduate with a 3.9 in EE

>>2038148
good, good, but it didn't have to be
https://www.youtube.com/watch?v=IzUoe-9bKa0

>>2038157
absolutely kawaii

Saw some youtube videos talking about IPC specifications and shit like that.
Is there study material available for me to get up to this quality of soldering and rework? I'm liftin' pads all the time.

Speaking of shitty rework, I think I'm either reaching the limits of my TS-100 or I'm just fucking retarded (maybe the tip's fucked, despite it still taking solder just fine), because soldering coax and shit like that just doesn't fuckin' work.

>>2038036
>>2038068
Thanks for the explanation, somewhere I knew that LED's operated on a gradient but in my autistic trashing I completely forgot.

Ill stop using LED's as a load when im trying to test CC shenanigans.

>>2038184
did you try adding an entire rossmann of flux to the work?
also, you could be reaching the limits of your power supply for your TS100, or the solder, or...
>study material
not really, there's an entire industry around soldering training. you can put together a good technique piecemeal but it takes a lot of watching and (especially) practice. the YouTube user jkgamm041 demonstrates some impeccable soldering skill and you can learn a lot by watching him

>>2038157
>I'm about to graduate with a 3.9 in EE
You know, it's possible to get all the way through an EE degree without knowing how to operate a scope at all. You're doing better than some students are.

>>2038157
I did that once and burned out my probe.
God bless the chinese and their shitty wires

>>2038184
Crank the temperature up 20C. Solder either melts, or it doesn't.

>>2038184
If you’re lifting pads, you’re in too hot or too long. Tip shape makes a big difference to how fast the workpiece gets hot and the kind of pressure you need to apply, I recommend a D24 or D12, depending on the size of the work. Conicals are bad.

Good quality solder and external flux are also important, as is convenient work holding. A TS100 should have sufficient power for most of what you’d ever want to solder, unless you’re running it off 12V. 19V is passable.

>>2037882
>current limit external GPS LNA bias through a resistor
bad design. Even in UBLOX reference designs they use a current limited source.

Ok so you know how the field effect moves the depletion region through a piece of properly doped semiconductor with an electric field? So when it's conducting it's got a straight channel of semiconductor with free charge carriers, and when it's not conducting it's got a solid piece of insulating depleted semiconductor with no mobile charge carriers. What are the optical characteristics of those? In silicon I imagine both would be opaque, but in something like SiC or PbS or whatever I'd be interested to see. Because if you had something reflective when it was conductive, and transparent when depleted, that could actually be a useful device. Like a reflective screen instead of an absorptive one, that can have a backlight.

>>2038100
Occurred to me I should clarify this: I'm talking about a little CNC machine for milling PCBs, not a company that makes PCBs on the cheap.

>>2038540
>What are the optical characteristics of those?
depends on the frequency. Si is transparent at infrared frequencies. besides I don't think the amount of (or lack of) electrons present in the material would change its transparency since doped Si is still 99.9% Si but I could be wrong.
Sounds like you're looking for something like electrochromic glass which is a little different.

>>2038552
>I don't think the amount of (or lack of) electrons present in the material would change its transparency
Well I'm just going by what I know about reflection. Conductivity of a material is required for it to reflect specularly, I think it reflects the EM wave just as microwaves bounce around the inside of a microwave oven. The conductivity means no electric field can persist in it, and the wave interacts elastically. Contrast this to white substances, which let the light penetrate some distance through them, hence the different nature of reflection.

I know of electrochromic glass, it absorbs light instead of reflecting it. I'm thinking something that never absorbs significant light either way would be more useful. You could make an LCD-style panel of it and use it in an SLA 3D printer or projector, without worrying about the heat buildup in what would be the LCD. Or make DLP mirrors with no moving parts.

>Si is transparent at infrared frequencies
That's interesting. A possible way of thinking about it is IR photons lack the energy to interact with the bandgap, though I'm not sure if that even applies. I wonder if red light goes through the semiconductor used for blue LEDs?

>>2037814
Anyone know of a 12V+ rechargeable li-ion pack with fuel gauging I can use to power my portable projects? I'm trying to power a 6V solenoid that pulls around 1.5A, A 5V microcontroller, an encoder and an LCD Screen, my plan was to step down the 12V to 5V with a buck converter for the encoder, lcd and microcontroller, and a separate buck converter to stepdown 12V+ to 6V for the solenoid. (The solenoid may be changed in the future so I thought an adjustable voltage maybe better). I'm trying to make a project to time how long it takes for something to stop with the encoder and I can't find a power source where I'm measuring it. I looked at texas instruments reference design and it's like a fucking atmel datasheet for each one? I tried using boost converters but 6V 1.5A is 9W which most portable boost converters struggle with.

Thanks!

Wanna convert most of the lights on my moped to LED but I don't want the pulsating voltage of the alternator to be visible in them. It's already visible in the stock incandescent bulbs (especially at idle) and I suspect it won't be as simple as a massive filtering capacitor. Any tips?

>>2038830
fuel gauges are an application-specific thing and often demand calibration. besides laptop batteries or similar, don't expect to find one off-the-shelf, and expect to have to hack on it and read big datasheets, maybe even calibrate it yourself
>I'm trying to power a 6V solenoid that pulls around 1.5A
you can PWM the input supply if Vrating/Vsupply isn't too small. idea: give the solenoid a hard pull start at 100% for several ms to get it moved quickly, then fall back to 50% or less duty cycle (at a few or several kilohertz) to hold. the solenoid's inductance will likely take care of all necessary smoothing. don't forget a catch diode. for development, consider a fuse or a safety timer that kills power if the sol is continuously switched on for more than half a second or so

>>2038843
Bridge rectifier + massive cap to build an unregulated supply, then use it to power LED drivers (constant current supplies).

>>2038571
>Or make DLP mirrors with no moving parts.
Maybe someone clever could use the electro-optical properties of https://en.wikipedia.org/wiki/Lithium_niobate

>>2038846
there's already a rectifier from the factory, since some were optioned with a battery and electric starter.

So just a big ol cap and an LED driver?

Oh and would one LED driver work for all? Cuase the indicators and brake lights won't be on constantly and I don't want those to cause the other lights to dim.

>>2038843
>I suspect it won't be as simple as a massive filtering capacitor. Any tips?
supercapacitors are good at that sort of thing, but they're also overvoltage-sensitive so you would need to somehow protect them from the notably hostile electrical environment of the automobile, including spikes and load dump. unfortunately, I haven't seen any chinkshit modules for supercap charging, only for balancing, so engineering would be required. hope you can solder QFN packages
in any case, you would then place a buck LED driver(s) on the supercap to power the lights

>>2038853
>Oh and would one LED driver work for all?
no, LED drivers provide a fixed constant current. if you disconnect one set of LEDs from the driver, the current will be shared among the remainder. you need separate drivers for each separate control

>>2038853
Yeah, a big cap to provide an unregulated DC supply. Mind the voltage rating. There are some electrolytics designed specifically for power supply use, but I would not really worry about these at first.

Each set of LED that you want to drive as a group should be wired in a string, and each string needs its own driver. You can imagine the big cap as a reservoir of energy, fed by the rectifier when the alternator voltage exceeds the cap voltage, and the drivers as valves allowing a controlled flow from the reservoir through the LED strings.

>>2038859
>but they're also overvoltage-sensitive so you would need to somehow protect them from the notably hostile electrical environment of the automobile, including spikes and load dump.
The charge circuit won't exceed 16 volts or so at any point, and it hovers around 12 at idle. Most supercapacitors can handle a lot more than that, no?

>>2038876
>won't exceed 16 volts or so at any point
oh, if you're tapping directly off the battery, you have some pretty good filtering and might not need to worry about the alternator's voltage regulation being relatively slow. look up load dump to see some of the extreme conditions that could happen on the alternator output
>most supercaps
almost all supercaps are rated for 2.7V and assembled in series to provide higher working voltages where required

Dumb question on spending $$$:
I've contemplated this for like a whole month by myself and I can't handle it anymore. Please give me some opinions or thoughts ohmbros. You can alternatively call me a colossal fucking retard too.
I'm about to receive $2000 from my college (I'm a 2.5 year EE student), no strings attached and I'm thinking about setting up a home mini-lab. I live in an apartment (in the states) but it's mostly just me and my /diy/-tier brother so we have a decent amount of free space, no garage or anything, I also don't have to worry about bills.
I can think of a lot of benefits if I go through with this decision, however, I'm not sure if I should pull through with the decision because it's admittedly a lot of money.
I unironically don't have a fulfilling hobby and the only times I did have fun was with my EE lab class. I can see having my own mini-lab equipment at home as a way to be 'closer' with my major as the only times I get hands-on with it is during extremely limited lab classes which are further limited by the corona and just solving circuit problems 24/7 on paper isn't really that interesting or fun.
If I do buy lab equipment, I won't be buying keysight/keithley levels of equipment. Probably total around $1500 max: a decent oscilloscope from siglent, a decent power supply, a decent function generator, and a decent bench multimeter.
It's either I buy something that would likely become a hobby or I buy a bunch of stuff that I very likely won't ever need (I did this in the past).

>>2038880
Problem is it doesn't have a battery, it was just an option. Right now if I were to hook it up with nothing else it'd look like the tail light in this video (my 'ped is mechanically identical), which is what I'm trying to avoid. It doesn't need to be flawless but this is unacceptable. https://www.youtube.com/watch?v=FuWxP3G9WDk&ab_channel=toped73

>>2038904
Not to mention that even if it had a battety, with the mods I'm planning to do to it I'd have to ditch the battery regardless since there'd be no room for it anymore. What if I were to put a bunch of fat (non-super) capacitors where the battery would have gone? That'd probably work pretty decently right?

>>2038907
The higher the load, the higher the required capacitance. So I'd consider using another (single diode) rectifier to isolate the LED circuits from the rest of the circuits, and use a capacitor after that. There are capacitor ripple voltage calculators out there, though all you need is something to stay above the minimum dropout voltage of a constant-current LED driver.

If your LEDs have built-in current regulation like resistors, then you could just use voltage regulators instead. Not sure what automotive LED lights are like, but I see people connecting them to straight 12V so I assume they have some regulation.

>>2037826
they sell rotary hex/BCD switches in a similar tiny form factor. you can use them with a summing amplifier to create a variable voltage with discrete linear thresholds. more retard-proof than a trimpot or dip switch, if cost isn't an issue.

>>2038911
Automotive leds are 4 LEDs in series with a resistor and or a bridge rectifier. Connected to 12v they will see less than 3v each

>>2038902
don't forget components and storage for them. unless you're limiting yourself to repair and analysis, you will also need breadboards to hold the components as assembled, and the appropriate tools and consumables to assemble the components onto the breadboards according to the mode of construction you choose: solderless breadboard, regulated soldering pencil, solder wire and paste flux at the minimum, also solder paste and tweezers and magnifier and regulated hot air hand tool and/or oven if you decide to work with SMT. the hot air tool is also useful at low heat for exploring temperature effects on circuits
sometimes you'll need at least two meters, to watch two quantities simultaneously. bench units tend to be a bit overpriced unless you're actually working with microvolts on the regular. I get the impression EEVblog Dave doesn't sell anything he wouldn't have on his own bench

>>2038907
supercaps are fine with lots of current in or out, in fact, DIY spot welders often utilize them for energy storage. but they are much less tolerant of overcharging, so you have to promise some things. if the alternator output seems tightly and promptly regulated, and desu you would probably need a scope or a spec manual to know for sure, a chinkshit 6S supercap module from alibay would be worth a try. just know that 100ms is a long time for an electronic component under stress

>>2038911
>If your LEDs have built-in current regulation like resistors
that's not regulation. this is regulation https://www.onsemi.com/pub/Collateral/NSI45020-D.PDF

>>2038914
>more retard-proof than a trimpot or dip switch
but less retard-proof than clipping out discrete resistors, if one-time programmability isn't an issue

>>2038100
unironically pic. cheap, cheerful, but effective

>>2038911
The LEDs in question are all Aliexpress stuff so I'm gonna assume they're entirely unregulated. I'll probably invest in a couple of 12v LED drivers then, thanks.

>>2038938
>storage for them
What do anons here recommend? I've got a cardboard box full of little china baggies with all my shit in them.

>>2038938
>Sometimes you'll need at least two meters, to watch two quantities simultaneously. bench units tend to be a bit overpriced unless you're actually working with microvolts on the regular. I get the impression EEVblog Dave doesn't sell anything he wouldn't have on his own bench
Whew, I've been looking at some handheld multimeters (87V) and they cost just as much as a bench one. Any reason to get one over the other, other than the portability/convenience argument?

>>2038994
>I've got a cardboard box
pic related for ICs and through hole components
a binder with plastic sleeves for cut-tape SMDs
a spool holder for reels of SMDs

>>2039001
Get a portable DMM and a decent ESR meter.

>>2039001
and definitely get a quality desoldering station/gun. Worth its weight in gold.

>>2038994
same, with 1L zipper bags to group little baggies
also picrel are extremely useful for SMT passives. buy pre-filled, or buy empty and fill them from LCSC. it works out about the same either way. you probably won't need all that many different small inductors unless you're doing radio so just stick with R and C in the sizes you would be using

>>2039001
>Fluke
oh, no wonder
an EEVblog forum user compiled a fairly comprehensive comparison table of DMMs. https://www.eevblog.com/forum/testgear/multimeter-spreadsheet/ UNI-T is a pretty good value brand at the high end. more than 4.5 digits is, imo, a waste unless you've got other gear to accompany it

Could anyone give any advice on how to design my own 4-bit microcontroller for a calculator? Where do I begin? Any similar *well-documented* and simple chips to draw inspiration from? How many registers do I use? How long to make the instructions (8-bit?)? How many nibbles of RAM is enough for a basic 10-digit calculator? How deep to make the function stack?
Currently trying to write some algorithms in C to get familiar with BCD math algorithms.

Reading about instrumentation amplifiers and it says
"Yet another common application problem occurs when
designers try to drive the reference pin of an in amp with a highimpedance source. Typical values for the impedance of the reference input in many popular in amps are 20 to 125 k. If a
low-impedance source, such as an op amp,...."

How is an opamp a low empedance source? In some applications i see a voltage follower between ground and the ref pin. i dont really understand what this does. Wouldn't low impedance between the ref pin and ground (or whatever reference) just mean a very short wire?

>>2039039
>How is an opamp a low empedance source?
It's an amplifier, that's what they are designed to do. Provide a strong output that can hold whatever voltage at the necessary current to do so.
Low impedance doesn't just mean "very short wire."
Voltage follower buffers a voltage from a weak source, which can't do much useful, to a strong source emitting the same voltage, which can do something useful.

A high impedance source could be a microphone or soil ph sensor or pressure sensor. It provides a voltage signal, but is not strong enough to provide any serious current. Low impedance sources like those need to be amplified to do something useful. The output of the amplifier is considered low impedance.

An analogy might be a car battery vs a stack of watch batteries. The car battery (low impedance) can supply 100s of amps current and the watch battery (high impedance) cannot.

>>2039027
http://www.intel4004.com/4004_original_schematics.htm
My advice is write an emulator for your 4-bit processor, write an assembler, then write the firmware in assembly.
If you can get that far, then you can design the ALU and learn how to read instructions from memory / registers and write back to registers.

>>2039065
Thanks, >>2039044 also recommended 4004.
##electronics at Freenode suggested HP Saturn and TMS1802. I'll check them out.

>>2039027
what technology are you designing in? discrete TTL? discrete 4000 series CMOS? CPLD? FPGA? discrete NMOS?
>Any similar *well-documented* and simple chips to draw inspiration from?
these were used in some early MCU-based calculators http://bitsavers.informatik.uni-stuttgart.de/components/ti/TMS1000/TMS1000pgmRef_1975.pdf
these were used in things like microwave ovens and timers http://datasheets.chipdb.org/National/COP400/NS_COP400_Family.pdf
>How many registers do I use?
maybe none, maybe just one accumulator/working register and a condition code register
>How long to make the instructions (8-bit?)?
8-bit is a good size, there's room for a lot of of instructions with 4-bit immediate operands, but you might have to leave yourself room to do some tricks and you'll probably be bank-switching. also remember to deal with long jumps in your instruction encoding, including any skip instructions you might implement
>How many nibbles of RAM is enough for a basic 10-digit calculator?
far less than you can buy as a packaged component tod- oh, Jameco has those 74S189 fuckers NOS for $4 each. anyway. 64 nybbles would be ample, even extravagant. room enough for the display, a memory, a scratchpad, and to memory-map the keyboard
>How deep to make the function stack?
depends on how you structure your code and how big your program store is. 4 levels might be enough, 8 is more than enough. you'll find out for sure when you write the program

>>2039039
in amps don't have external feedback in usual configurations. they just figure the difference between the two inputs, gain it up by an application-specific but controlled multiple, and offset that by a reference input (picrel at R6). not the same thing as op amps
anyway op amps are a low impedance signal source because their output currents are relatively high and R=E/I

>>2039060
I think i understand somewhat. But i thought that adjusting the ref pin was to compensate for any minute imbalances in the internal resistors of the inamp chip that reduces cmrr. i dont really understand how an arrangement like picture related would help

>>2039073
>what technology are you designing in?
Relays. I've got a 1050 5V SPDT relays, but I can buy more. I hope it doesn't get too hot, I might need to replace them with 24V ones. RAM is VERY expensive, every flip-flop counts.
DIP switches for ROM.
Hopefully I don't have to resort to cheating with chips for memory.

What's a decent beginner project that I can do at home? I'm hoping that it can employ capacitors/conductors and no more than that because that's all I've learned in a class other than resistors...

>>2039140
*and op amps

https://www.youtube.com/watch?v=ySuUZEjARPY
In this video, the claim is that the energy, the whole kaboodle, is contained in the FIELD surrounding the fucking conductor, and that the conductors simply act as waveguides for the energy wave emitted.
This sounds fucking stupid. Someone fucking beat my dumbass brain into shape and tell me how this doesn't spit in the face of the electron and its motion being the primary transfer of fucking elecfuckingtricity.

>>2039080
not 100% but i'm pretty sure it makes it so the internal resistance of the ref pin have no effect on the rest of the resistor network which sets the reference voltage.
Like, so the internal workings, different tolerances, and input signals of the instrumentation amp can't bully around the reference voltage.

>>2039022
>UNI-T
>High end
Lmao no, it’s middle-tier at best. Compare the CAT ratings, and open them up to see what sort of protection circuitry they have. Check their drift specs. If you don’t care about jabbing them into HV sources they’re potentially still fine to get, but I’d want to have something to calibrate them against on occasion. Be it a better meter or a precision source.

>>2039080
Because the input impedance of the inst-amp’s reference input is comparatively low. If your voltage reference is a 10k resistor and a reference diode, then if the internal resistance of the inst-amp is on the order of 10k or less, then some current will flow into or out of it, changing the reference voltage slightly. In an extreme example, a 1k input impedance would mean the inst-amp pulls 10 times more current through the 10k resistor than the reference wants, loading the voltage down a bunch.

>>2039081
Don’t tell me you’re going to try wiring them together on those dinky breakout boards? Or are you that 3DP anon?
Anyhow since relays are more complicated than transistors, there might be ways to make gates or FFs or SRs or whatever with less equivalent transistors. Since you’ll want freewheel diodes in order to preserve contact life, I recommend seeing if introducing other passives to your gates will improve them. Like DRAM. Or maybe magnetic cores.

>>2039140
Learn more. Mess about with a breadboard and simulator.
A crystal radio is about the simplest project out there, though it can be finicky to get it receiving properly.

>>2039148
Oh op-amps too? In that case you can make 80% of all analog audio circuits, in one form or another.

>>2039080
the ref pin is "just" another input to the in amp, but one that isn't multiplied by the amp's set gain. what you're describing is like the offset-null pins of early op amps, back before laser trimming was cheap and easy

>>2039081
you will need a hell of a power supply
just 1 Form C? that sounds rough. some 2 Form C contact units would help with RAM and math and multiplexing
so at bare minimum you'd need 30 nybbles for 10 digits each of result, input accumulator, and one scratchpad (mul/div), and a handful of small scratchpad registers and counter registers which you'd have to get a bit further into the design to quantify
>DIP switches for ROM
better: plug diodes into a matrix of pin sockets. you'll need the diodes on the switches anyway to keep 1-bits from backfeeding through other switches. diodes will be happier at lower current therefore higher voltage is preferred
I would try to make certain tasks as hardware-driven as possible, and reserve the program store for cycling through the big number arrays, normalizing results, etc. maybe the entire process of inputting a stream of digits into a storage register could be handled without CPU intervention. halt-until-event processor instructions were pretty common in mechanical calculators as an implementation technique, in discrete CPUs on mainframes and minicomputers as a cost-saving and virtualization-assisting measure, and in portable devices today as a power-saving measure. the whole project would look less like the sort of general-purpose processor we're used to, but it would be a lot quieter at idle

>>2039140
you need a cookbook. look up "op amp cookbook" and you'll find a Nuts & Volts article series, a book published by Sams, and a TI app note, among many others. then go to the roll page in the OP and get an application to assemble them in

>>2039171
it's so the internals of the inamp don't distort the intended ref voltage

>>2039172
I meant the high end OF UNI-T, ~equivalent to mid-low Fluke

>>2038938
>unironically pic. cheap, cheerful, but effective
How do you keep the copper from bridging?
last time I cut w/ an exacto I needed to do three passes for each cut, at angles.
Sounds tedious, or do you just cut a grid of squares?

Can someone double-check my comprehension of diodes real quick?
I'm trying to figure out why we need so many different parts to a diode/transformer thingy.
Because a practical diode (specifically a halfwave and a fullwave center tap) has too much variation in voltage due to its wave-like nature with occasionally reverse+forward (for DC), a way to reduce that voltage variation is to add in a filter, and the filter is done through a capacitor that 'ups' the minimum voltage from zero to a certain point above zero where it discharges to when a practical diode would've otherwise been zero because of the sin-wave thingy. And then to further simplify it, we add in a voltage regulator which is done through the form of a Zener diode?

>>2039325
Rectifier, yes. Filter, also yes. There’s a relatively simple equation for getting the output ripple voltage of a power supply like this, derived from the I =C*dV/dt and assuming constant current. Simply put, higher frequencies, lower load currents, and higher capacitance means less ripple voltage.

But barely anybody uses zener diodes any more. Most common is a simple fixed 3-terminal regulator. Like an L7805 +5V regulator. 7905 for a negative version. These regulators are series regulators, not shunt regulators, meaning they don’t dump as much current into the ground rail when it isn’t required by the load. They’re also rather well regulated, staying within maybe 10mV of their target voltage, depending on the input ripple. Internally they’ve got a voltage reference (sorta like a zener) that uses a small amount of current, and a circuit to compare the output voltage to the reference voltage, with a fixed factor or two in there for convenience. Then there’s a pass transistor that only lets as much current through as the feedback circuit will let. This is in the form of a negative feedback loop. If you haven’t read about electronic feedback loops yet, definitely do so.

>>2039325
>Can someone double-check my comprehension of diodes real quick?
2/10, transformers and diodes are distinct devices. although poorly drawn, pic related may help you visualize what's going on in the early part of the basic transformer power supply, esp at the diode-capacitor junction
pretty close on the rest, except the regulator stuff which other anon covered. the unfortunate fact about zeners and other shunt-type regulators is that they operate by diverting the excess, like overflow holes in a basin or tub (if you need visual aids, try >>>/b/), so their main use anymore, after protecting sensitive circuit inputs against high voltages like ESD, is in providing a stable reference voltage to an amplifier which controls current to the load. then that whole assembly works a bit like a toilet tank filler's float valve, but with a less abrupt action

>>2039317
well, a metal straightedge is also a great help. I generally just cut out troughs to separate big lands

>>2039325
Regarding rectification: half-wave usually isn't acceptable because it causes the current draw to have a DC component. Transformers don't like DC as it causes core saturation, which in practical terms means you need to use a larger transformer, and a larger transformer is a lot more expensive than using more diodes. Using the centre tap as ground with a single diode on each end doesn't have this problem, but it does mean that you're effectively only using half the transformer at any given time, so you need twice as much copper as if you'd used a bridge rectifier (four diodes).

Regarding filtering: when the AC output from the transformer crosses zero, the rectified output is bound to be zero. No arrangement of diodes can change that. (Note that this isn't true for three-phase AC, but you don't get three-phase out of a domestic wall socket). So you add a capacitor to store charge during the peaks and supply it during the zero points.

For some applications, rectification and filtering are enough. But if you need a stable DC voltage (which most electronic devices do), you also need regulation.

>>2039374
>half-wave usually isn't acceptable because it causes the current draw to have a DC component
>Transformers don't like DC as it causes core saturation
Oh really? I didn't know that. I recall that single-diode rectifiers are used in low-power SMPS designs, which makes sense since they're not going through any magnetics beforehand. I guess larger capacitors are cheaper than 3 more diodes, in some circumstances. I wonder if EMF chokes and such handle DC loads differently?

>>2039357
>Simply put, higher frequencies, lower load currents, and higher capacitance means less ripple voltage.
Got it! I'll note this down asap.
>But barely anybody uses zener diodes any more.
Damn, I need to put this down as well, from how my professor made it sound, it seems like they are used a lot but I guess not.
>This is in the form of a negative feedback loop. If you haven’t read about electronic feedback loops yet, definitely do so.
Yeah, sorry, I didn't get the previous part before this about the negative feedback loop but I'll go and read on it! It's really fun to learn about how these stuff actually work in action. Thank you anon!
>>2039368
>2/10
C-can I get a curve...
>transformers and diodes are distinct devices.
Yeah... I kept on confusing the two because I didn't learn about transformers yet, he started the course assuming we knew about it--though that's not an excuse, I'll have to do better!
>pic related may help you visualize what's going on in the early part of the basic transformer power supply
Yeah it helped! Thank you, it explained the thing way better than I ever verbalize.
>esp at the diode-capacitor junction
Does that just mean what happens when a capacitor is added? Sorry, the word junction fucks me up.
>the unfortunate fact about zeners and other shunt-type regulators is that they operate by diverting the excess, like overflow holes in a basin or tub ....... then that whole assembly works a bit like a toilet tank filler's float valve, but with a less abrupt action
Thank you anon, these analogies really help! And heh... /b/... It was the first board I saw and I never went there ever again.

>>2039374
>as it causes core saturation, which in practical terms means ...... more expensive than using more diodes.
I'll write this down! And then figure out what core saturation means!
>but it does mean that you're effectively only using half .... so you need twice as much copper...
What the hell, this is neat. I don't think my professor would ever talk about the costs and other external factors that could influence decisions like this.
>(Note that this isn't true for three-phase AC, but you don't get three-phase out of a domestic wall socket)
So new stuff to research...
>For some applications, rectification and filtering are enough. But if you need a stable DC voltage (which most electronic devices do), you also need regulation.
Thank you anon, all of these replies are super helpful in just expanding my general overview of the topic. I really appreciate it!
Sometimes 2000 characters isn't enough and I had to break down the reply into two, sorry... But thanks again bros!

For the next thread, can someone remove the extra "png" from the roll list?
>https://github.com/Rocheez/4chan-electronics-challenges/blob/master/list-of-challenges.png.png
the png.png causes an error, but once you take it out it will work
>https://github.com/Rocheez/4chan-electronics-challenges/blob/master/list-of-challenges.png
this one will work

Anyone else finds it weird and kind of wholesome that you'll probably spend a decent amount of your life here? At least I'll be here for a while.

>>2039390
>I recall that single-diode rectifiers are used in low-power SMPS designs
the isolated flyback converter is, roughly, what you get when you replace a boost converter's inductor with a transformer. the primary switch builds the magnetic field by gating power to the primary, the secondary winding(s) and their diode(s) accept the energy from the field's collapse

>>2039403
>curve
for a diode question? absolutely, fren
your latest replies show a better understanding than a few hours ago, I call it a strong 3. don't be disheartened, there are a LOT of subtleties in electromagnetics to know to get that perfect 10. my own understanding of electromagnetics is a solid 5 on a really good day
>he started the course assuming we knew about it
that's a bit of an ASSumption on his part, but I imagine it knocks out the ones who are absolutely lacking drive to research and explore independently. you, on the other hand, sound like a survivor
>Sorry, the word junction fucks me up.
I'll say "node" if that helps. waveform #3 is what you would see if you stuck a scope probe on that node where diode and cap meet. remove the cap and it's waveform #2. note that current only flows in the diode when the ac voltage exceeds that of the cap
>Thank you anon
no problemo. the hydraulic-electric analogy does work for the new student, but it's imperfect and shouldn't be taken too far

>>2039429
welcome to forever, anon

>>2039390
> I recall that single-diode rectifiers are used in low-power SMPS designs, which makes sense since they're not going through any magnetics beforehand.
The current is going through the transformer which steps distribution voltage down to the supply voltage. A fraction of an amp isn't going to matter there, but most jurisdictions place limits on DC current draw. This mainly affects thyristor/triac zero-crossing "dimmer"-type circuits (which are required to switch whole cycles rather than half cycles), as those can be used with relatively high-power loads. Smaller DC currents can be an issue if you're running off an inverter or UPS.

>>2039403
He's just referring to the node, the intersection, the electrical connection point between the diode and the capacitor. And also the load I guess. When you measure the voltage of a circuit, you have to measure it at a certain point. You could measure it before the transformer, after the transformer but before the rectifier, or after the rectifier. You also have to measure it relative to something, because voltage measurements are the electrical potential between two points, but usually that's relatively obvious. Usually you're measuring with reference to whatever 0V or ground point is the most applicable.

>>2039413
Done, don't know how that got in. Other additions are welcome.

>>2039429
I do really like this place. Hope some memetuber eceleb doesn't get us overrun with newfolk.

>>2039430
>isolated flyback converter
I was referring to the mains line rectifier, not the post-switching rectifier. Obviously the magnetics in most SMPS' are running with a net DC current through them, and are designed to handle it.
>boost converter
I thought they were more like a buck? Not relying on inductive kick to get a higher voltage, never letting the magnetic field collapse fully, I mean. A look at some google image graphs is thus-far inconclusive. Some show that triangle-wave style weighted sum of P and S currents, others show more periodic patterns that go to 0.

>>2039440
that's pretty bizarre, most I've seen have a full bridge on the ac line input. it would have to be an exceptionally low-power unit for a half-wave input rectifier to be sane
>flyback
>more like a buck
flybacks and boosts transfer power to the load when the switch is off. bucks and, iirc, forward topologies transfer power to the load when the switch is on. my oldfag memory is fuzzy but istr flyback transformer cores are gapped so that much more energy can be stored in the gap, which allows a smaller core per unit power
continuous vs. discontinuous operation is orthogonal to all that, and I don't really recall the details. I think the main detail is that discontinuous operation is less efficient because of diode drops, but there might have been an exception. still looking for an opportunity to get some design experience on an offline switcher, because I haven't had enough things blow up in my face lately. when I get some time I'll read the Pressman SMPS book again for serious, it's been decades

>>2039140
>What's a decent beginner project that I can do at home?

home-made electric chair.
been meaning to make one myself.

>>2039140
>What's a decent beginner project that I can do at home?

home-made electric chair.
been meaning to make one myself.
charge the kids a buck a try.
https://www.youtube.com/watch?v=G6RFHx_TRmk

>>2037814
i recently got a drill/driver set from work for free because the battery charger wasnt working, i opened it up and checked it out and nothing seemed blown/burnt out, any idea what it might be otherwise? im not a super electronics minded guy but ive done some soldering/wire work in the past, but this one looks clean, so im not sure what the issue might be and dont want to fuck around and blow anything up

>>2039511
>i opened it up

what did you open up? the drill or the charger? what do the stickers say about the battery: NiMh, Nicad, Lipo? take a pic.

>>2039517
i opened up the charger. i tried to get everything pertinent in one image i hope this helps

>>2039523

it's a switch mode power supply, so, even with instruments, it'd be difficult to fix. and impossible without instruments and drawings.
about the only thing you can do is check the fuse visually.

>>2039523
there are two batteries with it that both have power, for now, so i know everything else in the set works. when i plug a battery in to the charger the lights on the charger do not come on. thats the only indication that something is wrong. i havent had them long enough to know if they still charge but just no charger lights come on

>>2039527
is that even worth doing or should i just buy a new charger?

>>2039532
>is that even worth doing or should i just buy a new charger?

yes, buy. but spend 15 minutes looking at the solder joints with a magnifying glass, so you can console yourself with the idea that ''well, i tried my best''.

>>2039523
Looks difficult to troubleshoot.
First, check the simple things. Fuse, rectifier diode continuity, primary and secondary capacitances, check any power resistors too. The sorts of things that can be measured without powering the circuit on, so long as you know what kind of effects the surrounding components have.
Then inspect the board for any delaminated traces or cracked solder joints. Just to be sure, reflow solder joints on the transformer, chokes, battery connection terminals, anything that has a mechanical load on it.

Then consider a simplified diagram of the circuit; mentally split the PCB into primary power, secondary power, and control segments. Should be reasonably obvious given the trace sizes and isolation slots. Also make a mental note of the 0V and GND rails on each side of the isolation slot, these should also be obvious. Start with the primary stage, plug it in and turn it on, and see how far the voltage gets on its way from the line input to the transformer. Find the datasheet for the line converter IC and ensure it's got the right power going to its pins.
Then measure voltages on the secondary power side in a similar manner, going from the transformer to the. Then read off the main IC's name and find a datasheet, to see where its power pins are. See if there's voltage across it, without shorting pads together. Do this for any voltage regulators too.

If that fails to give you some idea where the problem is, then your only bet is to trace out the entire PCB. If you narrow down the issue to one place or another, just trace out the afflicted area. I find it's best to overlay the front and back image in GIMP (properly warped) and draw the electrical connections on a seperate layer. Changing the opacity of one of the two PCB layers as needed. In this case it's borderline whether you'll be able to backtrace the control side, as all that silkscreen around the passives can be a real pain.

>>2039533
>yes, buy
Even if he's not autistic enough to retrace the whole circuit, I think I'd still see about making a replacement. With a laptop charger to avoid line fuckery, and some sort of electronically variable switching controller. Probably just an MCU with a FET driver, though an external reference switcher could make the code simpler. Even use a fully analog circuit in the case that the cell chemistry doesn't warrant anything complicated. Maybe you could even use something like a TP4056 or some other dedicated linear charge controlling chip to give feedback to the switching controller, that could be fun if you could handle balancing with it.
Anyhow, set up your circuit to follow the appropriate charging curve, with balancing as needed. Neat little intermediate-tier project. Beginner-tier if you're one of those madmen who design extremely low-noise power supplies, or multi-stage differential analog transistor circuits.

Course, you'd need to know what you're doing. If the description above is over your head, then you should just buy one.

>>2037814
How do I increase the bandwidth of a transimpedance amplifier? I am using it for a photodiode. Ideally I'd like to use cheap shit only and regular opamps

A-anyone bought the siglent function generator / power supply before?

>>2039672
they are ok

>>2039696
ok time to buy them... $800 total...

>>2039672
function generators are the lamest lab equipment, buy something more useful

>>2039758
like what senpai, I genuinely don't know, so far I have an oscilloscope, a handheld multimeter (one of those expensive fluke ones), and a power supply lined up, or is that all I need? I have the breadboard and the components

>>2039762
a quality hot air gun or a 3d printer (great for making brackets/holders/etc for your projects) might be better investments if you want to spend the money.

>>2039765
>a quality hot air gun
I will definitely be getting this sooner or later to be honest, sounds like a crucial part to the part after breadboarding.
>3d printer (great for making brackets/holders/etc for your projects)
ANON STOP, YOU'RE MAKING ME EVEN MORE INDECISIVE.
Huh... a small 3d printer (~200-300$) is way less than a function generator...

Power supply is busted, I've run the piano off my lab power supply and it powers up and runs self diagnostics fine so I'll be using a generic PSU to replace this.
NC is not connected but what is TTL o. Th piano output pin 5? Piano seems to work fine without it and internet tells me it's transistor logic gates

>>2039767
if you are interested in function and are willing to sacrifice resale value, consider asian brands. UNI-T function generator on sale for $120 on aliexpress rn. or, DDS "function generator" boards for $10. depending on what work you'd be doing, you might not even care about 1% imprecision or 1% THD
be careful when assembling your shopping list, and set aside money for things like probes, random cables, hand tools, etc

>>2039804
it could be lots of things, from a 5V output to a power-on input to a data bus of some sort, photos of both sides of the PSU board could help

>>2039823
I'll see if I can disasemble further tonight.
The way I see it it will probably be a power-on input since there's no other way to keep the power supply from being constantly running. But other people advertising ghetto pianodisc power supplies don't have anything resembling a relay in their builds.

>>2039829
>relay
none needed, even older switcher ICs have enable/shutdown pins

>>2039823
Hey... this one looks much better than the shitty one on amazon by koolertron for 99... It's this one that you're talking about right?

https://www.amazon.com/Koolertron-Generator-Precision-Dual-channel-Arbitrary/dp/B07211YWMK/ref=sr_1_3?crid=1ZFG6I1WOLT2K&dchild=1&keywords=koolertron+signal+generator&qid=1614190802&sprefix=Koolertron+signal+%2Caps%2C146&sr=8-3

My only issue is that it'll take a while to come here as with all aliexpress products, I'll just buy a big bulk then! Thanks anon, going to be spending a total of $2-$2.4k by the end of all of this but I'll literally be set for a long amount of time by the end of it all.

>>2039838
yeah that one. fwiw there are 60MHz versions of that Koolertron on ali under different names for like $76. they look shitty but they generate the signal and that's the important thing
I mean, even with somewhat express shipping you'll likely still come out ahead. beware that customs starts to take interest in incoming shipments whose declared value is $800 or more
don't forget your components bro, esp if you don't have a brick and mortar store locally

>>2039853
It's okay anon, I'll take your recommendation with this UNI-T model. I'll definitely not go over $800, and components of course, I've been looking for an excuse to go buy from aliexpress since I have a bunch of components lined up. I also need to buy a suit because I don't look professional at all and it's kind of awkward admittedly being in a call full of people with suits and I'm there like an in a button-up.

>>2039862
>I have a bunch of components lined up.
don't underestimate the power of actual distributors like LCSC. don't overestimate the price, either. genuine, quality passives should be pretty cheap and available rn, while chips are expensive and dear
>a suit
have you ever tried zoom without pants?
t.Jeffrey Toobin

>>2039767
lmao never used hot air in my life, it's hardly a necessity unless you're desoldering or trying to solder parts too small to hold with tweezers.

>>2039865
>actual distributors like LCSC
I actually didn't know about distributors for just resistors. My go-to is amazon and now currently aliexpress...
>genuine, quality passives should be pretty cheap and available rn, while chips are expensive and dear
Noted!
>have you ever tried zoom without pants?
I never wear pants at home, I'm only ever in my boxers so I have to make sure they don't see me in my pants when I stand up. Kind of awkward when I either stand up or slide on my chair out of camera.

Dumb question, do you know if the UNI-T comes with an american plug? I've been trying to search the comments to no avail, this was the contents of the box for someone who got it shipped to america, aliexpress comments have small pictures.

>>2039870
See the two little plates on the picture, to the top-right of the right-most manual? One has US prongs, the other has EU prongs. They slot interchangeably into the power brick that's sitting between the device and the cable.
Also it's just 5V2A with a USB on the cable so you could power it off literally any decent USB power supply.

Also consider getting a better BNC probe. Get one or two more standard BNC-to-BNC cables too.

>>2039870
>distributors for just resistors
oh, absolutely, those books posted above are nice to get organized for SMT
it's been a while since I've worried about storing collections of THT resistors, last time I used large poly bags for each power rating, tolerance, and decade, inside which lives either cut tape or the cute little bags from a long-ago Digi-Key order, pic. people who use them more frequently might prefer a parts drawer system as depicted earlier
LCSC ships 90% of line items in static bags anyway, could be worth an order just for the reuse value

>>2039879
>LCSC ships 90% of line items in static bags anyway, could be worth an order just for the reuse value
Sounds cool anon! That'll be a distributor I'll list down for when I need higher quality components for more serious projects. Those bags do look more stable than the flimsy resistor bags I get which are already poked through. Thank you!
>>2039875
>One has US prongs, the other has EU prongs. They slot interchangeably into the power brick that's sitting between the device and the cable.
Ahhh, that's neat and smart!
>Also it's just 5V2A with a USB on the cable so you could power it off literally any decent USB power supply.
For some reason, I keep thinking that these industrial/testing equipment takes up as much watts/volts as my computer itself. I should get this habit out of my mind.
>Also consider getting a better BNC probe. Get one or two more standard BNC-to-BNC cables too.
Do you think I should get them from here (the states) or should I get them off aliexpress?

So far, the components I have listed are:
ceramic capacitors
1/4 metal film resistors
bunch of LEDs
I'm pretty sure I'm missing something, I should... be fine I think... I bought a massive Arduino kit (like 63 parts or something) a while back and never got to playing around with it since I didn't like arduino that much...

>>2039890
I just made a big order of components. As well as resistors and capacitors, I got NPNs and PNPs, N-ch MOSFETs, and 3.3V, 5V, and adjustable regulators. Also as well as assorted resistors and capacitors, I bought 100 or so of values I use often, like 1k,10k,100k,1M, and 100n,10µ.

>>2039867
i use mine at least as much as my iron. even if you never once need to remove a soic you put on backwards, reflowing with flux will instantly make all the nasty joints on your hand soldered 0402s look gorgeous. also practically required for doing qfns without kys.

>>2039661
Use a dual monolithic op amp package and have one of the op amps as active feedback.

First op amp does the usual stuff but has the secondary in the feedback loop. The secondary should also have RC filtering on it's own feedback.
Either try it a bit or read some TI application notes about active feedback because instability can creep up on you.

>>2039523
If nothing looks broken visually, use a multimeter to check diodes.
If they're ok, use the beep beep continuity check and work backwards from the output pins which you know deliver the voltage to the battery (just look lol).

Next, find a nice, convenient ground on the circuit so you can test voltages easily.

Power it up and work back the exact same way (from power pins back up the circuit) but checking voltage. Often, you hit a point where a broken component is blocking the voltage. You clearly know it should be power line, but something (transistor?) has full voltage on one side but not the other.

This is usually a quick way which avoids complete reverse engineering, since most faults are no power on a line because muh transistor gave up.
Otherwise, a switch mode has you up against the wall.

>>2039081
Dem solder joints tho

>>2039148
You covered op amps but not transistors?
Look at the MFOS page. The small "lo-fi noise boxes" are fun to make when you start out

>>2039914
Nice nice! Going to have a bunch coming in April!

>>2040001
If... If I told you that I completely forgot about transistors would you disappointed in me

>>2040042
With a few exceptions, analog low-power transistor circuits are less useful than op-amp circuits, so if that’s all you’re learning then there’s nothing too wrong with forgetting them. But for higher power or higher speed analog circuits, learning how to make transistor circuits becomes important. Just like there are a series of simple op-amp circuits (invamp, non-invamp, integrator, differentiator, ideal diode, etc) there are also a bunch of simple transistor circuits (common emitter/collector/base amplifiers, cascade, long-tailed-pair, current mirror, etc.). But because transistors are less ideal than op-amps (both impedance-wise and bias-wise), you often can’t just chain these together willy-nilly. They’re also definitely harder to model, in case you’re doing calculations to figure out the response of your circuit.
Also transistors are used heavily for switching purposes, such as in PWM, class-D amplifiers, power supplies, and logic circuits. There’s a bunch of different types of transistor too, with their upsides and downsides.

doing some desoldering on a 70s stereo receiver PCB. Non-plated through hole, single sided. Normal sized components from the era (to-220 etc), lead solder.
I have a fr301 gun (350C-500C) and i'm wondering if 400C the best temperature for this type of board?

Any opinions for what I should pick I have years to decide but I can only pick 6 (Electronics II is already taken desu).
The only thing to note is that VLSI Design requires Microprocessor which requires Digital Computer Systems... I definitely want to take Electric Power and Elements of Power Systems, the rest is undecided...

Any opinions for what I should pick I have years to decide but I can only pick 6 (I will be taking electronics II so it won't count to the 6, I want to be with my bros here messing with electronics and dying desu).
The only thing to note is that VLSI Design requires Microprocessor... I definitely want to take Electric Power and Elements of Power Systems, the rest is undecided...
I had to delete my previous post because apparently, the curriculum wants me to take Digital Computer Systems which is a prereq for microprocessors! So many options but so few choices...

>>2040125
>i'm wondering if 400C the best temperature for this type of board?
That's way too high. 200C (~400F) is more like it.

>>2040140
thanks, just realized I misread the manual, it suggests 400C for "plated through hole" and 350 for single sided pwb. Not sure if it can go below 350C, will look into it.
The melt point is ~183C, figured 200C would be too low in regards to "dwell time", no?

>>2040125
>>2040140
to contradict the other anon, i leave my 851 shamelessly cranked up to 400C and it works great. any discoloration is just due to flux and will come off with some isopropyl and a coarse brush.

>>2040148
It's easy to lift traces and pads on old boards so keep it low. Cranking the heat to decrease dwell time only makes sense for large parts and thick ground planes. Add more solder to the joint if you have trouble getting it all in one pull.

Do bench multimeters use up a lot of electricity? What about oscilloscopes? Trying to figure out if I can use an outlet extender or if I must plug them directly into the outlet if that makes sense.

>>2040207
Weird question.
Yes you can use a power bar and/or extension cord.

>>2040216
L-listen, my computer uses this outlet extender/power bar or whatever it's called, and when I plug in more stuff, sometimes everything (monitor, computer, some other stuff) turns off for a second before turning on so I started getting self-conscious over power usage...

>>2039930
>also practically required for doing qfns without kys.
this anon knows what's up

>>2040219
devices sometimes pull a lot of inrush current when first connected, but if your wiring from the computer to the service panel is not about to set the building on fire, there should be no issue

>>2040219
What country are you in, just curious because I'm in canada and never have I needed to be self conscious about power usage.

Part 1 (I need to post two pictures):
Long story short:
I'm a retard. I need to make up for my lab class. I missed out on the hands-on experience so I'm going to college early to rush through previous week's lab class.
I converted my circuit diagram to the breadboard arrangement (we use a breadboard in the lab)
Is this how I measure the voltage across a resistor in a series -> parallel circuit for V_r1, V_r2, V_r3 and V_r4?
I also need to make a picture for the current version as well heh

>>2040232
>devices sometimes pull a lot of inrush current when first connected
Yeah but if it's to the extent that his computer or monitor are shutting off, he's got some problem at least. I'd recommend having exclusively computer stuff on one socket strip, with a second one from an adjacent outlet for electronics stuff. Though just getting a heavier-duty extension cable should fix it too. That's assuming the problem is in the cable to the socket strip, not in the wall. If it's in the wall, you're fucked.

>>2040235
Anyone could get this if they plug a relatively high-power desktop into a relatively anaemic socket strip or extension cable. Might be able to get it from oxidation on the prongs too, though I've never seen it before.

>>2040237
Yes that's right for voltage.Though you can swap the leads around without much issue, you'll know to just swap the polarity of the measurement. Could also just make all your measurements with one probe on GND, and subtract them from one another, though not if you're not taking KVL for granted.
As for current measurement, you'll need to seperate the two resistors you're measuring between, and put the ammeter between them. Might be difficult without some kind of DMM leads that plug nicely into a breadboard. Been meaning to get or make me some of those. If you're taking KCL for granted, you only need to measure current 2 or 3 times; in series with R1 and in series with R3 and/or R4. If you're taking ohm's law for granted you can just calculate it from the resistance values and the voltages you measured in the first part.
Though I can understand if you're trying to use these measurements to validate those equations that you might not want to just infer the measurements with those same equations.

>>2040237
Part (2): Current Version
I don't fucking know how to use the current measurement AAAAAAAAAAAAAAAAAAAAAA.
On the manual, it says "the meter must be placed in series with the resistive elements so that the current that passes through the circuit also passes through the meter." What the heck does this mean?! I'm going to try and see if I can draw it.
>>2040246
>Yes that's right for voltage.
Thank you, I swear I was banging my head against the wall.
>not taking KVL ...and ... KCL for granted
What...? Sorry... I don't understand...
We're trying to compare theoretical ohm's law to actual measurements so I think that's what you're saying.
Did I fuck up the drawing? I feel like I did. I know that I_1 = I_2 but I decide to draw them anyway, but then I came to a stop with I_3 and I_4, how do I measure those...

>>2040235
In 'murica, I'm just afraid of what might happen to my devices if they consistently shut off randomly so I try to avoid it by limiting my power usage.

I CANT FUCKING TELL IF MY RESISTOR'S BAND IS:
SUPER DARK GREEN OR GRAY
WHAT THE FUCK

is he /our/guy?
I have been enjoying his videos a lot lately

>>2040251
KVL and KCL are Kirchhoff's voltage and current laws. They're useful, though in the end they're just derived from Ohm's law. KVL is: voltages add in series, all voltages in a circuit add to 0V (+12V from the supply, -5V from one resistor, -5V from another, -2V from the third). KCL is: currents add in parallel, all currents going into a junction add to 0A (100mA in one side, 40mA out one side, 60mA out the other). They're a good way of analysing simple circuits like the one you've drawn.
If they aren't working out nicely, then you might need Millman's theorem, which too is derived from Ohm's law.

>Did I fuck up the drawing
Yes. Should be picrel for measuring R4, same for R3. R2 is correct.

>>2040271
>I have been enjoying his videos a lot lately
Because you're 12. They're ok for understanding simple concepts for those new to the field, but they don't really teach that much beyond that. Compare them to Big Clive's, who examines how something (an LED light bulb for example) works, takes it apart, retraces the circuit diagram, and points out what each component is doing. Often also pointing out ways it could be modified, or other uses for some of the parts.

>>2040279
>Yes. Should be picrel for measuring R4, same for R3. R2 is correct.
Thank you! I went and practiced a little based off your work, hopefully they're all right! Also how the hell do you draw circuits so well? Or maybe it's time to admit that I can't draw for crap and that if my resistor drawing is just a little worst, they would look like inductors!
>KVL and KCL are Kirchhoff's voltage and current laws.
Ah I meant I know them already I just didn't mean the taking them for granted part, I really appreciate the review though! Really helpful for this lab class since it's been months since I last touched them...
>Millman's theorem
What the... now this is something I have to study on. Thank you anon, for going out of your way to help this retard.

>>2040287
Those all look correct.

If you were trying to prove Ohm's law, then you wouldn't just measure one voltage V1 of a voltage divider and assume the other is Vcc-V1, because that assumption relies on Ohm's law. But in 99% of other cases, you can assume ohm's law is valid (take it for granted) and proceed like that. In your case, I'd not take it for granted.
In the same vein, you might not want to just assume I(R1) = I(R2), and measure both at seperate points. Maybe. Be a bit like measuring V(R3) and V(R4) separately, which does sound pretty silly. But if you have no idea how electricity works at all, it's not that stupid.

I just draw them a lot, with my laptop trackpad.

Millman's theorem is not something you're likely to come across unless it's intended by your instructor. It's a quick way of dealing with parallel voltage/current sources with resistances in them.

Last warning, ensure that your DMM is set to volts mode, and with the probes in the volts slots, unless you absolutely know you're doing current measurements. Once you're done with currents, pop both the switch and the plugs back into volts. Accidentally placing 12V across your meter when it's set to current mode means putting 12V across the 0.1Ω or whatever internal resistance of your multimeter. That's a lot of current, will tug your power supply down and blow a fuse in one or both devices. Everyone's done it once or twice, but it wastes time, especially if you don't have a spare fuse handy. And some DMM fuses are expensive.

>>2040290
>you might not want to just assume I(R1) = I(R2), and measure both at seperate points. Maybe. Be a bit like measuring V(R3) and V(R4) separately, which does sound pretty silly.
Got It! And yeah I honestly don't have a clue on electricity works, this will be something I'll study today when I wake up! I'll make sure to measure them separately tomorrow in person too since the lab takes a while to start.
>Once you're done with currents, pop both the switch and the plugs back into volts. Accidentally placing 12V across your meter when it's set to current mode means putting 12V across the 0.1Ω or whatever internal resistance of your multimeter. That's a lot of current, will tug your power supply down and blow a fuse in one or both devices.
Ooooh that sounds nasty... I'll remember to put things back in place, hopefully I won't forget.
Thank you anon, for all of the help. Have a good night!

Speaking of big clive, I have this LED light bulb that I want to repurpose. It uses a rectifier, a somewhat undersized filter cap (20Vp ripple), and a linear current regulator set to 30mA or so that's dropping ~100VDC. I've already decided to desolder the warm white LEDs and replace them with 2700Ks, but I'm wondering about implementing some form of dimming.

The IC is a BP5131H, I think. There's about 0.6V across its programming resistors, which I suspect will remain constant regardless of the resistance, within reason. I have a few options, that being putting a HV MOSFET on the string itself, shorting the LED string, or doing something to the current-sense pin.
I think shorting the LED string will just cause the IC to try and shunt 30mA through itself with the full 340VDC across it, and it's already pretty borderline at 3.5W or whatever for an ESOP8 (the SOIC8 with the pad beneath it) on an aluminium PCB.
Switching the HV string with PWM should work, but I'm not sure how the IC's own startup time will effect the duty-cycle, plus it might alias with the 50Hz ripple and cause some strange effects.
So I'm wondering how easy it would be to alter what's on the programming resistor's pin. Ideally I'd be able to cycle the current from below 5mA up to 25mA or so. Not above the existing current. I think the best method is to replace the existing resistor with a larger one to set the minimum current, and shunt across it with a transistor (with a series resistor for maximum current). I'd send PWM from an MCU through a low-pass filter and into the transistor. With IR remote.

But that's assuming that the current is being shunted through the resistor, as opposed to through the GND pin with the resistor just being used for setting an internal voltage. Might want to do some more voltage measurements to ensure that the majority of the current is going through the resistors. Can't see any other method, since 30mA through that resistor is about as much as that IC can probably take.

>>2040294
Like this here. Looks reasonable, right? I checked the power consumptions and they're all fine. Current goes from 5mA to around 30mA. Actual voltages and values subject to change, the highest voltage BJT naturally in LTSpice was only went up to 150V. Thankfully my control transistor will never see those voltages, I think. Now I just need to figure out the 5V/3.3V supply, capacitive dropper + zener might work, so long as it doesn't interfere with the existing bridge. Else power resistor + zener will have to do. I'll want to minimise quiescent current, probably using an ATTiny13. But the TL1838 takes up to 1.5mA.

Does this look like how this sort of current regulator chip would be designed?

building a router+modem for 4g LTE internet. The m.2 modem is a quectel em12-g. I want to put a heatsink on it for good measure.
Based on the figure, i'd put the sink on the silver square area with "quectel" printed on it, in line with that copper plane visible on the back? Or do I put it on the copper plane directly?
Also, would the sink+pad combo pictured at bottom be a viable selection? Or am I cheaping out?

i give up i can make a op amp square wave generator thats single supply on a single 3.7 lithium ion cell. as far as i can tell its not even enough voltage to make sound on a small 8 ohm speaker.

i see it as at best 0.5 volts which sounds like to little to do fuck all with but the reality is worse. i tested my computers audio jack output and found it to be way less than 1 volt. maybe 0.3 volts . cant illuminate a led. internet says 3 volts but its not. its like 0.3 volts at maybe 1 ma. again i say maybe because this is a collection of others data and my own observations with what it can and cannot do at full volume.the led should cut on at 1 ma and 1 volt but not brightly . no light not even a dull bead of light

i managed to get 3 kv out of a bad wiring of the single supply op amp square wave generator on bread board... maybe thats enough of a accomplishment. no inductor no transformer no current pump but still multiplication of voltage. i forgot to wire up the gnd and i still got a output

>>2040321
The pad on the bottom will probably be more thermally coupled to the components. The silver box is just a shielding box that won't be thermally coupled to anything, unless specifically designed with that in mind. If you boot it up and touch both sides, you should be able to tell pretty quickly. But putting a heat-sink on the back wouldn't be terribly easy if it's sitting flat against its masterboard. If you come up with some sort of mounting solution that leaves the back open for a heat-sink then it's perfectly viable.

As for heat-sink selection, you need to do the thermal calculations: how many watts of power are being produced, what's the maximum allowable difference in temperature between ambient and the components, and hence what's the maximum thermal resistance between the components and ambient. Then you'll need to select a heat-sink and thermal bonding method that get below this value. A small heat-sink like that should be enough, but I'd be wary of those thermal adhesive pads; they're not nearly as good as proper heat-sink-compound, which itself is inferior to an indium pad or well-clamped thermal epoxy.

>>2040326
There are two issues.
First is getting your op-amp to work properly at 3.7V. They often don't work within 0.7V of each rail, sometimes even more. Meaning any voltage into or out of the op-amp should be below this. Read the specific datasheet for details. You should aim to get a low-voltage rail-to-rail input-and-output op-amp for this purpose, one that specifically will run on voltages as low as 3.2V (li-ion at minimum charge). The LM6132 will run on voltages as low as 1.8V and will work for this purpose.
Also considering an op-amp oscillator is actually a schmitt trigger oscillator that doesn't use any negative feedback at all, you might as well just use a comparator instead. Will have a higher slew-rate, don't forget the pullup.

Second is the output impedance.
>not even enough voltage to make sound on a small 8 ohm speaker
It's not voltage, it's current. An 8Ω speaker cannot effectively be driven by any common op-amp. At max 4.2V you'd be driving such a speaker with 260mA, while the aforementioned LM6132 can output no more than 2mA. 8Ω loads are relatively low impedance and need a suitably low-impedance driving circuit. Normally I'd suggest a class-B darlington stage, but that's 1.2V off each side, which is a big deal here. An LM386 won't work either, though something similar like a TDA2822 will. The TDA2822 will work on supplies as low as 1.8V, and can drive 8Ω loads. Can drive a single load differentially too.

>computers audio jack
That will not power an 8Ωspeaker. 32Ω yes, but not much lower. You need to feed that into an amplifier circuit before powering 8Ω speakers.

>3kV
Probably static or atmospheric electricity, problems that occur due to improper grounding, and can have strange or damaging effects on sensitive electronics.

>>2040346
the rail to rail is actually something i considered after looking around but i would have to order more parts. 30 lm358s and after i bought them i found nobody likes it . it was $6 for 30 all other types were $10 for 5 and i thought it was like bjts where its not going to matter as long as i dont try to do 500khz or drive 1a or 10a


but the inverter i got 2 schmit hex inverters that might be a good option though im probably going to use the led blinker circuit. sad but its the only one i know how to use to make a specified frequency range (its ball park) that doesnt require a transformer or inductors and has usable output

>>2040321

the copper pad is purpose made to act as a heat sink, so....
sticky thermal tape is not as good as quality compound, but it's gonna stick a lot better. so, which is more important to you?

>>2040326
>not even enough voltage to make sound on a small 8 ohm speaker

general op-amps arent made for delivering power. they have like 100 ohms output impedance, which is like putting a 100-ohm reistor in series with your 8-ohm speaker, which destroys 93% of your signal.
all 1,000,000 solutions to this problem can be googled.

>>2040299
that doesn't look unreasonable. linearity is an open question

>>2040361
LM358-type opamps are fine at ground, but not near Vcc
imo every experimenter should have RRIO op amps on hand, but read the datasheets first to make sure they really are

Upon further inspection of 4004, TMS1802 and HP Saturn I increasingly suspect that I will have to try to roll my own ISA like >>2039044 suggested, specifically tailored to minimize RAM usage as possible and keep ROM with 256B range.
Like >>2039065 said I'm currently "emulating" the ever-evolving ISA in a text editor in my mind, writing small programs and observing problems that arise (switched from JUMPx4 to relative jump because of too many NOPs for example). As my demo programs grow I will inevitably have to write a real emulator, probably in Python with Tkinter to save time.
I also have to lower my goals, perhaps 8 digit precision instead of 10, and maybe other concessions along the way, perhaps 256B is too little to fit a +-/* 8-digit BCD calculator with floating point, I'll find out along the way. Even if that's the case, other cool programs (generating primes, calculating roots) should be possible.
>>2039172
>Don’t tell me you’re going to try wiring them together on those dinky breakout boards? Or are you that 3DP anon?
Screw terminals everywhere. Inside boxes, between relay modules. Inside bigger boxes, containing smaller boxes with relay modules inside, between smaller boxes.
I'll grow my designs and if it becomes a problem as the projects scale I'll have to switch to PCBs
>Since you’ll want freewheel diodes in order to preserve contact life
How bad do I really need them? How much would the contact life be shortened without them?
>>2039180
>you will need a hell of a power supply
I will have to switch from my 10A 30V lab power supply once the projects go past 100 relays. Each relay theoretically 360mW (datasheet), let's say 500mW, 10A @ 5V is 50W, so the lab power supply can support 100 relays.
Maybe there are higher-current lower-max-voltage lab power supplies available, I'll look for them when the time comes.
Otherwise there are 5V 80A+ meanwell supplies, but I don't think you can short them and they cost a few hundred bucks.

>>2039180
>better: plug diodes into a matrix of pin sockets.
Really would like to avoid semiconductors, I think semiconductors break the spirit of the project.
>at bare minimum you'd need 30 nybbles for 10 digits each of result, input accumulator, and one scratchpad (mul/div), and a handful of small scratchpad registers and counter registers which you'd have to get a bit further into the design to quantify
Perhaps I'll be able to get away with A = A + B, see if that's possible with multiplication and division later on.
Currently trying 32 nibbles in registers, and 9 nibbles on the display, I'll see if that works in the long run.
>you'll need the diodes on the switches anyway to keep 1-bits from backfeeding through other switches
I don't think backfeeding will be a problem.
I'm thinking of a
8-to-256 decoder --> 256 8-bit switches -> single 8-bit bus
solution.
>halt-until-event processor instructions were pretty common in mechanical calculators as an implementation technique
Currently trying to use a WAIT instruction for input, the CPU clock is ignored until user presses a button.
The CPU pinout would be:
>power(2)
>data in (5, 4 data, 1 enable)
>data out (5, 4 data, 1 enable) (for 7SD display shift register)
>ROM address (8, output)
>ROM instruction (8, input)
>clock(1)
looks like 29 now

>>2040453
>I don't think backfeeding will be a problem.
Now that I actually thought about it with any depth, it does seem to be a problem...
I'll have to work it out later. A diode is a single relay, hopefully I can figure out a way to only need 256 of them for 256B of ROM memory.

>I don't think backfeeding will be a problem.
Now that I actually thought about it with any depth, it does seem to be a problem...
I'll have to work it out later. A diode is a single relay, hopefully I can figure out a way to "only" need 256 of them for 256B of ROM.

Is there any book resources to learn electronics from 0?
Like this:

>What is an electricity?
>Electrons Move
>What's energy
>Newton
>How to generate electricity
>Magnets and how to make one
>Technically you can take poop and make a poop magnet
>IDEAL VS HYPOTETICAL

etc... like a highschool book or something, thanks.

Oh and, are potentiometers like resistances? They come in various "sizes"?

Repairing a SONY WM-DD1 Walkman biard. The part I'm pointing to is a horizontal slide switch (two positions, stable, not sure if it's six or eight legs because it seems it has an additional leg on each end other than the six downwards ones but I can't tell if they serve any purpose other than anchoring the part in place)
Is there any direct replacement for it once the switch finger is broken or am I stuck fixing the finger?
Very unlikely to find a replacement donor since it seems these switches weren't in many things.

>>2040474

dont be a faggot perfectionist. just hard-wire it in one position and call it a day.

>>2040474
Is it a 3PST slide switch?
https://www.mouser.com/ProductDetail/Alps-Alpine/SSAG230400/?qs=YMSFtX0bdJC72mBb4dhwYw%3D%3D

>>2040474
Sony part number: 1-571-277-11
https://sony.encompass.com/model/SONWMAF42

>>2039162
There is a big difference between how DC current goes through a conductor and how high frequencies travel thru conductors. Starting somewhere in the Kilohertz range the energy does indeed travel on the outermost part of the conductor (skin effect)... even if the conductor is you (look at Tesla's demos, he understood this, hence how he could pass low-current Kilovolts across his his body harmlessly). By the time you get to the GHz range you are well to the point where the tiniest changes to the geometry of a conductor has huge effects on the behaviour of a circuit.

>>2040480
I probably wouldn't mind if it was the dolby switch but it's the tape type selector
>>2040504
two position
>>2040523
thanks, I was hoping somebody would know how to get that part number
no shipping to my country unfortunately, but maybe after corona
for now I think I'm gonna try to melt some nylon onto the stub and maybe manage to file it to shape, i had moderate success with that before, maybe it will do as a temporary solution

>>2040125
With experience, you will learn what I learned, and what rossman and others preach too... fuck the temperature settings for actual practical soldering / desoldering use, set it to max and leave it. In real world, you aren't going to ge useful results from setting a tool to a temperature to match the melting point of an alloy. you're dealing with large copper traces, component leads, and the ventilation system drawing the heat away, and you're fighting against time, because less heat over greater time is more likely to damage the board substrate / bonding between traces and substrate than maximum heat for a tiny amount of time. In this one case anon, the best options really is to "just fucking send it".

>>2040575
If you're fixing a broken off end of a plastic rod, try plas-T-pair, you can mold and cast parts with it, and it bonds well to most plastics. It's basically finely ground plastic powder and some kind of toluene / xylene solvent. Comes in jars with eyedroppers.

>>2040584
I'll look for something similiar, never saw that in the EU

King of like a Geiger counter but with a mesh membrane

>>2040361
>nobody likes it
They're great for what I do. ±9V supply for audio effects. Couldn't ask for anything more. Considering that the only low voltage source I ever work with is 5V from an SMPS, the noise alone means I likely won't be doing any audio stuff on it, so I never really need a RRIO op-amp. LM6132s shouldn't be that much more expensive, and there are likely cheaper RRIOs, but even the LM358 should work down to 3V so long as you ensure that the outputs and inputs remain within the part's capabilities to handle.

>2 schmit hex inverters
If they'll run on 3.2V, sure. Many 7400s won't do so, but most 4000s should.

The comparator schmitt trigger relaxation oscillator and the logical schmitt trigger oscillator are both common oscillator methods. But a 2-transistor circuit should work fine on low voltages. Might be more reliable if you made it with logic-level MOSFETs (BSS138). A standard NE555 won't work, but some variants (like a CMOS one) might do so.

You'll still need an amplifier regardless of what you do.

>>2040422
>linearity
Considering what I assume to be 8-bits of PWM, I think I'll have sufficient range to squash that 256 values while correcting for linearity. Not sure what kind of remote I'll have just yet. Could send PCM like RS232 across on the IR, or I could send PWM across, or something else entirely.

>>2040447
>Screw terminals everywhere
Please use pin headers and sockets instead, so they slot together onto a motherboard. You don't want to be screwing and unscrewing thousands of connections as you troubleshoot.
>How bad do I really need them
Do a test and find out. Turn one relay on and off with another relay at like 10Hz for a couple of hours. Have two of these setups, one with a diode and one without. Pretty sure it will be night and day.
Also what are you using for a display if no semiconductors?

What is this? It's inside a device I bought broken to see if I couldn't repair it.

>>2040632
>Please use pin headers and sockets instead, so they slot together onto a motherboard.
I was worried the connections would not be reliable with them. Clamping the wires down with a screw seemed more secure. Perhaps you're right though.
>You don't want to be screwing and unscrewing thousands of connections as you troubleshoot.
I was hoping to keep troubleshooting to minimum by keeping it very modular. Making a module (multiplexer, adder, whatever), testing it, and using the tested module in a bigger box to make a bigger module.
>Do a test and find out. Turn one relay on and off with another relay at like 10Hz for a couple of hours. Have two of these setups, one with a diode and one without. Pretty sure it will be night and day.
A couple of hours? The datasheet says 100000 minimum, I expected the actual durability be much higher?
Anyway, I will do the tests.
>Also what are you using for a display if no semiconductors?
I will print my own seven segment displays with E10 light bulbs in them.

How do you guys have so much experience? I feel like even if I had twenty years I wouldn't know enough.

Anyone here re-uses lipos from dead laptops? How do you do it?

>>2040643
It's corrosive poop. Give it a sniff.
>>2040655
wat

>>2040655
Reusing laptop batteries? An interesting concept but sounds like a great way to fuck something up

>>2040662
>>2040657
people usually throw computers away for one reason or the other. I misremembered the name, they are lithium ions. You can get around 6-16 18650 cells from a single battery.

>>2040670
>16
Damn that would be a massive battery
My laptop's is huge with a mere 9

For anyone interested, Marco Reps posted a video about an open-source 8.5 digit voltmeter design from CERN, that I think he's building himself. Looks interesting, I'll watch it over breakfast.

>>2040651
>I was worried the connections would not be reliable with them
If the contacts are spring-loaded, they'll work.
>A couple of hours? The datasheet says 100000 minimum
100k switchings is not even 3 hours at 10Hz. If the lack of a freewheel diode does anything substantial, you'll notice it in 2 hours. Have 5 of each experiment at the same time, just to be sure.
>E10 light bulb
Interesting thought, but high-power and difficult to get the desired shapes. Considered VFDs? Or even nixies?

>>2040654
I lurk and watch youtube videos and study what's interesting to me. Then try putting my knowledge into a project, then test my knowledge by making it or calculating its efficiency or having anons here laugh at it. Took me a while but I'm at the stage that I can post >>2040299 with no major faults.

>>2040655
Yeah often even if the battery is dead, you'll just have one or two dead-cells. The remainder will likely need to be trickle-charged carefully though, and likely won't have nearly as good performance as they did when new. IIRC people do the same with EV batteries, though usually to refurbish and put back in an EV. Never done anything of the sort myself though.
The difference between lithium-ions and lithium-polymers is subtle, but they are distinct.

>>2040690
I just want to use them as a rechargable alternative to my projects. Is using two fully charged ones in series safe(or do I need a battery manager IC)? I am not making packs or anything just powering sub 5W shit and blinkies. The plan is to charge them separately with one of those pajeet lm317 constant current circuits.

>>2040690
>Interesting thought, but high-power
I'll try and see if I run into problems and if I can solve them and how. Worst case scenario replace them with E10 LEDs in like 5 minutes. I think somehow wouldn't feel too guilty about it because they're not actually influencing any relay logic, they're just emitting light, although I would prefer to keep it without LEDs, if I can help it.
>and difficult to get the desired shapes.
Not sure what you mean about shapes. I only need seven segment displays, and I can print whatever shape I want, pic related. You must buy really opaque PLA though with enough pigment and print thicker walls otherwise light will bleed through walls.
Heat might be another problem, might soften PLA, fans may be a solution.
>Considered VFDs? Or even nixies?
Wanted to keep everything as primitive as possible.
Relays, light bulbs, wires, batteries were invented before 1850, VFDs and nixies came over a century later.

178 lm/w genuine OSRAM LEDs arrived.
Time to make Dubai at home..
Thing is, I didn't expect them to be this small... How do I even solder them? Shit with solder on PCB, put sand in cooking pan and reflow (using rosin, which doesn't evaporate this fast)?

>>2040710
>two fully charged ones in series
I wouldn't trust them to remain balanced for long, if that's what you're asking. But if you set the voltage cutoff to something higher than usual, you might be fine with no BMS on board, and just balanced/single-cell charging externally. If you don't have any method for low-voltage cutout, then a cheap BMS is likely a good option anyhow.
>pajeet lm317 constant current circuits
FYI, I think a TP4056 costs less than an LM317.

>>2040714
If you're 3D printing an enclosure then shape isn't an issue. But heat is. Consider bending some thin sheet metal into those shapes instead, possibly using a 3D-printed bending die.
>pre-1850
They had LEDs back then, albeit in the form of cat's whisker diodes that emit a tiny pinprick of light if looked at correctly.

>>2040754
Personally I'll just crunch the old LEDs off and hand-solder them individually with tweezers. But mine are 3x4mm, so more manageable. Haven't yet thought about the aluminium-back of the PCB, but I also haven't yet bought the replacements. I'd buy a ceramic hotplate if I had something to hold it with.
>178 lm/w
That's pretty good.

Based

>>2040770
what in hell am i looking at?
is he distilling turpentine or something?

>>2040770
Nice clock, Ahmed

>>2040447
you don't need a lab supply. an ATX power supply would be perfect if you were using 12V relays
>you can short them
oh, you can, you just might not like the consequences

>>2040472
close, they're actually resistance *values*, as distinct from physical sizes or voltage ratings
can't help with the book tho

>>2040671
not that exciting if they used 16500s or something smol. there is also an energy limit for batteries carried onto airplanes, which is an important laptop use case

>>2040710
>lm317
you can't charge them completely with a simple constant current without damaging them or worse. use a TP4056 or other application-designed charger instead, and then use a switcher on the output to get the voltages you're interested in for that project. don't connect Li+ batteries in series unless each one has under/overvoltage protection (can be provided by a single BMS board for all, or individual protectors per cell)

>>2040714
>opaque PLA
print cavities in simple shapes, line with foil tape, let the bezel do the segment shaping for you
you could also print clear PETG light pipes for the segments, and pretend it's glass if anyone asks
>heat
neon bulbs don't generate much heat
>program
I would strongly suggest looking into other modes of program control that won't be as power- or component-intensive. I'm only half-joking when I suggest using punched paper tape as a microcode ROM. since in that case you can't jump as easily, you should probably make every instruction conditional like ARM and make the decoder VLIW-thin
instead of breakout boards, try picrel, only $2 per set. note these are also 2 Form C
there are ladder logic simulators. mostly marketed for industrial control, which I'd suggest using to plan and verify your design before construction

>>2040770
that's one way to clean and test your fuel injectors

>>2040777
My friend is custom building an LS-swapped car, these are high flow rate fuel injectors that he needs to match and pair

So I made a circuit with adjustable duty cycle that drives a set of Mosfet drivers, driving the solenoids. The microcontroller runs the test for 15 seconds, and then shuts the test off. Diagnostics are displayed on a little OLED display

>>2040769
They are like 1,4 mm x 4 mm
So idk, I already broke 3 out of 500

>>2040690
>try putting my knowledge into a project, then test my knowledge by making it or calculating its efficiency or having anons here laugh at it.
That sounds like the motto here in the OP, based anon, I'll do just that.

>>2040802
that's pretty big, as SMT components go. a 2mm bevel tip and a hot plate on low would be your best friends rn

>>2040831
Thing is, I don't have proper tip. I have chisel 5 mm. And here you want hakko type K tip, so shit is massive (because copper traces would be giant) but also thin

>>2040796
>custom building an LS-swapped car
>high flow rate fuel injectors that he needs to match and pair
Wow that's some real high-end stuff there. Have the breadboards been reliable? At that level I'd want to be using some sort of soldered prototyping method. Either perfboard or manhattan.

Hello, newbie here.

I mean to fix my fender blues junior amp and in doing so learn general tube amp maintenance.

It turns on but only churns out a very fuzzy and quiet signal.
My first thought was that it was the power tubes as that amp is notorious for burning through em. New tubes and no change.
The tube sockets are mounted directly to the pcb, so alot of stress on those solder connections so I thought it might be shorting out there somewhere.
Proceeded to do the ol' tap test see could I identify any loose solder connections or microphonic components.
The tap test didn't reveal anything obvious unfortunatley.

Apart from checking continuity all over the board to check for shorts (after having discharged the amp ofcourse), I'm not sure what else I should do to troubleshoot as, at this point my knowledge on amp maintenance runs short. I've only been able to hypothesise thinking maybe it could be an issue caused by the cheap input jack or output jack that connects the amp to the speaker. Maybe it's the output transformer? If so I'm not entirely sure how to troubleshoot for that. I know the ribbon cables connecting the main pcb and the tube socket pcb are also notorious for causing problems but checked the continuity on those and they seemed fine. I also checked the preamp tubes and they seemed fine aswell. The speaker also works fine as I've been using it as a 1x12 cab for the last while. Furthermore the amp was serviced under a year prior to it failing with, notably the filtercaps having been replaced amongst a few other caps I believe. I'm quite eager to sort this out on my own without having to bring in to a tech so please abide from telling me to do so, even though I'm sure it's quite clear I don't really know what I'm doing, gotta start somewhere and all that. Any pointers and sources that could help on this and more importantly keep me from getting zapped would be much appreciated. Apologies for the long post. Cheers.

>>2040846
well, by "high flow rate" I mean he took some injectors from an LQ4 chevy engine and then ground off the restrictors on the tips. That's why we need to match them

the breadboards have been an absolute PITA to work with but they've been working fine and the circuit has been working well with only a few hangups. It's mainly the mosfet drivers that have been giving me shit, but a few pull-down resistors seem to have sorted that out.

Is it possible to kill yourself with just a arduino starter kit?

>>2040916
Don't eat the tarduino.

>>2040971
Not that anon, but genuinely curious, what's wrong with Arduino? I only ever spent 40 minutes before giving up on it because it was pretty boring so I don't have much experience to form an opinion on it.

More cells in LiPo batteries just gives a higher voltage, right? There's no corresponding increase in power density, i.e. a battery with 4 cells will be (roughly) equivalent in capacity to one with one cell of the same physical size?

>>2040655
Yes, but usually it's not possible. By the time a laptop battery is "used", it's usually trashed.
Newer batteries are flat polymer cells, older ones have 18650s. In both cases the tabs are usually spot welded. It's possible to cut them, carefully, and even solder to what's left of the tabs.
This one is from a Chromebook battery, I thought it might be toast, measured 1.9v when I found it, but it holds its full rated 3AH capacity.

Slashing up laptop batteries is hardly the best way to get 18650s, but it might be a decent way to get flat cells.

If you want to use the entire battery, you can implement the Smart Battery SMBus protocol using an arduino or something.

>>2040975
Power density and energy density should remain equal between batteries of equal volume and chemistry and assembly method. But rated capacity/charge (measured in Ah) and output current will change inversely proportional to the change in voltage. Such that current * voltage = power = invariant, and that capacity * voltage = energy = invariant.

As for assembly methods, some cells are made with higher discharge currents, higher lifetimes, or higher capacities in mind. High charge currents might also factor in. There's a compromise between these three elements. Lipos typically lean further towards the high discharge current end of the spectrum than lithium ion cells like 18650s. So just because you have two lithium ion cells or batteries, doesn't mean that it's an apples-to-apples comparison.
Then there's slight alterations on the standard chemistry, like LiFePO4s and such, which have their own balance of compromises. FYI, lifepos lean towards lots of charge-discharge cycles, lower current, somewhat lower capacity.

>>2040974
>what's wrong with Arduino?

Little, really. But there's a perception that it doesn't teach "real" electronics because you can so easily find snippets of code and libraries written by others, rather than having to figure it out for yourself. Plus the whole ignorant "throw an entire Arduino dev board into a finished project and have it do nothing other than blink an LED" thing.

To an extent, this is valid, and there are those who just hack together bloated, barely-functional implementations of basic features and think they're "1337 h4xx0r", but they're few and far between. And, obviously, having lot of low-level code abstracted into function calls means you won't really have a good idea of what the microcontroller is doing on a cycle-by-cycle basis. While, obviously, that is not necessary or even desirable for a lot of things, it helps greatly to at least have that fundamental understanding there in order to aid in troubleshooting, optimization, and general problem-solving.

But a lot of the "hate" is mostly just looking for meme cred. There's nothing inherently wrong with using a dev board or copypasting code into the Arduino IDE for a quick-and-dirty solution, nor are there so many crowing about how great they are for doing so that it's unbearable.

>>2040836
Ok, found thermally balanced iron with cone tip (bent) and it works nicely.

>>2041067
And yes, rosin piss is good for soldering, and plumbing flux is a good tip cleaner

>>2040294
>>2040299
Did more measurements, looks like the LEDs are 18V ones and there's only 20V across the controller. But the resistors themselves are 300Ω in parallel with 68Ω, (55Ω total), with 0.5V across them, suggesting that the current is only 10mA. Since the net DC voltage across the whole linear array is 308V (with 10Vrms of ripple), this implies a total power of about 3W. But this flies in the face of it being a 9W bulb on the package, that draws 8.6W when turned on as per a wall meter. Maybe it has an internal current mirror or something, to multiply the measured current?

>>2040974
From my experience, I was using AVRs since before arduinos came along, so a beginner-friendly platform didn't catch my attention. I already had a good dev platform (with a debugger, too, which I consider indispensable.)
It's a meme in a very real sense, an overwhelming marketing success that spread fast and wide. Atmel couldn't have dreamed of that kind of marketing success, because atmel would have just marketed it towards embedded developers instead of dumbing it down as an educational creative tool for the masses.
It's this generation's BASIC Stamp.

>>2040858
I won't tell you to leave it to a professional, but man, you should really find a mentor irl if you can. you will be measuring lethal voltages
generally you'll need a service schematic of the amp, which will have normal voltages annotated on each node. make a photocopy of it on paper and measure voltages at those points wrt the 0V bus, then write them on the copy. a component that handles ac poorly will usually also handle dc poorly and the voltages around it will deviate from the norms

>>2040916
do not put in mouth or rectum

>>2041075
18V at what If is the spec?
internal current mirror is plausible. any way to measure LED current directly, at the string? maybe pop out one of the LEDs or lift a diode pin?

>>2041106
same. also a lot of the code that's out there for duino is kinda shite, drivers too, check out the tinyMatrix problem itt. the Apple II boasted 10000 programs in their marketing material. most of them were written in BASIC and some of those had syntax errors

>>2041141
not this thread, but on /mcg/

>>2041141
No clue what spec, they’re 2835 LEDs and 18V (9 series chips) is a common arrangement. The actual measurement was a little lower. They’re warm white, so I can probably dig up a similar enough data sheet for them on digikey or wherever. Size-size they should be able to handle 0.5W on an al-pcb, if that helps.
Only way I could measure series current proper would be to pop out an LED as you said, and short it with a resistor. Might also be feasible to solder alongside an existing LED, but the exposed pads are pretty tiny. I’ll buy a replacement today anyhow. My only shot at getting an LED off non-destructively is to use two soldering irons simultaneously, one on each side. It’s got a monolithic bridge rectifier, so that isn’t going to cut it. But considering I’m measuring the power at the wall and the voltage across them, I can’t think of any alternative. As far as my plans, it using a current mirror or not doesn’t really effect them. I’ll just use the existing resistance as a maximum current, and something 10 or 50 times higher as a minimum. But since I’ll be desoldering the LEDs one way or another, I’ll be sure to test it.

I think I’ll buy another one, to implement my PWM dimming method on before my 2700K meme LEDs arrive.

>>2040472
you want electrons, study chemistry. you don't need that shit for basic circuit design

>>2040472
start here
https://archive.org/details/treatiseonelect02maxwrich/
https://en.wikipedia.org/wiki/A_Treatise_on_Electricity_and_Magnetism

>>2040981
>By the time a laptop battery is "used", it's usually trashed.
people throw computers away for all sorts of reasons and not all cells go bad, it is usually only a couple.

Nobody makes piezoelectric delay lines anymore. Can I get the same results by connecting a large number of oscillator crystals in series or am I being retarded?

>>2041251
This is true, but often the one or two bad cells will drag the voltages of the other cells down too. This can damage the cells and cause them to have less charge-discharge cycles less, and less apparent capacity, and more likely to fail. So I wouldn't use them for anything close to a critical application, but hand-matched and properly tested, they're probably fine for messing about.

>>2041256
You're being retarded. Crystals can act like filters, and filters can act like delay elements, but they'll be delay elements with a nonlinear frequency response and frequency-dependant phase shift. The main reason for making delay-lines limited by distance and the speed of something is so that it doesn't distort your signal like RC and LC networks do.

>>2041258
>making delay-lines limited by distance and the speed of something

Isn't that exactly what's happening inside a crystal though? Electricity causes the quarts to contract which travels through it as a mechanical wave. When you put this in a feedback loop it produces an oscillation with a period of however long of a delay this causes. Shouldn't I be able to use this delay by itself instead of as part of an oscillator.

>>2041268
Yeah but each crystal has a resonance frequency. Unlike a piezo delay line, the signal goes both forwards and backwards by the same amount. The whole thing will ring even without a feedback loop, just a step impulse is sufficient for a good few oscillations, possibly hundreds considering their high Q factor.

>>2041256
Can't you just buy some used delay lines?
They're not rare, even I have a few lying around.

>>2041256
What frequency and delay you looking for?

>>2037814
so i got all these stupid lm358s because they were cheap.they are dual op amps i figured i would get 2 times as much use from them but they arent good at much it would seem. i could use some latches to store single bytes. i had a vision of a latch circuit that would use them like they were inverters in series. you get 2 per unit soo loop the output from the second to the first. then control the sr latch with the inverting pins of both. 1 to set. 1 to reset. in my mind its that simple but in reality it will be more annoying because the damn things bleed. so tying the non inverting to ground wont be enough. i dont want to fuss with my bread board right now . pretty sure i will just kill another one of these damn things if i just slap it together on bread board

does the following make sense or is falstad being a faggot? initially i was using a p channel where the pnp is to reset the byte but then decided to use both inverting pins. but oddly if i do that it seems to think i need a pnp there i dont understand why other than shaving off some current but resistors dont work in its place. the n channel is just there to allow it to be read

>>2040796
So you are only testing the injectors to see how they perform?
What Mosfets/gate drivers are you using? What is your power supply (both for the digital parts and the injectors themselves)?
So long as you don't exceed the current handling of the breadboard itself, its perfectly fine to use. I would recommend cleaning up the wiring from your circuit to everything external to it. Maybe slap together a board that has screw terminals all together and that board connects into the breadboard. Fuel injectors are current hungry inductive loads and can generate a ton of noise on the load wires.

>>2040974
>what's wrong with Arduino?
For getting into electronics and hobby use, there isn't much wrong with it. Its a great starting point and it teaches coding (software) with pre-made hardware that is known to work.
The issues that typically arise is from so many hobbyists being able to blink an LED or spin a motor, running around acting like they are experts. Granted, the percentage of those in the Arduino community who do that are small, but they vastly outnumber anyone else.
Its also not a good platform to learn hardware. You have dev boards that include most everything you need plus 'shields' to just snap on and everything is pre-configured. Even the libraries for the shields dumb things down so you only need one line of code to execute initialization processes. Not that everyone needs to understand how you initialize a device, but this tends to hurt those who want to use chips and devices that don't have a pre-coded library because the ecosystem is designed to hold your hands and make things easy. Again, that isn't a bad thing, but it forces you to stick with pre-made stuff.
It wouldn't aggravate me so much if the community didn't get so crappy. Its a wonderful teaching tool and it does let those interested dip their toes in the microcontroller world without being too overwhelming. It has created a large group of people who think they know electronics without realizing how much they actually don't know.

>>2041194
>But considering I’m measuring the power at the wall and the voltage across them
ohhhh, the power factor on rectifier power supplies is actually pretty shite. I'm not really sure what you are measuring in that case

>>2041278
358-types are pretty undistinguished opamps
>use them like they were inverters in series
1. opamps are not comparators
2. opamps are not logic gates
3. y tho

>>2041278
358s are perfectly fine. you are just stupid.

>>2041251
>people throw computers away for all sorts of reasons
Correct, you have much better chance of finding a good battery if you pull it from a complete computer rather than just finding a battery.
>and not all cells go bad, it is usually only a couple.
Also correct, but even cells that aren't trashed lose their capacity over the years. If one or two cells are totally dead, the rest are probably worn out.

My power supply is coming tomorrow, is there any "DO NOT DO THIS" advice you guys can give? And christ, 22 pounds, some people's power supply looks like it only weighs 4-5 pounds (lifting it with one hand) but this one is 22, hopefully, the deliveryman isn't a jerk.

>>2041400
It's a linear LED driver. Surely it isn't 30% PF right? And they'd mark it as 3W not 9W if that was the case, right? Pretty sure my wall meter measures actual power too.

>>2041455

if you have a scope, set if to 10V, switch it off then on. is there a spike (like 15V) when you turned it on? repeat but with a load attached.

>>2041455
do not drop it on your feet

>>2041400
>not a comparitor

https://www.youtube.com/watch?v=5I2ZRxHbLoo

https://www.youtube.com/watch?v=k9zQjEaKtfk

https://www.youtube.com/watch?v=icFo5Zeydqg

https://www.youtube.com/watch?v=y0Q0ERSP24A

i mean you say that but on paper the damn thing is supposed to compare between 2 voltages right

transistors arent logic gates either but you can make them from transistors. logically i could make a latching circuit i can shut off and have effectively a sr latch. anyways with some real world testing i found that the lm358 will not act as expected and MUST have noninverting biased at all times so that increases part count and nulls any reason to use it for this purpose. you do not get to tie its noninverting to ground it will not do as the circuits found anywhere tell you 9/10 times.

as for people not hating it. well you have to look around. i think many made the same mistake in believing a non rail to rail would be fine and you get so many for $6

>>2041460
>if you have a scope, set if to 10V, switch it off then on. is there a spike (like 15V) when you turned it on? repeat but with a load attached.
My scope is coming Wednesday, fuck. What do you mean by putting a load on though?
>do not drop it on your feet
I'll drop it on my head instead

>>2041470
>What do you mean by putting a load on though?

>>2041462
> i mean you say that but on paper the damn thing is supposed to compare between 2 voltages right
Op-amps are designed around the use of negative feedback so that the two voltages are almost identical (i.e. close enough so that the output voltage stays between the rails). If you use them as comparators (driving the output into saturation), they're much slower than actual comparators.

>>2041470
>What do you mean by putting a load on though?

put a 12V lamp, or something like it, across the terminals, to suck a substantial bit of current. anything that sucks current acts as a load on the electronics.

Cheers, I'm re-posting my question from the Generator General as it's somewhat well.. general.

In short; we have a diesel generator used to top-up our off-grid solar battery banks. In the interest of engine longevity we run it at much lower RPM than you'd typically want to (3000RPM for 50Hz output, we run it at ~2400RPM for roughly 40Hz).

To achieve this lower RPM while maintaining output voltage we've swapped the exciter capacitor from the stock 40uF up to 140uF.

This worked perfectly fine for 60 hours of ownership but a few hours ago, under virtually no load (1kW draw at best) the stator blew out its magic smoke.

Is this due to our tinkering and swapping of capacitors or could it be written up to the natural decay of insulation as the generator is several years old (I'd estimate a 2004-2007 vintage)

>>2041462
>anyways with some real world testing i found that the lm358 will not act as expected and MUST have noninverting biased at all times so that increases part count and nulls any reason to use it for this purpose.
Yes anon, the 30 year old circuit (still in use) is wrong and you are right. Please blind us with the light of truth. You sound ultra pajeet

>>2041487
>To achieve this lower RPM while maintaining output voltage we've swapped the exciter capacitor from the stock 40uF up to 140uF.
why do you think this would lower RPMs and maintain the output voltage?

>>2041492
That's the apparent result.

On the stock 40uF capacitor, if you drop the throttle down to ~2700RPM you'd get 160V output at best, by stepping up to 80uF the output voltage is back to 220V
Dipping even lower down with the throttle required a greater step-up in capacitance.

>>2041283
the mosfets are IRFZ46N, the drivers are TC4420.
We were worried about the breadboard but I made sure to use a shitload of extra/redundant power and ground connections to all the high current points to distribute it better. It seems to be working like a champ.

The entire thing is powered from a 6 amp adjustable bench-top power supply running at 14 volts, you can see it in >>2040770. The digital board piggy backs off the mosfet board; I have a small 5 volt regulator that the 14 volts feed into that powers that board (not super efficient but I usually use it for 9 volt batteries and 18650s anyway). Both boards have some really chunky caps to stabilize the voltage, since the initial voltage drop from the injectors turning on would completely nuke the board before I put the caps on.

I'm just really happy this thing worked. It's definitely my most complex thing so far and that it works almost as I had planned is a real boost. I know I shouldn't get too happy since I know there's almost certainly some braindead boneheaded shit I've done from naivety

>>2041497
Were you monitoring the voltage and current output when the thing happened? What burned, the rotor or the stator?

>>2041462
>you do not get to tie its noninverting to ground it will not do as the circuits found anywhere tell you 9/10 times
Sure you're not working with split-supply circuit diagrams?

>>2041472
wtf

>>2041487
It's possible that the lower frequency AC caused the magnetic core to saturate somewhat, which would make the windings heat up more. But it's a long shot.

>>2041502
And this should't need to be said but generators are usually synchronous machines with two poles. if you want your whatevers to be in 50hz or 60hz they need to run at 3000 and 3600 RPM respectively

>>2041502
Stator is visibly blown, though I haven't torn it up entirely to get a good look at the rotor yet.

The output voltage was in the 220V range as it always is & the power draw reported by a wall socket meter was around the 1 to 1.1kW mark (generally if the battery banks are emptier it'd be from 1.5kW to 1.7kW).

The only appliance drawing power from the generator at the time was the battery charger.

>>2041505
As the only appliance actually ran by this generator is the mentioned battery charger, running at 40Hz doesn't seem to present any issue.

>>2041503
>It's possible that the lower frequency AC caused the magnetic core to saturate somewhat, which would make the windings heat up more. But it's a long shot.

That's mainly what I'm wondering & is the only thing my initial web searching points to.

>>2041508
just run them at the speed they where designed to run. I do not think the capacitor matters that much.

>>2041511
>just run them at the speed they where designed to run. I do not think the capacitor matters that much.
Running at lower RPM is there to (hopefully) positively impact engine longevity. For the diesel engine in question, 3000RPM is WOT. The specific engine is much more often utilized in agricultural applications (roto-tillers and whatnot).

>>2041485
>anything that sucks current acts as a load on the electronics.
Holy fuck so that's what load meant. I'm in my electronics class and my professor talks about putting a load on and I'm like, "huh, what load?"
Thank you anon.

>>2041503
no im sure i searched for single supply and found very little. its more disappointing really that i have to use a p channel to use a op amp. i have 2 and they are both surface mount...

the following should work as a sr latch single supply and only use a potentiometer 2 fixed resistors a op amp and a n channel and p channel. i can probably cut the n channel and just tie it directly

it should self feed when given a pulse and then the p channel cuts the feed and the fixed resistor to negative side of the battery drains the current that was trapped

>>2041514
well you just managed to make your engine longevity drop to 60 hours. Congrats

>>2041524
I'm not particularly happy about that either. I'd like to know the definitive cause for that to have happened.
I'm not trying to be a smartass or go against your advice, I'd just like to get a better understanding of excitation mechanisms.

Before doing modifications we did consult a person who professionally deals with electromotors (specifically performs re-windings). Though I suppose he could've just been securing us as a re-winding customer.

>>2041426
>but even cells that aren't trashed lose their capacity over the years. If one or two cells are totally dead, the rest are probably worn out.
they won't be new, but a cell that goes open or that is otherwise flagged by a smart battery controller as faulty and unfit to charge isn't going to ruin the entire pack. in the usual 3S2P configuration of an older laptop battery, a single failing cell will damage its complement in the pair. the other four should still have some cycles left in them, obviously not new condition or even the same condition as others in the pack

>>2041458
>Surely it isn't 30% PF right
that's exactly why I'm not sure what you're measuring. the number doesn't seem quite plausible. 0.5 is possibly realistic

>>2041455
do not put in mouth or rectum

>>2041462
>presumes to lecture me on the identity between the two IC types
>doesn't understand the practical difference between split supplies and single supplies in opamp circuits
bruh I was using resistors for pull-ups while mommy was still changing yours. you might want to lurk more if you can't get a simple hysteresis circuit to work
>>2041518
>its more disappointing really that i have to use a p channel to use a op amp.
I didn't, what's your problem?

>>2041518
This should definitely work. So long as you respect the op-amp's input and output ranges. Simple schmitt trigger. Replace the two transistors with pushbuttons, if you feel so inclined.

>>2041544
What is this and how do I make circuits like it

>>2041544
>do not put in mouth or rectum
I-I'm not that dumb, maybe.

>>2041548
It has positive feedback. Look up "hysteresis circuits". It is basically the op-amp version of an SR latch from digital logic.

>>2041548
circuitjs, link is in OP. it's not as accurate or precise as a SPICE but is a bit easier to use as a virtual breadboard

>>2041462
dumb pajeet, you can make SR latches with 555s

>>2041058
My problem isn’t in the software side, let people copy-paste all they want. If their code breaks or is inefficient, that’s on them. What I take issue with is it doesn’t teach electronics at all. Like ohm’s law, thevenin and such. I ran into many people with sufficient knowledge to bodge together code, but couldn’t figure out how to drive more than one LED with a single resistor, or couldn’t see why feeding 12V into an analog input and causing a huge amount of heat was a bad thing. You can’t learn arduino and expect it to teach you electronics, you need to learn both.

>>2041571
Depending on the state of the latch, V+ is either 0V or Vcc/2. When both inputs are 0V, and the output is 0V, I couldn’t call it latched at all, not to mention that would require a RRIO op-amp, I wouldn’t trust his LM358 to give a reasonable output with all 3 pins being held to the same potential as its negative supply. With another resistor or two you could amend these issues, but I’m not sure if it would be any better than my design in terms of part count. And what the hell is the resistor on the inverting input doing?

>>2041626
If he’s still running on a lithium ion cell, a 555 won’t work below 4V. At least not on the data sheet. Already mentioned it in an earlier reply to the guy

>>2041071
LEDs are super nice. OSRAM DURIS E4014 JELM2 4000K
This was first time I ordered not in china, and man, they are matched perfectly. Like you can connect them in parallel, push miniscule current they barely light up, and they light up with same intensity.
Also, they are 184 lm/w, not 178 as I said before... @ 65 mA. But at 20-25 mA you can get 200 lm/w no problem i think.

Oh, yes, that piece of FR1 from some TV is capable of dissipating 0.4W no problem. So I can use more LEDs per area I think.

No idea if this place is the place to ask for this type of issue but I'm going to try anyway because it's downright infuriating. Modem's downstream signal is blinking and upstream signal is completely off, do you think I need to go to the roof and plug cable back in (in the possibility that some guy pulled it off)? Never done it before.

Is there an easy way to create a symbol for this for eeschema for lazy cunts like myself?

>>2041673
No luck, apparently our coaxial cable wasn't unplugged (time warner cable would regularly come and disconnect everyone else's wifi and not plug it back in whenever someone called for internet help). Have a view of New York City's rooftop I guess. Guess this is fates way of forcing me to read sedra's microelectronics.

>>2041689
it would take you 30 minutes or less to draw it

>>2041555
timer checked. all outcomes are possible

>>2041654
>I wouldn’t trust his LM358 to give a reasonable output with all 3 pins being held to the same potential as its negative supply
it's not a 741. ground is fine, it's inputs near V+ that are problematic. besides, offset voltage is almost never exactly 0

>>2041689
learn the keyboard shortcuts. in the library editor, down down enter name tab pinnum enter
if you already have the pinouts in a machine-readable table of some sort, you can export a symbol and code up a quick awk/perl/sed/powershell/BASIC/batch command script to convert them into
>X (pin_name) (pin_num) 0 (pin_num*100) 200 R 50 50 1 1 I
format, whose output you plug back into the exported symbol's DRAW section and then reimport, moving and rotating pins as desired. great for FPGAs

Any place on where I can read up why we use different. materials on the same components?
>Metal Film resistors
>carbon film resistors
>ceramic capacitors
>electrolytic capacitors
and etc?
I just find it interesting because my professor once went over why we used silicon and not phosphorous/borium for semiconductors and that's been very neat to learn.

>>2041727
>electrolytic capacitors
When you need capacitance and do not care that much about parasitics (Rser specially)
>ceramics
The opposite. That is why you can see them in parallel in power supplies.

>>2041654
>I ran into many people with sufficient knowledge to bodge together code, but couldn’t figure out how to drive more than one LED with a single resistor

But that's not a problem specific to the Arduino environment. Microcontrollers, in general, let you substitute software for a lot of hardware. This is obvious on its face (it's the entire reason such chips exist in the first place), but I seriously doubt there are very many who genuinely believe they can solve any electronics problem just by dropping some atmegas on it. There are certainly incompetent designers out there who are satisfied with learning as little as possible and barely-functional solutions, but you can't blame Arduino, specifically, for their existence.

>>2041743
I’m not blaming engineers, I’m blaming arduino for making electronics appear far more approachable than it really is. Like it flattens out the very first part of the learning curve, and obscures the steepness that’s just been postponed. I met these people in a maker space, a “you can do anything you put your mind to” kind of place. Considering it also served to obscure the nuances behind 3D printing, sewing, CNC milling, etc. and just gave surface-level knowledge, perhaps I should be blaming maker culture, but arduino is undoubtably a part of that. It would be much better if the maker spaces specialised more, or at least had a lot more thought put into teaching advanced knowledge.

>>2041771
>nuances behind 3D printing
I'm going to take the bullet and ask what are the nuances behind 3D printing, I've never bought one or played around with one, all I know from my surface-level knowledge is that you download blueprints and print them.

>>2041727
any practical textbook will tell you. lread the OP, at least 3 of those books will.
IIRC
>MF res: good for many frequencies (values sag at higher freq, iirc "1k"@10MHz = 700ohm), high accuracy, very low temp coeff.
>Carbon res: good for <1MHz, low accuracy and 400ppm/degC temperature drift
>wirewound res: shit for high frequencies, can be wound with platinum wire to make +3500ppm/degC positive tempco resistors (a complementary temperature drift to diode junctions, -3500ppm/degC)

>ceramic caps: cheap, small values, low ESR, microphonic, temperature drift
>polyester/mylar (or poly-something) caps: slightly better dielectric/less leakage, S&H circuits can hold a steady voltage for weeks

>EL caps: cheap, large capacitances, high ESR (6-7ohm)
>tantalum: pricier, more compact, low ESR (I think), will explode spectacularly if reverse biased

>>2041784
Well I don’t own one (yet) but getting your printer printing smoothly and strong can be a very arduous task. Bed levelling/normalising is a main point of focus, so is ensuring there’s minimal vibration in the frame and minimal backlash in the motors. But there’s also things like filament moisture content, filament diameter changing, different chemistries of filament that like different temperatures of the bed and extruder, there’s settings for step-height and extrusion speed that can lead to under or over-extrusion, there are lots of different types of beds, heated and not, that aim to maximise print adhesion when you’re printing but not make it stick too much when you’re done, and there’s all sorts of different hardware to attain such goals. Then in the software side there’s different nuances to turning your CAD model into G-code, mostly layer-by-layer but there’s also non-planar printing. Theres also double and triple extruders for printing multiple colours or materials at once. There’s also some amount of electronic hardware in the stepper motors and drivers, which compliment the mechanical hardware design. As far as CNC machines go, they’re probably some of the hardest to characterise.

Talking outta my ass here, see >>3dpg for more info.

>>2041734
Ah, I'll have to research what parasitics and Rsers are, thank you anon.
>>2041798
Sorry, I've had this practical electronics book forever but never really touched it because my mindset is so stuck on ideal/theory stuff that they teach you in college that I've never thought to consult an actual practical text. Thanks for telling me where to look anon.

>>2041806
I understood none of that but I can tell that it's honestly that iceberg meme in action. It definitely agrees with your points earlier about making something more approachable than it really is, because I had literally thought that you plug it in, watch 2-3 youtube videos, and then just get shit done. However, I had forgotten that simply doing that is also just wasting 90% of the 3d printer's capabilities by not using software like CAD (with decent skill) and the like to tailor it to your specific needs.
>>3
Wow that's a wild ride, 2011 huh. I only came to ohm back in 2019 and lurked until mid-late 2020...
>Well I don’t own one (yet)
Let me know which one you plan on getting! I was planning to get one myself with the Ender 3 Pro but now I'm not so sure because with all the stuff you mentioned, it's probably harder to deal with all of those issues/maintenance with a $300 model.

>>2041660
>200lm/W
That's like 30-35% of the maximum possible efficiency of any light source imaginable. And ~70% of the maximum possible efficiency for a blackbody-style spectrum, as opposed to a monochromic 555nm source.
Are you measuring it somehow, or just quoting the datasheet?

>>2041808
>parasitics
Every single component imaginable has parasitic capacitance and inductance. Usually it's pretty small at low frequencies and DC, but at higher frequencies it becomes an issue. A capacitor will have some inductance in the legs, an inductor will have some capacitance between the windings, and a resistor will have both. Transistors and diodes and such also have them. In general, the better a component is at doing what it's designed for (high capacitance, high inductance, high drain-source current, etc.) the higher its parasitic elements. As with all things, its a compromise.
I think the other thing is meant to be ESR, which is equivalent series resistance. A ceramic capacitor will have a very low ESR, meaning you can pull much higher instantaneous currents from it, compared to an electrolytic. Inductors naturally have ESR due to the resistance of their windings, this can be as low as a few ohms and as high as hundreds or maybe even thousands, in the case of transformers with lots of windings. Thicker inductor wire means less room for turns but less resistive losses, this is a straight 1:1 compromise meaning any given size and geometry of inductor+core will have a fixed magnetic field per watt regardless of what gauge of wire you wind it with. Assuming you don't change what metal you're winding it with.

>>2041844
I bought a Prusa knockoff 2nd hand. Shipping is a bitch. Don't do what I did unless you want a really cheap printer.
Meant to say >>3dpg (picrel) but we'll see if it posts properly this time.
I've been here for 5 years or so, haven't really contributed properly until the last 2 or 3.

>>2041854
>(Dead) (Cross-thread)
wow that sucks

Reading electrical publications from the 1890s is fascinating. So many things that are normal and standardized today were developed in that decade.
This 1897 issue of The Electrical World has an article about revolutionary new junction boxes with knockouts as we know them today. They standardized pretty soon after, I have some light switches from the 1910s that still fit in modern boxes.

Other neat articles here about the designs of the first magnetic circuit breakers, the question of whether or not neutral should be bonded to ground, and the use of laminated iron instead of solid iron for transformer cores. And of course, the current war was still raging, with strong opinions on both the AC and DC sides.

>>2041855
It's not dead. /3dpg/ is the title of the thread but it's on the /diy/ board, same as /ohm/.

>>2041727
wikipedia is unironically good on basic science like that

>>2041734
also
>ceramics
cheap per volt-farad for lower values, expensive at higher values, dense, very low ESR
>elcaps
cheap per volt-farad for higher values, expensive at lower values, slightly less dense than ceramic, medium-high ESR so self-heats with ripple

>>2041771
>and just gave surface-level knowledge
yeah the toddler-esque certitude is annoying as fuck, but otoh when they have all eyes on some collective project they are able to work out the kinks each in their particular domain

>>2041784
have you ever used a hot glue gun? now imagine building quality sculptures with one mounted on a robot

>>2041798
only class II and (obsolete) class III cercaps are microphonic and drifty. but class I caps (C0G) are small per unit volume
>will explode spectacularly if reverse biased
or overvolted for long
>S&H circuits can hold a steady voltage for weeks
PTFE or polystyrene. greencaps, days, maybe

>>2041854
>Every single component imaginable has parasitic capacitance and inductance
including wire and pc board traces
>very low ESR, meaning you can pull much higher instantaneous currents from it
also cercaps smoothe ripple instead of self-heating with it
>fixed magnetic field per watt
weird, I thought field strength was determined entirely by current and turns and area of the loop

>>2041855
>>3dpg

>>2041855
>>2041875
It can't be that hard right, it's literally just
>>>/3DPG/

Okay this time it has to work
>>3DPG

>>2041855
>>2041875
kek
>>>/diy//3dpg

>>2041854
>ESR
I swear the god every new reply I get, I have to hit the textbook or Wikipedia--which is a good thing, don't get me wrong, I never learned any of this in college (I'm not kidding, 99.99% of the time we go over ideal stuff and only rarely do we learn practical stuff like in CVD diode, maybe because I'm only a recent 3rd-year, though I highly doubt that's the reason) and I don't think I will learn any practical stuff anytime soon unless I start by buying stuff myself (which I did) and learning with ohmbros. Pretty excited about the scope and power supply coming...
>In general, the better a component is at doing what it's designed for, the higher its parasitic elements. As with all things, it's a compromise.
I keep forgetting this is all a game with 'compromises' and it's kind of surreal in a way that the better a component is at its task, the higher their parasitic elements... Definitely, something to note down, thank you anon.

>>2041861
I know, what I typed in was the image >>2041854 but it got corrected to post number 3, plus "dpg" on the end. Just like >>2041875.

>>2041875
>I thought field strength was determined entirely by current and turns and area of the loop
Double the current and half the number of turns, you'll have the same volume of windings and the same power wasted.

>>2041890
Neat

>>2041927
>every new reply I get, I have to hit the textbook or Wikipedia
That's why coming here was such an educational experience for me in the first place. Don't even have to ask questions, if you see someone talking about something, give it a google. It might be way over your head, it might be just right. I just keep learning.
Hope you enjoy your new scope, as an opportunistic poorfag I managed to snatch an old CRT hybrid (i.e. both direct analog and ADC > storage > DAC) scope for free from my uni's physics department, and a ~30MHz function generator to go with it. Still no PSU though. Before that I was just using a $20 single-channel aliexpress scope kit (DSO138?), a cheap multimeter, 5V from a chopped-off USB lead, and a couple of breadboards. Now days I just simulate almost all of my ideas.

>it's kind of surreal in a way that the better a component is at its task, the higher their parasitic elements
It isn't just parasitic elements, it's all design parameters. If you look at similarly priced MOSFETs from around the same year, you'll probably be able to draw a pretty consistent power trend between gate charge, on-resistance, and breakdown voltage, all compromising with one another. The nature of pushing materials science to its limits results in these compromises. Then there's price compromise too, naturally. Only rarely do you get a technology that's objectively better in all ways, that's the sort of thing that brings about revolutions. Revolutions like the 2nd, 3rd, and 4th generations of computing, caused by the invention of transistors, ICs, and microprocessors respectively.

>>2040770

Awesome fuel injector flow meter setup!

Do you have links to the build or etc?

>>2041854
>Are you measuring it somehow, or just quoting the datasheet?
Quoting datasheet. And I don't think OSRAM would lie this badly.
And this efficiency is not exclusive to OSRAM. I think Nichia has similar specs on some LEDs, but idk, OSRAM was cheaper, so I got OSRAM.
>That's like 30-35% of the maximum possible efficiency of any light source imaginable.
Yes, LEDs are this efficient I think.
And this is not something special desu. There are commercial light bulbs with 150-200 lm/w efficacy I think.


As for feelings, this 6 LED @ 0.4W (at LEDs, driver is inefficient as fuck) on some scrap PCB contraption is about as bright as 1W LED from usual suspects. Which correlates nicely with fact that OSRAM is 200 lm/w (at this current) and chink LED is probably 80 lm/w.

>>2041936
>I just keep learning.
That's my favorite part desu, I actually get to see what I learn in application along with other materials not covered in courses so I have to do research (so every IC stuff heh...)
>scope for free
Wow, I'm honestly surprised that's possible. I've never tried since I don't know who to ask but I'm glad that you were able to get it for free. The only reason why I was able to buy the scope is because I'm a poorfag who qualifies for financial aid, so any excess unused aid goes to me (I don't dorm). I would've otherwise wasted it on dumb stuff which is why I decided to go through with the purchases. Plus, it's just more fun and interesting to be able to do stuff in your own free time when you're bored. Hope you get a power supply though!
>Now days I just simulate almost all of my ideas.
Oh, on SPICE? I never actually learned how to use it because professors go "they can cheat with it!" so they never taught it. I should start using them more often though, no excuses for me.
>all compromising with one another.
Gosh, that's so much factors to consider. Wonder what it'll be like in the future at this rate with the revolutions. Thank you anon! Literally have to write everything down heh

>>2042003
>I don't think OSRAM would lie this badly
I'd trust it. It's definitely within the range of what commercial LED technology is capable of.
>There are commercial light bulbs with 150-200 lm/w efficacy I think
The good ones, yes. I haven't taken a look at the wikipedia page for "luminous efficacy" in a while and was just surprised how good they're getting, that's all. Might have to rewatch Clive's video on the Dubai bulbs since I forget how many lumens they were on the box. They're deliberately underdriven for high lifespans and high efficiencies, fyi. Can do similar mods yourself, which I plan on doing.
Apparently underdriving LEDs by a large margin can be so efficient that they act as heat-pumps and cool down the surface they're thermally bonded to. I wonder if that's possible with commercially available LEDs. I imagine the phosphor in a white LED decreases efficiency by some amount, so you'd want to use a monochrome LED and spec for radiant efficiency, not luminous efficiency or efficacy.

>>2042006
>I'm honestly surprised that's possible
The scope was 30-40 years old, but capable of 100MHz and can pause traces. They'd already been upgraded and were lying around unusued, and were costing them money getting calibrated each year. So the scopes were lying about to be taken by any staff who wanted them. So I, seeing scopes stacked halfway to the ceiling, lurked about until I saw some people inquiring after them, and I tagged along with them.
PSU isn't critical, I've got a multitap transformer lying about, and some breadboard holder with some power rails built in, which is good for prototyping. When I make a project, I'll usually be making it to run on a power brick or 9V battery.
>I never actually learned how to use it
LTspice is reasonably easy to pick up. KiCAD is great PCB design software and has a spice sim built in too. There are plenty of alternatives, I think there's a list of good sims in the "Some guy’s list of electronics resources" in the OP. There's also the sims in the OP itself. Logisim Evolution is very good for digital logic designs. But sims aren't perfect. There's always inconsistencies with reality, like when my guitar preamp circuit started picking up AM radio. Some stuff is just better to build in reality, to see how noise and unideal power supplies and signals and such will affect them.

>Wonder what it'll be like in the future at this rate with the revolutions
Hardware revolutions are getting more and more difficult. In the past it was about making things smaller and smaller, but modern CPU manufacturing is getting close to as dense as possible. But there are other paths to take, like analog processing arrays for running neural networks, or memory topologies like FRAM, or biocompatible electronics. I'm not terribly hopeful for quantum computing myself, but having a dedicated quantum processor connected to your silicon CPU is something that I could see in ~80 years.

Ok, if I could get some help with this I would owe y'all big time.

I'm building the project here: http://www.grandideastudio.com/vacuum-tube-night-light/

Schematic is on the page. It is pretty simplistic, but if I've got everything connected correctly and using the correct parts, I'm stumped. Pic attached is what I cobbled up.

What happens is I plug it in and the magic smoke is released from the 100Ω resistor right after the fast acting fuse, which does not pop before the resistor does. Tube is an OA3, and definitely hooked up correctly as there is only one way for it to plug into the tube holder and it's really hard to mess up 2 and 5.

Is there something wrong with this schematic or is it something I've fucked up while making the point-to-point circuit?

I can provide clearer pics of the individual components if need be, but I think I have the right ones.

>>2041927
> the better a component is at its task, the higher their parasitic elements
This is a consequence of higher inductance/capacitance = physically larger = higher parasitics. You can usually get lower parasitics for the same core specs, but you'll pay more and other specs may suffer. E.g. for inductors, using a larger core gives you more space between windings which reduces the capacitance

This is why e.g. for filter capacitors you'll often see 1-10μF in parallel with 10-100nF. The smaller capacitor will be a type with lower ESR/ESL which handles the high frequencies better, the larger capacitor handles the low frequencies which are less affected by ESR/ESL.

For capacitors which need to handle high-current spikes (e.g. input filter capacitors in switching regulators), you need to consider the average I^2*R power dissipation. For anything measured in μF, aluminium electrolytics tend to be the cheapest and thus used by default; but they have relatively high ESR, and the power dissipation often isn't even considered until it pops within 30 seconds of operation.

>>2042023
>I'd trust it. It's definitely within the range of what commercial LED technology is capable of.
Well. At nominal current it is 184 lm/w. But at 1/3 it is about 200 lm/w.
Hard to tell, from datasheet. But they have a calc https://apps.osram-os.com/MyLuminator/
>. Might have to rewatch Clive's video on the Dubai bulbs since I forget how many lumens they were on the box.
They claim 200 lm/w
> They're deliberately underdriven for high lifespans and high efficiencies, fyi.
Yep. This is what I was going to do.
> Can do similar mods yourself, which I plan on doing.
Well. Not really. If LED is 100 lm/w turd, it will get maybe 150 lm/w at best. It is way more practical to use better LED in first place.
>Apparently underdriving LEDs by a large margin can be so efficient that they act as heat-pumps and cool down the surface they're thermally bonded to
this is some communism kek.
>. I imagine the phosphor in a white LED decreases efficiency by some amount,
It does degrade. I think 80% of output after 25 000 hours. Idk.
Same deal with fluorescent tubes...

>>2042006
>(so every IC stuff heh...)
that's why I started the tradition of >>2037826 years ago. it helps keep the caliber of posting up. also it's good to know what's out there in case you want to use it. glad to see others have kept it alive
t.recovering OP

>>2042032
that 100ohm does not look like a 1W resistor

>>2042038

Shit, I suspected I was sent an under wattage type since it looks way smaller than the other (albeit higher wattage) resistors.

Thanks for the heads up!

>>2042038
>t.recovering OP
Have any images for me?
t. 85% active OP

>>2042045
It might not be the only issue, it's popping pretty quick to be just a 1/4W issue. Though maybe it will subside as just being due to inrush current to the capacitor. Pretty sure you can take the cap out and it will still work, but it might flicker, maybe even make noise.

Looks like it's the vacuum tube equivalent of a zener diode, which is interesting. It has a drop of about 75V, though it needs at least 105V to strike. Assuming 170VDC after the cap, that's about 95V across the 2k resistor. The resistor will be running somewhat hot at 4.5W, and the current through it will be 47mA. Then 170V across 47kΩ is an additional 3.6mA. So about 51mA going through the 100Ω resistor is 0.26W, which is barely past the limit of a 1/4W resistor.

So it should definitely work without the capacitor (average current will drop), and doing so will eliminate the inrush current, but doing it prolongedly may reduce the life of the tube, so I'd aim to replace it with a 1W 100Ω in the near future. If you want, you could integrate the input energy minus the thermal radiating energy to see how many watts you need, but that sounds like a pain. If 1W was good for them, it will be good for you.
You could also consider using a capacitive dropper circuit to remove the hot resistors, but I'm not sure what the time/voltage profile of the tube will do to the capacitor current.

Pic for reference.

>>2042032
>does not pop before the resistor does.

that makes no sense. since they're in series, once one dies, the current stops in both.
possibly your ''helping hands'' thingy is causing a short circuit. those alligator clips will sometimes bite right through wire insulation, esp if things are hot.
pls lay down the circuit on a piece of cardboard, like god intended, with tape to keep things stationary.

>>2040271
must be 18 or older ot use this board

how do I get a constant solid 12 volts out of a supply that supplies a non-constant 12v and a non-constant load

(motorcycle, lights)

>>2042066
this one's pretty comfy

>>2042373
boost converter
>>2042375
machine warshable and dryable

>>2041961
unfortunately not but I've written up most of the design in other posts in here
Where do people usually put writeups of their projects? I would love to start a portfolio kind of thing for my resume

>>2040796
>>2040887
>>2041499

>>2042377
But the supply can hit upwards of 16 volts and that'd fucking nuke a 12v output boost converter

>>2042382
No it wouldn't. Boost converters have a wide input range.

>>2042388
All the ones I can find say something along these lines. Should I run it through a buck regulator and THEN through a boost converter, just to be sure?

>>2042391
https://www.analog.com/en/products/ltc3780.html#product-overview

>>2042392
>buck-boost controller
So that is a buck regulator and boost converter in one, right?

>>2042400
Yes, and it will do what you need it to do so long as the lights draw less than 5A.

>>2042401
I feel like with low beams on I'd draw a lot more than 5A, thanks though I'll see what my lights would draw

>>2042405

you still using caveman burning-wire bulb technology, like Edison 1000 years ago?

>>2042405

you still using burning-wire technology invented by the caveman Edi Son 1M years ago?

Bros, I still can't decide between a bench multimeter vs a handheld one.
1) Siglent Multimeter
2) Fluke 87V
Both are very good in their own right, I'm an electronics student so I honestly don't know which to pick.

>>2042456
If you're polling /ohm/, I vote Fluke.

>>2042456

i vote for a bench model, coz you know they didnt have to make compromises to make it small and low-power.
also like the fact that you can run it 24 hours a day, no reaching for the ON/OFF so you dont waste batteries.

Hey anons, need some help here. Trying to build a wireless LED circuit to try out induction coils
The current issue I've got is that the mosfet gets really hot very quickly.
My deduction through googling is that the IRF540 is not fully activated using the output from NE555, thus heating up due to the higher resistance.
Is my deduction correct, is the solution then to get a logic level mosfet instead? What is the appropriate fix here? I am not very familiar with mosfets
Also, what is the output voltage for the square wave produced by NE555?
Is it always 5V PWM being produced in this case?

>>2042373
the answers aren't going to get any better by re-asking the same question already asked at >>2038843 but even more vague. if anything they get worse and less relevant

>>2042381
hackaday project blogs, github pages, youtube

>>2042476
>induction coils
for heating? the resonant frequency, where power is transferred best, will change depending on whatever material is near the induction loop, so this design is inappropriate for that application. you need a self-resonant design. the Royer oscillator is simpler and proven in induction heating apps
>the IRF540 is not fully activated using the output from NE555
possible. 555s are specced to work at up to 16V, get that regulator out of there if your input supply is stable
>Also, what is the output voltage for the square wave produced by NE555?
high is about Vcc-1.2V, low is about 0.2V
have you tried simulating the circuit? MicroCap has a nice parts library and it's free now. it beats the hell out of guessing while
>trying out
not really having any idea what you're doing

>>2042466
>>2042458
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA, I hate you guys. I'll eventually buy both so I'll just go with the bench one for now and when I get an internship (hopefully in the coming months) I'll get the fluke for the "on the go".

>>2042555
Do you want to drag everything to the bench or have the mobility to test wherever you need to?

Hello I'm new to this board so bear with me if I'm retarded. I would like to DIY my Dualshock 4 controller to be hardwired instead of having a usb port at the top. Pic related is an example of the part that can be easily purchased and replaced in the controller, but I suppose my question is if it would be feasible or even possible for someone like me to turn it into having a cable firmly attached. My guess would be desoldering the current usb female port and then soldering one end of a usb cable onto the chip? I'd love to hear input.

>>2042577
>desoldering the current usb female port and then soldering one end of a usb cable onto the chip
Yes, you could do that but your biggest problem will be strain relief on the cable. Figure that out first.

I'm developing a test for a control PCB that takes a +15 / -15V supply and I'be basically got everything I need down besides one thing.
I've been supplying it with a bench power supply while prototyping but was planning on just getting an AC to DC PCB mounted module for the final test rig.
The board has alarms for high and low supply voltage though (They trigger when it gets above 16V or below 14V).
What would be a simple way to get the 16V?

this is a xenon driver which I can't find its schematic.
it's made by valeo, france, and it takes in 12v power and a signal that the key is at the ON position to bring the light into the horizontal position.
I used a multimeter to check all the components, igbts, diodes, etc, and then I ended up checking the transformer.
I didn't remove it, so my results might be off.
I checked all the pins and the resistance was about 5 MOhm from left to right terminals. The left terminals are connected together (denoted by the + sign).
I am looking if there's a replacement part of that circled transformer, but I get garbage search results.
I can get more pictures if it's needed with clear part numbers if anyone is interested. It's a cool project to reverse, as it's a simple converter but I am too dumb with electronics.

>>2042595
Oh hell that already sounds a lot more insurmountable. Thank you though.

trying to diy a generator
Would someone mind identifying the terminals on this alternator for me? I can't find a diagram for this specific one from a Chrysler sebring, Denso 04606

>>2042476
You need a proper MOSFET gate driving circuit, not just an NPN with a 1k pullup. The 555's own output can probably sink and source more current than that if it were run on 12V, albeit only to Vcc-1.5V. I'd recommend a monolithic solution like an IR2184. Many of them are also half-bridge drivers, with proper high-side N-ch bootstrapping, which is probably what you want for driving an induction coil anyhow.

As for logic-level MOSFETs, you want to be using a 10V+ power source for this anyhow. It's better to drive the FET at 12V or even 15V then replace it with a comparable logic-level FET, logic level FETs have slightly worse specs.

As the other anon said, you want a resonant drive circuit of some sort. No expert on them myself.

>>2042597
You want to figure out how to detect if the power input rails are over 16V? I'd use a TL431 circuit with BJT base in the current path, and an appropriate divider to get 2.5V from 16V. Basically acting like a zener diode, though I'm not sure if you'll need to include the b-e voltage in that or not. Use an online calculator to figure out which values make that easiest. I think the same thing can be made for both low and high-sides, with minor differences.

>>2042601
5M is a lot, first try reflowing the pins on the transformer in case a solder joint is cracked. It's possible that you were just measuring between normally unconnected pins and the 5M was from external components. In the case that the transformer is broken, I'd look on digikey or mouser or wherever and see if I can sort by dimensions, to see if you can find an identical package. Then look at all the transformers in that package. If you can't find any replacement, then desolder it and see about measuring all the intact coils. With any luck you'll be able to infer the ratios of any broken coils from the surrounding circuitry.

>>2042625
I'm thinking this over now. I wanted to stick an industrial fan blade to the rotor but...
>only outputs specific current at specific RPM
>wind speed is highly variable so the RPMs and thus current would also be highly variable
>would either need to somehow control fan speed or design circuitry that could easily handle large variations in current
>also need to find a way to mount it, calculate how to counterbalance it so it doesn't just tip over in a stiff breeze, and how to let it rotate with the wind
it might just be easier to hook it up to one of those cheap $50 chinese scooter motors from ebay like I originally intended

>>2042631
>You want to figure out how to detect if the power input rails are over 16V?
No, the board I'm testing should do that.
I'm looking for the best way to apply the 16v.
Im assuming it's not simple enough enough to do it from a 15v supply without some kind of inductor circuit but it's hard to find +/- 18v ac to DC modules

>>2042649
You know an alternator can dynamically change its Kv, right? All you need is a feedback circuit to change the rotor coil current in order to keep the output current or output voltage or something constant.

>>2042655
>the best way to apply the 16v
You're looking for a power supply, but can't find the "AC to DC PCB mounted module" you need? I'd consider modifying a 15V or 18V or whatever one if you can't find one on mouser or wherever. The feedback circuit should just have a voltage divider you can change, assuming the switching controller isn't designed for only one output voltage.

>>2042655
Could you use a 36V center-tapped transformer?

What's the difference between a voltage regulator and a voltage converter? As I understand it, a voltage regulator takes some input voltage and attempts to stabilize it to a fixed output voltage - how does that differ from a voltage converter, or a boost converter?

Anyone else gets disheartened by failure?
Just have to keep moving on...
I'll build this fucking temperature-controlled fan eventually...

>>2042747
At risk of sounding like a faggot, that's how you learn. It's only failure if you give up, otherwise it's only an error.

>>2042742
There's no hard and fast distinction, but the term "regulator" usually implies that the input voltage is slightly or moderately higher than the output voltage. If the voltage is increased, or substantially decreased, or converted between AC and DC, you'd normally call it a converter.

Linear regulators can't increase voltage and they're inefficient at decreasing voltage (input current equals output current, so the voltage differential translates to power which is discarded as heat). A "switching regulator" generally refers to a buck converter with similar applications to a linear regulator, i.e. taking in low-voltage (possibly varying) DC and outputting a lower, stable DC voltage.

>>2042577
>hardwired instead of having a usb port at the top

the gadget you showed is for charging only. the controller will still use wifi, so it's not really hard-wired.

>>2042602
>a lot more insurmountable

you're kidding? all you gotta do is wrap a tie wrap or two around the cable, so when you pull on it, the solder connections wont be the only thing holding it back. or you could loop the cable thru one of the holes on the pcb of the gadget. as a last resort, you could glue the cable to the gadget with super glue, or epoxy.

>>2042655
how much current? you could boost a single dc voltage to ±24V or more and use a power op amp similar to pic to drive each DUT rail to whatever voltage level you like

>>2042768
>all you gotta do is wrap a tie wrap or two around the cable
during a hard game, that will slip

>>2042577
>hardwired instead of having a usb port at the top

the gadget you showed is for charging only. the controller will still use wifi, so it's not really hard-wired.

>>2042602
>a lot more insurmountable

you're kidding? all you gotta do is wrap a tie wrap or two around the cable, so when you pull on it, the solder connections wont be the only thing holding it back. or you could loop the cable thru one of the holes on the pcb of the gadget. as a last resort, you could glue the cable to the gadget with super glue, or epoxy. as a second-to-last resort, keep the USB connectors, and solder them a lil bit at top and bottom. this is the ''least effort'' solution.

>>2042776
>the gadget you showed is for charging only. the controller will still use wifi, so it's not really hard-wired
does it make a difference if it's intended to be coupled with a cable as in this image? I only selected that image that didn't include the cable because it was clear and shit

>>2042781
Actually I'm certain there must be something to this effect because I use it exclusively with a PC that doesn't have a wireless connection

>>2042783

the product description only talks about charging, but it could be wrong, i suppose.
''PS4 Controller Charging Cable , 10 ft, Compatible with Playstation 4 DualShock"
" Great cable that allows to sit comfortably away from the PS4 while charging your controller. "

>>2042781
sure looks like there's data lines

>>2042807
the cable can be a charging cable (data not connected thru)

>>2042776
>the controller will still use wifi
It uses BT, right? Doesn't that imply serial data?

I'm several shots deep at this point so further apologies but I'm the one who initially asked the question about the PS4 controller and I'd like to reiterate that I use it solely without any wireless or bluetooth connection, just a USB cable to my desktop

>>2042853

>>2042858

>>2042859

>>2042863
lmao

>>2042748
No bro you don't sound like a faggot, thank you. I needed that, going to get this thing to work! I desperately need a temperature-controlled fan because the weather here is abysmal (go to sleep being hot as hell ~50F with the heater on, wake up with 20F), and buying one is ehhh-tier when I'm broke.

Can someone call me a retard? I just learned that semiconductor isn't a device/component but is a fucking type of material.

>>2042828
>imply serial data
only at the on-air side of the controller. the signaling between the controller and the host processor can vary

>>2042863
grade A

>>2042877
it's both, halftard