/diy/ - Do It Yourself

tin whiskers

>>2725181
foot fungus?!

>>2725172
somewhat on topic: i found out that these scotch brite wheels work very nicely for cleaning up shitty, carbonized soldering tips.

>>2725172
I don't know where to ask, I've been trying to find out what kind of emag this is, I can find round ones, I can find square ones, but I cannot find ones that look like this.

All I know is that some russian dudes put this here, that it is 12v and that it loves to overheat after 12 hours of continuous usage, but I cannot say where they got it from, I just need a replacement because this one's cable has torn off due to a decade of chafing on the insulation.

>>2725305
Where is it installed? What is its purpose?

>>2725386
It holds a latch shut for a riddle in an escape room. When the players solve the riddle, a microcontroller switches a power MOSFET which then switches low and stops powering the magnet. An Ikea Spring overcomes the residual magnetism and pops open the latch to reveal another piece for another riddle.

It's been built by literal gopniks, but it has worked out for a decade or so.

>>2725239
I find your attitude overly abrasive.

>>2725389
What if you used a spring-loaded deadbolt type mechanism that was controlled by a solenoid? Not as big as a regular deabolt, maybe just a pin. Then it would only need to be powered long enough to release the panel.

>>2725397
I typed deadbolt but I meant door latch. Forgot to drink my Ovaltine.

>>2725389
Kind of sad it doesn’t have a connector that is hidden in a recess. Maybe they built it.

The first thing I would try is to dremel out around the connection and expose a bit of the wires, solder new wires on the stubs, test it, then drop a blob of epoxy over the area where your wires join up.

I’ve done this with those brick power supplies that are heat welded shut, and the power cord has come off right at the case. Looks ugly but gets the job done.

>>2725305
just image searched it

>>2725432
Those are square, not rectangular.
They are also permanent magnets and not electromagnets. There’s no wire.

i might be posting in completely wrong place and this is probably not easily solvable but here goes.

I have a moon lamp, like this guy, only mine doesnt have remote control, its only touch controlled
https://www.youtube.com/watch?v=KEHZRbQsxmk

>touch it the first time, and the side lights (underlined in red) turn on, its just white weak light
>touch it second time, and the middle RGB lights (underlined in green), they glow red
>touch it third time, and the middle RGB lights (underlined in green), they glow blue
>etc, the middle RGB lights cycle 7 modes
>touch it 8th time, and it turns off.

my problem is that i would rather lose the colors if I could gain the ability to
>touch it - it turns on
>touch it again - it turns off
so i dont have to cycle to turn it off.

i cannot destroy the middle lights, cause the cycle will still go through them. I think i roughly understand how this whole thing works, the black chip in the corner sends the signal through those lines, its some kind of AND/OR thing... is there anything i can do and what tool do i need?

P.S. yes, there is an on/off switch at the bottom

>>2725437
trade it for a better lamp

>>2725437
>>touch it - it turns on
>>touch it again - it turns off

you're disrespecting this craftsman's 8-way design with your contemptible arrogant round-eye 2-way attitude.
well, lets see you program your own modified chip to implement it, then, smart ass.

>>2725437
Soft latch switch toggled by a capacitive touch sensor in the metal base that you're going to make with copper or aluminum tape.

>>2725437
Trace the board to get a schematic, that should tell you what you need to change. If the capacitive touch controller is the same IC as the LED controller like I think it is, you'll have to ditch that controller for another capacitive touch controller of some sort. Like an adafruit module.
I had an RGB COB that just annoyingly cycled through the colours, I just desoldered the control chip and added my own PWM circuit.

>>2725397
>>2725400
I don't have any time for fancy modifications, I just want to order the part, give it to someone else so he can install it and be done with this.

>>2725402
If I had the time, I would do this. Sadly I don't.
Anon, the people who built this weren't exactly ... sophisticated. There's another room where a puppet is hoisted into the air at a certain time to scare the guests. The jeffries' tube behind it (maintenance space) is a 30 cm wide crawl space / death trap where badly cut / chipping, thin plywood is mated to razor sharp metal studs that will cut open your flesh to the bones, chains hanging around and the counterweights being cold-welded gym-plates. Absolute death trap and I have no clue how I made in and out of there multiple times with only minor injuries.

>>2725432
Has to be an emag though.

I'm playing around with sotft latch switch and a boost converter for a little project. The latch thingy is controlled by an MCU and works as intended. The boost converter also works by itself, raising voltage from 3.7 V to 5 V.

However, when I connect the boost converter parallel to the MCU, so I can use the switch to turn the converter on/off, the voltage at VIN1 drops to about 2V. The whole thing draws 5 mA from the power supply when I hold the button down, and current drops to zero when I release the button.

Any idea what might be wrong?

>>2725500
No inductor?

>>2725504
or filter caps on the input/output?

>>2725504
L1 is the inductor for the converter. The converter works fine when I'm not trying to power it through the soft latch circuit.

>>2725507
Separate the EN pin from Vin and connect it to your soft latch circuit.

when I hook up 120v to the large winding side of a transformer, I get lower voltage on the other side but when I hook up the 120v to the small winding side to get a larger voltage, I set off the breaker. Why would it work on one side but not the other? Wouldn't the magnetism created work either way?

>>2725172
Check out reverse engineering content here and on the related discord:
https://wiki.recessim.com/view/Main_Page

>>2725305
You know the real answer here. You get one of the round or square ones that doesn’t look like that at all you screw it down with plumbers tape and you call it a day.

>>2725546
>Wouldn't the magnetism created work either way?

the secondary has fewer turns, thus smaller impedance, so it takes way more current than the primary.
it's a good thing the breaker blows, or else you'd get stinky smoke, and eventually fire.

sure, you can apply a voltage to the secondary, but not 120V, use something like 0-12V

there's no reason why the ADE7913 wouldn't work for single-phase (120VAC mains), right? https://www.analog.com/media/en/technical-documentation/data-sheets/ade7912_7913.pdf

>>2725562
can I apply a large resistor to it to limit the current, or would it melt the resistor

>>2725437
best and cleanest way is to reprogram the microcontroller to only toggle on/off, how hard that is depends on the microcontroller and the tools at your disposition. The next best and probably easier way is to get another microcontroller and interface it between the original mc and your LEDs, with the MC acting as a decoder for 8 light modes to 2. Since they cycle through each mode the same way, you can map all odd modes to "on" and even modes to "off". You could also technically ditch the MC and do it with transistors, logic gates, FF, etc., but that would take more space.

>>2725604
Your problem is that the secondary doesn't have enough turns. If you place a resistor, you'll limit the current through the transformer, and the amount of flux is reduced, which reduces output power. Furthermore, reduce the current too much and your transformer is no longer operating efficiently. The best option you have it either to get the right transformer, add a lot more turns to the secondary and primary, or increase the frequency of your signal.

If your resistor is a high value resistor, you won't be burning it but you won't get what you expect. If the resistor is too low in value, it will definitely burn up though. Also, what you are trying to do is dangerous, don't flip the primary or secondary, since if it's a 12V output, with 120V input, you'll be getting 1.2 kV if you actually were successful in your attempt to swap the two.

>>2725620
thank you anon, it's hard to see responses like yours these days. Appreciate it.

>>2725500
Is this your full schematic? U1 enable is connected to Vin, which means your converter is always on. If you try to pull down enable, like using the MCU to do so, you're shorting Vin to ground via your microcontroller. Fix your schematic, have it all on kicad instead of to different ones with properly labeled nodes, your circuit doesn't make sense and I have to assume some things. For instance, on the left you have Vin and on the right Vin1, I'm assuming they are the same nodes.

>>2725627
Remember, inductance can be seen as a frequency dependent resistance, where the Impedance is Z = jωL,where ω is 2π*f, f being the frequency. L is dependent on the amount of turn (as well as other things), so more turn, higher L. When you connect 120V to the primary, because the primary has a lot of turns, L (call it L1) is high, and the impedance Z1 = jωL1 at 60 Hz is high enough to limit the current to a safe level where the windings won't burn. On the secondary,you have a lot less turns, which means L (call it L2) is lower. If you try to connect the 120V to the secondary, now the impedance is Z2 = jωL2. The current through the primary (using ohm's law) is I = V/Z, since Z1 > Z2, I1 < I2. So this is the reason why you need a higher L2. To do that, you can increase the amount of turns, but then your transformer turn ratio changes, so you need to also increase the turns on the primary. Either that, or you remember that Z = jωL, and keep L as it is and instead change the angular frequency ω. Since you cannot do that for 120V mains, you need a different L.

Adding a resistor in series with L will limit the current through the secondary. According to Bio-savarts law, the magnetic field B generated by a wire depends on the current I flowing through it and the distance from the wire. By coiling it up (like in an inductor or transformer), you can apply bio-savart's law (or ampere's law which is easier) and you realize that the more turns or current you have, the stronger the B field is because the wires coiled up add their B field together. The stronger the B field, the higher the magnetic flux is. The induced emf in the other transformer side is directly related to the magnetic flux passing through the core of the transformer accoriding to faraday's law. So by putting a resistor, you're reducing the flux and what is generated on the other side, so that won't work.

>>2725500
Just run your MCU directly off the battery voltage instead. Not like basically anything these days requires 5V to run.

>>2725611
I think it would be easiest to desolder the old MCU and find an MCU with the right features to replace it. Preferably the same footprint, but deadbug is ALWAYS an option.

I was thinking about picking one of these up, and modernizing it a little bit. If it's in working condition, I would want to upgrade to a rechargeable lithium battery setup. If I can find a broken one nearby, I would hollow it out and replace everything if I can't get it working. How difficult is it to do something like that? I've done basic automotive wiring and have the tools (never done audio, however). I don't know anything about circuit boards, volts, and watts in the context of what a speaker system needs. d-batts in series to equal so many whatevers. battery life and all that.
Sorry if I phrased this all dumb

>>2725647
I thought of that, but I don't know how the circuit works apart from turning the led on. By replacing the microcontroller, you need to interface it with all those sensors which might be a pain to code for. By putting a second microcontroller, all you need to do is decode the signals, which can be easily done with an arduino.

>>2725652
12V li-ion battery pack + BMS
12V li-ion charger
12V power supply

>>2725673
That actually sounds pretty simple. What is the BMS?

>>2725678
Battery management system (BMS)
https://en.wikipedia.org/wiki/Battery_management_system
It keeps your battery from spontaneously incinerating your face.

>>2725680
I KNEW IT
thanks, anon. :) So all of this would be pretty simple to wire up. Would I need a transformer or anything to step up or down current(I don't even know if current is the right word)? can I wire up more than one 12 volt battery pack while keeping it 12 volt?

>>2725682
If the battery pack voltage is too high when fully charged then you might need a buck converter to keep the battery output at 12V. If the radio isn't drawing too much current you could use a 12V or 9V regulator instead of a buck converter, which would be better as far as ripple (circuit noise) is concerned.

>>2725684
Ahh, I see. How should I interpret these numbers in picrel? I don't actually know the draw yet. I might increase it, and I guess I would have to measure the draw when/if i do.

>>2725687
Well, that's some confusing shit right there. Luckily the power consumption (43W) is labelled. You don't need a buck converter or a regulator because 10 D-cells is 15V and your fully charged battery pack won't rise above 13.6V.

>>2725687
Download the user manual and service manual.
There are a few different ways this can be powered. Plugged in to mains power (the AC 120V), 10 D batteries (DC 15V), or from a car battery (DC 12V). You will want to supply 12V to the car battery input.
The 3 AA batteries (4.5V) are used for battery-backed storage of presets on the radio, you don't need to care about that.

>>2725687
43W/12V = 3.6A
43W/15V = 2.8A
The radio apparently has 4 amps inside. The AAA powered one is probably for the headphone amp or tweeters only.

>>2725699
Anon wants to put a rechargeable lithium battery pack inside so it should be connected to the 15V battery input.

>>2725700
>43W/12V = 3.6A
>43W/15V = 2.8A
That's maximum current draw (full volume).

>>2725702
The internal battery and external battery connectors both supply power to the same rail.

>>2725707
Then the DC jack should be re-routed to the lithium charger so Anon doesn't accidentally explode things, and the AC input needs to be isolated or switched.

>>2725707
I'm having trouble reading that schematic because 50% blind, 50% retarded. Is that a relay in the center?

>>2725713
It's probably a mechanical switch that isolates the battery from the AC transformer + rectifier when something is plugged in to the jack.

>>2725714
Isn't j1902 on the bottom right a switched DC jack?

>>2725682
and yes, you can use more than one battery pack as long as they each have their own BMS, and both packs are charged to about the same level before connecting their output in parallel.

what is the best way to create a big "pad" linking other pads in KiCad? I want to join all of the Vb-net pads into one physical rectangle to ease hand soldering. I tried using the rectangle tool on the copper layer, but I couldn't assign a net to the resulting shape. I don't really want to modify the footprint just for this one part.

>>2725637
I dont like this, inductance is inductance. Inductive reactance is frequency dependent. Resistance != Reactance !=Impedance

>>2725684
Probably not an issue, lithium ion cells have a smaller full/empty voltage ratio than alkaline cells. Alkalines are like 1.6V fully charged and go down to below 1V when discharged, so 10xD cells gives 16-10V. A 3S lithium ion pack gives 12.6-9.6V, which is fine. A 4S lithium iron phosphate pack gives 14.6-10V or something, which is also fine.

>>2725713
I think the two circles in the middle is the AC power socket, while the switch above it is a manual switch that's physically close to the AC power socket. The switch is meant to swap between AC power and battery power, though there's also some funky stuff with that switched DC jack. I'd draw out a schematic that uses rails and maybe some labels instead of spaghetti.

But be sure that current isn't backfed from either AC or DC source back into the batteries.

>>2725727
Best off modifying the footprint itself with one big pad joining them. You can also add another footprint (e.g. testpoint pad) and alter its size and shape in the PCB editor, I'd do this with a symbol linked to it so you don't have to manually mess about with net assignment and having your DRC yell at you. There isn't a good way of assigning nets to rectangles.

>>2725742
>Best off modifying the footprint itself with one big pad joining them.
You were right. I was under the impression you had to save it and add it to a library and all that jazz, but it seems like it saves it local to that object in the PCB editor, took like 2 seconds. Thanks.

>>2725744
If you plan on using those footprints often it's best to save them as a duplicate of the original footprint, but for one-offs yeah I'd do it this way. Just watch the DRC doesn't complain.

Coil gun
>Steel whetstone rod ammo
>Put spring at end of chamber
>Fit snuggly in a tube/straw (learn how to >3d print?)
>Strip wire before coiling it?
>1 long coil across tube or multiple coils?
>Safety switch probably
>a dpdt trigger or even a regular switch i guess since its 1 shot to reverse polarity
Am I missing anything I would need to know?

>>2725772
>Am I missing anything

yep.
a google search for ''youtube coil gun''
so you can learn all you need from other, more experienced, violent psychos.

>>2725772
>Am I missing anything I would need to know?
yes

have a look at some existing solutions to see what design decisions they make, and why. thorough write-ups will probably be better for this than 10 minute youtube videos aimed at low-attention-span zoomers

some pointers:
>V = L * dI/dt
>B field is proportional to I
>enamel-insulated wire exists
>number of turns determines speed at which its magnetic field builds and releases, which is a compromise with regards to the speed of the projectile
>you need some way of switching the coils on to attract the projectile, and off once the projectile has passed their centre
>sensors to find the projectile's position
>power source needs to handle very short spikes of very high current
>multi-stage is definitely best
>perhaps pull the energy back out of the coil and into the caps/batteries using more transistors or diodes
>perhaps trigger a voltage spike from an ignition coil into the force coils to build up the magnetic field faster

>>2725727
bro youre hardly going to notice a difference when soldering. if anything itll be harder because you fot a giant copper heat sink ruining your solder adhesion. also creating many different footprints for the same component is common practise in industry.

>>2725745
>just watch drc doesnt complain.
this is the cuck move. the chad way is to modify the drc rules list to ahere to much wider/easier rulesets

yo guys, so i got this hugeass speaker enclosure and it came inside with a giant transformer which is a mains>30-0-30VAC, and by the size and weight id say 15A. i guess this was for making it not just a cab but actual amplifier, but what for, what chips or design would go good with such voltage and current output?

pic semirel

>>2725814
>what chips or design would go good with such voltage

too lazy to get out my calculator but an amp in the 100-200W range will need those kind of voltages to drive 8-ohm speakers.

>>2725804
I would think it would be easier, because of a lack of thermal separation between pins should make the solder wet them easier.

>>2725727
make a copper pour on the pcb joining all of those together. You can assign the pour to a specific net, look it up on kicad's website on how to do copper pours, it's very easy and they have great tutorials.

>>2725737
>inductance is inductance
It's my fault, i'm not always great with terminology. I should have said
>Remember, an INDUCTOR can be seen as a frequency dependent RESISTOR
One thing to note (for ideal components), that a purely resistive (R) load, resistance = impedance or purely reactive (L/C) loads reactance = impedance. The concept of a frequency dependent resistor is used heavily in the analysis of AC circuits and is also mentioned in
>Horowitz and Hill, The Art of Electronics
As long as you understand the limitations, there is no harm done.

>>2725855
nah mate. the extra metal sucks away the heat it takes longer for the pad to rise in temp such that the solder will flow and stick. you notice this most when a pin is directly connected to a larger copper pour or plane

Need a "publication quality" (read: pretty) schematic for printing, any good software for this? my actual schematic is no good.

>>2725985
If you're autistic enough you can try using CircuiTikZ with LaTeX. Another option is to redraw it in inkscape, but you'll have to make your own symbols (you can just trace over existing schematics and copy styles you find in datasheets, books...).

anyone know what size of passive heatsink is ballpark necessary to run 3 x 10W LEDs? i bought pic related. haven't cemented the LEDs do it yet but i have a feeling it's nowhere near adequate. i only need to keep the LEDs cool for about 15 minutes since they are for intermittent use (365 nm LEDs for PCB photoresist.) the LEDs have a metal base ~20mm x 20mm each.

>>2726007
something like your picture would probably be fine, with a fan pointed at it.

>>2726007
You can use smaller heatsinks by running them at lower current. I'd like to try using a XL4015 for driving all three.

>>2726007
>>2726032
Don't forget about total surface area of the heatsink being measured by combining both sides of all the fins surface area.

>>2726007
First of all you need to figure out efficiency of the LED.
With white LEDs it is basically (lm/w)/(315 lm/w). Or in other words, shitty aliexpress special white LED wastes 70% of energy on heat, while high quality LED wastes 36% of heat.
But with monochromatic LEDs - idk. Some manufacturers, like OSRAM do state LED efficiency in the datasheet, sometimes.

But I think assuming 30% efficiency you'd be pretty safe.
That heatsink looks pretty big, comparable to Slot 1 CPU heatsink, and those could dissipate 40W no problem with fan (ofc). So you would be fine.
Maybe it would be fine even without the heatsink, but LEDs might get hot, 80-90 degrees hot, and since this is not for general lighting, I think 90C is perfectly fine since reduction in lifespan (100 000 hours -> 15 000 hours) won't do shit in this case as it would probably never ever get close to said work time.

>>2725772
2000 J capacitor bank, IGBT or Thyrisotor stolen from local train and inductance big enough to prevent thyrisor from exploding

>>2725985
>any good software for this?

are you a sissy boy, or do you have the balls of steel needed to download torrents?

>>2726110
these niggers allow igggames to exist, i wouldn't trust any of their .exe's without an airgap involved

Can anyone please tell me how to remove rust from charging dock contacts on picrel? I've googled and all the articles are either "teehee vinegar" or "bro just dump it into sulfuric acid and electrolyse".

It's not just some discoloring, it's a solid pile of hard rust.

>>2726280
that's not rust, but a failure of the nickel plating, exposing the bare copper underneath. if you're autistic you could re-plate it. I would just ignore it.

>>2726283
So it shouldn't impact charging? The robot vacuum cries that it cannot charge properly.

>>2726290
scrub it with some steel wool until it looks bright again, if the robot still complains, that's not the problem. if the robot stops complaining, the oxide layer is blocking conductivity.

>>2726294
Great, thanks

anyone have a PREFERRED CHINK VENDOR for very low melting point alloys used with desoldering, like knockoff chip quik? i tried searching aliexpress but their search engine is beyond fucking terrible and i can't find anything, which i find hard to believe.

>>2726378
Chipquik kit = $16
hot air station = $50
chinese mystery alloy = negative value

>>2726280
mechanical cleaning would be the easiest, like sandpaper. but like >>2726283 says, it looks like exposed copper. it will quickly tarnish again.

I was asking about replacing control board in garage operator in the last thread. Got some good info from smarter anons than me but had more question about the Arduino uno I bought. It says inputs are 5-12v I'm pretty sure the motor controls use 24v AC as signal. Do I need a different board that will take 24v? Should I try and rectify each input for the Arduino or do the input go to the relay module and everything will be fine?

>>2726280
Salvage some contacts and shape them to fit or use electroless plating to restore the nickel.

>>2726392
Use a full bridge rectifier with smoothing caps to get clean DC output, then either buck it down, clamp with a 5V zener, or use a regulator to maintain 5V to power the Arduino and relay board. The I/O going to the relay board inputs must be isolated from the AC side which is what the optocouplers on the relay board are for. The AC side of the relays don't need anything else as they are just a switch.

You could also use a USB charger wired in parallel with the transformer primary 120VAC to power the Arduino.

>>2726280
Make a 15W EPP wireless charging pad/dock.

I need your help, /ohm/!!! Best I can tell, this broken part (thermistor?) is an inrush current limiter on a piece of music gear. The hot side of the AC supply hits this thing, then the little breaker it's epoxied to. Do I need to get the exact same part, or can I use any thermistor that's advertised as inrush current limiter? What do?

>>2726460
25-Ohm 3A NTC thermistor

>>2726463

>>2726466

>>2726460
Does it still work? Clean it off and put a blob of epoxy or jb weld (which is epoxy) over it.

I have an old generic 12V power supply brick that is glitchy. It appears dead but then I unplug it and plug it back and the LED comes on and it starts working again. What could be the reason for this?

>>2726558
If it's old and/or heavily used, it's probably bad caps, but it could be several other things. Open her up and take a gander.

Anyone have any conductive paints and the means and inclination to measure its ohms per square? I'm developing some cheap stuff, approximately 36.5 ohm/sq after about an hour and a half of drying, I'd like to see how it stacks up against the commercial stuff, most of which provides volume resistivity measurements instead of sheet resistance.

>>2726569
Hmm, I opened it up, it looks like a pretty sophisticated high end device with a lot of parts. What could go wrong?? The big cap looks a bit bulged up? What is that DIP-8 part, 555?

>>2726580
yeah the big cap is fucked. the dip-8 is probably a comparator but there are a few dip-8s you'll find in these things occasionally, so who knows.

>>2726580
Definitely bad caps. I suggest replacing all of the electrolytics. Only see four in the pic so it's not going to be time-consuming or expensive. Just make sure you buy Jap, US, or German caps because the others are shit. Match the capacitance rating, voltage, and physical size. They are radial leads, not axial.

>>2726580
and clean the electrolyte off of the PCB and surrounding components (it's corrosive).

>>2726588
do you mean that brown discharge?

>>2726590
Yes. We call it "e-smegma" in the biz.

>>2726586
>I suggest replacing all of the electrolytics

disagree.
the output filter cap is the one which gets all the mistreatment and goes bad.
the small input caps have an easy life and will outlive the unit.

>>2726597
By the looks of the unit, it's been run hard and hot, and the caps are sitting between two heatsinks. I'm not saying you're wrong, but if it belonged to me I'd change them all.

>>2726599
>it's been run hard and hot,
Yes, it powered a monitor several hours a day for years.

oh wow looks like the big cap is optional? i soldered it out and it powered right on!

>>2726626
Also what's the point of a 400V cap in a 120V unit? Why is the fuse only 2A if the supply is rated 4A according to the label?

can more of you retards post pictures of your current projects so i have something to look at please

>>2726626
>the big cap is optional?

definitely not.
dont plug that into your monitor.
the output voltage is gonna be all whacky.

>400V cap in a 120V unit?
120Vrms is 340Vp-p or approx 340Vdc.

>2A if the supply is rated 4A
output current is 4A.
the input current, where the fuse lives, is smaller.

>>2726637
>120Vrms is 340Vp-p or approx 340Vdc.
I don't understand this at all. Are you saying 120Vrms rectified = 330Vdc? Why would we care about p-p if AC is rectified into DC? The big electrolytic cap across the diode bridge is 100V. The blue non-polarized cap is 400V.
Aren't caps rated in Vrms which should be equivalent to VDC? I thought the whole idea of rms is so that both AC and DC voltage could be treated as the same units.

>>2726630
I don't have a picture but I just had a great idea after I again left my soldering station on for a few hours: auto-shutoff. Either time based or better yet motion based: an accelerometer senses no-motion for 20 minutes and power is shut off automatically.

>>2726671
could probably do a PIR sensor if your station is pointed at you

>>2726558
Often they shut off when they detect an overload condition, and come on again either after a delay (could be a long delay if it’s thermal-based) or after they detect a sufficiently small load current. It may detect such a state either from a current sensor or by measuring temperatures of internal components. If you aren’t overloading the power supply, try finding the sense circuitry in it. If it’s current-based, it’s possible the amplifier or it’s passives have gone funny and need replacement, or maybe the output caps are shorting. If it’s thermal, check the thermal compound, if it’s fine then the transistors or diodes or thermistor may need replacing.

It might also be detecting some other fault condition that’s sufficiently intermittent.

>>2726575
Can’t you just calculate based on paint thickness? Also integrating the infinite resistance grid problem.

>>2726627
Peak mains is 1.4 times that, which is what those caps will charge up to. It’s probably also 220V compliant, so multiply that by 2. Round up to 400.

>>2726671
Chinky T12s and TS100s and such have had timed auto-shutoff for years, they have a tilt switch inside the handle, though you could also certainly do it by measuring thermal loading. They go into a standby mode at a user defined temperature (150C for me) after five minutes, and drop to 0C sleep after half an hour or so. Can’t imagine spending over $60 for a soldering station without it.

>>2726694
>Can’t you just calculate based on paint thickness?
possibly, but I don't have a wet film gauge to know how thick my film is and they're like $120.

>>2726699
Figure it out by volume or by weighing it.

>>2726701
it still wouldn't be accurate because a brush doesn't leave a film of consistent thickness, you need a wet film applicator* for that kind of test. I'd rather just use ohms/sq

any recommendations for a cheap (or even diy/kit) spot welder? For 18650s of course

is there an all-in-one lithium battery charge+protection+whatever the fuck IC?

>>2726694
>Can’t imagine spending over $60 for a soldering station without it.
I have an old analog Weller WES51 ($120-$150) and it obviously doesn't have it and a new yihua 939d+ (about $70) and it doesn't have it either. Another one that I know is very popular Hakko FX888D doesn't have auto power off either. And it's $115.

>>2726709
yes, TP4056. you can order clones of the bare chip, or you can buy little modules with it already on them. The modules are handy if you want to stick a rechargeable battery in something that doesn't normally support it. Only good for 1S devices because no onboard BMS/balancer.

>>2726719
I forgot about the TP4056. Thank you.
Now, if I wanted more cells in series (with a focus on portable applications), what's the standard way of managing that without having to deploy a full BMS? Or, is that the way? Do they make, say, 4056-equivalent ICs, but for 2/3/4...S?

>>2726725
standard way is probably just using a BMS. I think there are 4056-based boards, but not entirely sure. If you could figure out a way to flip a switch and put them all in parallel, you could charge with a single 4056 (albeit slowly, because it's limited to 1A iirc)

>>2726728
If you want your battery to last a long time, you trickle charge it. Also don’t let it get below 20% or above 80%. Don’t discharge it hard, and keep the cells cool at all times.

>>2726776
All bullshit placebo garbage. It makes a 1% difference altogether at best.

>>2726719
TP4056 only has overcharge protection. For overdischarge protection you need a DW01 and dual MOSFET IC.

>>2726785
Keeping a standard cell at 4.2V constantly (i.e. for months on end) is bad for it. I’ve had two devices die because they didn’t have the firmware to prevent that when they were left constantly plugged in to power, within 2 years they had significantly below 50% capacity, while a cell stored at 3.7V should stay good for at least a decade.

>>2726787
I have never had a battery go bad in less than a decade despite never giving a shit about meme shit like muh 20/80% or muh trickle charging or muh temperature.

>>2726785
Found the cordless battery shill already.
It’s basic chemistry and physics.
I’m not replacing my entire device every 2 years with it’s glued-in battery, nor am I shelling out $100 every 2 years on a new drill, whether you like it or not.

>>2726788
> the 80/20 rule, heat and trickle charging
All come from Panasonic themselves.
Feel free to keep your own batteries continuously charging at 200 F, and run them down beyond 1 V daily. Don’t forget to tell us you’ve been doing that for decades, and have never purchased a new battery for anything.

>>2726792
>>2726791
Once again, I've done this shit for the past 40 years on all kinds of batteries. Never had a problem, they've all lasted 10+ years above 50% charge.

tfw you'll never get paid to drink and fix tube amps in your comfy home shop

>>2726794
you think he groks impedance?

>>2726795
he's actually pretty cool for a boomer
will do full period stuff when customers want it but also makes pcbs to make the hv power supplies and filter caps easier to install.
Not to mention PTT boards for old ham radios an the like.

>>2726793
I smoke 2 packs a day for 40 years, and I know a guy that knows a guy that smoked and lived to be 98 so it must be good for you. Guhvment can’t tell me whut to do.

>>2726808
>don't believe your lying little eyes, goy, just bend over backward for no reason to increase my profit margin instead

>>2726794
Ironically, the oscilloscope STAND is probably worth 3 times what his oscilloscope is worth. Looks like a 20 MHz TV repair man special. They heyday for oscilloscopes.

>>2726812
When all you watch is audio generators that's all you need

>>2726709
No, but you can use TP4056 or whatever clones to charge, and DW01 for discharge.
DW01 is quirky as it uses MOSFETs as shunt, so choice of mosfets matters.
>>2726785
Depends how shitty your cell is. Some cells would die if kept at 4.2V, while others are absolutely fine with that.
>>2726788
80/20 rule is a scam. You move exactly the same amount of watt-hours through the battery. Only benefit is that now you can store cell for longer without it self-destructing (as you'd cut off charge at 4.0-4.1V).
Rate of charge and discharge kinda matters, less current - better.
>>2726793
NiCD and NiMH allow more abuse than Li-ion. Li-ion (the 4.2V variety, NMC, NCA, etc) are fragile bitches.

How can I start learning electronics? Can I just buy breadboard, some parts, pick a book and just learn how to do schematics by practice? I only know the basics which part does X thing (capacitors, resistors, etc.), and I don't really understand how do you build anything that looks like shown on any scheme, it looks vastly different when built properly. Should it be enough to not to destroy my very first attempts at doing schemes?

>>2726839
I'm down to help you out, dude. cometoverman on discord. Glad you're interested in the hobby

>>2726839
>How can I start learning electronics?
First you learn ohms law, kirchhoff's law....
Then you figure out there are inductors, capacitors, they are differential equations, you try solve them, remember that you forgot how to solve differential equations, and get depression, become alcoholic, and then jump infront of the train on tracks or run off the cliff...
>Can I just buy breadboard, some parts, pick a book and just learn how to do schematics by practice?
Kinda. You also should get couple tools, like a multimeter, a power supply (in the beginning you can probably get away with a batteries)
> Should it be enough to not to destroy my very first attempts at doing schemes?
You learn by inhaling the semiconductor smoke. Dont worry about it, parts cost cents.

Well, first you start with resistor. Connect them in parallel, series, measure voltage, current, etc. Figure out Thevenin equivalent circuit.
Then you connect resistor in series with an LED. Figure out dependence between current and resistor value.
Then you make a circuit with a BJT. First as a switch, then a multivibrator to make LEDs blink, then an amplifier. Figure out how to amplifier properly.
Then you make differential amplifier out of discrete transistors... Then you figure out op-amps. Figure them out. Make oscillators, headphone amps, equalizers and filters. Maybe a simple synth.
Then try using P and N MOS in order to make a logic gate. Get some logic ICs figure out what they do.
Then try making a class D amplifier, or a PWM controller using op-amps.
Then idk.. Possibilities are endless.
While you're at it, learn to solder. Breadboard is nice and all, but what if you want a semi-finished product, like idk, LED blinky christmass tree ornament? Figure out PCB design software, ordering PCBs and/or making them at home.

>>2725172
Ive converted a retired 3d printer into a pcb mill and the isolation routing works great but Ive burned some pads in my attempt to solder to the board. I have copper/nickel plating solution and wanted to know if I could nickelplate the board before routing and if that would serve the same purpose as tinning.

>>2726893
>Ive converted a retired 3d printer into a pcb mill
What printer? The bedslingers I see are nowhere near rigid enough, but an ender 5 or corexy printer might work. I've got a snapmaker original 3-in-1 myself.
>Ive burned some pads
Do you mean you've delaminated them from heat? A solder mask layer will hold the top copper layer on a bit better. You can mill a solder mask layer using one of those spring-loaded bits.
>plating
The common method is to electroless tin plate a board (nurdrage has a video on this), but in your case that won't do anything structural. Electroless copper plating is used in industry to make the copper thicker, and in your case you could definitely electroplate it before milling it, but I'd want to redissolve the copper you mill off. Would be doable to put the copper chips on an MMO electrode lying flat and electrochemically dissolve them in HCl or H2SO4 or whatever. Sodium bisulfate would probably work. The same applies to nickel, but I'm not sure how easy it is to solder atop nickel, so I'd start with copper.

>>2726897
>he same applies to nickel, but I'm not sure how easy it is to solder atop nickel
very easy, it's commonly used for battery terminal strips.

>>2726897
Its a bed slinger. Im having no issue with isolation. Cutouts arent as nice as Id like though. I figured a soldermask would help but Im also having a much harder time soldering to them then Im used to. I assume its because of oxidation/corrosion on the copper. Nickel is corrosion resistant so I figured it would help here. I cant do electroless plating though. Just gonna plate the whole board and then route.

>>2726893
>>2726905
Just clean copper with 1000+ grit sandpaper before soldering. Use additional flux.
Soldering nickel-plated stuff sucks. It doesn't wet as good as pure copper, gold or HASL.

>>2726905
For ease of soldering, the traditional method is tin plating, because that wets perfectly every time. It might actually be easier to electroless tin plate than electroplating it, but idk. I've never tried electrochemically dissolving tin but it's probably faster than just leaving it in HCl for a week (you want to exclude any oxygen so it can't oxidise the Sn2+ into Sn4+).

Before that though, I'd try using abrasion or chemicals to get rid of any oxides. HCl works well, so does any etchant like ferric chloride or some sort of persulfate, though for etchants you have to actually care about the length of immersion. Copper polish should do a good job, and I suspect citric acid will work well too. For abrasion, I just use the kitchen scotch-brite pad. Gives a faintly scratchy look but you can't make me care about that.

>>2726857
Will write some time later, thanks for that!
>>2726859
Thank you very much for advices!

>>2726911
No problem!!! Hope you enjoy it here!!!

how can i control a 120v 1.6A AC motor's speed with a pot? I had a 50W rheostat but it burnt up obviously. I have a triac but not a diac, and all the circuits in google images seem to require one.

>>2726644
RMS = root mean square (average over time)
Peak voltage is different. It's a measure of the maximum voltage at any point in time.

>>2726703
Spin-coat

>>2726706
Salvage a MOT and build one.

>>2726839
Yeah that's how I did it.
>I only know the basics which part does X thing (capacitors, resistors, etc.),
My experience has been that I truly didn't know. For example I only learned by practice that a BJT doesn't actually let current through if the ratio of base to source collector is unfavorable. I had only ever learned "current goes to base -> current flows from collector to emitter".
Another example is that you'll only ever learn how bad chinkshit is, and how you can't possibly tell anymore which is real or fake based on reviews, by getting scammed, so I recommend getting scammed at least once and then only going for reputable vendors even though it's 10-100x the price for some parts.
In fact, don't pick a book at all. Just choose a project and go for it.

>>2726930
AC motor speed depends on frequency. You can modify frequency by having a capacitor in front and using a pot to modify voltage to cap, thus modifying charge/discharge time, thus frequency.

footpedal broke, so now it's a hand control

>been fucking around with AA/AAA NIMH recovery
>timer falls in the middle of my usual comfy sleepy time
>take an unmarked duracell out of the torch somehow mix it up with another and put it into my old dumb charger for a discharge cycle
Guessing this one is forever ever gone at this stage. Came out after a bit of a charge at 0.3v in one smart charger, then charger dropped to null, refused to charge, charged it up a little.
Goes from 0.8v to 1.00~ at 0.5c, switch over to 0.1c and voltage starts plumetting, falls to 0.3 and the charger rejects it, put it in another smart charge, fuck it, 1c, comes out at the end, chinesium smart charger lied to me.
BACK IN THE slow DUMB CHARGER since I don't have another battery to force the meme Sanyo dumb charger to work, give it an hour or two, back at a comfy 0.3v again nice nice, back in the smart charger, 0.8 looking good looking good, 0.9! 1.0!

0.3
NULL

okay. 22 hours it is.

>>2726960
Give it a kick with a gas grill igniter. lmao

Learning electronics, made a 12/0/-12 power supply. Is there something that could be improved or something absolutely retarded in this circuit?

>>2726986

Since garbage is collected soon, I'm doing small cleanup. I have two old-ass PCs that I could throw away. Cases go for sure as they take lot of space and are empty, but I'm salvaging shit like drives for motors and lasers, fans, PSUs because it's easy to find some other use for them. I have problem with motherboards and electronics like RAM, chips or CPU though: is there any nice use for them? Or should I rather ask in microcontroller general?

>>2726986
>something that could be improved

it's 10x more useful if it's variable.
pic is dual polarity, variable, 0 to 30V.

>something absolutely retarded in this circuit?

all the diodes are wired wrong.
prob you wired them correctly, but drew them wrong.
coz, as shown, they've would've smoked and died.

>>2727001
Yep I only made the schematic for the post and fucked it up. If I want to get +5V too, does it make sense to add an L7805 to the positive side or use Zener diodes to drop the 12v output?

>>2726991
>is there any nice use for them?

you can pull CPU + RAM + cooler if and only if you intend to repair some similar computer in the future.

useful silicon: the BIOS (usable as an erasable memory), and the FETs used for generating various voltages.
other: coin cell, coils next to FETs, hookup wire, connectors (audio, USB, PS/2)

>>2727002
>add an L7805 to the positive

yes
you can get 1A current but you'll need a big heat sink.

>or use Zener diodes

no.
not enough current capability, and weak regulation.

>>2726987
yes, the rectifiers are wrong in your schematic

>>2726986
>something that could be improved

one more thing.
dont use TO-220 packages on your breadboard.
coz it stretches out the springs and leaves them permanently loose.
i solder a short length of #24 wire to each leg.

where can I get those large, flat lithium cells that powerbanks use

>>2727012
correction: where can I purchase them without having to take apart a powerbank

>>2727009
I've noticed some of the springs are fucked on the boards in the picture but they're cheap 1e/piece so I don't mind. I have a better quality breadboard too and it's been fine with TO-220 jammed in.

>>2727012
keywords: high density prismatic cells

>>2727013
>where can I purchase them

everywhere.

>>2727016
fuck me running, I couldn't for the life of me get them to show. Just specialty custom-order battery packs for things like model planes

>>2727004
>you can pull CPU + RAM + cooler if and only if you intend to repair some similar computer in the future.
There is no use for CPU like it is for other chips that can be reprogrammed? No use for RAM besides scrapping gold from connnector rail?
Also what's wrong with cooler? Fans are piss easy to find use for and heatsinks... well, I don't know where I'd use heatsink.

>>2727019
>There is no use for CPU...

note i said ''intend to repair''
which meant: keep these items only as replacement parts for repairing similar computers.
you failed at reading comprehension today. slap yourself on the back of the head.

>>2727021
It's 20-25yo hardware, it's hard to find such shit today so I doubt I'll ever want to repair something like this. I'm seeking for alternate uses.

I wish kicad allowed more "modular" designs. I have a bunch of basic circuit blocks I use all the time and I would love to be able to just hyper-optimize each circuit block independent of one another. Then I could just import whatever I needed into any given project and have a corresponding "optimized" PCB layout without having to individually route the same circuit every fucking time. Then if need be I could custom tailor the boilerplate layout to fit whatever I needed afterwards

>>2727012
Huaqiangbei, Shenzhen, China.
Just search on aliexpress. Li-po size. But thing with lithium batteries is that shipping is always a mystery, since chink doesn't have un38.8 certificate for cell and has to wrap it in resistors so xray cant see shit, so if you can find local re-distributor it is better.

I bought a soldering iron to fix a trip conputer display in my old jalopy and I was met with this fucking microchip tha needs a resolder. How do I do this? I don't have hot air nor any soldering experience aside from 6th grade 17 years ago (I made a door buzzer and reaction 'duel' game. We even etched or own pcb's for the latter which I think was hella neat).

>>2727092
>How do I do this?
shitload of flux + bevel tip
drag solder technique

To remove the chip you use hot air.

>>2727100
Holy shit that is nothing short of magik.
I'll grab a bottle of flux (only weller lw25 available locally, best be good), the brass tip cleaning thing and a roll of braid when I come home from work and try it out.

>>2727107
Sounds good, Anon. The only other advice is to keep the iron moving and use light pressure. If you have trouble, let the chip cool down for a couple minutes before you try again so you don't de-laminate traces or kill the chip.

>>2726575
found a kind of shitty way to estimate the volume resistivity: using a layer of packing tape (0.1mm thick), into which a stencil is cut, then a straight edge drags the paint over it to form a pretty consistent layer. Doing this, my latest formulation calculates at approximately 1.1 ohm-cm, which is pretty good if I do say so myself. Sheet resistance measures ~28 ohm/sq when applied by brush.

>>2727126
Actually I was wrong, my tape is 0.6mm thick. this calculates out to approximately 0.66 ohm-centimeters, which puts it on par with MG Chemicals' 838AR carbon paint.

>>2727133
Is it tempera and graphite with a laquer finish?

>>2727142
i read this as tempura

>>2727145
cram it weeb

>>2727145
Have some panko.

>>2727142
it's a mixture of black dollar store paint, which is itself slightly conductive, likely due to the use of carbon black as a pigment, [additional] micronized carbon black, micronized graphite flake, and glycerin.

I, for one, order my resistors with a shot of black

>>2727172
Interesting. Try mixing some thermochromic pigment in for shit and giggles.

>>2727186
or maybe a layer of thermochromic after the conductive paint dries.

To record only water for ten days

>>2726007
update: i ended up putting two LEDs in series on this heatsink. without a fan it gets quite hot at just 500 milliamps, 40-45 C per the thermocouple i taped to the hetasink. i tried cranking it up to 900 milliamps (rated spec) and it hit 55 C before i shut it down. pointed a small tabletop fan at it and now it sits at a comfortable 37 C @ 900 milliamps.

>>2727199
did you use thermal paste?

>>2727201
thermal plaster, but yes

>>2727202
Isn't thermal plaster an insulator?

>>2727206
i don't think so. it's not as thermally conductive as thermal paste, but it is used on consumer electronics e.g. motherboards used to have thermal plaster on north bridge heatsinks.

>>2727092
You can also use chip quick removal alloy, but it's kinda expensive. You only need a tiny bit.

hey fellas I want to buy a nice soldering station for my dad's christmas present, any rec?
last year I got him a nicer soldering iron (picrel, it was reccommended to me by my ex electronics teacher) and he liked it, any chance there's a soldering station where he can plug it in and discard the probably worse iron that comes with the station?

>>2727405
How nice are we talking? If you're rich, get him a JBC station. If you're not, get him a T12. I have a Pinecil V2 and it's my favorite iron, but it's not a station.

>>2727406
nah not a 200 dollar thing, more like a 30-40 bucks, I'm also getting him a bench power supply and we're strapped for cash.
yesterday he broke his soldering station and built this abomination, he swapped the soldering iron of the station for an extension cord so he could plug his nicer iron in, and since the board was busted he bypassed it so the only thing this thing does is turn on and off, the dial does nothing

>>2727411
You can get a Pinecil V2 in the US for $40, or from overseas for <30. It's a great iron, just need a USB PD charger and flexible C to C cable.

>>2727413
>>2727411
if you get him a pinecil, consider also buying an M2x4mm thumbscrew so he can quick-change the tip if he buys more than the one it comes with.

Just as my cheap ass was agonizing over having to spend on brand name tool just to drill holes to PCB I stopped by Lidl and found this USB charged dremel clone for 25e.

>>2727444
You still need carbide bits. FR-4 chews threw HSS like nothing else. The chinky bit sets with the colored rings are fine, they're just normal bits that have been resharpened after they snapped.
Consider building or buying a drill press like fixture for your rotary tool, helps massively when drilling.

>>2727444
Run it unloaded for 5 minutes to seat the brushes before you use it.

>>2727446
I tried with the included 1mm drill bit, it did make a hole relatively easily. I some bits for PCB from electronics shop, likely they'll be way better. Drill press fixture sounds good and I think I remember seeing some in price range fitting for the Lidl drill.

>>2727447
Thanks for the tip.

>>2727455
normal HSS bits work totally fine, they just dull after like 5 holes when drilling fiberglass. carbide stays sharp, usually the only thing that kills them is snapping, which is very easy with PCB drill bits because they're so tiny.

I built this on a breadboard after simming it. It's from the TL08xx datasheet (minor changes to component values to suit what I have on hand), sims okay but doesn't oscillate on my breadboard. Op amp is a LM741C, input voltage is 7.5V. The LED takes about a second to come on the first time, then 1/2 a second later dims a tiny bit, and sits at that intensity until I shut it off. Any ideas?

I've got:
- ESP 32 Wroom
- analogue microphone
- piezo speaker (it says 927~9 on it, is that the impedance?, nothing else on it)
- breadboard

Can I make a shitty walkie-talkie, granted I have an unlimited supply of the above?

Talk to me like I'm a retard

>>2727462
What's the oscillating frequency on sim? That configuration may be acting as an integrator unintentionally and pulling the output down. Could be wrong, trying to understand how the oscillation is actually being generated here.

>>2727464
>>/mcg/

>>2727465
Approximately 3Hz in sim, but it should be 1Hz by the formula in the datasheet.

I've got:
- ESP 32 Wroom
- analogue microphone
- piezo speaker (it says 927~9 on it, is that the impedance?, nothing else on it)
- breadboard

Can I make a shitty walkie-talkie, granted I have an unlimited supply of the above?

Talk to me like I'm a retard

>>2727469
well fuck me, phoneposting is a pain

>>2727469
>>>/mcg/

>>2727468
get a voltage reading on the breadboard, my best guess is that its reached some sort steady state on low and won't hit the lower voltage threshold to actually switch.

>>2727468
>>2727474
This section of the datasheet I think is relevant, that 0.2V differential. Maybe try a potentiometer going into that negative input on your op-amp

>>2727478
no change from 10k down to 8 with a pot configured as a rheostat replacing the 3k3 resistor

>>2727480
sorry anon, op-amp circuits were never my forte
>t. CompE anon

>>2727411
Buy him chinese station with hot air and iron. Maybe also set of tweezers.

What is the most ghetto lab bench power supply one can make?
Im thinking of using random chinese DC-DC module (I think it is based around XL4015) and 19V brick.
Which means I can probably get like 1.25 - 17V @ 3A+
But I am kinda concerned about filter capacitors on the output of dc-dc, wont they burn stuff?

Think there's a market for op amp testers? little box with ZIF sockets that can test any DIP/SOIC 1-2-4-channel op amp for basic functionality, but not things like GBW or slew rate?

>>2727520
The absolute most ghetto is a normal ATX power supply converted for bench use

>>2727523
But it is just fixed output and no current limiting functionality. And altering circuitery of ATX PSU is too spicy.
After all, with proper lab bench u can charge li-ion sausages with CC_CV and idk, test leds to find Vf vs I.

>>2727523
This. Should at least fuse the 12V and 5V rails, but better to buck/boost it with a DC-DC module like pic related. Could also use several XL4015s.

>>2727525
>But it is just fixed output
that's what LM317 is for
>no current limiting functionality.
fuse

I can power a MCU via usb or from a L298N motor control board. Both supplies are 5V but there's no guarantee that both 0V are actually the same. Currently I use DIP switches I have to manually open if both are plugged in. A mechanical solution would be to move the dip switches with the usb cable. Is there an electronic solution?

>>2727530
>fuse
Each fuse is like 1 cent. Unacceptable.
>that's what LM317 is for
LM317 is rated 1A and has insanely high drop-out voltage or like 2.7V IIRC?
You're much better of getting couple op-amps and making your own control circuit for the transistor. But it still would be linear, but now you'd be at least able to pick 3.3V line to get 1.8V instead of 5V

Other flaw with ATX PSU is that it is just 12V. Need more than this.
>>2727528
Hmm, maybe i should have got the buck-boost converter instead of buck, but whatever order is placed, still would be better than whatever i have right now

>>2727520
Ghetto? Pic related circuits with 19V brick.

>>2727533
>Is there an electronic solution?

a simple solution is to add a diode in each path.
load will get power from whoever is supplying a higher voltage.
Schottky diodes will drop maybe 0.3V to 0.5V.
if that's too much, you can use P-FETs in place of diodes.
they drop essentially nothing.

>>2727462
>It's from the TL08xx datasheet
>Any ideas?
Look closer at supply voltages. Notice how they use +15 and -15 V supplies. In their case + input moves from -13.3V to 13.3V. In your case LM741 needs at minimum about 1V of headroom so it can't really pull to ground. This means that your thresholds move from +6.66V to 0.888V. Now here is the main problem. How are you going to reach 0.888V if output can't go below 1V? It can't discharge the capacitor to cross over 0.888V threshold so it stays stuck at some low level output. You'll have to change the circuit (cont.).

>>2727548
Pic related is designed to work with single supply. Let's say opamp starts up high at 7.5 V. Voltage at + input is at 2/3 * 7.5V = 5V. Now capacitor starts charging until - hits 5V. When that happens opamp slams to ground (or as close to it as it can). Now divider changes to 1/3 ratio and sets bottom limit to 2.5V. Once that limit is hit it once again turns output high and sets limit back to top 5V. This makes the circuit oscillate.

>>2727540
okay, this is more ghetto than i expected lol

>>2727446
> FR4 chews through bits
FR2 is is paper instead of fiberglass, and good for 99% of everything you’ll ever do, and won’t kill HSS bits.

Wear a mask when drilling with a high speed dremel, you don’t want micro-fine epoxy or phenolic resin in your lungs.

A lot of people are faster freehand instead of a drill press.

>>2727594
>A lot of people are faster freehand instead of a drill press.
faster sure, but you break a lot more bits.

>>2727528
you’re better off with at buck-only converter on the device you are powering or at least close to the device with short leads in a separate box. the long leads are the 12 V into the atx psu. especially with something like 3.3 volts due to the voltage drop at low voltage. this is my modus operandi.

>>2727530
> lm317
again, separate box. based solution.

>>2727525
> li-ion charging
I think that’s nuts, I charge everything with actual chargers normally, because you know your wife is going to call you to fix the new dent she put in the car and you’re going to remember you left them to charge for a month and now you’re afraid to go look.
Recovering a cell that won’t charge (Ni-mh/li-ion etc) just requires a resistor and it will eventually get it back up enough to charge.

>>2727533
Jerri E had a short bit on reverse polarity that i thought was good.
https://www.youtube.com/watch?v=QZ0WXQWND-I

>>2727541
I suspect I will get 8V if the two voltage sources in pic related happen to be equivalent to 5V and 3V in series.

>>2727521
never had a problem with op amps but maybe the chinkshit buyers would appreciate

>>2727606
>I think that’s nuts, I charge everything with actual chargers normally,
Normal charge is all good for 1x18650, but when you're trying to repair an e-scooter battery which has idk, 4-5 cells in parallel it is annoying, as idk. Normal charger for li-ion cells, let's say Liitokala lii-500 is limited to 1A, some shitty TP4056 board is providing 700 mA... this is too slow.
Other usecase is when you have some device, like a phone, that has dead ass battery at idk, 2.3V and no charger would charge it, and u need to bring it up to 2.8V so built in BMS accepts charge, and you don't have time to wait to charge it properly.. What you do? Push all PSU can suffer for 1-2 minutes lol. at 5A? 10A? All she would suffer amps :D
And main benefit of proper PSU is that you don't need to use resistors (which u dont even have because who needs 1 ohm resistor), or idk, find 317 regulator and make it output CC-CV...

Problem is that I don't have the lab bench PSU. And stealing one from uni doesn't work because they knew students would try stealing shit, so they have picked the biggest and heaviest linear PSU.
But what I have are 19V chargers... I stole from uni recycling bin. And there were some OEM chargers and only fault they have is broken cable. I tried dicking around with feedback circuit, but after exploding 2 output capacitors, and then an input side circuit somehow, I decided that it would be much smarter to just buy a DC-DC converter, which would be cheaper than a proper heatsink, power transistor, bunch of op-amps and circuit board...
That said, I'm pretty sure chink DC-DC converter sucks, and produces fuckton of noise on the output, and smartest idea would be to add linear regulator that drops just idk, half a volt on average?, so output is smooth and noise-free, but that's too much effort for something that would be used to push insane currents into batteries and LEDs.

>>2727520
That's similar to what I've done. I have a panel meter and an LM2596 module with its main two CC and CV pots desoldered and replaced with panel-mounted pots. The module itself is just floating around in a box. It works fine, but those modules aren't really rated at 3A in a box without heat-sinking, doubly so if they're the fake chips that oscillate at half the frequency.

>>2727521
Yes I'd say so. Considering you'll probably want to have an MCU doing the probing and measurement, it would probably not be too much trouble to make it measure slew rate too, at least for audio-speed op-amps. GBW would require fast ADCs and DACs.

But the hardware wouldn't be trivial. Even if you have a different IC socket for each pinout style, and some way of measuring all the op-amps in a given chip, you'd have voltage levels to care about. Many RRIO op-amps can't handle a supply of more than 5V, while most JFET-input op-amps won't work properly with a supply voltage less than 9V. So it would also need an MCU-controlled DC-to-DC converter, plus the DAC and ADC would need to be able to scale up to the higher power rails, probably using higher voltage RRIO op-amps but that complicates the measurement process by adding its own nonlinearities.

>>2727541
That simple gate-to-ground PFET circuit will work to prevent reverse polarity, but it won't stop reverse current between two positive supplies. For this you need a comparator-driven "ideal diode" PFET circuit.

That said, in this instance the drop and leakage of a schottky diode shouldn't prove an issue.

>>2727548
is there an explanation for the 3.3k resistor? I don't get it.

>>2727560
Yeah this is almost certainly the problem. A schmitt trigger circuit needs the positive feedback to be pulled to somewhere between the two power rails. Common trap for young players is to not recognise the difference between split-supply circuits and single-supply circuits.

>>2727617
Any reason you can't just short the two 0V rails together? If they're not floating and they're not equal, how would they vary with respect to one another?

>>2727633
>is there an explanation for the 3.3k resistor? I don't get it.
In general you would use resistors on the input to either
>balance input bias current (not applicable here since it's a JFET amp)
>limit current through clamping diodes present on the inputs
You must consider what happens when you remove power from the circuit. Since you have a large capacitor it will hold charge even after you disconnect power. If you have capacitance on power rails it can happen that it will discharge faster than timing capacitor. If that happens then charge from timing cap will flow to power cap. If too much current flows it could damage clamping diodes, so it's best to limit it with resistor.

>>2727092
>desolder
Wick. There's some guy on YouTube (MrSolderFix?) that shows bending metal rods laid across the pins, solder added and then heated up. Looks simple but who knows how many attempts it took.

Testing random scrap LG 18650
>resistor + excel sheet
500 mAh, discharge rate 2.3A ~ 1.6A
>liitokala NOR test (500 mA discharge)
1700 mAh+...

What the fuck. I thought those cells were trash, but they apparently still have the capacity. ESR can go this high? Or Liitokala lies

>>2727633
Shorting the 0V rails might break things? Pic related would be infinite current to bring the 3 down to 0.

>>2727637
>limit current through clamping diodes
In the TL08x there aren't any such diodes, and its P-ch input JFETs can survive having a higher voltage than their other pins without any problems. Though with the capacitor to the ground rail, it's quite possible to get the input going lower than ground in the original split-supply schematic I guess. Not an issue with OP's single-supply circuit.

>>2727703
Yeah but why would the power supplies be connected in series like that in the first place? If you're not planning on connecting them in series, then it doesn't matter. If they shouldn't be connected like that but it's a faint possibility, connect the grounds together and use a fuse or three. If you do need them to be possibly connected together, then use isolated power supplies or isolated DC-DC converters.

>>2727762
Thanks I will connect the grounds but limit the current with a fuse or one of the current limiting ideas in >>2727616 's link.

>>2727697
Not even the li-ioooon shills will tell you true, but I will now.
They make different li-ion cells for different purposes. They make high-drain, long self-discharge times (but low drain) and high-cycle (but low capacity) cells.

Over time, there has been thousands of SKUs for all these things across different brands. Like when another company steals panasonic’s dialectric astro-glide, panasonic eventually gets around to test it, files lawsuit and they get import banned, and the company switches up to use sony’s dialectric KY.

>>2727624
I opened a battery pack from a scooter and it had like 8 leads coming out of it. I wouldn’t know where to put the supply on.

Also, it sounds like what you really need is more time.

>>2727776
>They make different li-ion cells for different purposes. They make high-drain, long self-discharge times (but low drain) and high-cycle (but low capacity) cells.
This is neither.
Originally it should have been 2.6 ah I think, and with max discharge rate of 2C. And I didn't exceed it, and I would expect a bit more capacity with fast discharge.
Sanyo cells (trash too) they measured 1.6 ah with my test, and 2 ah with liitokala. Which is reasonable I guess, because they are 2C discharge rated too.
I think it is just old abused shit with ESR measured in ohms.

Other thing I also measured e-scooter cell (SINC 18650) and it measured 2 ah instead of 2.6 and idk why, maybe i did bump it and charger decided that it fit the floor. It has the shittiest datasheet ever, on website they tell you it is 3C discharge rate, but datasheet (which isnt published on their site) they tell u it is 2C cell.
As for sanity check i decided also to test LG M26 cells I have laying around, they should be 10A discharge and 2.6 ah.

Why the fuck did I buy liitokala again, i will test all cells in my house out of autism now.

>>2727782
Follow wiring into the BMS. Count number of cells and total voltage.
36V battery should have 11 cables going to BMS, 1 ground and 10 positives. So in order to fast charge it, you get paperclip into the connector and push amps. Or blow up a charger lol, that's what i did. With lab PSU that would have not happened. Probably.
>Also, it sounds like what you really need is more time.
Time is limited. Faster is better.

>>2726987
your schematic has the primary shorted and fbr all fucked up.

Is there such thing as a "potentiometer multiplier" circuit? I'm getting tired of having to buy 10k, 1k, 20k, 50k, 500k, 1m pots etc. for every meme project.
Is there a way to convert a 10k pot into a 1M pot using BJT's or something? Like a VBE or capacitance multiplier sort of thing?

>>2727775
>limit the current with a fuse
you can't limit the current with a fuse, a fuse is a protection component, it will blow and then you get no current at all. If you need an actual current limiter, a fuse is not it. If you do have a current limiter circuit, a fuse might also still be appropriate (it always is)

>>2727812
the real question is why would you need multiple values of potentiometers in the first place? If all you need is a reference voltage, then you could try a simple buffer.

>>2727787
Okay, new liitokala is measuring capacity correctly. I have bunch of shitty 700 mah cells and they measure 700 mah.
So I guess yes, that LG cell is just on its last legs, and cells may die not only by losing capacity, but also going high ESR like capacitors.

Any tips for desoldering through hole components? Ive been trying the copper braid saturated in Flux but I can't get all the solder off the pins, there's always some left that I just can't get

>>2727835
Use flush cutters or fingernail clippers to clip the leads then de-solder what's left, or get a Soldapullt.

>>2727835
For 2-3 legged components, just heat it up and pull.
Everything else - solder sucker, braid, stainless needles (important not to use any flux when using stainless needles), even hot air.
Alternatively, you can cut of legs and remove 1 by 1 using tweezers.

But the end solution is not to use through hole and stick to SMD. I have got allergy for through hole components when i was making PCBs at home, and now, after discovering out chinese PCB manufacturers, I still avoid through hole components where possible, sometimes I even surface mount through hole components lol.

>>2727835
Try a solder pump instead of wick if you have it. Add more solder before pumping it out, it helps.

>>2727839
THT is good for point-to-point prototyping. SMD is aids without an etched circuit board.

>>2727843
Yes, but it is more fun to design circuits on paper and test with SPICE instead of prototyping IRL.
Also trying to point-to-point with SMD is good for your nerve system, if practiced a lot, your hands would shake less and eventually you can consider becoming a surgeon or dentist or something.

This might be a retarded question, but is there a cheap, common type of cable/connector with at least 7 conductors you’d recommend for transmitting 5v signals at around 200hz over a 5m distance?

I’m trying to hook up some light controllers that have 7 GPIO pins on them, but they’ll need to be physically connected/disconnected frequently so I’m trying to decide on a connector.

I was thinking maybe Cat6 Ethernet cables but I’m not sure if the twisted pair configuration would fuck up any signals as all references mention using a balanced transmitter and receiver. The signals are 1ms pulses evenly spaced at 200hz, with random control pulses a few times a minute.

I will Regret this Purchase

Do any of you guys use any scientific FEM or field solver tools for circuit board current or heat or RF analysis?
I don't know much about these tools but they seem like a cool thing to learn
I'm working on a capacitance based proximity sensing at work, any way to simulate the sensing without having to manufacturer the parts will be appreciated

>>2727881
>I will Regret this Purchase
That's like 99% of electronics purchases

>>2727881
Maybe. There are fake TP4056 that barley can handle 1A and don't charge correctly (CC-CV? nope, 10% cutoff? nope)
>>2727887
>Do any of you guys use any scientific FEM or field solver tools for circuit board current or heat or RF analysis?
For PCB heat I use this
https://www.heatsinkcalculator.com/pcb-temperature-calculator.html
It is somewhat pessimistic, but was OK for an LED light board.
You can also try solving copper dissipation manually using magic of numerical anal isis, but it gives me PTSDs.
Probably there is a plugin for freecad, but it doesn't account for emissivity iirc. And emissivity is important, which is why you want to hide your cooling polygons under solder mask.

>>2725172
There was an anon in a robotics thread trying to make an autonomous lawnmower, but the thread has disappeared.
I'd like to attempt similar.
I need a way to regulate the speed of a 2x 24V motors (for turning).
Can I use a GPIO pin on an microcontroller to do this? I'm thinking something like DAC pins to the PWM modules to control the wheels.
Is this a sensible way to do it?

>>2727914
Controlling motors is tricky. You need to limit current and speed. Even for simple brushed motor.
Just PWM won't work, you have to understand what frequency to use in order not to exceed current capabilities of mosfets.

>>2727914
If you need to be able to reverse the motors (you almost certainly do) then you'll need to find H-bridge motor drivers that suit your motors at your battery voltage. Then you'd just PWM those drivers directly from the MCU. As the other anon says, you'll likely want some sort of feedback, so I'd want to pick a motor driver that has built-in current sensing, preferably automatic current limiting as well. If you're using brushless motors, then most ESCs will have those features built in, plus they'll naturally be able to tell your MCU each motor's speed. You can technically measure the speed of a brushed motor from the current ripple frequency, but it's a really noisy signal so if you're going brushed I'd recommend having sensors (hall effect or optical) for speed feedback. Sensors on brushless isn't a terrible idea either, especially if you're operating the motors without a gear reduction.

>>2727914
PWM on the IN1/ IN2 of L298N doesn't slow a 32GP-31ZY much unless the frequency is low (say 200Hz) which is noisy and probably bad for the gear/motor life.

>>2727762
>In the TL08x there aren't any such diodes

>>2727887
It's probably going to be too much effort for what you'll get out of it. Problem with simulations in general is that you need to calibrate them so that results aren't just a pretty picture on the screen. If you still want to play with this you can try using student version of ansys or go FOSS route by watching videos on
>https://www.youtube.com/@TuxRiders/videos
There open source tools are mostly just PDE solvers, so some math is required. You need to re-write your problem from PDE to weak formulation. Elmer has some pre-configured equations and GUI tools so it might be easier for beginner.
>http://www.elmerfem.org/blog/

>>2727835
it all depends on the component and its pin layout. also depends kn what youre tryinf to achieve. are you removing the part to replace it or are you salvaging/dismantling. but in general you need massively high heat and a tonne of flux. you gotta break down oxidisation of shitty lead free solder thats most likely connected to a huge heat sinking copper plane.

or, you could use a heat gun set at 450 C, heat it till its all liquid, then bash the pcb on your desk and send the part and molten hot solder everywhere. this is the patricians choice. make sure you got safety glasses on though. i do this same thing again, except with an iron to clean out the holes for a new component.

>>2727876
usb 3.x cable. use shielding for gnd you dont habe room. you have 5 + 4 pins available for data assuming you dont need a vcc. even then its 4+4

>>2727980
oh
but why?

>>2728048
I think they're meant for ESD protection. If there were nothing there FET would just get destroyed if you touched the circuit after walking on carpet.

>>2727983
>It's probably going to be too much effort for what you'll get out of it. Problem with simulations in general is that you need to calibrate them so that results aren't just a pretty picture on the screen.
Ngl that's half the reason why i wanna try these tools, to see pretty colours on my screen and feel like I'm smart

I've been playing with arcs and long story short, I want to make hotter arcs. However, I don't want to invest in a more powerful hv supply. I have heard before of circuits that use a high voltage low current system to strike an arc and then activate a low voltage high current system once the plasma reduces the resistance of the air. My question is how do you go about this? Could you use diodes to stop the 10,000v system from destroying the 28 volt system without blocking the 28v system? Do diodes with that kind of operating range even exist? I have diodes to make a rectifier for the hv so everything is DC now.

>>2728072
Use an old cfl ballast.

>>2727941
why don't people just use DC control?

I mentioned earlier that my monitor power supply was acting up so I replaced it with another one, same thing: it works for a while then turns itself off. What's more likely, I have two bad power supplies or the monitor is doing something they don't like? Both are rated 3-4A which is way more than the average monitor would ever draw. And it is unlikely that a perfectly working monitor would produce momentary spikes in current consumption and suddenly draw > 3A instead of a few hundred mA, or somehow momentarily shorting the power input, right?

>>2728168
Either bad caps in the monitor or dying backlight.

>>2728172
*most likely

>>2728172
could either of these result in a short circuit seen by the power supply?

>>2728178
Caps can fail short or near-short. Is it a LED or CCFL backlight?

>switch out dead fans from rtx3070 to 120mm ziptied ones
>other one is not spinning and reports ridiculous numbers like 750000 in hwmonitor
I checked the wiring and it's identical to the spinning one, possibly a bad contact in pwm/tach line? help me anons, /g/ is useless as usual.

>>2728184
Swap the fans first.

>>2728139
You can use DC control, but you will waste a lot of energy as heat.
What is the power of the motor?

>>2728186
goddomnot, more soldering. well, I won't waste heat shrink until I make sure it works this time. at least I have extra fans if it's faulty.

>>2727876
This was common place a while ago.
Printers used parallel ports. Also the “Centronics” connector, which is still in use today, but not for printers.
My thrift store usually has about a dozen of these.
5 meters is a bit of a stretch though, but I have run them from front desks, up through ceilings and down to a back office where the printer is just fine.
This was in the 80s.
You can get USB to parallel ports too, but I don’t know how they work exactly.

>>2727844
> More fun to test with SPICE instead of building
Wat?!? First time I’ve ever heard anything like that.

>>2728168
>Both are rated 3-4A which is way more than the average monitor would ever draw

wrong.
my 24-inch Dell, for example, is rated at 110W.
assuming your supply is 18V, it would draw 6.1A.
https://www.displayspecifications.com/en/model-power-consumption/ffd8759

>>2727876
I'd recommend rj21 because ypu can get board mounted connectors and also make you're own professional looking cable with a rj21 crimp/splicing tool.

>>2728205

>>2728180
No idea, it is an old Dell monitor but it is as thin as modern monitors so I am guessing LED. However it is odd that it has an external 12V power supply which is rare. And it works fine for hours until the power supply turns itself off. If the capacitors failed, how could it even power on?

>>2728241
Is the power jack intact? Try re-flowing the pins, and check the power button too.

>>2728072
build a chamber to arc in and flood it with argon (less expensive) or helium (hotter)

look at these little 6.3x20 100uF 63V caps I got, isn't it just the cutest thing?

>>2727876
HP-IB/GPIB was designed for this exact sort of thing. You just have to decode/encode the wire protocol at each end.

>>2728291
Those look tiny for the capacitance. Sure you didn’t get chink’d?

>>2727941
>>2727957
>>2727972
Thanks for this. So the search term I was after was "motor driver".
I haven't worked out what one I'll go for yet.
The mower won't be big - it'll likely be 2 x 12V LiFo batteries, so not too heavy either (does 250W motor per back-wheel sound like overkill?).
I'll probably go shit brushed-motor setup for now (just because it's easier/cheaper to setup and allows me to experiment with approach).
Longer-term, the intent is to actually train an ML algo for it to drive around, obstacle detect and find grass it can handle. GPS will likely set hard-boundaries.
>for training data, I'll probably RC it with a camera and gyroscope
Most difficult thing for me is working out the motor for the blades (and how to mount). I suspect that needs to be quite powerful (most mowers apparently run at ~3000RPM).
Other autonomous mower-anon was driving blades with a belt - and I'm thinking that's probably the safest approach.

>>2728341
You could use 2 front-wheel hub motors for propulsion. A 48V system would be better so you could fit an off-the-shelf electric lawnmower motor.

>>2728341
>The mower won't be big - it'll likely be 2 x 12V LiFo batteries, so not too heavy either (does 250W motor per back-wheel sound like overkill?).
250W is enough to move 100 kg at 25 km/h. So it is kinda an overkill, but you can use hoverboard motors which are dirt cheap.
>Brushed
Now you're adding complexity of some weird chain drive or idk. gearbox.

>>2728291
>63V
>100 uF
Did you test them?, this shit might just be fake.

>>2728412
>>2728319
They're made by Ymin, a relatively well-respected company as far as chinese caps go. Arguably better than Rubycon.

>>2728341
Reel mower blades cut at low speeds. A vacuuming robot might be an easier less risky first project.

>>2727525
>But it is just fixed output
3.3, 5 and 12V should cover most of your use cases. You can also get other voltages if you connect between the rails.
>>2727530
No need for a fuse, any self respecting ATX PSU has built in overcurrent shutdown.
>>2727537
>it is just 12V
You can get 24V if you connect between +12V and -12V rails but you're limited to 1A.

>>2728052
A bidirectional TVS setup makes more sense for that. Or a unidirectional TVS diode from the input to ground, since the input shouldn't be sent below 0V. I think they're clamping the input voltage to the positive rail because its convention and they're anti-fun.

>>2728072
Have a look at a schematic from a high-frequency-start arc welder. The usually have a low-winding-count transformer in series with the output and drive a really high-frequency wave through its other winding. Old welders would do this with a spark-gap oscillator, which is naturally high-frequency and high-voltage, but your local hams will hate you. If you get up into the MHz you can get really long-range feelers. Lots of protection circuitry is a requirement to prevent voltages or currents going where they don't belong, but because of the big difference in frequency you can do most of this just with chokes and caps.

>>2728191
>he didn't use connectors

>>2728515
>You can also get other voltages if you connect between the rails
ATX supply rails aside from GND and -12V can't sink current. You couldn't get 7V by jumping between 12V and 5V.

Anyhow, a reliable fixed multi-rail PSU, a low-noise split-rail PSU, and a variable CC/CV PSU are all quite useful benchtop power supplies to have. It's more convenient if they're all the same PSU but it's hardly a requirement.

>>2728515
>No need for a fuse, any self respecting ATX PSU has built in overcurrent shutdown.
That overcurrent would trip at such high current most wires would melt. Especially on old PSUs where most of the power is on 5V rail.

This made me think, is it possible to modify feedback circuit of a welding inverter to supply DC :D
You know, it would be nice to have 70V max, 100A max PSU

>>2728568
>That overcurrent would trip at such high current most wires would melt
Yes, but it trips in milliseconds or less. Even really thin breadboard jumper wires have no time to heat up noticeably, let alone melt.

>>2728668
>it trips in milliseconds or less

not really.
i just checked my 700W unit.
the 5V rail is rated at 22A.
so, suppose short circuit protection kicks in at 30A.
it means it's perfectly happy as long as the load is more than 0.167 ohms.
the combined resistance of wires + connectors + breadboard can easily surpass that.
so you could have 150W of heat being dissipated continuously into your breadboard.
or onto a reversed diode, or backwards chip.
so, no fuse = you need the fire dept on speed-dial.

>>2728696
Nice real-world demo, Anon. I'm gonna build an easy bake oven now and bake some shitty brownies.

>aliexpress ICs come in
>apply flux so as to solder them to breakout boards
>label/engraving comes off with the flux

>>2728750
Hello friend our IC is real and not fake, problem is soldering skill, try resolder IC. No refund.
$0.00

Replacement panasonic caps off digikey test at 2.2 mF
Originals (aside from one which must be broken - untestable) are 1.0mF
Is this expected?

>>2728463
>A vacuuming robot might be an easier less risky first project.
You're not wrong, but I have no use for an automated vacuum.
Because the intent is that it's AI driven, if something goes wrong and the blade mechanic kills someone, it might be the catalyst that triggers the war between Humanity and AI too, which will be worth it imo.
>>2728358
>You could use 2 front-wheel hub motors for propulsion
Thanks, took a quick look at these, but a bit pricey to find where I am. I think my younger brother has worked on mounting motors to wheels, so I might see what options he has.
>A 48V system would be better so you could fit an off-the-shelf electric lawnmower motor
Looking at what's available where I am, most range from 20V to 40V. I should probably clarify on the use-case though:
We have a 42" petrol mower (Hustler) and live on acreage. We mainly just want to be able to keep the grass down around our living area (helps prevent snakes, etc), so it'll just be used to maintain that area and the long-grass will act as a barrier of sorts (hopefully recognised by the AI).
>250W is enough to move 100 kg at 25 km/h.
Didn't realize that wattage would translate to that amount of power. Even if I use 250W motors (would prefer less), will definitely dampen these (I want to keep it below 10KM/h.)
Postage becomes absurdly slow over Christmas here, but will start ordering bits and pieces and report back in another thread in a month or so.
Thanks for the tips everyone, appreciate it.

So I've been watching probably unhealthy amounts of tube equipment repair on youtube and there is really two main camps in that area that I can tell. Guitar amp repair and everything else repair. Boat anchor ham radio repair is cool but very niche and doesn't pay the bills. Techs usually replace most caps and overall troubleshoot the circuit to get good performance.
Guitar amps are a whole nother ball game. They'll literally try to keep using capacitors from 50 years ago because of the "tone". The users commitment to originality means they'll wait till their bias caps go and fry their tubes and still insist the tech do the least amount of repair possible to get it functional. I just don't understand it. One of the people I was watching was rare in that he insisted upon recapping the bias circuit of any amp he worked on. People still complained in the comments about it. Guitar players are literally worse than audiophiles.

>>2728811
Just lie to them. Put new caps inside of the old asbestos caps and charge them 10x-100x the price because it was a Herculean effort to find the old asbestos caps during these trying times of biowarfare-induced supply chain disruptions.

>>2728811
There are tests you can find on the internet that show this isn't actually baseless. It's because of the manufacturing process changes.

>>2728820
like actual scope/waveform tests? Because sounding "better" is subjective

>>2728822
Actual waveform test on the output, yes.

>>2728820
also it just seems silly for them to risk their NOS $$$ tubes and their unique tone for the little bit of extra tone that these old ass caps give them.
Kit amps are at least logical and you don't have to worry about amp collectors thinking you spit on Leo's grave for daring touch their 60s wavy eyelet board genius

>>2728820
I'm finishing up my prototype audiophile-grade capacitor recipe as we speak. The magic is in the internal oxygen-free RGB LEDs that brighten the tone by flooding your ears with rainbows, and the 7/0 gauge solid gold conductors with pure braided silver cable shield.

>>2728186
I had it wired wrong. Traced it back to the connector, used my head for a moment and resoldered. works 5/5 now.

>>2728785
capacitors usually have a high tolerance, 20%. 2.2 mF is a bit high though. I don't know if this will be an issue in your case, it might be if the inrush current is too much, i'm not familiar with flash circuits.

>>2728811
>get customer’s amp to repair
>characterise the frequency response
>characterise the distortion
>recreate with DSP algorithm
>replace guts with DSP feeding a class-D amp from china
>sell old tubes and caps to audiophiles for megabux
>profit
Problem?

>>2728844
> replaced everything with a simulation
It’s missing the point. The point is to do a minimum-effort fix.
I’ve had lots of sprague and mallory capacitors that still test perfectly fine—they’ll probably out-live civilization, so I only re-cap things that are absolutely necessary.
Usually, it means the capacitors were over-run so they got hot which can kill them in the long run.
Check out shango066. Although, if I were fixing old CRTs, today, I’d probably put one or two PC fans in them to keep the sets cooler.
> go to doctor with a mental problem
> replaces your brain with a DSP and a 16 Kb ROM.

>>2728785
> 2.2 mF ≇ 1000 μF
That’s crap, and would throw things like filters way off—farther than an adjustment trimpot would allow for.

Also, because new caps are smaller than their older counterparts, usually I go up one voltage (or whatever will fit, within reason), might as well get 400 V if it will fit.

>>2726957
this was dogshit so i bought a Singer 4C-316B pedal for sewing machines and rewired it. works great.

>>2728878
interesting side note: the new "electronic" speed control PWMs a triac to control speed instead of a rheostat, and this makes the motor sound different.

>>2728871
>> replaces your brain with a DSP and a 16 Kb ROM.
Id become smarter and have more memory. win-win.

>>2728785
Try pulling them up to 24-34V and holding them there for a few hours for dielectric absorption. Then measure the capacitance.

>>2728898
I'd be a lot better at vector calc if nothing else.

>>2728889
> triac
yeah, it’s the standard light dimmer control.
makes a lot of emf noise on the line, too.
eventually burns out any electronic thing you put on the load end, I’ve killed several wall warts, even transformer-based ones with dimmers. upgrade to 4007’s from the 4001s helps.

>>2728912
It works the same way in sewing machines, and they don't die.

>>2728924
> sewing machines don’t die
I’d argue the motor is ‘electric’ and not ‘electronic’
Likewise It won’t affect heating elements, incandescent light bulbs, charging batteries, etc.

>>2728927
Ah. My flex shaft pendant motor is an ordinary AC "universal motor", so not much to fail.

What's a good JBC compatible soldering iron. I can't take it anymore. Every time I use the JBC at work it's gold, and then I come back home and my cheapo iron feels like dogshit.

>>2729002
Get a JBC?

>>2729003
but it's like $300

>>2728912
Why would you put a triac dimmer before an SMPS or transformer? SMPSs basically always have an input X capacitor, which is going to mean big current spikes and so bad longevity for the parts involved. I'd only consider it if I had substantial normal-mode chokes before any X caps, even then there's no fucking reason to. Unless you're in a strange situation where you've got a 110VAC SMPS and only have 220VAC mains, and so tune a triac dimmer to peak at 170V or less, which is what some cheap travel adapters from india do. Transformers and induction motors also work kinda strangely on triac dimmers.

The only things you ever want to run on triac dimmers are brushed universal motors, incandescent lamps, and some other LED bulbs that are designed for them. Maybe heaters if you can find a triac dimmer that can reliably pass 17A without cooking.

>>2729002
Aixun T3A with the T245 handle. Aixun also sell some T245 stations with AC transformers instead of switchers if you care about reliability to that extent, but they're a lot more expensive. I found a listing for the control board and screen for such a station, so you could get that and put it in an enclosure with a plug on the back for a name-brand AC-to-DC adapter. Or a lipo.

I'd like to see a 200W pinecil with T245 tips and an XT60 on the back.

finally unsoldering the piezo buzzer from my chink bench supply
don't know what the chinks were thinking making it beep for every digit change

>>2729097
for penny pinchers, they sure do love their loud, shitty, unnecessary noises

>>2729005
how fucking bad are you with money that you're employed but can't afford $300

>>2729097
Audio feedback for squinty eyes. Real humans use pots or rotary encoders with detents. Bugs prefer piezo buzzers because it massages them through their exoskeletons.

>>2729104
the worst part is that is has rotary encoders with detents. They slapped the buzzer right next to them right on the front board.

>>2729103
It's 2023. Only top-tier jobs let you afford that kind of leeway after taxes and cost of life.
A cashier would not be able to even so much as afford to live.
A programmer with 5 years of exp would barely be able to save 500 a month.
A mech eng or EE makes half to 70% as much as the same programmer.

>>2729107
lmao
Just for the hell of it, look inside for a weight like a container of dirt or compressed radioactive waste with yellow tape wrapped around it so it looks more official.

>>2729112
nah it's all legit
wanptek

>>2729115
That's not bad at all. Wonder if they had overstock buzzers taking up too much space in the warehouse.

>>2729118
It's been pretty good. I did a whole review on /csg/ when I bought it from aliexpress. Load tested it and all that jazz. But the beep everytime you did anything drove me crazy.
Blew up the skytoppower beneath it being dumb and trying to run a cigarette lighter out of a mercedes on it. Obviously the current protection didn't work.

>>2729103
I'm unemployed. I'm a neet freelancer building a lab.

>>2729005
Steal it.

>>2729103
Not that dude but I net 500$ a month. I learned electronics as a necessity because a 5$ soldering iron and 10$ worth of components is infinitely more affordable than buying new shit.

Is it worth trying to save a soldering iron tip? I must have left it on and the tip has no solder on it at all and none will stick. I don't have any rosin to try and save it with.

>>2729155
If you like punishing yourself you could clean it up and electroplate it with copper, then nickel.

>>2729155
Plumbing flux. Brass wool
Avoid wet sponge

>>2729163
I use a sponge dampened with de-ionized water and it has worked out well for Hakko tips.

>>2729115
the chad form over function layout

>>2725172
What's the best tutorial to build a headphone amplifier? Should I use op-amps or BJT?

>>2729175
That's just ancient chinese philosophy of PCB design.

>>2729176
It depends on what your performance goal is. Amplifier itself is probably the easiest part. Making a good power supply is another problem. Making sure you don't mess up with ground loops is yet another problem. How will you feed signal into your amp? Do you want to use optical cables with S/PDIF and use onboard DAC?

>>2728820
Usually, defective caps sound "better" because it drives the amp hard into class A - the distortion may sound nice, but the overload will eventually burn out tubes, plate resistors, transformers, you name it.
From the factory, they sounded the same as being recapped with modern components. There are non-destructive ways to recreate the sound of strained equipment, but in true audiophile fashion, they're dismissed because the placebo affect tells them it's not as good.

I'm planning on replacing a battery of an old device that has 3.7V with 660mAh
However, there are no exact replacements. The only thing that matters is the dimension and the volts I'm assuming? Since there are no more batteries that are exactly 660mAh. There is 750mAh and a 1200mAh. I'm assuming the latter is better, but would that cause any damage to the device?
Also, the method of replacement will be to remove the new battery from the controller it comes with, and attach it to the device's controller, since the device's controller has three wires, and the replacement new battery has two (red and black)

>>2729011
> why triac dimmer before before transformer/smps
SMPSs don’t do well at all.
Off the shelf transformer-based PSs are sometimes good for years without modification.
Application: sealed high-reliability/longevity 12 VDC evaporation fans. They’re nominally 12 V on “high” but are dialled down to 8 V for “low” speed and controlled by a microcontroller, and they have a speed feedback wire. The fans are good for decades, but the electronics package is fucked up in so many ways it’s not funny, so I put them on manual speed control with a wall wart and a dimmer. The fans normally run about 90% of the time anyway.
When you replace the diodes and get a bigger and higher voltage capacitor for the “wall-wart” (we don’t really use wall-warts) it’s good for another 20 years.
One day, I’d like to build an filter module to take the spikes out completely, I’m sure some spikes are still getting through. Maybe add a smaller capacitor to catch the higher-frequency spikes and perhaps an inductor.
I’m sure this would work on PC fans as well. I’ve tried it on similar fans, but only with two wires as they look like PC fans but are in a magnesium housing.

>>2729262
> 1200 mAh ok?
Probably. They lie about the mWh and mAh rating all the time anyway. Plus it *starts* out as good, and then goes down to 20% of it’s initial capacity rating until you get annoyed enough to replace it. As you are at now.

There are *some* devices that rely on the internal resistance of very small batteries to limit the current but these are in only the cheapest-of-cheap devices. e.g. Flashlights.

I often replace those pouch cells with bigger 18650s and everything has been fine.

>>2729267
Alright I guess only the volts and the size (dimensions) of the battery matter in this case in a way. To be more exact, it's a Zune HD, the lithium battery bloated out, and I looked around and found a leddit thread of modders. It's pretty impressive that people have been modding these, even making a bluetooth module for it.
The battery the guy used isn't for sale anymore, but I guess getting the one that has a higher mAh will work just fine, as he used something higher than the initial 660mAh.
I just don't want to damage the entire board, since they don't sell them anymore.
>until you get annoyed enough to replace it. As you are at now.
Weirdly enough, this device lasted a very long time, it's just been sitting in the car all the time, and maybe the sun had something to do with it recently
>devices that rely on the internal resistance
Looking around the thread I found, they also added a thermistor, I don't know if that's gonna do any support, but I might have missed something while reading it.

>>2729270
Batteries that fit will have a slightly higher mAh rating because of advances in battery technology.

The thermisistor would probably be for “fast charging” so the cell doesn’t get too hot. You can solve this problem by limiting the charge current and/or trickle charging.
Some devices—especially older ones—can charge on less than the 500mA they get through a standard USB so you could limit it there.

What kind of gear does /ohm/ have?
Power supplies, scopes, soldering irons, etc

I know very little about electronics.

I have a shoe druer (blower + heating element) that has stopped working. Ive taken it apart, but i cant see anything burnt or broken.

The circuit seems very very simple, with no components other than a pair of diodes (1n4007 I think) put in series.
How do I check if the diodes are broken? When measuring with a multimeter, the resistance is either "infinite" or 698 ohm, depending on the direction.
I haven't been able to find whether that value indicates whether the diodes are broken or not.

>>2729277

>>2729275
>Power supplies
None yet, but I will make one using DC-DC buck and a laptop charger
>Scopes
None. Thinking to buy "toy" 200 KHz stm32 oscope from ali, while i search fb marketplace to score a good deal on anything more real
>Soldering Iron
Chinese Hakko 900 clone + hot air.

>>2729275
>Power supplies
Oltronix B200 linear 40V 3A
>Scopes
Fluke/Philips PM3384 quad channel with memory expansion and math co-processor
>Soldering iron
Some $5 shit from local electronics store
>etc
Philips PM2534 6.5 digit true rms meter

>>2729277
What does this plug into? AC? It only has two diodes so it might as well be running in a half rectifier which is super cheapo. Where do the wires go?

>>2729286
It plugs directly into AC.
It is, however, no longer a mystery why it isn't working. While dicking around with the multimeter, trying to measure a bit of everything, I accidentally shorted the circuit, resulting in a bunch of Blue Smoke getting released.

So it sure as fuck isn't working now, even if I could find the original fault.

I had hoped to open the drier and see like a burned out resistor or a loose connections or whatever, so I could make a simple swap/resolder.

>>2729290
How did you short it?? Did you destroy your meter in the process? Anyways, given its only two diodes in that circuit you might get away with replacing them directly to some of the wires with adequate heat shrinking all over. Make sure they are correctly rated. Though personally I would retrofit this entire deathtrap with a transformer

>>2729293
One of the ends of my meter slipped, and accidentally bridged the circuit.
The meter itself seems fine, but one of the electrode tips now has a "crater" in the middle and a melted point.

It's not like the drier is that expensive (30 something dollar), so I'll just get a new one. I tend to leave it on overnight to dry my daughter's boots, so I won't risk using the current Smokeless one.

But just out of curiosity, what is the resistance supposed to be in 1n4007 diodes?

>>2729293
> deathtrap
using diodes is a common and cheap way to basically cut the AC voltage in half (and, hence, the power, into a resistive load) as we just kind of discussed with triacs and light dimmers, except as some anon mentioned, things less than 1A. That’s still 100 watts of heat from a 220 outlet.

It’s probably the exact same device in china, and you’re only getting 50 W of heat in north america, but whatever. Maybe it’s safer.

In the olden days, I’d get 240 V light bulbs from the UK and they would (albeit dimly) last forever.

>>2729304
> resistance
You’d have to know the voltage that your ohm meter sends through the diode in order to know what “resistance” the ohmmeter would show.

Many such meters, including my fluke 8020B have a diode setting specifically to test diodes.
Mine cannot test blue LEDs because their turn on voltage is too high.
Diodes drop a voltage. Your 1N4007 drops .7 volts. A blue LED might drop over 3 V.

Anyway, if you get some resistance one way, and then reverse the ohmmeter and test it the other way, and get nothing (open) then usually the diode is good.

>>2729275
T12 soldering station
Half finished DIY reflow plate
Hitachi 100MHz CRT analogue+digital storage scope (with RS232 HPGL waveform output)
0.1Hz-30MHz analogue sine/triangle/square function generator with sweep and modulation
Shitty LM2956 module-based CC/CV PSU
fixed voltage split-rail Heathkit breadboard power supply with LEDs and switches and a clock source
Modified Ender 3 V2
Snapmaker 3-in-1

do any of you bother etching your own PCBs for small prototypes
I'm running into an issue of all the parts I need are SMD and it's impractical to breakout board them all, and I don't want to wait two weeks for an expensive batch of PCBs to arrive from china
so home etching is looking mighty attractive to me right now

>>2729275
Wanptek power supply
LG DM-441B (DMM)
Yihua 8786D soldering iron with hot air (supposedly)
Hantek scope

>>2729435
> make own pcbs
Hardly ever. I can usually do SMD parts on veroboard. The PCB route, I’ve decided, is the low-skill solution unless you’ve got a BGA or something like that.
You can mask off, then spray down insanely tiny details with tape, stencils, and latex masking compound; spray it down with whatever paint, nail-polish, or thin strips of cricut vinyl knock-offs from the dollar store.
the etching is the easy part, dunk it in ferric chloride or ammonium persulphate.

I’ve also glued wires on a substrate, screeded some high temp epoxy on top of it, and then sanded it down to expose the copper wires and soldered on top of that.

There’s no end to the possibilities for doing something so goddamn simple.

Most of the PCB technology, including SMD was all done for quick and cheap factory process when you’re making a million of these things.

>>2729435
>home etching is looking mighty attractive to me right now

home etching is the least attractive option possible.
some noobs have a harebrained fetish for PCBs.
they need to realize PCBs only move electrons from one place to another.
which can be accomplished many other ways.
for simple one-offs, perfboard is less work, faster, less expensive, and without STAINS on your pants.
even for SMT stuff, when you buy the right board.

>>2729476

better pic.

>>2729435
I do it, but only after working out a streamlined workflow involving using a laser engraver to ablate away an etch resist. Even then it’s a pain to get holes drilled with correct alignment, so in future I’ll be using my Snapmaker laser+router instead.

>>2729435
>do any of you bother etching your own PCBs for small prototypes
Hell yeah brother, if you've got a cheap resin 3D printer and a dremel it's neat, easy and fast as fuck.
I learned it from this video:
https://www.youtube.com/watch?v=RudStbSApdE
Basically you just export your PCB design as a DXF, extrude it in a 3D modelling program and set the exposure time to the photoresist time, etch it and you're done. I made a little agitator out of a stepper motor and a plastic bucket to automate the etching, so the whole process consisted of slapping the board onto the printer, waiting 90 seconds and then dunking it into bucket.

Drilling out the through holes is easy, just put the boards in a vice and dremel them out. The copper sort of helps guiding the holes since it has less friction than the bare fiberglass, so I had 0 issues centering. Obviously you don't get plated vias, but you can just solder a wire through the hole on either side if you that.

>>2729478
so what do you do when the SMT package is small enough that it would have multiple leads lying on one copper square

>>2729435
I made a few boards, but it's just so labor intensive. do >>2729476 instead. You could also do something like pic related, just flip your SMD ICs upside down.

>>2729494
Mmm RF

>>2729491
For sot23 transistors, I just place them diagonally. Fits perfectly on .1” pads.

oh how the mighty have fallen

>>2729525
i liked this guy's videos when he compared DIYing to purchased products, this is just embarrassing.

>>2729537
what becoming a "full-time youtuber" does to a motherfucker