/diy/ - Do It Yourself

>>1995379
lads help me out here

Noob here. Trying to implement PoE on a PCB.
Am I making any sense in pic related? 48v coming in through the magjack, regulated through the buck converter before going into the PHY chip. Any help greatly appreciated

Reposting
Other than Analog's DSPs are there any out there with an easy no programming config tool?

>>1995422
>MFW when I will never have an army of tiny waifus to assemble surface mount boards for me

>>1995422
Should she really be touching both ends of a big capacitor like that?

>>1995442
I also am curious about this.

>>1995448
They're new caps so I doubt they're charged and I also doubt the circuit is above the voltage required to shock you under most circumstances.

>>1995316
Depends on what you're doing, but usually most of the way.

>>1995353
>I wouldn't put too much effort into it
I know, it's just for messing about. 16 bit DACs are somewhat expensive compared to this.
>get a DAC that supports I2S or basic PCM style input
Any reason for that?

>>1995371
That sounds perfect for a simple SDR.

>>1995444
You can always make a DIY pick&place machine.
Then print waifus on it.

>>1995474
and put stockings over the ballscrews instead of those concertinaing pieces of plastic or cardboard

>>1995427
Rolling average might be easier, depends on whether the peak current is able to be handled by the ACS712 or not. If the peaks are too high, then yeah you'd need to use a filter.

>>1995432
That looks right assuming you are running 10/100 in "mode B".
If your PoE source is running "mode A" then your V+ lines run mixed with your Rx pins and V- on Tx pins.
Gigabit is different too.

>>1995459
>I know, it's just for messing about. 16 bit DACs are somewhat expensive compared to this.
You can get 16/24-bit audio DACs for $5 or less.
>Any reason for that?
I2C is a protocol designed to have a lot of slaves on one dat/clk pair. Its an order of magnitude slower than single-slave protocols such as SPI.
If you are dealing with audio, you need as much data throughput as possible, and the I2S protocol is designed specifically for that. With it, you start up transmission and use the dedicated lines to stream data to the DAC. I2C isn't designed to constantly stream data because it deals with bi-directional communication and packets with ACK/NACK responses.
tl;dr: I2S is designed to stream audio, I2C is not. Audio is very data hungry.

>>1995487
>Audio is very data hungry
At 44kS/s, it's 528kb/s for 12 bit, 1.056Mb/s for 24 bit. Double that for stereo. Not familiar with I2C or I2S limitations.
>I2S (Inter-IC Sound)
Ah, it's literally designed for this.

>>1995482
Alright cool, thanks chief

>>1995495
CD-quality is 16-bit, 44.1khz for each channel (2 because its 2.0 stereo)
That's 1,411.2 Kbps.
I2C is normally run up to 400 Khz, but you can go higher if you know what you're doing. I2C also has overhead - chip addresses, sub addresses, start bits, stop bits. Those all add up.
I2S (as you said) is designed for audio and although it uses 4 lines, it can run up to very high speeds with no overhead.
Yes, it is a bit more complex, and timing has to be exact, but when you're dealing with that amount of data, you need to be.

>>1995422

Add CNLohr to the YouTube channels.

>>1995518
He needs to submit an application first. Please allow 6-8 weeks for processing by the Supreme Council of /ohm/lets. The decision will be notified by snail mail. Thanks for shopping S-Mart.

>>1995606
Sallen-Key is the best filter topology.

>>1995656
>Sallen-Key
Defend yourself

>>1995663
Two poles from just one op-amp, high input impedance, low output impedance, customisable Q factor and gain and frequency, need I say more?

>>1995759
Yes

0603 or 0805? For my first SMD components.
I'm leaning towards 0805 because 0603 scares me.

I'm the guy from the last thread trying to figure out phase-locked loops
I very conveniently found a tab I opened months ago among the 2385 tabs I have open that was an application note for the CD4046

using it, I tried my best to implement a bare-bones PLL in falstad: (iny tay)url.com/y43rzu99

Would one of you guys who are much more proficient than I, tell me if I actually did it right? I mean, the circuit does... something. The input and output waveforms look similar. They sound similar. It ignores noise, to an extent. If I change the pitch of the input, the output pitch changes as well.
But, I'm now realizing I don't really know what the output of a PLL is supposed to look like (or what they're supposed to really be doing - I thought they just forced a waveform to align in phase with a reference signal but it seems that was a misnomer???)

https://www.google [.] com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&ved=2ahUKEwj2n4SKqYbuAhUswlkKHQJrBpAQFjAAegQIBRAC&url=https%3A%2F%2Fwww.st.com%2Fresource%2Fen%2Fapplication_note%2Fcd00003774-stepper-motor-driving-stmicroelectronics.pdf&usg=AOvVaw024GyI5lH0gpfzCojqUj7f

Looking for a stepper motor, You're welcome.

>>1995806
0603 is fine so long as you've got some nice tweezers. I recommend 45° ones, but most important is that they match up perfectly at the tip. My current ones have a ~0.5mm mismatch, which is fine for most cases, but SMT soldering isn't one of them.
IIRC 0603 is more common than 0805

>>1995828
I think it's locked, but to some arbitrary harmonic, judging by the varying duty-cycle after the XOR. If it's acting properly, the duty-cycle of the XOR's output (once at a steady-state) should be constant, and with twice the frequency of the reference wave.

Replace your frequency reference with a straight square wave, the actual discrimination of an analog wave into a square wave for the XOR is a whole process in itself (schmitt triggering), and I've no idea how that XOR model will be doing that. Especially considering your reference wave looks to have extra zero-crossings each cycle.

No clue what your VCO is supposed to be (some sort of sawtooth?), looks kinda janky, consider swapping to a 555 astable with the phase-comparator going into the CV input if it proves to be a problem.

>>1995828
>>1995833
Oh, note that it is in some sort of harmonic lock because the XOR's output is a repeating wave. If it was failing to lock completely, you'd see a much more sporadic wave.

>>1995834
Or just:
www.st.com/resource/en/application_note/cd00003774-stepper-motor-driving-stmicroelectronics.pdf
learn to trim a link

>>1995833
the VCO is just a generic two op-amp VCO like on the front page of the lm358 datasheet. It has a triangle output and a square wave output, and the XOR is hooked up to the square wave output.
I'll give the 555 a shot

Sorry if this sounds stupid, but I thought the other input to the XOR (I think that's what you mean by the frequency reference) was supposed to be any arbitrary signal. How then does switching to a pure square wave do anything useful if it would just cause the VCO square waves to converge to the frequency reference's phase?

if that sounds too retarded just let me know. I think I still need to hit the books on PLLs hard, it doesn't feel like I'm understanding what's going on with the frequency reference at all.

What is the type of amplifier called where there is a gain stage for both the positve and negative signal channel,

Phase matching is used on random noise common to both pos and negative channels and then subtracted before gain?

>>1995479
The peak current is lower than ACS712's limit. Do I need RMS or just regular averaging?

>>1995843
Balanced input?

>>1995656
>>1995663
You both have shit taste.
MFB is where it's at.

>>1995838
>the VCO is just a generic two op-amp VCO like on the front page of the lm358 datasheet
Just the use of a MOSFET to GND when there's a split-supply sounds a bit fishy, if the drain voltage is below the source voltage it will flow via the body diode.

>just cause the VCO square waves to converge to the frequency reference's phase
It won't converge on the phase. In steady state, there will be a constant phase shift between the input and output waves. If your PLL can operate from 1kHz to 100kHz, then the phase shift will be different at 1kHz than at 100kHz.

>How then does switching to a pure square wave do anything useful
There are three intrinsically useful features of a PLL.
The first seems to be what you're doing here with an audio output, FM demodulation. By monitoring the control voltage, you can demodulate any changing frequencies; the opposite of a VCO. It's best to convert the input wave into a square wave for this, because logic gates are digital devices.
The second is frequency multiplication/division. By putting a div-by-N counter between the input reference wave and the XOR, your VCO can operate at a frequency N times higher than the input wave. And precisely so too, if you're running it off a crystal oscillator. A counter between the VCO output and the XOR on the other hand divides the frequency, so by using programmable clock dividers you can get nearly arbitrarily precise clock speeds, which can be useful for generating a radio frequency or something. IIRC this is what a DDS clock synthesiser does inside itself. It's best to have the input wave as a square wave for this, since counters are digital devices.
The third is more nuanced, but say you have a variable frequency oscillator that outputs a square wave, but you want a sine wave instead. If you choose a VCO topology that outputs a sine wave, then you can use it in a PLL to convert the input square wave into a sine wave. Or a triangle wave, or sawtooth, or whatever you want.

>>1995843
A differential amplifier with differential outputs. I've never seen such an IC, but in internal circuit diagrams I've seen them a few times.

>>1995870
Depends on whether the load is ohmic, constant-current, or something else. Simple integration over time is what you'd do for CC, true-RMS is what you'd do for an ohmic load, nasty shit is what you'd do for something like an LED.

>>1995828
You know better simulation software has the 4046 spice models in their libraries right?

Why would you use falstad ever? It's so much extra work for really shit results. FFS you can't even adjust simulation speed or set a timebase on your scope or v/div, you get no frequency readout, and any parts more complex than op-amps you have to hand assembly an approximation of the part with basic parts with shit models. They don't even have super basic shit like voltage regulators. You'd literally have to build it in yourself. Why would anyone ever subject themselves to this? It's terrible. It might be better if it was ever updated but it isn't. It's just shit.

>>1995906
>Depends on whether the load is ohmic, constant-current, or something else. Simple integration over time is what you'd do for CC, true-RMS is what you'd do for an ohmic load, nasty shit is what you'd do for something like an LED.
it is for charging a battery and sometimes an external load is connected to the battery, what do you think?

>>1995915
Well, what do you want to know the current for? If it's to measure the current going into or out of the battery, I'm pretty sure that a straight time integral is most useful with regards to a battery. Or you could just measure voltage and current and integrate instantaneous voltage multiplied by instantaneous current to get a cumulative energy value, that way you'd have no ambiguity whatsoever.

>Well, what do you want to know the current for?
Now that I think about it, it is for CC. I'll just do it simple then.

>>1995960
Pretty sure you need a linear current going into a battery, not PWM, if it's for charging. I'd put a bigass LC filter on it, if a linear current regulator isn't an option.

>>1995961
It's the output of a buck converter actually, need to control it so it would lower the duty cycle when the battery reach a certain voltage

>>1995962
just use current feedback for the buck converter itself

How am i supposed to make isp header work when the chip is connected to other periphetals on the board? Do i have to like use jumpers and diodes and shit?

>>1995973
the buck converter is a uC, I need the "current feedback" for the uC to control the duty cycle.

>>1995983
In Circuit Serial Programming?
If you can't leave the programming pins dedicated to just programming, you'll have to use jumpers so you can disconnect the lines from other connections.
You don't want any diodes, caps, resistors, ect... on icsp lines, unless the datasheet specifically says that it is OK to do so.

>>1996075
>If you can't leave the programming pins dedicated to just programming
is that what big boy pcb makers do?
i don¨t really want to add pin jumpers since my pcb is tiny and jumpers would be sticking out like the WTC before 9/11. maybe i could add small smd resistor footprint as a jumper and connect it via solder bridge

But the mail problem here isn't really the programming pins it's the power pins, the chip is on the same +5V rail as a bunch of lights so if i connect ftdi programmer it's +5v pin will power both the micro and the lights (several amps) so it will melt super fast

>>1996084
I use to put a jumper on the VCC pin of the ICSP
U can actually not use an VCC pin on your ICSP and program your IC while powered from your main power supply

>>1996090
>U can actually not use an VCC pin on your ICSP and program your IC while powered from your main power supply
really?? Not even when you connect grounds of the main PS and of the programmer? How can that be?

>>1996091
You need to connect grounds so that both sides are using the same reference. Beyond that, the chip doesn't care *how* it's getting power so long as it is actually getting power.

I mean, when you connect two devices, the connection usually includes a ground line but doesn't usually include a power line; each side has its own power source.

If I connect a potentiometer between the -12V rail and the +12V rail, does the wiper output become a 0-24V voltage divider?

>>1996164
What happens when you attach a load?

>>1996164

yes, but only if the output is referenced to -12V as its only ground.
if the output circuitry uses a common ground, it becomes a -12V to +12V divider.

>>1996168
It works fine if the impedance of the load is significantly larger than that of the pot. If it's significantly smaller instead, then it will act as some sort of current regulation. But you could buffer it with an op-amp.

Please help if you can. Someones listed this H.G. Palmer chord organ, but it has no power lead included. They sent a picture but I can't tell if it's an 8 input lead or three pins. Obviously I'm going to ask for more info, but I haven't been able to find out anything online.

>>1996501

>>1996502
That’s 3 pins.

Though I’m unfamiliar with both that era and that genre of appliances, I think it’s somewhat likely that it’s a split-rail DC input. Say, 0V, -12V, and +12V. While probably difficult to get a replacement male connector, let alone a cable, it’s a reasonable idea to replace the socket with a newer standard. Not many prolific 3-pin split-rail socket standards out there though, might find yourself going for a low-current (30A) BLDC connector, basically the 3-pin version of an XT30. Pretty sure they are indexed/keyed, even though they don’t need to be for BLDC purposes.
In this case you’ll need to make or buy a PSU for it. I think direct AC transformer PSUs are easier for split-rails, and they’re also lower noise for this kind of purpose, but you’ll probably need some beefy caps.

It’s also possible, if it was made before the IEC-C14 standard was popularised, that the three pins are LIVE, NEUTRAL, GROUND. In which case you definitely want to replace it with a C14.
Either way, you’ll know once you open it up whether it wants AC or DC.

The third janky possibility is that it’s for an external centre-tap transformer, to be turned into split rails with internal rectifiers. Not likely though.

>have Samsung cell phone with busted power button
>check part and decide it's easy to do
>desolder old one, place new one, everything seems ok
>assemble everything again
>it won't turn on
>it's not recognized in a USB port
>connect a charger
>it only draws 53 mA
any guidance on how could I troubleshoot this?
everything was fine before this, there's no visible damage on the board

>>1996591
Was the new button broken during soldering?
Perhaps the old button wasn't broken but rather the SMD pads were delaminated/broken due to pressure on the button. In that case soldering a new one wouldn't fix the issue at all and you'll need to make some fine bodge wires.
>it's not recognized in a USB port
What, the charger doesn't recognise the phone as something it can charge?

>>1996084
>is that what big boy pcb makers do?
I don't know what the big boys do exactly, but if I was making the call, I would buy the minimum-sized chip that allowed me to debug & develop the board in the field. You can share pins only when it doesn't interfere with development. IMO it's better to spend the extra $0.25 on a part with enough pins than to aggressively minimize part cost. The Art of Electronics suggests a philosophy of overengineering, rather than being pennywise and poundfoolish.
If board space is a concern make your own adapter jig with pogo pins or a small-pitch smd connector or anything with a minimal footprint.
Using ICSP you'll never go back to the caveman technique of socketing chips.

>>1996626
>adapter jig with pogo pins
Saw pogos and pogo jigs recommended to me on ali, it's neat tech. All you'd need is a series of test pads, and a second PCB with the pogos soldered onto it held in some sort of vice.
>the caveman technique of socketing chips
Still required for things like EEPROMs and the like, so I still don't mind having the hardware to program an MCU too. It's kinda aesthetic, in a stupid way.

Also some madman did some insane programming to fit space invaders (including graphics and display driving) onto an ATtiny10 (only 6 pins, including power and reset), was really impressive. He didn't have the liberty of extra ICSP pins, I think he basically soldered it onto a breakout board and socketed that for programming. Not really in the spirit of making a tiny PCB with a tiny MCU, but that wasn't his goal. The programming was interesting enough that I watched the full 2 hours / 2 episodes without getting bored whatsoever. Lots of neat tricks.

>Mlx90640
Why aren't they popular?
Everyone I look at is using shittier but equally expensive amg8833/amg8834

I'm trying to design a PCB in KiCad for the first time, having never designed a PCB before. I'm trying to make a custom footprint but there's not enough space for me to place 9 pads in a row. I can't find any way to increase the size of the work area, only how to increase / decrease the grid size which still keeps the same work area size. Any suggestions?

>>1996691
>he buys name-brand electrolytics
Is this one of those low-ESR things? For audio, or switching PSU filtering, or computer stuff? IIRC the "stacking multiple values of capacitor in parallel" thing isn't a scam so long as you're careful how to do it.

>>1996692
>I can't find any way to increase the size of the work area
Pretty sure zooming like in eeschema should work just fine.
pic 1/2

>>1996692
>>1996697
pic 2/2

>>1996697
>Is this one of those low-ESR things?
not the same anon but MLX90640 is a thermal camera, anon

Does this connector have a name? pic related.
It's from a yamaha pw-x ebike motor.

I want to sniff the CAN bus, but i need a spare cable to access it, but this shitty cable costs 26€ on ebay. If i can get just the plain connector I'd solder my own quick.

>>1996712
pic related now*

>>1996713
Found it. It's the JWPF from JST.

https://www.avdweb.nl/solar-bike/hub-motor/yamaha-pw-sx2

>>1996621
>Was the new button broken during soldering
i tested continuity on the pads and it was ok
>delaminated
possible
>the charger doesn't recognise
nothing recognises it, the pc, powerbank, charger as it only draws 53mA

there's no reaction from the phone at all not even the charging screen

>>1996732
>there's no reaction from the phone at all not even the charging screen
Is it already like this from before you changed the switch or not? Were you able to charge it before you did the repair?

>>1996591
Did you remove the battery before soldering?
Its also possible it got popped from static while you were working on it.

>>1996642
>The programming was interesting enough that I watched the full 2 hours / 2 episodes without getting bored whatsoever. Lots of neat tricks
You got a link to that sempai? Sounds like it would be worth a watch.

>>1996084
>is that what big boy pcb makers do?
Typically, yes. The mcu's that I work with normally have 3 different ICSP pairs to choose from, even on a small 28-pin package.
Since you are using an FTDI usb-serial converter, for programming, you might be able to get away with some resistors on the dat/clk lines, but you'll have to keep the resistance values high and possibly lower the programming speed, if thats an option.
You can multiplex the ICSP lines with other IC's inputs if you need to. Typically an IC's input won't strain the line, so it might be ok depending on the input impedance.
And of course, anything that can go into a high-Z mode would take stress off the line as well. diodes, capacitors and inductors are always a bad idea.

>>1996697
>>1996698
Updated my Kicad to post-5.0 and it works now.

made a ventilation controller for my earth cellar.
worked pretty well during summer as the environment stayed cool while the fermentation gases got sucked out.
now during winter it's hell though... forgot about dewpoint and everything inside is overgrown by mold. even the fucking glass bottles! didn't even know they can mold, kek

welp, long story short, had to add another sensor + wiring and updated the program.
still not sure if it's not just easier to just use a dehumidifier and call it a day.

>>1996752
yes I did remove it, my hopes are thinner by the minute and I fear static damage, would that fry the mainboard to the point of not even charging?

>>1996750
the phone was working fine the only problem was the button

A question on UV LEDs: Do they have very narrow spectrums compared to halogen bulbs?
Say if I want 365nm-405nm, I'd have to buy like 4 wavelengths no?

>>1996806
Take the phone apart again. Try to power it on outside of the case.

My first attempt at designing a PCB. Apart from the general mess, is there any glaring errors (e.g. should I keep the traces away from the holes?).

>>1996808
about 5nm below and above the spec peak

>>1996894
Spread out. There's no need to have your traces run so close to the pads. Don't serpentine the traces between pads, use a different layer or use some vias. Acute angles coming off of trace junctions can form acid traps. This will not be a problem with most modern pcb manufacturing processes but it's best practice not to have them. Consider using a plane for power/ground. Not having to route power/ground traces can save you a lot of space and time.

>>1995453
Shock (you) as in shock you the normal sized human or a tiny anime girl size of max 3 capacitors

>>1996297
you might use this method for voltage references, but OpAmps dont generally output more than 10s mA of current.
if you want more than that just use a controlled buck converter.

>>1997033
>tfw ywn be a tiny human on an anime girl's circuit board
>tfw ywn slave away trying to fix her vibrator in the 20 minutes she gives you before stuffing it up her anime lady parts

>>1995422
Cute

>>1997043
Based picture, but it would've been more based if they had been working on a CD40106UB

>>1997074
wait i fucked that up
I mean the CD4069UBE
Schmitt triggered hex inverters got me slippin and slurrin

>>1996754
From Ben Heck Hacks:
Part 1:
https://youtu.be/PZsWqOuJFKI
Part 2:
https://youtu.be/aq30EdgSpOY

>>1996808
They have pretty narrow peaks, yes. But you can't just buy a 360nm, 370nm, 380nm, etc. UV LED, since the wavelengths are determined by the semiconductor bandgap itself. This is mainly a function of the semiconductor being used (silicon = 0.7V = 0.7eV photon), but IIRC the doping can push it about by 10%. Look around on digikey, maybe UV LEDs have a wider wavelength variance.
Things like gas discharge tubes aren't any better, they also have narrow peaks.
To get a wider band, you'd need some sort of phosphor, and I don't know if anything of that sort exists.

Why do you need a wideband UV source in the first place? Some sort of janky spectroscopy?

>>1997129
I want to cure UV resin but have no idea what wavelength it cures at.

>>1997129
Also why exactly can't I just slap together LEDs of different wavelengths?
There's this 4 chip one that has 2x 365nm and 2x 405nm, I thought it'd cover enough but apparently there will be a big gap.

>>1997149

Any alternatives to proteus ISIS for circuit design that allows me to use Arduino emulation? Downloaded 2 cracked torrents and they're not really working, i would rather use something open source too...

>>1997163
You say you're a fan of ultraviolet light, but you can't even name 3 wavelengths?

>>1997074
>>1997081
CD4093 is fine too.

>>1997149
>>1997157
They cure at a minimum frequency, anything over it will work. (conversely, at a maximum wavelength, anything under it will work). It's not like just one particular wavelength is required, the photons just need to be above the energy threshold to interact with the resin molecules.

Read the resin's datasheet for specific info, but I'm pretty sure common 405nm LEDs will work just fine. UVC (~250nm) mercury tubes will definitely cure the resin, though might have destructive properties by interacting with unintended parts of the molecules' structures. These resins are probably specifically designed to have as low an activation energy as conveniently possible such that the process of curing them won't also damage them. You could also look into what common UV curing boxes use.

>>1997214
>you can't even name 3 wavelengths
Huey, Dewey, and Louie. What do I win?

Pic related is from the data sheet of an LED I'm working with. The maximum listed forward current is 30mA.

If I supply it with 6V from batteries, can I get away without a resistor?
It looks to me like the LED will draw ~22mA at 6V, which should be safe -- is my interpretation of these charts accurate?

>>1997231
You still need a current limiting resistor. Try 120-130 ohms.

I'm interested in building panels like this (pic related). Is this something I can realistically learn how to build in my small apartment? Or would it be something that I need to have manufactured in a factory?

>>1997236
You could easily manufacture them yourself. It's just pots and switches on a PCB/perfboard attached to a panel.

>>1996819
after thinking I'm pretty sure this has to do with the PMIC or charge controller being fried

>>1997236
Yep, it's quite simple.
Build a box. Cut holes for whatever devices you want to attach. Button it up. Done.

>>1997235
Look at the image again. It implies that you want 20 mA, and at 120 ohms that would be 2.4 volts. If he actually has a 6 volt supply, then the led will not even get 4 volts, and that one needs at least 5.

He needs a current limiting device and more than 6 volts, ideally, but if his battery actually is 6 volts then he should use it directly with no series resistor.

>built CNC plasma table
>want to make sure the wiring to the steppers is protected from EMF
Any good pointers you gents could give on wiring this thing? I'm using nema 34 steppers with an ebay mach 3 controller board. I plan on shielding the computer inside a steel case but I'm lost as far as shielding all the wiring. is it as simple as grounding every possible thing?

I'd also like to build a torch height controller for the z axis, which would work off of the voltage of the torch to the work piece, which means voltage divider, I'm definitely a noob on electronics so if there's any wisdom for feeding the voltage from that into an MCU without frying it that'd be great.

>>1997245
>He needs a current limiting device and more than 6 volts, ideally
Terribly sorry for my probably stupid question, but why is this the ideal scenario? If it really is ideal, then I can go ahead and do it that way; it doesn't matter to me how many Vs I have to throw at it. I'm just new to electronics and don't understand these things yet.

Thank you both for your replies.

>>1997245
I'm clearly retarded. Pardon my drooling.

>>1997231
No, because of thermal runaway. That forward voltage will drop the hotter the die gets. Unless the batteries have sufficient ESR.
Also 6V sounds a bit high for a 4.5eV photon

>>1997236
Getting a professional-looking faceplate might be difficult, I think the easiest way is to order a PCB with black solder mask and white silk-screen, with the holes pre-cut in it.

>>1997249
Coaxial wire helps, but if you're using twisted pair it should be fine. Cat6/6a is shielded twisted pair with 8 wires, which should be great for signals, but for power lines you have less options. Clipping ferrite chokes on the wires is a good method IIRC.

As for that voltage feedback idea, you're right to be cautious. I'd use a voltage divider, some TVS/zener diodes, maybe also an LC filter to protect against particularly fast spikes. Then before going to the MCU, you could use an analog optocoupler or vactrol or whatever. Would take some calibration, but that's not a big deal.

>>1997245
>If he actually has a 6 volt supply, then the led will not even get 4 volts, and that one needs at least 5.
Incorrect. The LED will draw significantly less than 20mA and have a voltage drop of about 5-6V.
>if his battery actually is 6 volts then he should use it directly with no series resistor
Also wrong. The fact that the forward voltage can range from 5-7V at 20mA doesn't imply that 6V will be safe if you want to run it at or below 30mA. That forward voltage will vary with temperature. On a cold day it might be safe, on a hot day it might not be.

>>1997254
The safest way to drive an LED is to have a voltage source with enough headroom for a constant-current supply. A switching supply with current limiting is ideal for a battery-powered solution where you care about power loss.

>>1996894
It seems like you are relying on the auto-suggested routings of kicad too much. You should think about what you are doing instead of letting the computer think for you. Also as >>1996894 said, you are routing everything too close to the pads. If you left more space you would have an easier time figuring out the connections.

1. Why are you zig-zagging your traces in between pads? You aren't making good use of your two layers, you should try to make connections as direct as possible.

2. You are routing multiple red traces through the connector even though they aren't used in the connector. You should move these traces up or down to make routing easier. The loop originating from pin 5 of J2 should be eliminated with a red wire and a via.

3. These acute angles are ugly. Just do 90 instead.

4. You are designing a multiplexed display yes? These elements have many lines shared between them and you combined them into a sort of bus. You need to take this concept further. You don't seem to know about vias, they let you transition between copper layers. The can let you connect a wire from the red layer to the green layer as I did in 2. You can construct a wide bus running to each of the displays on say the green layer and make the final connections to the displays using red wires and vias. That would make routing much easier.

>>1997236
There was a post years and years ago on hackaday.com about just this topic, but I can't locate it right now. The gist of it was that you can do toner transfer by printing a negative of your label/panel onto a slick paper using a laserjet printer, then ironing it onto a piece of aluminum. You wet and remove the paper, leaving the toner pattern fused onto the metal faceplate. Next you immerse the part into an etching solution for long enough to etch a relief into the face, followed by filling with some type of heavy paint (maybe automotive paint?) then sanding the toner and raised sections of the plate clean with sandpaper.

Here's a link to an instructable which demonstrates a similar process: https://www.instructables.com/How-to-etch-aluminum-panel-labelsdesigns-with-a-r/

It's very doable in a home environment and doesn't even take that long to achieve professional results, but it is probably not sensible unless you need small quantities of high quality parts on a budget.

Explain why haven't you ordered one, hacked it, and become a magical sushi girl yet /diy/?

>>1997346
hacked it to do what, play shitty transformation scene jpop?

>>1997347
To do whatever you want buddy. The sky's your oyster.

circuit bros, i studied a little bit of electrical engineering in college but quickly switched to software because i honestly could not visualize how circuits work. that shit went way above my head. is there any tool that lets you visualize it nicely?

>>1997373
falstad

>>1997377
No. Falstad is absolute garbage. Stop promoting it. There's much better software.

If I want to start playing with robots and know absolutely nothing, is it better to focus on learning electronics and leaving the programming for later, or the opposite?

>>1997373
hydraulic analogy
works for everything
absolutely everything

>>1995379
have you tried reading the fucking datasheet

>>1997373
What goes in must come out. It all boils down to Equivalent Exchange. Wa La!

>>1995422
i need to make sure a li ion doesnt discharge too low. i have a micro controller with free pins and an adc. should i just measure the voltage and if it gets below a certain value go into sleep mode and check once in a while till the voltage jumps back up then get out of sleep. is there a better way to do this?

>>1997401
opposite. most dweener stuff is plug and play
though honestly, mechanics will likely be your biggest hurdle

I suck at understanding electronics, so please bear with me.

I want to use a 3 volt signal to turn on a 9v circuit. At first I thought I could use an AND gate - but since they share ground the voltage is averaged, correct? This would damage my 3 volt components. I feel I should be able to buy a cheap transistor that does this type of thing, yes? What would it be called?

>>1997496
boost converter, $3 off chinkbay
little more complicated than a transistor though

>>1997496
by the sounds of it, the 9v and 3v side have their seperate sources right?

then you're looking for a "high side switch" ( which will probably be using p-ch mosfet or just a simple pnp transistor, etc, plenty of example circuits. )

if you're looking to POWER the 9v circuit off the 3v supply, then yes use boost converter.

>>1997496
Where is the 3V signal coming from, where is the 9V signal going? How much current does it need to be?
Do you already have a 9V power supply?

>>1997519
IIRC a common-gate NMOS will work as a level shifter.

>>1997448
>read muh datashit!
are you a retard? read the question first before throwing stupid shit like this

>>1997454
yes, just put a BMS on it. it will usually cut off at 2.8v

>>1997533
>2.8v
Sounds pretty low, IIRC it was better to use them from 3.2-4.2V. Not sure what a smart battery charger (e.g. B6AC) has as a cutoff.

>>1997506
>>1997519
>>1997529

<--- This is what I'm trying to do. I have purchased . I've ordered this: https://www.banggood.com/JHEMCU-Electronic-Switch-Module-30A-3_7V-27V-1S-6S-Universal-For-RC-Drone-Car-Receiver-p-1524397.html

However IF I had to build it myself. First I would use a low pass resister to turn the PWM cycle into voltage - I could then turn that on and off using the controller.

But once I get that voltage on, I couldn't figure out how that would allow a second circuit to start. I can't have them on the same circuit because the voltage gets shared ( I think ). I thought a transistor would do it - doesn't that allow power to the Emitter, when the collector is given power? But that is as far as I got before I realized I had no idea what I was talking about. Especially hitting the PNP and NPN shit. That just reminded me I needed to think about the ground, and then I gave up.


>>1997373
>>1997428
I have the same problem. Hydraulic analogy doesn't do it for me. I don't know why. Maybe when it comes to dealing with having the positive and negative the analogy breaks down for me. In particular the concept of everything having to connect to ground I find frustrating when I'm thinking of multiple things going into a circuit.

Also things like something "pulling" too many amps. The load doesn't "pull" shit - it is acted on by electricity. That's like saying a water wheel is pulling the water past it too quickly. Or the leaves are pushing the wind too fast.

>>1997496

simplest is the low-side switch at the top using a single transistor and resistor (1K should do it).

if the 3V and 9V circuits share a ground, then you can do the slightly more complicated high-side switch at the bottom.

>>1997541
>Sounds pretty low,
correction, it is 2.9v as most of chinese BMS uses 2120CB which has 2.9v undervoltage and 4.28v overvoltage protection. I've tested this with my own battery and it cuts of around 2.88v-2.9v

>>1997569
ok thanks. This is where my head started hurting . I really appreciate your help - I'm going to review this in the morning.

>>1997566
to do what? pwm to binary on off?
take pwm through low pass filter as you say, a capacitor will do, then use an opamp configured as a comparator, use a voltage divider to set the on/off point. if you need to drive a load from it then add this on the end. >>1997569
just the low pass capacitor might work on its own if you can get the voltage below/above the voltage it needs to turn on.
also you could pass the low pass filtered voltage straight into the transistor, if you can get above/below the turn on point (~0.7v typical per junction)

dont think about analogies, most people don't really understand fluid dynamics in the first place either so trying to apply it to something else is a waste of time.

I'm planning a project where I make a digital 16-step sequencer that can handle MIDIDIN and other things. I plan to use rotary encoders instead of potentiometers for the steps but I need a way to monitor all of them for input. Although I've never handled so many rotary encoders at once, my thoughts were:
>2x 16 bit parallel in, serial out shift registers
>One register tied to all A inputs
>Other register tied to all B inputs
>shift register produces 2x16-bit binary values
>fed into a slave ATTINY (or other mcu)
>Constantly compare the input values to detect changes
That should work, right?

>>1997033
The human body has an average impedance that doesn't vary with size much. Even if it was charged with 12v, it wouldn't discharge through her.
Imaging grabbing the terminals on a massive 1000+ amp 12v car/marine battery. It won't discharge through you under normal circumstances. Same goes with a toddler that is 1/10th your size grabbing the same 12v battery, the body doesn't have a low enough impedance to conduct.
If your hands were wet, the chance increases, but that is because the water on your hands changes the impedance across your body.

asassa

Alright anons im working out the 12v onboard electronics system on my boat, right now i've got the following setup
>thicc wire from + side of the battery runs to main switch
>main switch has wires running to fused switch panel and some other things
>negative wire from switch panel runs to ground attachment on engine
Now this works, however i cant really fit a lot of attachments to the engine ground so for this i was thinking of making a central ground rail out of copper and using bolts to attach all the loose negative wires to it and then having one thicc wire attached to the engine ground bolt.
Now for this i have a strip of 1mm thicc copper plate of about 5x20cm that i want to screw down on a wooden plank bolted to the floor. Now my questions for this are:
>is it safe to use a copper plate like this for a 12v 60A DC current? I imagine a thin strip like that could melt or at least get very hot when under high amperage
>what should i use to protect the wooden plank from moisture? A boat is a humid environment and i can imagine the wood getting soft and moldy if i just put it in there raw, i was considering linseed oil to keep it nice and water proof however that will burn like a Dresden in case of overheating and i think a layer of varnish will lose it's benefit pretty quickly if i drill holes in it.

>>1997637
1mm isn't very thick and while it might work, I wouldn't do it.
Home improvement stores (Lower or Home Depot) should have a beefy solid metal strip with holes and bolts to bolt down lots of wires. Kinda like common rails in a home electric box. Run a thick wire from battery ground to the rail, then another thick one to the motor. That rail is now a common grounding strip to use.
Or go grab some car audio distribution blocks, they can get pricey, but are designed for this stuff.

From the pic related calculation, I got the final P as 2.45W. How do I convert this to farad? I am dumb so bear with me

it's not pretty, but i can now program embedded chips via ISP pads on the pcb

feels good

>>1997634
Holy shit anon thanks for serious answer. Did not expect that, cheers

>>1997388
such as?

>>1997825
the iraq

>>1997701
noice

>>1997701
What's that hot-glued shit on top?

Noob here. Got a vacuum motor wired up and it runs. I got a dimmer from the vacuum too and it also works. Now I want to make use of them.
Trying to save my lungs from welding fumes. How difficult(dangerous?) would it be to make something like pic related? I already have the tube and and I can make the housing out of plywood, mdf, or aluminium sheet. I can buy a scuba mask. The biggest difference is I'll have the motor outside so I won't need to carry it or use a filter. I'm using this video as reference.
https://www.youtube.com/watch?v=lmNIEDDCqSk
Bad first project?

I'm taking a course called "Linear Circuits I" this semester, pic related.
According to the Rate My Professors page for my specific prof, he has written his own textbook and it's apparently god awful.

With that being the case, does anyone here have any recommendations for a textbook that would be good for this class? I already have TAoE. I also have a book called "Fundamentals of Analog Circuits" by Floyd and Buchla (got it for $6 on Ebay lol) which seems to fit the bill well.

>>1997388
I don't use falstad, all that I know is that falstad will visualize current and voltages and that might be enough to help the other poster.
Posts like yours are extremely annoying when you say there's much better software, but don't take an extra second of typing to name it.

>>1997869
Who's your prof?
I used "Foundations of analog & digital electronic circuits Book by Anant Agarwal" which is free online, last I checked.

>>1997873
https://www.ratemyprofessors.com/ShowRatings.jsp?tid=490350
wish me luck lol
he's *still* the highest rated professor for this course

>>1997878
>>1997873
also I have a digital copy of Agarwal's book, I remember going through it a few years back. It was pretty good - I meant to mention that in >>1997869. I'll keep it handy going through the course.

>>1997701
>it's not pretty

no shit! you need more acrylic in your life, dude.

>>1997566
>Also things like something "pulling" too many amps. The load doesn't "pull" shit - it is acted on by electricity.
Yes but those are colloquialisms that are understood well in industry by people who've done this stuff every day for years. The idea of pulling a voltage up or down with a resistor showed up in datasheets and books for decades and isn't controversial. Also push-pull in amplifiers or motor drives, etc etc.
Vacuum cleaners technically don't "suck" dirt either - they use fans to push air out of the unit which causes a pressure drop, and the only reason a vacuum works is because atmospheric air pushes the dirt into the hose as it rushes by to equalize pressure. This is why a vacuum can only hold "pull" a water column up so high - the vacuum isn't doing any pulling, the atmosphere is pushing the water up, and it can only go so high because there's only so much pressure.

Now I don't want to presume what you might think next, but if you thought "but that's just what sucking means!" then by analogies can develop the same intuitive sense of electronics when people say something pulls a load, or pulls a voltage high or low, etc.
Also it would get really tedious if every time people said "provide the circuit conditions necessary such that current flows due to EMF..." :)

What would cause a 30V solar array to drop to 5V when you connect it to a controller, even with no load?

I'm working on a small solar project, the thing used to work fine, the panels would be at wound 27 V under load, then one day i may have accidentally overloaded it, and now this happens.

The panels still show 30V in open circuit, so did something burn out in the controller or whats a likely cause? I dont even know how to diagnose this

>>1997532
<lads, is it possible to sense the current of a 100KHz PWM signal with ACS712?
>frequency bandwidth 80KHz
maybe you should learn to read faggot.
Here's the answer for you since you are unable to figure it out yourself.
The cutoff frequency for the chip is 80 KHz so no, it cannot read an AC current of 100 KHz with any degree of accuracy.
A low pass filter will not work because the current fundamental frequency is 100 KHz. The only thing you would measure is DC with that.
Averaging will help you either, you've already lost half your power past the -3dB point so what ever current you "think" you are reading is fucking wrong.
You need a higher bandwidth current sensor or just make your own with a precision current shunt and op-amp into ADC then do RMS calculations in SW.

>>1997566
Water wheels and wind and such are not good uses of the hydraulic analogy. The analogy needs to be used with fluid flowing within conduits, otherwise pressure isn't as good an analog as it could be. A hydraulic or pneumatic circuit driven by a constant-pressure pump is the best place to start. If you have a valve in series with a turbine, then with the valve closed/off, no current can flow. With the valve on, the amount of current flowing is determined by the resistance-to-flow of the turbine. This is what's meant by the load "pulling" the current; the reference situation is one where no current is being pulled at all, unlike the wind or a river where the reference is where there's no resistance to flow. Constant-flow hydraulic systems also fail to make a convincing analogy for the same reason.

Charge conservation is another way to think about it, Kirchhoff's current law comes from this.

>>1997905
The controller is probably fried and is acting like a load when you connect it to the panels.

>>1997373
I like the program called "everycircuit". There's a web version and apps for android/ios.
It costs $15, but it's a one time fee and it works pretty well.
It's not like spice, you wouldn't use it for engineering real things, but it's fantastic for visualizing current flow and circuit components and such.
The only caveat is that on the web version you have to enable Nacl to run it (google chrome execution environment, you can enable it through chrome://flags).
It's a lot smoother and more polished than falstad. That said, people shit talking falstad are retarded.

>>1997388
falstad is perfectly fine, youa re just dumb

>inner jew wins out, finally make an order on lcsc instead of digikey
>realize how cheap this shit is
>order bulk quantities of passives and discretes i'll never use
>end up spending more money overall
i blame you, /ohm/

I'm going to do an electrical mod to my car (energize the heater relay through the cig/radio fuse instead of the gauge fuse, to allow the blower motor to run with the key in the ACC position).
I will use crimp connectors since that is what was originally used on the car and I heard solder can vibrate and break. I have a crimping tool already. Are there any tips for making good crimps that don't come loose and fuck up?

>>1997653
what part?
in my experience, the datasheet would have more equations that give you how to calculate the capacitance for both input and output.
(assuming its a brand part like Analog devices or TI and not some chinkshit)

>tfw the next step of prototyping is fine tolerance mechanical design
>tfw the final working model will need either lathe + indexing/dividing head or 5-axis machining
Where do I even start. I’m thinking of getting a 3D printer for ensuring my CAD fits and works, even then stuff like press-fits for a bearing will likely be different depending on the material, and I basically have no room for even a small printer.
No clue who I’ll ask to get the machining done by, don’t have any machining friends, anyone else had to do this?

>>1997496
pic related is called an inverter. it turns a high signal into a low signal. it can be done with either NPN transistors as pic related or with N-channel mosfets.
Mosfets turn on when the gate-source voltage is over a threashold (1-2V) for most devices and BJT's turn on when the base-emmeter voltage rises over 0.7V and the current is over a few microamps.
you can probably see now that with these two thresholds it is easy to control either transistor type with a 3V signal,

but wait, now your highs are lows and lows are highs. not to worry, just put another inverter IN front of the first to invert the inversion and get the original logic.

>>1998099
meche here getting anything machined is monstrously expensive, if you ask them to just ream a fucking hole they'll ream you instead. you should run a whole business case and call for quotes before proceeding with your project. 3d printers are great though.

>>1997496
Just use a relay

>>1997825
>>1997870
>>1998050
Jesus Christ, why do I always have to spoon feed retards.
IT'S PSPICE. GET ORCAD PSPICE.

If you're a little retarded then multisim is an acceptable alternative. falstad's component models are hyper simplified and generic and therefore virtually useless. It doesn't have basic components like voltage regulators and there are few tools available to analyze circuit behavior. You have a basic voltmeter, ammeter, and an oscilloscope so crude it shouldn't properly be called that. You have no manual control of the simulation time step you get no FFT or spectrum analysis or any remotely useful features. FFS you don't even get a frequency readout for periodic signals. Its worthless and the only people that like it are retards because it visualizes the current flow. Unfortunately they're too stupid to realize that doesn't even fucking matter.

>>1998181
this, i use multisim but i would not recommend multisim.

>>1998181
>You have no manual control of the simulation time step
you do
>you get no FFT or spectrum analysis
you do
>FFS you don't even get a frequency readout for periodic signals
you do
>falstad's component models are hyper simplified and generic and therefore virtually useless
no they are not. What components are you refering to?

PSPICE is for niggers. Real men use LTspice for serious things and falstad to quick test stupid shit of to do sanity checks. If you have not found those features in falstad YOU are the retard. Eat shit and die.

>>1998193
this just use ltspice
super easy to get working on linux too by just using wine.

>>1998193
And Psim for power stuff if you do not have the patience to create your own ideal switching element library and other crap for power electronics in LTspice.

>>1998181
Yaaay for Multisim! Me like Multisim!

>>1997967
>I like the program called "everycircuit"
Holy balls I forgot about that. Its been a while since I last tried it out and holy shit they've come a long way. Thanks for reminding me.

>>1998105
>getting anything machined is monstrously expensive
It’s a part of the prototyping process so I don’t mind dropping a grand or two on it to produce a functioning prototype. But iterations on it are going to be even more expensive, hence why I’m looking to get a 3D printer, as an investment. I’m actually surprised that entry-level resin printers (anycubic photon iirc) are only 1.5-2.5 times the price of FDM, so I’m considering getting one of those. They look a bit better for high-strength mechanical prototyping. If only I had a well-ventilated environment to use one in, besides my busy and messy stovetop.

It’s amazing to have a disposable income now, even if it’s from a supermarket job that will never save me anything towards a house or boat or student loan. Thank god for chinkshit.

>>1997896
I can only imagine how nerve wracking it would be to do this by hand on a drill press.

>>1998234
pardner, where I come from, "by hand" and "on a drill press" are two mutually exclusive things.

>>1998181
ltspice is fine too

>>1998181
what kind of dipshit gets mad about other people using a tool he doesn't like? Yes it's slow and limited, but it has its uses. I don't know of any spice software that runs the simulation and holds states as you edit the schematic, and for that reason it's sometimes worth tolerating its limitations.

>>1998181
>Jesus Christ, why do I always have to spoon feed retards.
>IT'S PSPICE. GET ORCAD PSPICE.
All you had to do was say it in your post and you could have avoided the rest of the shit in your post which I didn't read. 2 seconds of typing.

>>1996542
Thank you sm anon for your very comprehensive answer. You're the best.

>>1998193
But it doesn't have any of those features. Stop lying dipshit.

>no they are not. What components are you refering to?
Any of them, let's pick BJT for instance, there's exactly one adjustable parameter, beta. THAT'S IT!
Now let's look at multisim... oh wow look at that, like a proper simulator. How can you defend garbage web software that uses the most bare bones modeling possible. Falstad only works for modeling the most bare bones basic circuits and as soon as you try and do anything remotely more complex it becomes useless.

Would it be ok to use something like a tranny or fet to connect a data line? somehing like i2c or serial or isp

instead of using 10 jumpers every time i want to change shit i would like to use trannies instead and used them instead of jumpers to connect the traces together in a certain way

>>1998441
How many fucking times do I have to spoon feed retards in this general?

>>1998471
There are analog switches ICs made exactly for that. Internally they are FETs. The 4016 is an example

>>1998441
>only works for modeling the most bare bones basic circuits and as soon as you try and do anything remotely more complex it becomes useless.
It was not made that you dingus.

>>1998441
>moar parameters = gooder simulation

>>1998474
Nice shop. Do you like own falstad or something? Is that why you're so deadset on insisting it has features it doesn't actually have?

>>1998479
Yes that's exactly how it works.

>>1995606
so which op amp oscillator is the one where drake is getting a cum facial? that the 3 in one function generator where it can do square, sawtooth and sinusoidal ?

>>1998271
Jesus Fucking Christ, this is a blue board, keep that shit to yourself.

my doggo's collar is almost finished, all i need now is to program in some cool light effects.
man, he's gonna get so many bitches at the park

also i found out i can use single cell lipo instead of 2S, because the strip i'm using is 5V but actually works in ranges between 3.2 - 4.2V at reduced brightness, which is great, because i can use a really tiny lipo this way, which will be really light on the neck and i don't want to run at full brightness anyway, since it looks as if UFO is landing at night

you guys know where the best way to find buttons with nice action are? digikey i can find buttons but they arent cool. I could just steal switches from a old keyboard that has blue switches but itd probably end up be a pain in the ass.

Does anybody know what connector this is? It's the TV end of a LG TV power cord.
Need to make a longer one and would prefer keeping the old one.

>>1998603
Front view.
Pics are from some AV Forum. But I have the same connector.

>>1998594
Mech switches on aliex.
Vandal switches if you want a power button or something.

In addition to 0603 capacitors, which capacitor should I get for filters where I need higher capacitances? I'm looking into getting 1206 C0G capacitors, but maybe I should be aware of non-ceramic ones which would be useful for my application (filtering ac signal).

>>1998603
>>1998607
Just use an extension cord.

>>1998579
Put a small BT receiver, amp & speaker on his collar so you can talk to him and anybody near him from your phone.

>>1998706
Is that really a good idea? I don't think the dog would be smart enough to know what the sound was so all you would do is frustrate the dog. A haptic module sounds like a better idea; Then you could use reinforcement learning to get it to come back to you when it vibrates.

>>1998471
get a buffer with an enable line.
TXS level shifters from TI can be used for this.

>>1998474
>being able to edit the color of a component, time step, and voltage range is the same as being able to edit the full physics parameters of a device, Is, Bf, Re, etc.
anon you're a retard.

>>1998097
it's from https://e2e.ti.com/blogs_/b/powerhouse/archive/2016/11/18/input-and-output-capacitor-considerations-in-a-synchronous-buck-converter

>>1998716
That's a nice idea, but you can't make a panhandling dog with haptic feedback.

From pic, what is that Cff capacitor at the voltage divider feedback input for?

>>1998729
check the datasheet for the actual part you're using. that topic is nice for theory but doesn't help with design.
manufacturers have a vested interest in people using their part for designs since it means more sales so they always include a reference design. You should always start with the reference design or even better, cop the design from an eval board if available( the schematics and gerbers are freely available for TI and analog devices) then use that + read the relevant sections in the datasheet to tweak it to your needs. Pic related for some random buck regulator, as an engineer you decide what a tolerable ripple level is.

>>1998750
There is none in the datasheet, it's a chinkshit ic. I've checked LM2596 datasheet from ON and it explains there what a CFF is. Fuck chinks.

>>1998737
"ripple injector" for better full scale range frequency response.
Vout = 1.25(1+R2/R1) at DC levels since the cap is an open circuit.
as frequency increases, the capacitors impedance decreases which adds a frequency component to the above equation. as frequency gets really big, and Zc is really low, R2 becomes bypassed by the small capacitor impedance.
Then the equation at high frequencies becomes
Vout=1.25(1+0)=1.25V.
Basically what you get is ~100% of the HF voltage signal is fed into the FB pin. since FB is negative feedback the compensation circuit can negate ~100% of it to provide better regulation.

>>1998759
>chinkshit ic.
dont design with chinkshit for actual things. Fine for hobby shit but actual production that's a big no. to many fakes and poor quality control.
Plus proper parts are only a few dollars.

>>1998763
how do I calculate the CFF value?

>>1998766
it's for hobby and I thought I would take a look at it as my XL4016 module actually has pretty good efficiency and good enough voltage ripple

>>1998771
>how do I calculate the CFF value?
generally speaking, I would start with the reference design value unless I had a specific reason to change it.
Then during testing measure the HF ripple and empirically adjust as needed.

>>1998703
If I wanted to use an extension cord, I would not have asked for the connector.
Plus, an extension is not ideal in my use case.

>>1998850
Re-use the connector you have. Do you have a crimp tool and pins for your new cable?

>>1998850
cut and splice in an extension

>>1998866
>>1998878

Both are options I would consider, worst case.
But I would prefer to build a new cable to keep the original one untouched in case of a warranty claim.
I can't believe that this is a proprietary connector.
Maybe I can use female spade connectors. But a real crimp connector would be preferred.

>>1998900
It looks like a AMP/Molex locking connector, but I'm not sure.
https://www.molex.com/molex/products/group/power_connectors

>>1998900
If you can find quick disconnect terminals that fit the power header on the PCB, you could just crimp them onto your new cord and mold a plug housing out of epoxy putty or sugru.

You know it's really amazing electricity can do useful things at voltages and currents that don't instantly kill a person. We're quite fortunate that it worked out that way.

>>1999035
It isn't a coincidence that the forward voltages of semiconductors are similarly low to electrochemical voltages from batteries and the like. While current is what does the damage, and our current limit is arbitrary and set by our size, the currents that are required to deal damage to us aren't high by electrical standards at all. A 9V battery can supply sufficient current to interfere with a human heart. Skin is an insulating barrier, and so probably isn't very ohmic. Whether this effective breakdown voltage is something that scales significantly with size isn't something I'd know, but it would definitely be different for different classes of animal with different skins, be they vertebrates like lizards or birds, or invertibrates like arthropods or molluscs. So I guess we are rather fortunate.

Imagine an intelligent species evolving underneath an ice-shelf of a moon, and the kind of difficulty it would have developing industrial techniques like metallurgy or electricity.

I have four 5m strips of rgb leds i want to drive from an arduino. I figure i can connect all the data pins for that but obviously need a different source of power for them all. Each would need 5V and about 18A per strip maximum. With a 120V 15A wall outlet available, is there a way to get a power supply to handle this drop down for four 5V 18A connections from only one wall outlet connection? Bonus if the method is expandable to accompany more than 4 strips in the future.

>>1999213
5V @ 18A = 90W * 4 = 360W
Search for Mean Well Se-600-5.

>>1999226
That looks exactly like what I would need. Now I know this is all basic but I want to just double check my thought process on this. So I would want to wire everything in parallel to the output with a resistor to make sure that each section is limited in current to what its max should be. So that if I had an extra like 2 meter strip the strips would each still be getting the specific amperage they need due to the resistor limiting? So in this case, R=V/I, a strip needs 18A and the power supply has 5V for the parallel connections, so it would be R=18A/5V = 3.6 ohm resistor between each of the strips and the output of the PS. And if I added a 2 meter strip that demands 5A then its a 1 ohm resistor I would add in to the parallel connections on its output?

>>1999245
LED strips don't need current limiting resistors
also R = V/I, and you're supposed to set V to the limiter's dropout voltage, not the LED's voltage, which doesn't apply for a 5V LED strip

99% of LED strips already have resistors on them

>>1999252
So then for 99% of cases I should just be able to hook these strips straight up in parallel to the output of the PS and their resistors will handle dropping current to their needed values?

I have a bunch of devices (monitors, hard drives, wifi router + switch), each with their own individual adapter, which are all bulky, so I end up having two power bars just to power some peripherals.

On a scale of 1 to 10, how bad of any idea would it be to try to build my own PCB that steps power down to 12V / x amps; one adapter where all of these devices could connect to a single, more compact unit?

>>1999245
>>1999255
A couple of extra notes:
Don't exceed 2m per circuit.
Run multiple conductors for each strip, connected on both ends (2x 5V, 2x GND).
Fuse the output from the PSU just in case.
You may need to filter the PSU with a choke and caps.

>>1999258
You could probably power all of it with an ATX PSU.

>>1999255
>I should just be able to hook these strips straight up in parallel to the output of the PS and their resistors will handle dropping current to their needed values
Yes. They're designed to be easy to use like that. Shorter runs of 5V LED strip can be run directly off a 5V USB output (10W max).

>>1999258
>build my own PCB that steps power down to 12V / x amps
What, from mains? Making your own SMPS is a tricky thing at the best of times, but line converters are even worse.
It's not a bad idea to buy a PSU block and run a bunch of cables from it, but I'd have overcurrent detection on each individual output. Something like 1 comparator, one sense resistor, maybe 1 op-amp, and one pass-MOSFET per output. Probably add some extra circuitry for it to latch, requiring manual resetting (two universal logic gates), if not just use an MCU for all of it.
Not as in-depth as a SMPS design, but still could be fun. As the other anon said, an old computer power supply is going to be a good thing to run it off.

>>1999160
>Imagine an intelligent species evolving underneath an ice-shelf of a moon, and the kind of difficulty it would have developing industrial techniques like metallurgy or electricity.
If that were to happen, then it would need to either be the case that this species evolved a much less energy intensive form of intellect or there would have to be a sufficient energy density in this sub arctic ice biome to sustain the high energy demand of intellect.
Either situation would be fascinating.
I bet the latter, energy dense, case would involve thermal vents of some such.
Then the question becomes, what technology evolves as a result of this thermal vent focus?
Okay I'm done pontificating.

>>1999259
So are you saying to run 2 connections like in the pic I tried to make, and with fuses there to just be safe?
Also do you mean not to exceed 2 meters per circuit? why would that be the sort of limit per circuit?
Final also, I figure at the end you are talking about capacitors but what would a choke be?
I am pretty new to this and was hoping this would be a pretty simple start so I just want to be sure of these things.

>>1999282
>are you saying to run 2 connections like in the pic
Yes
>do you mean not to exceed 2 meters per circuit?
The wires and traces can only handle so much current before they turn into a cigarette lighter.
>what would a choke be?
It's a noise filter made of a coil of copper wire wrapped around a ferrite core usually. So you don't get flickering in the LED strips.

Help
my MC34063 is outputting ~190 mV, instead of -5 V
I corroborated all my calculations with an online tool and it shows the same results.

>>1999293
I'm doing it in a breadboard btw

Am I retarded? Or is inductance literally just negative capacitance?

>>1999279
Well, intelligence has to be naturally selected for, there needs to be an advantage it gives. Tool use is about the only way I can think of, but it needs to be more than just banging a shellfish with a rock, it needs to be a mammoth-hunting, brick baking, farming, brewing, trap-laying delayed gratification type of intelligence.
Hydrothermal vents are the most likely primary energy source for any sort of ice-shell life, if there was enough sunlight to be harvested then the ice would melt and the water would evaporate into an atmosphere and get whipped away by solar wind. But whether life there would be in great enough volume that they could stray away from the vents in order to hide from predators, or travel from vent to vent looking for prey.

>>1999296
It isn't, though it is an inverse form of it:
>V_inductor = L*dI/dt
>I_capacitor = C*dV/dt

>>1999298
so -j*(something) isn't just negative j*(smth else)?

>>1999293
>>1999295
You have a shitty connection somewhere, your power source sucks, or your chip is bad.

>>1999299
Read about audio crossover circuits for a better understanding of how power inductors and caps compliment each other. Capacitive reactance, high-pass and low-pass filters are good subjects for you to investigate further.

>>1999305
*are also good subjects

Do I need a flyback diode in an LC filter?

>>1999299
Yes, because the something ≠ something_else. One scales proportional to frequency, the other scales inversely.

>>1999322
Depends on whether the input of the filter is push-pull or not.

>>1999324
>Depends on whether the input of the filter is push-pull or not.
If the input comes from a buck converter, does that mean that I don't need a flyback diode on the LC filter as there is already one in the buck converter right?

>>1999326
There's already one in the converter. Sorta. If the MOSFET in the buck converter had a pull-down FET next to it, which pulled the input to the filter to 0V when the main FET was off, it would do exactly what the diode already does.

>>1999330
>If the MOSFET in the buck converter had a pull-down FET next to it
What do you mean by this?

>>1999333
Green is what currently exists
Red is what could exist instead (except for you know crossover shit)

>>1999348
ooh, so a synchronous buck converter. If the lower side MOSFET has internal diode that means even if it's turned off it's still act as a flyback diode, correct? This means there is no need to put another flyback diode on the LC.

>>1999252
>LED strips don't need current limiting resistors
wrong.

>>1999362
Sometimes an actual diode is added as a fail-safe measure. The body diode is a P-N diode (which has a much higher voltage drop than a Schottky diode), and its maximum current may be well below that of the FET. It's not needed if you're confident that the synchronous rectification can be relied upon.

>>1999373
>Sometimes an actual diode is added as a fail-safe measure.
That would be a good idea, I would do this then
>It's not needed if you're confident that the synchronous rectification can be relied upon.
I don't know if 5A for 680ns is enough to destroy the diode of IRLZ44N

>>1999368
lmao

>>1999378
Don't actually design a synchronous buck unless you're real keen, the anti-shootthrough requirement will mean having a dedicated half-bridge driver IC. The extra efficiency likely isn't too big.

>>1999391
>Don't actually design a synchronous buck unless you're real keen, the anti-shootthrough requirement will mean having a dedicated half-bridge driver IC. The extra efficiency likely isn't too big.
I'll use IR2111 for the driver as that's readily available for me

>>1999324
>One scales proportional to frequency, the other scales inversely.
that's fine, you just won't be able to make it resonate

>>1999395
> IR2111
Ah, that's where your 680ns is coming from. I guess if you're already using it as a high-side driver for your main FET there's no reason not to. 420mA is a little wimpy though, I bought a bag of IR2184s, which have almost 10 times that.

The IRLZ44N has no allowable operating range for its body-diode, and while you can probably extrapolate a reasonable one from the thermal characteristics of the forwards channel and the forward voltage of the diode at 5A, it likely wouldn't be accurate. In the ns domain, thermal constraints aren't a factor of how much heat the package or even die can dissipate, but rather how much heat the junction itself can dissipate.
Even so, I suspect you'll be fine, because you're barely over the continuous rating for the package, and the nature of a buck converter means there will be a delay before the reverse current starts flowing anyhow. I'd still do the calculations.

Anyhow, you're basically making a class-D amplifier.

>>1999406
>420mA is a little wimpy though
I know, but I just don't have the patience to wait a month from LCSC, I'll order them anyway then for future use
> In the ns domain, thermal constraints aren't a factor of how much heat the package or even die can dissipate, but rather how much heat the junction itself can dissipate.
Is it a good idea or waste of money if I also add a diode in parallel with the lower side MOSFET?

What's the most common barrel jack size?

>>1999413
3.5mm

>>1999412
>add a diode in parallel with the lower side MOSFET
I'm no expert on designing any kind of SMPS, let alone this arguably obscure variant of the buck topology. I'd consider it, though it might actually be cheaper to just stress-test a spare MOSFET with an arduino and a load overnight. It's not a bad idea at the very least, you could just leave some pads to populate if desired.

I'd see if there are any appnotes or datasheet examples on doing this with similar transistors.

SEE A TRANSISTOR
>take capacitor
>make resistor from pulp
How do humans even work if electronics doesn't

>>1999424
>though it might actually be cheaper to just stress-test a spare MOSFET with an arduino and a load overnight. It's not a bad idea at the very least, you could just leave some pads to populate if desired.
great idea, I'll run 5A from SD of IRLZ44N then, if it's fucked I'll add the diode

to charge a lipo with bench power supply should i set it like this?
3.7V lipo 200mAh 1S

power supply 4.2V, max current 200mA and just let it rip right?
since i'm charging at 1C temperature measuring should not be an issue

>>1999490

https://www.eevblog.com/forum/blog/eevblog-919-how-to-charge-li-ionlipo-batteries-with-a-power-supply/

>>1999490
>>1999496

actually, this other one is better coz it does in 2 mins what Babbling Dave does in 14 mins, coz of the 1000 useless tangential comments.
https://www.youtube.com/watch?v=0yhbof6_s64

>>1999496
>>1999499
thanks frens, well the aussies says to charge a .5C and the afro dude says 1C so i guess i will go with 1C since it's faster and the battery is completely cold so far

>>1999503
it's usually OK to charge modern lipo with 1C

>>1998965
Yeah, I think I have matching connectors. Good - and very obvious - point on the housing.
I'll just throw something together in cad and print one.

aaand it's done

the doggo will have to stay away from airports but it's great for finding him at night, in fact if i put it on max brightness it actually puts out so much light he can sprint around a pitch black forest at night
i programmed in various cool effects too, this is so much better than the shitty collars i saw at the store
now i just have to encase it in silicone to protect it from water and damage and it's complete

>>1999554
shit, i wanted to do a test drive outside but the doggo instantly started to divebomb the snow and the collar was soaked in snow, so i had to cancel the field test until i pot it in the silicone.
the only issue is that i have to leave the battery connector exposed in order to be able to charge the battery and i worry about moisture shorting it and thus turning it into a neck bomb like in pic related

>>1999695

- replacement dogs are cheap. you can even get 'em free at the SPCA
- water has a high resistance, like 5K, so shorts will carry just a few milliamps

>>1999713
>water has a high resistance
yes clean water, but snow has often shit in it, like for example what do they use to melt snow on the roads? salt. and what is the one thing salty water is know for? its excellent conductivity, so one dive bomb into a street snow and it's one way road to headless city

>>1999718
>dog runs outside while you're grabbing the mail
>massive explosion 13 seconds later

>>1999554
clear nail polish also works

>>1999718
>salt. and what is the one thing salty water is know for?

just measured
- 4 ounces tap water, resistance 11.5K at 0.1-inch distance
- same but with 1 teaspoon salt, 4.9K

result: no exploding doggos, unfortunately

>>1999725
depends on the salt concentration, but your measumerents seem off compared to some research

>>1999734
Is this Ohm per Meter or what?

>>1999734

so worst worst case is 185mA of current.
hardly explosive results.

>>1999695
Couldn't you put a fuse on it?

DIY plasma CNC anon here again, I have another question.

If I were to use shielded CAT5 cable for the data lines on my steppers and used seperate, non-shielded wire for power delivery what would I be looking at as far as filtering for EMI? Am I right in that I'd need a choke, and what size choke would I be looking at? Should I even worry about filtering the power lines?

>>1999746
Ferrite beads in the motor lines. You can put both the supply and return wires through the same core.

>>1999746
https://www.mouser.com/pdfDocs/schaffner-engineers-guide-designing-emi-filters1-13-17.pdf

>>1999440
Pretty sure 5A continuous will kill it without a heat-sink just from the heat, so do it at a 10% duty-cycle or lower.
Measuring the I/V curve both before and after the experiment.

Does anyone have recommendations for or experience with a relatively cheap (~$100) diy phono pre-amp kit? I really like the look of this:
https://www.tindie.com/products/skrodahl/muffsy-phono-preamp-kit/
but it's sold out.

I've re-capped an 80's amplifier to use with my speakers but don't have much experience otherwise. It doesn't necessarily have to be a kit, I can order parts separately but I do appreciate the convenience. Just getting into electronics and I've been enjoying it a lot

>>1999769
https://sound-au.com/project06.htm

>>1999746
The main thing is to put adequate capacitance at the driver end, so that the spikes are supplied by the capacitor and the cable only supplies DC current. Also, consider using differential signalling on the data lines.

Wait, what sort of noise does a DC arc make?

>>1995422
Hey, I have a question. I was working on something and I realised I have no idea of what the fuck are the op amps power supplies supposed to do. I thought they just limited the output voltage but that's not the case apparently. So what do they do? And I've seen them connected with for example -5V, 5V and 0V, 10V what's the difference, are they model related? Also I don't know anything about what's inside an op amp, (I will read about it in the future) so please avoid refering to that if you can cause I won't understand, I just want to know what do they do.

>>1999802
The inputs of the op amp don't actually flow through to the output . The output voltage is generated using the power supply.

What are the best materials for a DIY mat which bleeds static charge from my bench and is also acting as a good working surface? Because placing a bare metal sheet/mesh on my table is rather impractical.

>>1999802
Look at a simplified transistor diagram of an op-amp, say an old 741. There's a long-tailed pair on the input, and a totem-pole of some sort on the output. The power rails source the current to the totem-pole, limiting the output voltage range.

As for how you actually use the op-amp, if you're not doing a unity-gain-buffer, you'll need a ground reference. Common op-amps do not have rail-to-rail inputs or outputs, so chances are you'll be putting it halfway between the voltage rails. But sometimes you have an actual ground rail (dual/split supply) and sometimes it's implicit by just having a resistor to V+ and to V- where you'd have one to GND (single supply). Have a look at op-amp circuits on google images, most will be split-supply but some will be single.

this is the schematic of a 5v relay for arduino

i'm not understanding how in1 and in2 works, so you have +5V and GND, and a signal (in1 for relay1 and in2 for relay2) but how that works? i thought you'd only need to power the coil to work the relay

I want to use pic related for a button array. I still need pull down (or up) resistors for every button, right?

>>1999870
I'd hazard a guess that the current required to produce a a magnetic field capable of flipping the relay switch is either too much for the arduino to output or would damage the arduino. The optoisolator is pretty similar to the relay in that it can be used to drive a larger signal that's isolated from the signal source.

>>1999872
Not necessarily but you're more likely to have false reads if you don't tie it so it's good practice to.

>>1999802
>>1999814
>>1999840
Hmm... let me be more specific. I am designing a low pass filter (see pic rel) and so I have set both resistors to 10 kOhm (since I don't want any gain) and the capacitor to 100 pF to get the frequency I want. The problem I have is that when simulating it, it has a gain of -30 dB in the region that should pass (then it gets lower). I have tried increasing the specific resistor to increase gain but it doesn't go higher than -20 dB. What could be causing this? I thought it could have something to do with the power supplies... or maybe I have messed up something in the simulation... what do you think?

>been using Xpedition Enterprise to do PCB design for a week
>License costs like $70k
>has crashed around 10 times so far
I made a few changes to my schematic and wanted to forward annotate it to the PCB. Not even a lot of changes, I swapped around 4 ICs. Crashes every time I press the button.
How is this not robbery? You'd expect software for this price to be able to encounter errors without just shitting the bed entirely.

>>1999890
kicad

>>1999908
You think I paid out of pocket? It’s for work and I don’t have a say
I’m just ranting because it makes no sense to me that it’s this bad considering the license cost

>>1999884
What sort of filter topology are you using? A LPF should have a capacitor tied to ground so that it creates a voltage divider with R1.

>>1999884
Are you generating a bode plot? You might be getting a lower corner frequency than expected when simulating. If you're just spot checking at the given frequency then it could be that your capacitor needs to be adjusted so that the corner frequency is higher (lower capacitance).

>>1999884
>>1999938
>>1999939
Never mind, I found the problem. I had messed up the power supplies in the simulation. To answer your questions however, for the topology see the pic in my previous post and I was generating a bode plot it started from -30 dB instead of 0. I had conected to the op amp the wrong power supplies (I belive the specific model requires symmetrical power supplies and I had given it asymetrical. Once I converted them, it got fixed. Anyway I appreciate your help

>>1999939
>>1999969
>it started from -30 dB instead of 0
Sorry, I am still unclear. I know what you are talking about, I meant the whole plot was moved down by 30 dB and it had the expected shape

>>1999934
Then ask the chief pcb software guy to tell you how to do it. If he also complains about how bad it is, you might have good grounds to ask them to swap to KiCAD or Altium.

>>1999969
Why are you using that topology in the first place? It’s only really useful for being inverting and having gain (might also be a little more stable), without requiring gain you should just use a passive RC filter and an op-amp buffer if required. Doing so doesn’t require a split-rail power supply either.
If you want a better filter (2-pole instead of 1-pole, -40dB/dec instead of -20), make a sallen-key.

Ordered a 3D printer the other day, hopefully I’ll be able to strap a laser to it and get precision PCB making.

>>2000006
damn you almost got the get
>>2000000

>>1999986
>Why are you using that topology in the first place?
Because professor said so lol. We are using a chip that has 4 op amps so I guess he wants to use all of them? But yeah thanks for the advice. How is it that the op amp buffer does not require split rail power supply? And since we already have split rail power supply for the other things, are there any other benefits to the RC and Buffer idea? Because he said we could also do that.

I want to mix 3 audio signals (3 outputs of AY-3-8910/YM2149 chip). Can I just connect 3 of those signals into one signal that goes to the audio amplifier?
Wikipedia says that the simplest additive mixer is a resistor network, showing an example where each audio signal has it's own resistor before being connected to one signal.
Those resistors are only for volume adjustment between channels, right? I don't need them if the volume of each channel is similar, right?

>>2000039
monophonic signals?

>>2000055
Yep, 3 monophonic audio signals.

>>2000062
yeah dude the resistor network will work. You could use an Opamp for a good result as well.

>>1999884
as a general rule of thumb, if you have no experience, most opamp circuits are made to run on dual supplies. For single supply operation you need to bias the input signal somehow. Even stuff like hysteresis oscillators need that.

>>1998721
>you do not have frequency counter
>no fft
>no way to change timestep
>you actually can do all of those things
eat a dick and choke

Hi /ohm/. This is a guitar amplifier. But what kind of amplifier type is this? And how do check how much watt is has?

>>2000103
I do not know why falstad shows "edit component", the correct menu name is "other options" but it still does what it says.

>>2000084
Thanks, tested it on the breadboard, looks like i can connect all three channels with a single resistor into the audio amplifier, no need to use a resistor for each channel.

>>2000105
>But what kind of amplifier type is this?
An old one.
>And how do check how much watt is has?
With an ammeter and a frequency/sweep generator, or by multiplying input voltage and current if it's printed anywhere.

>>2000114
>An old one
How could you tell?
Any way to tell if it's A or B or A/B?

>>2000112
Will THD be a factor in your configuration?
>>2000117
I don't know.
http://www.tungsol.com/html/faqs11.html

>>2000125
>Will THD be a factor in your configuration?
Frankly I don't know even what THD is. I am a noob when it comes to AC signals and audio, but thanks for giving me a topic to study.
I am mostly playing with microprocessors/microcontrollers, interfacing them with other logic ICs, building computers etc, so I could get away with knowing almost nothing about AC or audio. But now i picked some old Yamaha frequency modulation synthesis ICs to play with, they are all really easy to interface with. I made some hardware music players with them, but the audio side of things is really stitched together, for audio amplification i just implemented the example circuit from LM386 datasheet without really understanding how it works.
I just couldn't be bothered since I needed it only for audio output anyway, but I really should pick up a book and learn the theory.

>>2000143
Basic rundown on THD :
https://www.youtube.com/watch?v=Xof9fvKzXr4

>>2000035
If you already have the spit rail supply, the inverting topology your prof gave you might be a bit more stable. Since a unity gain buffer doesn’t need to change the amplitude or polarity of the signal with respect to some central voltage, you don’t need that reference to make the circuit work.

Also you might be able to make a non-inverting circuit similar to the inverting one you already have, with the bonus of higher input impedance.

>>2000039
Two voltage sources in parallel is an invalid situation. Consider the current flowing between two signal sources I = (V1 - V2)/R. As R is decreases, I increases. If R is removed completely, you’ll fuck something up. Chances are it will be acceptable or at least non-destructive because the signal sources have their own output impedances, but you can’t rely on them being the same, which may cause differing effective amplitudes. So use resistors that are larger than those output impedances. If you want lower output impedance, put a buffer of some sort after the network.

>>2000039
>I don't need them if the volume of each channel is similar, right?

yes, you do. it's not wise to just collide signals without some way to smooth the addition, coz one will be stronger, or have less source impedance, and so one signal can backfeed and distort the other.

>>2000105
>And how do check how much watt is has?

sticker says 180VA, which presumably is input power, so expect output to be around 120W.

>>2000039
>I don't need them if the volume of each channel is similar, right?

yes, you do. it's not wise to just collide signals without some way to smooth the addition, coz one will be stronger, or have less source impedance, and so one signal can backfeed and distort the other.

>>2000105
>And how do check how much watt is has?

sticker says 180VA, which presumably is input power, so expect output to be around 120W.

>>2000188
An inverting summing amplifier is even better for minimising distortion.

>>1999763
I know, I'll time it for 1uS pulse so 10% duty cycle PWM at 100kHz

I want to charge my watch without the magnetic dock.
These 4 pins should basically be just usb cable connections right? so if i connect +5VDC as marked it should charge, is that correct?

>>2000445
Not necessarily, but it's a good bet. Measure the voltage on them before potentially messing with it. And Use pogo-pins instead of soldering.

>>2000446
what is the forst scenario here? it won't work ir it will kill the watch? if i set cc to like 100ma then it should be fine right?

>>2000446
>pogo-pins
i want to hotglue thin wires there because i want to wear it with them, basically i want to be able to charge the watch while wearing it because i HATE having to take it off every day for charging and this way i can make some small magnetic connector and charge it while sitting at my computer which i do like 12 hours every day anyway

>>2000105
>how do check how much watt is has?
You'll have to do a frequency analysis on the output to see what frequencies it passes. From there you'll need to input a constant frequency sine wave and measure the voltage/current output at the speaker.
The power transformer says 180VA, so assuming a class AB amplifier, your output power wouldn't be more than 120W if its a really efficient AB design.
>>2000117
>Any way to tell if it's A or B or A/B?
Based on the TO-220 ICs on the heatsink at the bottom, I'd guess its a class AB.

>>2000117
>How could you tell?
it's all THT and those caps and resistors it uses haven't been in use for decades

>>2000117
Not him but perf board and "point to point" assembly" is not used in consumer products anymore. It takes too long to assemble and you need to hire a dude to do it.

>>2000445
Make another band with the charger and wear the watch over top of it.

I want do build a dimmable lamp with some design constraints and I don't know how to go about it. What I want is:

- 50 Hz (mains) powered or equivalent so it flickers (unnoticable to the human eye)
- Very low noise, both radiated and fed back into the mains supply, hence why I'm hesitant to use a triac like a normal dimmer

Best I can think of is a variac but they're huge (at least I haven't found a small enough one yet), expensive and then I'd have to worry about hum and magnetic fields

>>2000611
What kind of lamp? Tungsten, LED, gas?

>>2000612
Irrelevant as long as it fulfills the other requirements. It would be nice if it was rather effective and not very hot

>>2000616
It is not irrelevant you retard. Different lights have different dimming techniques

>>2000560
>>2000572
>THT
It is a DIY amp

>>2000617
I'm sorry, I meant that the choice of lamp is not a design constraint - I'm open to using whatever (within reason) as long as the system meets my requirements. I aware that you don't dim a tungsten lamp in the same way you dim a LED

>>2000622
Create a mechanical contraption that turns the switch on and off at 60Hz

>>2000621
Are you sure? Old regular stuff looks exactly like that. But whatever, you can use the xformer as a top limit for the power rating and call it a day.

Guys I have several questions:

Do I need gate resistor if I use MOSFET driver such as TC4420?

To calculate gate resistor value for N-Channel MOSFET, I need to find the gate current by doing Qtotal / trise, but for P-Channel MOSFET do I need to calculate it by doing Qtotal / tfall ?

>>2000705
What does the datasheet says?

>>2000715
Which datasheet?

How many miliamps can they take?

>>2000721
the driver and the mosfet
>>2000742
usually 300mW or thereabouts

>>2000752
I probably got the answer

1. TC4420 has internal high and low output resistance, I probably don't need to add gate resistor, pic related

2. Depending if I need gate resistor value for discharging or charging the MOSFET gate

Let me know if I'm wrong

>>2000742
the datasheet says 0.5W

>>2000460
>100ma
More like 1mA.

>>2000611
Why do you want it to flicker? LEDs can turn off much faster than incandescents, meaning they invariably flicker if they're being turned on and off at 50Hz.
Best option in my opinion is to run an LED array to a variable constant-current buck converter running off a AC-DC power brick. But that wouldn't flicker. Replacing the buck converter with a PWM dimmer would give you your flicker, and likely be completely visible.

>>2000831
>Why do you want it to flicker?
Anon might be doing stroboscopic effects.

>>2000838

50/60 Hz strobes are no fun. you gotta move as fast as Sonic to see any action. 5-15 Hz is good for dancing. 15-30 is for watching things like water flowing from a faucet. higher than that and the eye smooths over the flashes.

>>2000845
What about tachometers or video effects?

>>2000849
tachometers have variable frequency duh

>>2000849

for tachometers, you need to vary the frequency to match the rotational speed. so a 1500 RPM motor would need 1500Hz flashes.
in old turntables, they had a 60Hz light flash onto the platter, but in order for it to work, it needed 100's of equally spaced painted dots painted around it, coz speed is so slow.

>>2000838
>stroboscopic effects
Considering he said he wanted:
>50 Hz (mains) powered or equivalent so it flickers (unnoticable to the human eye)
I'd wager not.

>>2000859
>a 1500 RPM motor would need 1500Hz flashes
rpm ≠ hz

>>2000849

for tachometers, you need to vary the frequency to match the rotational speed. so a 1500 RPM motor would need 25Hz flashes.
in old turntables, they had a 60Hz light flash onto the platter, but in order for it to work, it needed 100's of equally spaced painted dots painted around it, coz speed is so slow.

>>2000831
Anon >>2000838 and >>2000859 are right on the money.
I wanna build a lamp for my turntable shelf. Some of them use strobe discs (without any built-in lights).
Proximity to the turntables (as well as a few tape recorders) is why I'm anal about electrical/magnetic noise. The signals coming out of the pickups are only a few mV and the preamps are very sensitive to all kinds of interference, which is why I'm skeptical towards using a triac or something of the sort which would introduce transient noise and whatnot back into the mains and along conductors

(disclaimer: I barely know what I'm talking about)

>>2000878
In that case, you don't need something terribly bright at all. A single powerful LED should be sufficient. The proximity to the turntable could be vastly expanded by using a lightpipe on the LED. I'd also put the flickering circuit inside a grounded metal box and have all wires coming out of it be twisted pairs, possibly with shielding. With plenty of input filtration too.

>>2000878
Maybe you could etch copper tape with a pattern like a rotary encoder and wrap it around the edge of the turntable. Then use optical feedback to control the motor.

>>2000883
if it's a dc motor, he could literally make it a phase locked loop, with the dc motor being the vco

>>2000881
Oh yeah, I should have probably mentioned from the get-go that I don't need a floodlight. The lightpipe idea is neat, I never would have thought of that, thanks.
Unless I am misunderstanding you seem to be under the impression that I want the flicker to be visible? I want it to appear constant to human eyes but allow the stroboscopes to function.
I've thought about just making some sort of flickering circuit and just shielding/filtering the shit out of it, but unfortunately there's no protective earth on the outlets in my shitty rental apartment (which I probably also should have mentioned)

>>2000883
>>2000886
Panasonic engineers did this in the 70s, coincidentally one of my TTs is a Technics direct drive which does just this

Anyways this is besides the point, making my own TT from scratch is a project for when I grow up and get my own machine shop

>>2000151
Thanks, very interesting, I can see there is a lot to learn in audio. I will need to work out the basic issues first though, like the radio interference or the hum from power supply. I suspect that at least the radio interference should get reduced when i solder everything to the perfboard, those long breadboard wires certainly don't help.
>>2000165
>>2000188
Thank you, now I understand, makes sense since the chip is basically a black box.

>>1999872
nah, you only need to pull down the pin being read.

>>2000892
In that case, I’d make a simple little power supply out of a toroidal transformer, step it down to 4VAC or so. Toroidal means less noise. Have some class X noise-suppression capacitors on either side of a couple of chokes, probably both common mode and ordinary/single. A fuse wouldn’t be a bad idea either. Then use a bridge rectifier to some sort of linear current limiting circuit. Use an optocoupler before the rectifier to get your 50Hz signal, might need some extra circuitry depending on what kind of flashing you want and to stop the option from dying.

>>1995422
does anyone know a single inductor buck boost converter ic (>5A or supporting external mosfets) that is available cheaply on chink websites? (XL60X9 wont work)

>>2000878
>The signals coming out of the pickups are only a few mV and the preamps are very sensitive to all kinds of interference, which is why I'm skeptical towards using a triac or something of the sort which would introduce transient noise and whatnot back into the mains and along conductors
Put them in a metal box. The noise a triac would generate with such a small load would barely be noticeable.

>>2001023
Does LCSC count? Because just search for the specific category (switching controllers) and narrow down using its parametric search function.

>>2001037
I think he’s more worried about the TRIAC’s mains ripple, not it’s EMI. Still, so long as there’s a line filter before it, with a small load it should be fine. Using a TRIAC dimmer on a mains LED is a bit of a shitty solution though.

>>2001123
not really when shipping is 5x part cost

>>1995422
Will something like this work for switching battery and output to a buck boost converter?

>>2001130
What’s the free shipping threshold?

>>2001201
too much

>>1999798
>Wait, what sort of noise does a DC arc make?
not much. A DC arc (between carbon electrodes) makes a hiss.

>>2001288
I meant what kind of noise on a spectrogram. It's DC, so what determines the range of frequencies?

So I finally got myself a wire crimper since I don't want my wires to slip out of terminals. Problem is, after looking into it, it appears that my ferrule colors are abso-fucking-lutely not adherent to the color standards they should be, or at least not any I can find. To add to the confusion, pic related says 0.5mm^2 wires should be put in the white ferrule, and the box says that white ferrules should be used with 22AWG, but the charts I've found say that 20AWG is 0.518mm^2 and 22AWG is 0.326mm^2, which is definitely undersized for that.

What do? I've got 18AWG and 20AWG wires I need to crimp and no fucking idea what size ferrule I should be using for them.

>>2001374
>>2001374
>>2001374
NEW THREAD

>>2001130
stop being poor and order in bulk

>>2001392
but i dont need to build 50 dc-dc converters

>>2001516
order all your resistors and capacitors and other shit you'll use in the same order
if they sell tools then get some of those too

>>2001721
i buy resistors locally for cheap. everything else is available on aliexpress for really cheap with free shipping.

Lads, help me out.
So I have this U-229 to kenwood connector (basically from radio to headset: microphone, speaker, push-to-talk). After I opened I noticed something weird. I have to this day connected all the pins to correct places with few resistors if needed but Ground and Microphone ends hand some weird things going on.
Might this be a amplifier circuit or some audio quality circuit or what is going on?
Can capacitor microphones even be amplified without external power? I imagine the sound quality would suffer.
I made a shitty picture of the circuit cause the whole thing was covered with hot snot. I'm 95% sure it was wired like this.

>>2002154
GO TO THE NEXT THREAD:
>>2001374
>>2001374
>>2001374

(also use actual transistor symbols)

>>2000000

>>1997242
Check the charging port for obstructions and bad solder joints.

>>2002407
the other get was in the /rcg/ thread btw