/diy/ - Do It Yourself

>>1566896
this thread's digits were brought to you in part by the MAX6896 ultra-small power sequencing/supervisory circuit. yes, power is boring and tedious, but power is everything

>>1566896
I how do I get into aliexpress plc?, just want flow meter.

>>1566907
>>>1566899
Most, if not all commercial buck,boost, inverting and buck-boost stuff can`t be hacked. It`s just a special purpose IC + passives. All you can do is use the IC for something else, because they are made for several modes of operation.

>>1566844
Pic related is one kinda shit way to do it, probably close to what the CC/CV module is doing
another is to connect the LM259x in constant current mode, with the load -ve side connected to the feedback terminal, and a power resistor Rsense of (Vfb / Imax) from there to ground. then use a comparator to simply shut down the LM259x when the desired voltage has been reached
you could combine the two solutions to save loss in Rsense by choosing Rsense small and feeding the output of the current sense amplifier section of pic into the 259x's feedback pin, and using the other op amp as a comparator as above

>>1566923
stop posting

what crimp tool is the best one? should i get one that closes like pic related
or the one that looks like aligator mouth?
i want it for putting various terminals on various size of wires, all the way from bread board wires to outlet wires

>>1566937

>>1566937
that one looks like it's only for "bootlace" ferrules
you probably want one like >>1566938 illustrated. see if you can find one with interchangeable dies

I need some advice everyone I am very new in designing electronics and I made a dc power supply with a required regulated output, assume that my voltage regulator is LM317 because I don't have that transistor in Multitism and the variable resistors are in fixed position will they get the output that I wanted ?

>>1566940
can the blue one also do bootlaces?

>>1566942

Not without fucking up the ferrule, no.

>>1566944
This pretty much contains all terminals i need to use.
I asked ping pong which crimp tool he recommends and he recommended the red one

>>1566941
What output do you want?

Circuit looks reasonable. You'll need to add a big capacitor across the input if you want to draw much current, 0.1uF isn't much for the 120Hz filtering, something on the order of hundereds of uF is what you want. Some capacitance across the output will also help.

>>1566957
3v, 4.5v,6v,9v,12v dc output.

>>1566945
I don't see any "Dupont" terminals in there

>>1566941
bretty gud but there are some finer points you missed
you need a bigger value for C1. a lot bigger, like 1000µF is probably a reasonable minimum. it has to supply current to the regulator and load for up to 10msec
the datasheet recommends 240 ohms for R2, so that the 317's minimum load requirement of 5mA is dissipated by the voltage divider. otherwise your output voltage will be too high under light or no load
you also need an output capacitor. it doesn't have to be huge, the datasheet recommends no more than 10µF on the output unless you add protection diodes
if the rotary switch happens to fail to make contact, output voltage will shoot up to the voltage across C1 minus 1.2V. if you only change the voltage selector with the power off and measure the output voltage after changing it, that's not a big deal

>>1566941
The key is understanding that LM317 will always try its hardest to put 1.25V between its ADJ and OUT pins, and ADJ will draw practically no current. This is why they use the Vout = 1.25(1 + R2/R1), because R1 and R2 share the same current.
As for the load, LM317 needed a minimum load of 3kOhm if I recall correctly, so you would need to put a 3.3kOhm resistor in parallel with any load you want to use.

>>1566961
I really messed glad I asked this.

>>1566964
cheers. practical electronics design is 50% physical science, 50% library science

>>1566945

Who the fuck is ping pong?

Also, again, the 4-jaw crimp tool flat-out won't work on most of those terminals. I wouldn't be surprised to find out it doesn't even fit over the larger ring/fork terminals.

I got one of [pic related] from Home Derpo: https://images.homedepot-static.com/productImages/c3399dd6-0f14-4b67-a231-58c2b4b3207d/svn/commercial-electric-wire-strippers-ce140901-64_1000.jpg

Could probably do similar for cheaper online, but I had a job to do that involved a lot of crimped terminals and needed it quick. The cheap, thin, stamped-steel ones that come with every low-end terminal kit are uncomfortable to use and have a tendency to damage the insluation on the terminal. They do work quite well, at least, but, then again, you can crimp that kind of terminal with a pair of pliers in a pinch, even if it doesn't give a great crimp.

>>1566968
>Who the fuck is ping pong?
every salesman on aliexpress

>>1566844
I tried this with the LM2575-ADJ and it works very well as a switch-mode constant current source.

I improved on this circuit by replacing the "high-side" resistor/transistor network with a "low-side" sense resistor connected in series with the load. The sense resistor provides a voltage signal corresponding with the load current. The sense voltage is fed into an LM358N op-amp configured as a non-inverting amplifier with a gain of 2 or so. Its output then drives the FB pin of the LM2575. This allows to use a sense resistor with lower resistance value (lower power loss, better efficiency).

>>1566896
I see this is a smoking room again? You realized vapes are just as annoying?

>tfw waiting until I get done with this Disney trip to see if I have any money left over from my tax return to buy a Chinese hot air station

>>1566981
Ooops, misread the OP

>>1566968
I don’t care for much from Channellock, but I like the crimpers I got from them. The cutters are stronk too, the thing works so much better than those all in one crimper stripper deals like this Klein or the cheap ones that come with the terminals.

>>1566981
Only poor people smoke cigarettes

I want to use one of this MP1584 based buck converters to power my camera. How much can I trust them? I read that if the regulator fails it will send supply voltage through.

Does via stitching actually help with noise?

>>1566978
Why did you ignore the 5.4V limit? The $1.50 CV/CC modules can do both.

>>1567025
then use a zener

>>1567028
More for high freqs. (Designers don't stitch just because it's fun)

>>1566806
>>1566807
>>1566834
Oh that makes a lot of sense. I quite like this idea of thermal PWM.

>>1567050
or heat dissipation

>>1567050
But it is fun. I imagine it turns two ground planes into a single surface.

>>1566941
LM is not good at suppressing large ripple like the output of rectifier.
It's good at keeping the mean voltage precise.
You need much more capacitors, and preferably some ripple filtering before LM.

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

Also keep in mind the huge heatsink.

>>1566928
>changing resistors
Arduino general is on the other side my friend

Soooo for this, to keep current from flowing backwards when the panel isn’t getting enough light, would it be a Schottky diode? How do I know which value to use?

For this, to keep current from flowing backwards when the panel isn’t getting enough light, would it be a Schottky diode? How do I know which value to use?

>>1566899
This is how it's done. Now you can "hack the module".

>>1567109
Get a diode rated for your maximum current. That looks like a 500mA array at best so basically any DO-41 diode will work (1A max). You can quite easily look up the current tolerances for common diode packages, along with any other component package you can think of.

>>1567109
You want to use a MOSFET connected as a diode. You can use a P-channel like in the pic or a N-channel with the difference for N-channel being you'd connect the gate to the input power source (the solar panel) instead of ground.

A few things of note: a lower Rds(on) means lower power losses across the FET. Also since your input power source being a solar cell is presumably relatively low voltage you need to pick a FET with a low Vgs(th), low enough that your nominal solar cell output voltage under typical light conditions will be able to comfortably put it into its saturation region with some room to spare.

>>1567125
>not using the FET as an ideal diode
The guy probably isn't packing enough current to bother with a TO-220 package, but for larger projects it's a good suggestion.

>>1567125
>>1567132
Oh and surely the forward voltage of the substrate diode would be the cause of power dissipation far more than the R_DS(on), no?

hey guys
I want to replace this dumb mini b connector with something more robust, ideally a USB-C.
I can solder but some more technical stuff is beyond me. Any suggestions?

>>1567050
(I stitch just because it's fun)

>>1567135
mini USB on a board designed for it is gonna be more mechanically reliable than trying to stuff a C somewhere.

>>1567133
No because that diode doesn't conduct when the transistor is saturated. It's forward voltage is 0.6-0.7V or so, maybe 1V at high current. Meanwhile the drain-source voltage drop is on the order of 10s of mV or less with a suitablyblow Rds(on). All the current flows through the junction, none through the body diode.

>>1567137
Well the connector is already dead, and if I'm going to replace it I don't want to have to do it again. I could just hardwire it and be done with it.
One thing to remember is the keyboard came with both a USB cable and a PS/2 cable, so it might have some extra functionality.

Is there a "more robust" mini b connector I could drop in?

>>1567141
But unless you have a circuit that turns the junction on (you don't because the gate is commoned) the current will have to flow through the body diode.

The ideal diode circuit turns the junction on when the voltage at the panel end is higher than at the battery end, and lets the current flow through the opposite direction to what you'd usually do to a NFET; in the same direction as the body diode. But current can flow in both directions through this junction, so you need the circuit to turn the FET off when the voltage at the panel end is lower. The FET has to be reversed to what you'd usually use because a FET's body diode will conduct in the reverse direction. Pic related.

>>1567143
Hard wired and clamped well should be pretty reliable. The Mini connectors are pretty standardised, not gonna get something significantly tougher that still fits the board.

>>1567169
Grade of metal is likely to play at least a small part, perhaps the chinesium connectors are more likely to break from metal fatigue. Not that I'd know where to buy quality metal connectors. A bit of silastic could help, especially to support the surrounding PCB to stop it flexing with respect to the case.

>>1567135
That one is pretty robust when used with plated holes

>>1567190
>plated holes
yeah it didn't come like that

>>1567135
i have the exact same keyboard. i've repaired the usb connector twice. fuck rosewill.

>>1567190
>>1567194
It's a single-sided board, which I guess is the problem.

Soldering a USB socket pigtail to the existing wires is arguably a good method, provided you secure the wires on both ends to prevent fatigue (epoxy, silastic, hot glue, etc.).

>>1567125
>>1567124
Hmm ok. So with the MOSFET, that basically keeps the switch deal open until the voltage coming from the solar panels would be high enough to actually charge the cells?

Gotta go back and creep the Amazon for that DO-41. Would pic related be worth getting to have around for my steep learning curve?

>>1567198
where did you get the connector from?

>>1567268
the connector isn't what breaks, the solder joints crack from the leverage. a pad has lifted on mine too but i frankensteined it. here's the connector though:

https://www.digikey.com/product-detail/en/molex/0548190519/WM17115-ND/773802

>>1567271
also you won't be able to find a more robust usb connector with the same pin configuration that you can just solder in. the 5 pin is unique to the mini/micro for usb otg.

>find exactly what I'm looking for
>local distribute discontinued it
FUCK
Can anyone think of something similar to this? I need an rf transmitter/receiver for audio
http://www.cdt21.com/products/audio/wa03/default.asp

>>1567246
You mean to say it keeps the switch closed until the solar panel can charge the cells? In other words it keeps the MOSFET open as long as the current is trying to flow out of the solars, not into them. But for your purpose a simple diode (or MOSFET with commoned gate) should work without an issue.

The diode bundle is a pretty useful thing to have, but you'll never use half of those linear regulators. I'd just get a bunch of LM317s, 7805s, and whatever the 78XX thing is for 3.3V. Linear regulators like that are basically only used to get a nice smooth voltage for small-signal/digital electronics which usually only run at 5V or 3.3V. The 317s will solve any issue that doesn't fit into that. Though you might want to instead look specifically for "low dropout" regulators. With non-LDO regulators you need about 2V of headroom, which won't work too well if you're stepping from 5V to 3.3V, which is somewhat common.

>>1567275
Perhaps look for that on digikey, figure out what subcategories and specs it's under, and find similar replacement parts.

>>1567118
Thank you anon. That was precisely what I was searching for. I'll buy the chink module and integrate it into what I'm doing.

>>1567282
I found a diode kit with a coiple more values, only $7 so I’ll order that eventually. I was thinking about repurposing this little board from the original light, but it’s a surface mount diode and the power to the light is controlled by a daylight sensor.

I don’t know what I want to do. Those solar panels originally each charged 1x 3.7v battery, but I put them in series for some reason and now they’re putting out high 4v on my workbench but low current. I guess I was just playing around to see if I could trickle charge something off the workbench lights. Maybe with bright sunlight it could trickle charge my 50cc scoot.

>>1567289
Repurposing any board is a nightmare if you're not already skilled in circuit-bending.

When a circuit is simple but still won't fit easily on perfboard, what's the best way to make all those jumpers? Stranded wire? Thin or thick single core? Should the wires be floating over the top of the components or hard against the other side?

>>1567292
Those things are old, I ripped them out of the front yard because they were done. I’m not trying to reuse those boards, that’s why I wanted to figure out the diode thing.

>>1567311
They sell specific jumper wire spools. I think those ones are solid. I been meaning to get down on more silicone wire spools because PVC is annoying with small stuff.

>>1567351
Agreed, PVC insulation on hookup wire just melts if you spend too long tinning it.

>>1567311
wire wrapping wire, 30AWG, for most signal stuff
for power stuff you might want something thicker. bookmark this
http://williamsonic.com/WireTable/index.html

>>1567354
in answer to the perennial question "but my solder wire already contains flux, why should I add more?" some highly active flux is helpful for just this sort of thing

>>1567431
>highly active flux
What does this do to PVC? Dissolve it?

I usually just used the wikipedia page for awg, but that source is quite nice actually. I've heard other people use wire wrapping wire, but I already have some ~28awg solid core wire for breadboard jumpers, will there be much of a difference between the two?

>>1567438
no, it cleans the end of the wire more vigorously so that solder sticks to it more quickly
>28awg
any size should be fine as long as it bends the way you want and the insulation isn't too thick to fit

>>1567442
I don't find the melting insulation stops the solder wetting at all, it just leaves the stripped tip too long, renders the remaining insulation near the end too soft, and makes a bit of a mess. The first issue is quite problematic as it effectively strips the wire back past where it was twisted and tinned up to, leaving some untwisted strands free to bending and metal fatigue.

>>1566984
great tools as always. thanks for sharing

what are these small sdjustable components in a power supply? what function do they serve?

>>1566981
Yihua are pretty good, I got the 835D for $100AUD and it's been great, nice soldering iron with good array of tips shipped with it (Hakko tip compatible too!) And the hot air does the job with a good temperature range, comes upto speed real quick and auto-turns off when you dock it

>>1566981
>avatarfag bought more stuff than I ever did
>never did anything with them
epic, you should buy a book you dumb shit

>>1567462
I haven't a clue, can't find any.

>>1567462
in what power supply? May be voltage adjust, current limit

cant post pics from phone. 4chan is annoying as fuck on mobile

How can you tell if a board design is good enough EMI-wise or not? I'm 100% sure this would work with no problem because I've made a worse one and it works perfectly but how can I know how much interference it would pickup from stepper motors/brushless spindles/rotary tools, etc? The prototype, which uses a daughter board for the microcontroller, doesn't seem to suffer with any of that, but I don't have a scope know for sure.

BTW, in another day I was saying that EasyEDA was good. Well, for the better part it is, but the autorouter is garbage. On Eagle I can for example break out the pins in a small IC like this MCU, do some basic traces like power and certain signals and the autorouter will find usable a solution 99% of the time, but EasyEDA's will just time out.

>>1567524
>How can you tell if a board design is good enough EMI-wise or not? I'm 100% sure this would work with no problem because I've made a worse one and it works perfectly but how can I know how much interference it would pickup from stepper motors/brushless spindles/rotary tools, etc? The prototype, which uses a daughter board for the microcontroller, doesn't seem to suffer with any of that, but I don't have a scope know for sure.
Shielding, filtering, and routing noisy traces away from sensitive traces. By skillful design and confirmation with EMI/EMC test equipment. Henry Ott's books is required reading, it explains all this, including how to test.
Also get a scope.

>>1567524
by testing, of course. get your shittiest sparkiest Dremel out and run it near this thing for a while
compare your Vss/Vdd at 35/36 vs. the pair at 47/48. guess which one of those is horribly wrong
don't be too scared of hand-soldering a dozen or two caps to the other side of the board for decoupling. fast processors like it when you do that
don't try to use the on-board pullups as your only means of pulling an input up. try to make sure there's at least 1mA pull-up current, maybe more
give serious consideration to adding a couple of power planes on their own layers, or at least give more priority to keeping your planes whole by e.g. bundling your trace runs together. that flimsy little 5V trace running around the board scares me. if you really want to be a kinky little slut use your outer layers for the power planes and use vias to get out directly at your component leads
also get the Ott book
>autorouter
ahh that explains it

>>1567550
>get your shittiest sparkiest Dremel out and run it near this thing for a while
That's something I could try right now, even thought it doesn't represent the typical usage case in the very least

>compare your Vss/Vdd at 35/36 vs. the pair at 47/48. guess which one of those is horribly wrong
Neither of them looks wrong

>that flimsy little 5V trace
Well I didn't set a specific netclass for it so the autorouter just went with the minimum trace width but it's actually fine since the stepper driver internal logic doesn't use that much current at all (8ma at most)

>don't try to use the on-board pullups as your only means of pulling an input up. try to make sure there's at least 1mA pull-up current, maybe more
The chip supposedly has between 30-50k ohm internal pull up resistors, but I designed the board so that external pull ups could be used on the endstop sensors themselves.

>bundling your trace runs together
I've been recommended to do the opposite to minimize interference

>if you really want to be a kinky little slut use your outer layers for the power planes and use vias to get out directly at your component leads
That's a good idea to prevent reverse-engineering. Most hobbysts wouldn't bother delaminating the board or running it through a x-ray machine.

how do I git this gud

>>1567571
>put spaghetti in pocket
>drop it on board

>>1567475
>buying books
I bet you pay for porn

>>1567584
Thats no spaghetty, comrade, this is glorious soviet multi-strand wire in teflon insulation - MГTФ. Outstanding material for prototyping, shame it was made only in one colour. To git this gud you need to live in late 90`s-early 2000`s Russia|Ukraine|Belarus and be engineer too broke to buy sound card for your PC

>>1567472
Pic related is the one that some anon has shilled a bit. There’s a handful of them that are all about the same for $70-$100 so I’ll just get whatever has the best reviews and features when I eventually do it.

>>1567450
No problem bby

>>1567571
1. live before laser printers were everywhere and you needed to lay out pcbs with dry transfer sheets and resist pen
2. be too lazy to lay out anything but the ISA connection and a ground bus
3. etch it
4. do a messy job at cutting the pcb to fit
5. use a horrible mess of wires to connect everything

>>1567597
That one is pretty big in real life

>>1567561
Continued: ripup;, then I try to break out every important pin from the MCU, lay down the basic power traces (5v, 3v3), run the autorouter and... it misses a single signal. But it's getting closer, I think I need to space the vias a little bit more.

>>1567619
I think this was another one that has been mentioned. Has good reviews compared to others.

>>1567626
Why not just an 858D and a separate iron like the T12?

>>1567275
>>1567275
>>1567275
samefag here

would a regular RF transmitter/receiver like these work for audio transmission? does anyone know?
http://www.saelig.com/RW433/RW044.htm
http://www.saelig.com/RW433/RW034.htm

>>1567636
That was an option too, but people seem to say that these stations are decent and lots of them support Hakko tips and are almost as good once you do that. It’s only like $20 more to get the soldering iron in addition to the hot air, so if it sucks and I still want the Hakko or T## after, it’s no loss.

I still have yet to plug in this Weller that I bought months and months ago, that’s the medium duty one for like $25 from Home Depot. I wanted to abuse the shit out of the $4 iron and learn what I’m doing first.

>>1567642
>it's only like 20$
You can get an 858D for 30$

>>1567646
Hmm... there’s a handful on Amazon for $40-$45 shipped. I’m sure I could find cheaper on Ebay or Ali, bur Amazon is too easy. It’s something to think about, but I’ll probably end up with something like this >>1567626 because of the good reviews and extra heating element and lots of tips. I went back and forth on those Txx things and all of the programmable features don’t really appeal to me, the main reason I would want one is to build something to run it off a drill battery. And if the soldering iron on the all in one thing leaves me wanting more even with good tips, then I get a Hakko for $100.

>>1567524
Good enough as in works, or good enough as in would pass FCC testing? Very different targets.

>>1567028
>does reducing the ground plane impedance help with noise

>>1567028
>>1567663
Coming soon to adafruit!

>>1567619
>888A
it's not some little FX888D but it's not too bad
>30-50k ohm internal pull up resistors
for efficiency. that's about a tenth of what you would need if you have serious interference flying around
>Neither of them looks wrong
35/36 was good. 47/48 was not so good. directness is important
>do the opposite to minimize interference
particularly sensitive traces like to be spaced 4 widths apart, this is true, but not at the expense of hacking up your ground plane
also, that USB pair definitely prefers to be routed together and somewhat length-matched (see section 7 of the USB 1.1 spec)

>>1567641
nope too slow
any particular reason you don't want to use bluetooth audio?

How do I get 3.3v from 3v without a voltage regulator?

Alternatively why does a voltage divider that goes from 6v to 3.3v drop the voltage to ~1v when connected to a load?

t. Brainlet trying to power 8mhz atmega328 with two cr2430

>>1567707
An Atmega328 can operate with 1.8 - 5.5V.
You can directly use the 3V of the cells.

>>1567712
Forgot to mention that it has a screen that requires 3.3v to 5v

>>1567707
>How do I get 3.3v from 3v without a voltage regulator?
Schottky diode to drop 0.3V jej

>Alternatively why does a voltage divider that goes from 6v to 3.3v drop the voltage to ~1v when connected to a load?
Because any load you connect changes the characteristics of the voltage divider. Voltage dividers are only useful if you are driving a high impedance.

>>1567721
>Schottky diode to drop 0.3V jej
You got it mixed up. He wants to step up the voltage.

Easiest way is to just use a small module like this:
https://www.aliexpress.com/item/DC-DC-0-8-3-3V-to-3-3V-Step-Up-Boost-Power-Module-For-Arduino/32840571878.html

>>1567707
Voltage dividers work on the principle of the same current going through both resistors. As soon as you try to power a load in parallel with one of those resistors the current going through them both becomes different.

I'd put the cells in series to get 6V and use a 3.3V linear regulator. Though you can test to see if the system will run fine off the battery voltage, which it might. Regulators are handy shit so you should get some regardless.

>>1567738
>6 to 3.3
>linear regulator
You're pissing away half your battery life that way

Dumb question:
If I have 20 sensors all giving 5v to 20 individual pins at the same time, will I fry it?

Basically I have 20 rfid readers. They each have a RESET pin that I can read as HIGH or LOW, but only output LOW when a tag is read.
I want to read them all individually, but they all output 5V all the time.

I'd think 100V divided on the 20 pins would fry it, right?

>>1567746
Sorry, I'm using an Arduino.

>>1567746
Voltages don't just stack up when you connect them to an arduino
so no

>>1567765
okay cool, thanks.

Would this work?

>>1567778
Yes, it heats, fries and zaps.

>>1567778
lol nice bait

>>1567778
replace R2 with a 5V zener and it might sorta kinda work for a minute
btw they have ICs just for off-line bucking, e.g. ST's VIPer06/16/26. if you're going to kill yourself do it in style

>>1567746
You can look at the ATMega's effective input impedance diagram and see that it shouldn't cause an issue. More theoretically, it's like all the sensors are in parallel since they're all referenced to a common ground, and since each requires not much current (though you should check that) the arduino isn't overburdened. The arduino has a maximum current limit shared by the pins, so check that your RFID scanners' maximum current draw is lower than this limit.

>>1567748
I'm sorry too.

>>1567778
Probably, if you want a fucking house fire. Any reason you don't want to just buy a $1 flyback converter?

>>1567585
>scientific calculators
>hasn't learned ohms law
ok

>>1567778
google capacitive divider

Retard here, I've wound up some inductors with some randpm ferrite, I don't own a lcr meter but I have access to a scope. How would can I measure the inductance using the step response? My idea is
>feed square wave into a series LR circuit
>measure voltage on resistor
>find the point where it's 67% of the peak value
>that=time constant=L/R
>L=time*R
Is this correct? I don't own a signal generator nor can access one, but I have a pwm thingie I bought.

>>1567852
This is true.

Am I measuring voltage of the power source with no load, voltage drop, or voltage under load?

>>1567908
Both can be handy, though looking at the waveform under load is more useful.

>>1567906
it's 1-(1/e) actually, 63.2%

>>1567906
That will theoretically work, but there are two main problems with it. First is time: a DMM will usually only update a few times a second for whatever reason, so you'll need to have enough of these measurements over the duration of the measurement in order to pinpoint the 63% mark more precisely. For this reason I'd consider measuring after a set amount of time (in time with sample periods) and calculate how many time constants it was from the measured percentage instead. Second is the impedance of the DMM: if you construct an RC/RL circuit with a 1MΩ resistor and your DMM has an input impedance of 1MΩ in parallel with the resistor, the system will act like it's got only 500kΩ instead. You can take this into account in your calculations though.

Personally I'd make a relaxation oscillator with a 555 timer and the RC/RL circuit and use the period of oscillation to calculate the inductance or capacitance. This time measurement can be done trivially with an ardiuino or other MCU hooked up to your computer. Note that the ESR of an inductor needs to be much less than the resistance of the main resistor. I've done this only once before, but it worked relatively well. The period was tuned somewhere in the ballpark of 1Hz, then I measured the time for 20 oscillations with a stopwatch, since setting the frequency lower would mean very large resistances, and I couldn't be bothered busting out a dusty AVR.

>>1567906
https://www.youtube.com/watch?v=74fz9iwZ_sM

>>1567778

>>1567906
Can also try measuring the ringing frequency with a known capacitor.

>>1567992
excellent

>>1567990
>>1568042
Only really possible with a scope, if he has a scope then I imagine the time constant method will be pretty easy assuming he doesn't run into issues with the scope impedance.

>>1568047
>Retard here, I've wound up some inductors with some randpm ferrite, I don't own a lcr meter but I have access to a scope
>I have access to a scope

>>1568051
Oh, sorry for not reading then!

>>1567778
I don't see why not

>>1567619

>autorouter

>>1567707
>>1567731

Ok I have found this, it boosts from 0.9-5V to 5V, can I re-purpose it to make it output 3.3V?

>>1568103
Run it through a buck converter

>>1568103
Change the feedback resistor ratio until you get desired output voltage.

>>1568104
That's stupid. You're stupid.

>>1568103
just get a 3v regulator you mongoloid. if you have to ask the question if a surface mount pcb can be repurposed, the answer is almost certainly no.

>>1568105
>imblyign it was any more stupid than your idea

>>1568105
I guess I'll try this and if it doesn't work I'll order proper parts.

>>1568109
I don't want to wait a week+ to finish the thing Im doing.

>>1568112
>I don't want to wait a week+ to finish the thing Im doing.
patience is a virtue, boy.

>>1568103
Googling the "HXN-Xh" print on the IC, it seems to be a Chinese clone of the LTC3429.

By adjusting the voltage divider on Pin 3 you can change the output voltage 3.3V:
https://www.analog.com/media/en/technical-documentation/data-sheets/3429fa.pdf

>>1568103
Those R1 and R2 do look like a voltage divider for feedback purposes, I bet they're the divider in that datasheet. Looking at the values the threshold voltage is 1.21V (so it working down to 0.9V might be a long shot if it doesn't boost it's own power rail). To get from 5V to 3.3V, R1 would need to get smaller, which you could either do by replacing it with a ~26kΩ resistor (27k would be near enough) or bodging a ~58k in parallel with it (56k is probably good enough), which might be a better idea if you only have THTs and not SMTs lying about.

>>1567968
It's a scope, not a DMM. And the pwm thing

>>1567952
Yep, thanks man.
>>1567990
Thanks.
>>1567992
top kek
>>1568042
I think I'll go with this because it's more reliable than cursors.

>>1568177
I'd agree that the ringing method is probably best if you have a scope.

hi /ohm/
anyone know how i can monitor the electric signal coming out of a buzzer with an arduino ?
would a simple transistor be enough ?
i don't need to get the frequency or anything just know when some sound is outputted

>>1568103
You can "make" a regulator with a transistor, 3 resistors and a zener. Get a 4 V zener and you are set.

>>1568193
You can make a simple linear regulator out of that, but he needs a boost converter.

>>1568192
I'd put a diode detector circuit on one of the buzzer's pins, and use a common ground between the device with a buzzer and the arduino. You could also use an electret mic with an easy 741 amplifier circuit and feed that into the diode detector circuit.

>>1568197
No he doesn't. Is this a meme?

>>1568200
He has 3V and wants 3.3V.

Hello,

quick question regarding serial communication with Arduinos.


Iam currently building a model railroad using Arduinos to automate everything.

I have 5 Arduinos named A0-4.
A0 is the master.
A1 to A4 the slaves.
They are wired like this: A0 TX goes to RX on A1-4, A0 RX goes to TX on A1-4.
All Arduinos are connected to the same Ground and 5V.

Communication looks like this:
A1-4 are constantly listening for messages
A0 sends out a message using Serial.println.
The message is composed like this: the first two characters address the targeted Arduino "A1" for example. Then I add zeros and ones to tell the target what to do. The message is decoded by the target.
Then the target sends out an answer in the same way, A100100 for example.

This all works perfectly fine as long as only two Arduinos communicate.
When I add more Arduinos I still can send instructions from the master to the slaves but I wont get an answer.


Can someone tell me if this is a hardware based problem?

>>1568201
>0.9-5 goes to 5 V
>needs to step it DOWN to 3.3V (are you retarded?)

>>1568204
https://www.arduino.cc/en/reference/wire

>>1568221
Iam not using I2C.

>>1568220
Follow the replies, he has 3V lithium cells and wants to drive something that needs 3.3V.

Or are you suggesting he should boost the 3V to 5V and then step it down to 3.3V with a crappy linear regulator?
Just change the configuration of the boost module instead.

>>1568204
You really wired the TX of A1-4 in parallel?

>>1568234
Yes.

>>1568236
If you have a scope you can see what happens to the TX signal.

>>1568238
Unfortunately I do not have a scope.

>>1568239
then stop being so poor and get one

>>1567681
>also, that USB pair definitely prefers to be routed together and somewhat length-matched (see section 7 of the USB 1.1 spec)

Good catch. One of my USB data lines in an existing board is twice as long as the other one. The board works as intended, but I fixed this anyway for when I need to make more of it.

>>1568222
Then use I2C since it's made for this kind of usage. That UART is for only two devices, it's not really a "bus". You can have more than one receiving, but not more than one transmitting.

>>1568222
y not tho
https://en.wikipedia.org/wiki/Bus_contention
quick fix: put a 2k2 pullup on your master Rx line, put diodes in series with your slave Tx lines (cathode toward MCU), slow your serial link way down to like 9600 or less

>>1568251
>>1568252
I cant use I2C as I do not have enough analog pins.

>>1568256
I2C was made for this exact thing, and uses SCL and SDA pins.

>>1568200
Thx mate i will try it
can i use any diode or do i need this fancy one in pic related ?
i have also considered thoose mic board but im afraid it will add some response time i don't want

>>1568261
They're pins A4 and A5 on a ardwayne

>>1568270
only if you're detecting RF
for a buzzer input, a slow diode like a 1N400x is plenty

>>1568279
give it back, atmel

>>1568252
What if he uses some relays to do the switching?

>>1568256
stop stealing pins Massimo

>>1568308
what if? draw picture for maximum mockery

this is fucking weird
i use one of those chinks sources that converts 230v to 5v and i had mc connected to it which keeps a relay on to test how it works, and it keeps the relay on nonstop, but about once of twice a day i does a click as if there is not enough power to keep the coil engaged, but how can there be enough power 99.9% of time and then suddenly not?
whoever wrote the code for the simulation that runs our universe is a retard

>>1568239
>I do not have a scope
I learned that even a toy scope is better than no scope because it can tell you _why_ something doesn't work. Anyway, this is the suggested solution, don't know if the pull-up resistor is really necessary (internal pull-up?).

>>1568345
In a pinch, he could use also his sound card as a scope

what's a good variable power supply?

>>1568355
If you now run your stuff from batteries, any such power supply is good enough.

>>1568355
get old PC PSU for free, attach to an adjustable chink buck converter for $4, done

>>1568355
Fluke PM2811 digital bench supply. I got one used and refurbished, it's pretty good.
It acts as a constant voltage / constant current supply, or optionally will fault to zero output with overcurrent protection.

>>1568345
Thanks for that diagram.
Are you sure the diodes are in the right direction? Dont they block everything coming from the TX pins?

>>1568389
If a TX pin is 0, it'll pull down the voltage at RX A0. Otherwise the pullup resistor keeps it near Vcc

>>1568341
are you drawing too much current from gpios and causing mcu to shutdown?

>>1568345
built-in pullups are on the order of 30k. the stronger the pullup, the faster the possible data rate

>>1568355
one that honors the voltage and current limits you set
this anon >>1568379 has the right idea
>$4 chink buck
yeah, for ten pieces maybe

>>1568389
no, they're correct. the diode passes current only when the output of its slave pulls down (NB only one slave can talk at a time). when the slaves pull up, they don't pass current and the pullup resistor puts a high level on the bus. it's known as a wired-AND connection (term often used interchangeably with wired-OR)

>>1568379
>yeah, for ten pieces maybe
Not him but you can get this kit for under 4$:
https://www.aliexpress.com/item/0-30V-2MA-3A-adjustable-DC-power-supply-laboratory-power-short-circuit-current-limit-protection-DIY/32780330525.html

But it doesn't seem to come with a heat sink for the tranny.
And you'd also want a volt- and ammeter module (around 2-3$ on Ali).

>>1568393
>ten pieces
If you want it to handle any current that in a and not ma you need a good one with heatsinks and quality caps for stability

>>1568373
No, they avoid mutual loading of the TX ports. That's why I added the bits of a byte. The idle state is H=1 and if one of the TX outputs pulls the line LOW, the other ports are automagically disconnected. Be sure that only ONE of A1..4 sends at a time.

I can imagine that there's also a software solution (no diodes), at least I know how I would do it, but I don't know enough about the port set options yet.

>>1568389
No, they avoid mutual loading of the TX ports. That's why I added the bits of a byte. The idle state is H=1 and if one of the TX outputs pulls the line LOW, the other ports are automagically disconnected. Be sure that only ONE of A1..4 sends at a time.

I can imagine that there's also a software solution (no diodes), at least I know how I would do it, but I don't know enough about the port set options yet.

>>1568405
>at least I know how I would do it
Please tell me.

>>1568419
I would initialise all four TX ports to hi-Z (disabled, or maybe as dummy inputs) and only the bus device that wants to send enables its port as output (default high) and disables it again when done.

>>1568440
Yeah that is what I was thinking too.
I will try that.

>>1567135
USB-C connector pinouts are... Diverse.

Anyone can point me out in the right direction, I'm kinda stumped
>need a circuit that limits the current and has short circuit protection
>in the event of a short it doesn't reach the current limit setting, instead it puts out a very small ammount of current until reset

>>1568440
sure, if the port will let you unhook the transmitter like that
aside, the diode trick is a nice way to make CANbus or any other half-duplex shared bus work over a desktop range without transceivers

>>1568449
this is 4chan, you can say mentally ill here

>>1568471
like Pic related?

>>1568345
How about this one?

>>1568474
does that actually work? seems a bit wack. usually you'll see the base of q2 biased with resistors and a Rsc to turn on at Vbe

>>1568477
>>1568471
might help design better
http://www.eeeguide.com/foldback-current-limiting/

>>1568475
I have one, it's fine for audio stuff but your rising and falling edges might be too fast to be captured well. It's only single channel but I think it has an external trigger, along with a 1kHz frequency reference. It's also an illegal knockoff of a product that is being sold for like twice the price, but you already knew that.

>>1568317
I have a client that every now and then will non ironically ask me to design a board according to a schematic like pic related.

He'll also say stuff like
>it's a small relay, it doesn't need a transistor
>it's a small motor, it can be driven straight from the arduino pin
>make it single-sided so that the product can be much cheaper, no SMD components since they're more expensive
>those relays are basically free since I have a lot of them in stock (they're discontinued and cost over 10 dollars each to buy, so for every board he sells he's throwing away like 20-40$ by not reselling the relays themselves and using a more modern option)

The last one was like this:
>this transistor steals too much power from the motor, this is why it spins so slowly, so I want you to use a miniature relay because it's "cleaner"

He had a little board with a motor on the emitter of a transistor, then I built something on a breadboard with a much smaller/cheaper transitor but with the motor on the collector and then it would spin as if it was connected directly to the power source, i.e: with no visible power loss. He reasoned that since a transistor was like a faucet it didn't matter if the motor was placed before or after it. It does.

>>1568482
Forgot pic related

>>1568475
that could be surprisingly not bad for $25

>>1568482
kek
what sort of special hourly rate do you give him?

>>1568481
Why even a scope?
Sounds like what he needs is a basic logic analyzer which is cheaper and more useful than those shitty toy scopes.

>>1568486
a scope is the yardstick of civilization

>>1568487
And the posted scope is a tiny plastic ruler with faded markings.

Save your money for something actually usable.

any of you anons work with TFT/LVDS display interfaces? what do you use to generate your bias voltages? know of any neat integrated 1-chip solutions for bias voltage generation, especially such as that might be adjusted by an MCU?

>>1568482
>>1568483
That schematic belongs on the arduino hall of shame. My biggest pet peeve with the tarduino ecosystem is that it enables a know-nothing """maker""" culture to design utter crap.

>>1568499
Sounds more like some incompetent boomer.

>>1568499
meh. I can't say I'm not a grateful beneficiary of the convenience and ubiquity of breakout boards "for arduino"
let them design. I mean, it's probably better to have them put something together that does something of some kind, if that's what's going to pass for thinking in this enlightened 21st century. however, I start to have a Problem with it as soon as they start pretending they know something about electronics, rather than the cutesy little walled garden of mechatronics they inhabit

>>1568515
you're probably right. an incompetent """maker""" would have surely used a CD4066 or two

>>1568499
I could see using Arduino crap as a proof-of-concept for something real, then when you work out the details you design the real device using real parts, but using Arduino as part of a real project or product? Laughable.

>>1568490
It's less than $30, it's hardly a steep investment. I'd much rather buy one of those than a $150 scope that I'll want to replace in a few years. If the thing combusted upon turning it on for the first time I wouldn't particularly mind. When I want to buy a decent scope, I'll buy a fucking good scope. Until then, I'll settle for this.

>>1568490
Sometimes a shitty plastic ruler is the best option, because the other two options are nothing or a set of micrometers.

>>1568537
Even 30 bucks is too much.
That's already half of what a cheap Hantek USB scope costs. Which isn't a great scope but a million times more useful than this piece of shit.

If you don't care about the quality at all get a 1$ usb sound card and use that as scope. At the very least you'll get a better user interface.

>>1568542
>USB scope
To be honest I'd definitely consider getting one if it worked with my operating system and included sufficient isolation from my computer. The sound card method is quite nice aside from the isolation issues, but as I deal with both analog and digital stuff I prefer having the option to measure with DC coupling.

>>1568474
>>1568480
>>1568477
Thanks

>>1568551
Then again, I'm using a 4channel scope right now and I'm really liking having this many channels. Pic related, a RGB PWM prototype for a driverless RGB COB. I'm at my local maker space and might get a PCB made on their CNC router for this purpose.

>>1568475
No experience, mine is much older.

>>1568551
With those cheap sound cards you can just bridge the caps on the input to get rid of the AC coupling.

>>1568557
Oh really? I might make a "USB oscilloscope" as my next project with one of those soldered down onto a board with peripherals. I'd want to find a multi-channel USB sound card though, probably designed for stereo microphones. Might do something gimmicky with the audio output also.

>>1568529
yet, half of the 3D printers in the world today...

>>1568551
>operating system
OpenHantek
>sufficient isolation
looks like it's about $100 for a USB isolator that you could use for your logic isolator and whatever other test gear too

>>1568559
"USB sound cards" are often just an ASIC on a board with jacks and passives, see TI PCM27xx series for example

>>1568562
>about $100 for a USB isolator
Can't you just put bargain bin opto-isolators back-to-back?

>>1568565
no, too slow and not bi-directional
if you only need 12Mbps there are $10 boards with dc-dc converters and USB-aware monolithic transformer isolators (ADuM4160) but if you want something faster prepare your bunghole to pay audiophile prices

>>1568529
Even more so, you can load the shitty bootloader onto a DIP ATmega/ATtiny, program it on a cheap ZIF board, and solder it directly onto your PCB. Which is probably what I'd do for making less than 10 products if the bootloader doesn't take too much room. But that's also because I'm a lazy shitter who falls back on the piss-easy arduino documentation for easy sensor interfacing. There's basically no excuse for using an arduino board (especially an uno) as a permanent part of any project, except for the fringe cases where the board is not overkill and needs no peripherals.

>>1568559
Shouldn't it be possible to just add an additional USB sound card for every additional channel?
Dunno if any of the software supports that, though.

>>1568571
I imagine audio mixing software should have as many mic inputs as you could ever want, since recording studios often have a mic for every instrument and vocalist. Including a USB hub IC into my build could be viable, though not necessarily the smoothest method and it would require some semi-high frequency PCB design.

>>1568570
or just skip the bootloader and load the hex file?

>>1568572
they also very often run a master clock between all their interfaces
tbqh you may as well just use a bluepill (or design your own equivalent) and enjoy the two 1MHz ADCs onboard

how the fuck am I supposed to know how channel selection works here? this is all it says in the shitty chink documentation
http://www.saelig.com/supplier/airwave/awm630.pdf

any guesses?

>>1568589
first page:
>Channel selection is four channel and default value is ch4 (not enable). Other channel can be selected by pulling low to enable.

>>1568592
So I would supply 5v to those 3 channel pins and to change the channel from 4 I would turn off the 5v to a channel?

>>1568595
you would ground the one you want to select and leave the rest alone

>>1568554
>tfw everyone in the maker space refers to end mills as drill bits

Fellers I accidentally bought online some surface mount ICs, I need to test them as I am experimenting with some designs I have in mind.

What is the best way to test them? Do I need to make a PCB or is there an easier way? the shit's too fucking small and I wont be able to solder wires on it.

>>1568674
if you can wait for chinks, use these
https://www.aliexpress.com/item/35pcs-7value-5pcs-PCBenis-Kit-SMD-Turn-To-DIP-Adapter-Converter-Benis-SOP-MSOP-SSOP/32959444043.html
>too fucking small
if you're in a hurry, you could dead-bug them with wire wrap wire and hot snot, if your soldering rig doesn't suck

>>1568674
There are some sort of SMT sockets, though the only ones I've seen are like $30 each and for ~100+ pin QFPs. Have a look on alibay though, might find something. Theoretically one of those breakout boards (>>1568676) clamped to your IC with a fair bit of force should do ya well, though I'd just solder solid core breadboarding wires (28ga or so) to each pin, plug those into a breadboard, and desolder when done.

>>1566896
I'm trying to have a 4PDT switch, without the size+cost of a big toggle

I want each switch to throw between two different audio signal paths (1Vp-p max)

Is an "Analog Switch" what I'm looking for?

TS3A24159
pic rel

>>1568163
checks out.
nice job!

>>1568685
yeah that's probably fine; THD is important to consider but looks that chip has less than .001% up to 20k so no biggie

>>1568704
there might be a better alternative though; i'm not hip to modern audio trends

>>1568704
thanks, homie.

these are a much better solution. I was using a PIC, but I'm trying to make this cheap enough to manufacture.

>>1568705
no worries, I plan to do a deep dive to find the right part.
mostly I just want to know if an analog switch is the easiest way to do this. This is the first I'm I'm working with them and not an actual toggle or button

On my 3d printer I have this switch which gets pressed by fillament to signal to the printer it hasn't ran out, but sometimes i get a false positive that the fillament ran out even though it did not and when i checked with a multimeter, the switch was pressed and and closed, so how in the fuck did PI get an open reading?
the reading is done on gpio on a raspberry pi, which pulls an input pin high, and the endstop keeps it pulled to ground in low state and it filament runs out it opens and the the raspberry reads the pin as high (with nothing to pull it low)and stops printing
so with a simple system like that how in the fuck am i getting false positives?

Can you recommend me an entry level FPGA kit please? Is better Xilinx or Altera? Digilent kits look nice but they are really expensive.

>>1568858
I had to use altera in my university, the software is one of the worst things I've ever seen. Also
>intro class to digital logic
>forces people to use fpga (which are a subject of their own)
>waste more time dealing with the shitty software then actually learning digital logic
>each board was like 100$
They could have bough a truckload of CD4xxx and those 74h chips and a bunch of transistors..

>>1568918
unis are always retarded like that
why not just buy your own for $5 on ali and use one of the awesome free software like fritzing for designigs the circuits and shit like that?
I had the same deal with programming, we hard to learn these completely obscure retarded languages just because they were the only ones the prof was good at and was lazy to learn any popular ones and you literally never used them ever again after passing that class
honestly they should just sell the diplomas for like $100k a pop so you can just buy it and don't have to waste time with learning retarded stuff you will never use ever

>>1568923
Stuff on Ali is completely unacceptable for a school.
They need something with guarantees and someone to call when things go wrong.
It's not different in the industry, except that schools are way easier to rip off.

>>1568925
and from where do you think the people who supply the industry buy it?

From the same chink you are except the they put on a price 500% markup just for fun

>>1568947
People that have to stand for the quality of what they're selling and do actual QC and provide support. People you can hold responsible and sue.
The chinks on the other hand don't give a fuck.

>>1568679
you don't need much force at all, lest you bend the pins. a spring clothespin is more than enough

>>1568858
Lattice's iCE line if you're just starting out and fiddling around. probably not enough to fit much of a soft CPU, but plenty for an SPI/I2C/etc. slave, and anyway working within resource constraints builds character. the software is quirky but not terrible

>>1568952
>support
>for dev boards
really now

>>1568925
>students are way easier to rip off
fixed

>>1568960
>support
>for dev boards
>really now
Of course not, they'll just hand the bare board for you without any support.
What you want software to actually use it? Too bad. Download some freeware made by hobbyists.
Have questions or problems? Ask on /diy/.

>>1568962
>Of course not, they'll just hand the bare board for you without any support.
>What you want software to actually use it? Too bad. Download some freeware made by hobbyists.
>Have questions or problems? Ask on /diy/.
say that to my arduino's serial input not online and see what happens

>>1568838
do a percussive diagnostic

>>1568923
My problem wasn't the price of the boards, more like
>digital circuits class
>learn digital circuits class in a thing which needs it's own class to be used
>fucking VHDL
We were doing just finite state machines, sequential and combination logic. ALL of that could be done MUCH faster if we used discrete components instead of the altera dev board and that cancer software. The result is that a class that was suposed to be easy, and really cool, was a pain in the ass because of
>pedagogy

>>1568977
I mean, Apollo's computer was made with fucking discrete NAND IC's the size of my dick, my end project in that class was to simulate a Microwave oven.
>insert time
>choose mode
>door open (y/n)
>begin heating
>leds blink in heating pattern
>timer ends
>beep
>leds blink the ready pattern
>go back to state 1
I wasted an entire semester of my life in something that now I see it was trivial, because we were learning VHDL and troubleshooting with the vodoo software (it was called Quartus I think) instead of learning about digital circuits. I mean, even if we were messing arround with ttl logic gates I'd have enjoyed it because
>blinky LEDs

>>1568977
Makes kinda sense for FSMs, doing ROMs with discrete components is a pointless waste of time.

>>1568679
>though I'd just solder solid core breadboarding wires (28ga or so) to each pin, plug those into a breadboard, and desolder when done.

Yo that costs more than $30 to do

>>1568522
>would have surely used a CD4066 or two
But it's too expensive and it uses too much power! The relay is much cleaner. It's a small relay, so you can drive it straight from the arduino pin.

>>1568962
the software usually works just as well against a clone board as against the original. a clone Altera ButtBlaster is identical to the original. they don't care, they just want to sell chips
exception: Xilinx are douches

>>1568964
topkek

>>1568978
well that's how digital circuits are done these days. they're preparing you for industry, not lolduino. whatever shit they sublicense from Synopsys/Mentor and slap their own name on is the standard in industry. also
>VHDL
>2019
who is there left not using Verilog by now
>freeware made by hobbyists
unironically tends to have friendlier interfaces not burdened by 25 years of bad decisions
>pedagogy
but are you really going to learn much about timing constraints, setup/hold times, etc. with a bunch of discrete logic?

>>1568986
I ask again, how much of a stupid premium do you charge this guy

>>1568994
>I ask again, how much of a stupid premium do you charge this guy
I charge too little to be honest, like 40$ for a board design

>>1568994
>unironically tends to have friendlier interfaces not burdened by 25 years of bad decisions
Unless it comes from the Linux crowd, e.g: Pinguino

>>1568390
>>1568393
>>1568405
Not him but I still don't quite understand...
each explanation seem to contradict eachother in my head

https://youtu.be/bkqoKWP4Oy4?t=333
What program is this guy using at 5:33?

>>1568994
>but are you really going to learn much about timing constraints, setup/hold times, etc. with a bunch of discrete logic?
It`s a first semester class, we had other classes entirely dedicated to projects and project management.
>clunky interfaces and shit solutions
>development boards
>using the wrong tool for the job
>industry
lmao, where have you worked?

>>1568978
I did the same microwave end project
I think it's a set of requirements for engineering schools handed down from ABET.
microwave, elevator, traffic light, etc

>>1569025
Probably EasyEDA he has shilled for it in other videos

>>1569029
Thank you

>>1569018
the pullup puts a 1 on the bus. the idle Tx level of each slave is 1. when any slave Txes a 0, the bus gets pulled low through the diode. due to the diodes on this and other slaves, none of them can pull the bus high with a 1. they can only stop pulling the bus to 0 and leave the pullup to pull it back to 1
some multidrop protocols cast the 1 and 0 as recessive vs. dominant states, if that helps your understanding

>>1569026
>timing constraints
>project management

>>1568345
Thanks, I added the diodes and resistor and it works great now!

>>1568440
This however did not work.

>>1569051
>Thanks, I added the diodes and resistor and it works great now!
Does it really? How do you know when it's time for each module to start talking/listening?

Is using a relay in a CD spot welder ok? Won't it weld itself?

>>1569039
>when any slave Txes a 0, the bus gets pulled low through the diode
how exactly does the diode do that?
I thought diodes only let current flow in one direction and a forward voltage is required

>>1569109
The diodes prevent the outputs from sourcing current. With a diode in the path they can only sink. The forward voltage and the current through the diodes is supplied by the pull-up resistors.

>>1569053
iiuc it appears he's using a master-slave protocol with pre-assigned addresses for each slave node

>>1568983
?

>>1569076
Provided it's rated for that current it should be fine, it won't be though. The relay will open and the current will just keep flowing through an arc between the relay contacts, which will melt the relay. That's if you don't get contact welding in the first place.

>>1569076
Put a diode across the relay coil.
Put a 100Ω resistor in series with a .1µf capacitor across the contacts.
If the relay is an automotive starter relay, you should be fine.

Are there any arduino-likes for audio DSP? Zoom G1 Four uses a Texas Instrument chip and it sounds like 1800$ Kemper. Imagine there was an open source version of it that you could buy for 50$ from Aliexpress that you could just put inside a steel enclosure and program to sound like whatever you wanted it to

>>1569144
How much would you charge someone to solder a chip, assemble a breadboard, program it, and then dessolder it? With this little puppy you could spend only 1% of the time. That reminds me I should stock them for resale.

>>1569172
Of course you use one of those if you're doing more than 5 or so products, but the guy said he just wanted to experiment with designs, not program every single MCU he goes through.

>>1569169
you could get close to it with a bluepill + a codec, but I can't speak to latency
maybe they know about duinoshit here >>1567647
>but yeah would be kinda cool

>>1569185
After a bit of searching it turns out that there is something just like that. It's uses an ARM core from Itead + a base board with random components. You can reprogram it and it sounds pretty good. The only problem is that it's crazy expensive (i.e: more expensive than a Line6 Helix stomp) and comes from Germany, after starting in Brazil (something good coming from Brazil for once). I don't understand why they didn't move to Shenzhen instead since it would be so much cheaper to make it there and easier to sell it to all over the world.

Wanting to get a small, 2-layer (45x35mm) THT PCB fabbed, is OSHPark the one to go for? I only need 1 and the 3 option from them is cheaper than PCBway's minimum of 5. Thoughts? I might try to bundle in another larger PCB if it makes sense, but the global free shipping from OSHPark looks really appealing.

>>1569256
I guess the fact that there aren't knockoffs on ali answers that question? For now at least.

>>1569172
>How much would you charge someone to solder a chip, assemble a breadboard, program it, and then dessolder it?
You don't need socketed programmers, I haven't touched one in over a decade.
Every MCU on the market lets you program flash in-circuit, using (where applicable) ISP, ICSP, SWD, JTAG, and others I'm sure.

>>1569260
JLCPCB might still be running their $2 special. worth considering if you wanted to order some SMD components from LCSC along with your board and save a bit of shipping. LCSC doesn't exactly specialize in THT though

>>1569256
cute
if you didn't need the box or the kewl display, you could just get a pi and add a decent entry-level audio interface to it. even if you did, you could add pedals quite easily, as is well practiced and repeated by arduinans

>>1569153
>>1569144
I don't know exactly how many amps a Capacitor discharge welder can give, my fear is that in the nano second that a 5000A passes through the contact it'll do what it's suposed to do and weld two pieces of metal, albeit in the relay.

So, one year now in electronics. I made a linear supply like 10 months ago and it works, it's ugly as shit but works.
But I want to make something smaller and that doesn't get so toasty, so I'm getting one of those MC34063 (I know they are old and you probably have a 0.2x0.2mm smd IC with 40 pins with 90% efficiency to recomend)
.

But to get rid of noise and ripples and shit, I want to make a variable linear regulator on the output (basically something like the supply I have) in the output, in a way that both the switching and the linear reg, change voltage outputs together (keeping a 2 or 3 v headroom).

This way the linear regulator heat will depend mostly on current instead of the voltage drop (which will be constant-ish).

Now tell me, before I start soldering stuff, how stupid is this idea? And putting an LC filter with low corner frequency or a cmm choke in the OUTput of the supply is sensible?

Why is there no SPICE with a decent frontend?

>>1569300
Then you'd need to ensure that a significant amount of heat is not being generated at the contact interface; i.e. that the contact resistance is much less than what you'll be welding. I imagine if you use a starter motor relay this should be taken care of.

But why not just use an array of IGBTs or MOSFETs?

>>1569300
If you're worried about it, use MOSFETs.
The contacts in the relay I posted are larger than pencil erasers.
The contacts at the welding points are the size of pencil lead.
Current will be limited by the capacitors and wire but also by the welding probes.
You can't get more current through the relay contacts than through the welding tips.

>>1569304
What you described is basically what the high-end benchtop switching power supplies already do. Naturally, you need some form of feedback with a low-pass filter in it and some method of giving a ~2V voltage difference. If you already had a discrete PSU designed, incorporating this wouldn't be terribly difficult, but trying to do so with SMPS ICs like the one you posted might be more difficult. You could try looking up what ICs are used inside those expensive benchtop power supplies (dave reviewed a few a month or two ago) and ordering one and slapping it on a breakout board, if THT is too much for you. Else a few hours messing about on spice might result in a modified feedback circuit that works. There may also be hobbyist documentation of doing exactly what you're describing, so have a look on EE forums I guess. I was thinking I'd have to buy a PSU to do what you're describing, but if you figure out a good method (hopefully one that can go to 0V) be sure to mention it on this general.

How can I charge a 300v capacitor up to 300v?

>>1569341
PS: using a 3.7v lithium battery

>>1569341
use a boost converter with a fixed on-time controlled by a microcontroller (arduino), which will sense the voltage through a divider and shut off when the desired voltage is reached. you'll probably have to design your own but it's only ~10 components. don't kill yourself.

>>1569349
So with a simple voltage divider I can sense when the capacitor is at 300v?
After the capacitor is charged, how can I discharge it for spot welding?
I have some flash capacitors that I desire to use for this purpose.

>>1569354
BTW, I don't want to make a spot welder but a exploding wire detonator for fire works

>>1569354
>how can I discharge it
i'd recommend an SCR for controlled discharge, but you can just short the terminals.

>>1569304
nah but there are a literal fuckload of SOT-23 switcher chips for 10 burgercents each that all function well
not a stupid idea at all, *but* you're gonna have to work out a means of giving both the pre-switcher and the post-linear the feedback they're looking for
you can inject a small current into the feedback node of a 4-terminal regulator (as switchers usually are) to proportionally reduce its output voltage by replacing some of the current from the upper feedback resistor. if you use a 4-terminal regulator for the post-linear, you can use basic op-amp arithmetic to scale and offset the two control voltages, thence the control currents, thence each stage's output voltage

>>1569339
>>1569375
Thanks

>>1569375
>nah but there are a literal fuckload of SOT-23 switcher chips for 10 burgercents each that all function well
Can you give me some examples?

>>1569390
okay, more like 20 cents, from LCSC
XL2001 (oops, SOIC-8) for 40V Vin, 2A
MP2359 is good for up to 24Vin, 1.2A "peak current", has a high frequency so you can use a smaller inductor
if you want a lot of current, the XL4015 is a good choice, rated for 5A, but more like 50 cents, and in a D2PAK-5 instead of SOT-23
most of the rest are for point-of-load stuff so input voltage maxes out at 6V or so, or are rated for less than 1A

>>1567135
Not sure if anyone's mentioned it: just solder a USB cable on, and give it a little knot and/or hot glue to keep it secure inside the keyboard.

>>1567642
That soldering iron gets really hot, smokes the shit out of solder. Maybe good for high-temp stuff, but I have some lower-temp (not lead) stuff and the flux burns right off. It's an issue I have with many uncontrolled soldering irons.

>>1568838
Use two switches in parallel, if either is high, you have filament. Your switch is just bouncing. Add debouncing in software or hardware.

>>1568977
At our uni VHDL and the digital logic classes were separate. There was an urge to "reduce the credit load" or something because EE's were something like 140-150 credits for a 120 credit degree. Instead of better classes they just let kids cheat :(

>>1569304
mc34063s are really hard get efficient, with proper tuning you're talking 80-90% for buck/boost modes, but a normalish design could be 60-65%. A linear regulator would chop off another 1-1.5V, even low dropouts are like .7V. It's a decent idea to spread heat, but there should be a way to have the output drive a mosfet instead of passing current. I heard it's hard to get the darlingstons to do it properly, but maybe a certain resistor value would have good on/off times. Mosfet gate charge could be a problem here, since high gate charge might interfere with pushing the frequency, and would need lower valued resistors.

>>1569322
Because it's free software, obviously.

>>1569413
>Because it's free software, obviously.
But KiCAD is actually good. On that note, I'm using KiCAD to lay out a circuit for perfboard because I need this circuit done ASAP.

>>1569419
if you're already bothering with doing your layout in an EDA, just etch it yourself.

>>1569446
I'd do that if I wasn't flatting with only half of my personal possessions in a tiny room.

>>1569448
Oh and I also don't think this flat has any sort of mailbox, which is a bit of a pain, it's a slotless door opening onto a busy street.

>>1569413
> mc34063s are really hard get efficient,
No harder than any other buck converter. The main issue is that a npn/nmos high-side switch needs a base/gate voltage *above* the voltage being switched.

If you tie pins 1, 6 and 8 to the same supply voltage, the output transistor is never going to saturate. Like, you'll get 3 p-n drops across it and the efficiency sucks, particularly at low output voltages. If you want reasonable efficiency, you need >2V of headroom between pins 6/8 (Vcc and drive Vc) and pin 1 (switch Vc). This can be obtained either by stacking another (low-power) floating PSU on top of the primary supply rail, or (provided that you can stay within the 40V limit) by driving a voltage doubler from the output switch.

But then you need to take care not to over-saturate the switch, as that increases the turn-off time. If you want good efficiency at high frequencies, you need to replace the output stage with your own pass transistor and a decent driver (with active pull-down). That discards one of the main advantages of the 3x063 (its built-in 1.5A switch), but keeps the other one (availability).

What topology and switching controller can anyone recommend me for a 12V to 200V 20W step up converter? I've been struggling with this design for awhile and have never managed to get various boost/flyback/half-bridge/etc. converters to do more than 5W or so. I have on hand the MC34063, the TL494, and the MAX1771 for switched mode controllers. Anything else I'd have to order. Recommendations for transformer cores too if recommending a transformer based topology.

>>1569459
12V->120V inverter, 120V/208V transformer

>>1569053
Because the slaves only send information when they received a message from the master.
The master sends a set of instructions and every time the slave receives the instructions it replies with a set of information from the sensors attached to it.

>>1569471
Why an inverter as opposed to an SMPS out of curiosity?

>>1569476
I assumed you needed AC. If it's DC you could just rectify the 208 output and maybe drop in a floating regulator adjusted to 8V, idk. 12->200V at 20W is quite a bit, one of those "are you sure you're solving the right problem" kind of things

>>1566898
This is a very good birb, big thank

>>1569480
I need DC. I'm driving a large array of IN-9 nixie bar tubes and I'm looking at a minimum current budget of 150mA or so to max them all out at full scale. 200mA just to be safe. In theory they strike at 150V but I wanna go 200V just to have a bit of extra room to account for variation and plus I might use other tubes in the project the need higher voltages. The tubes cannot be multiplexed because not only the brightness is a function of current but so is the length of the glow in the tube. If you multiplex them you reduce the average current and so reduce the length of the bar as well. No good. They all need direct drive.

>>1569456
Why don't they use PMOS for these? It's not like you'll need a cutting-edge 2019 model MOSFET anyways so finding a PMOS with the same specs shouldn't be too difficult, right?

>>1569496
The MC34063 and TL494 both have internal transistors so would require a bulkier flyback transformer or some other arrangement, but the MAX1771 uses an external transistor so you could easily pick one rated for >200V. Plus it looks super easy to adjust the voltage on. Non-bootstrapped is probably the way to go.

You just need to spend some time figuring out which transistor, and inductance to go for. Inductance will be a function of frequency and output power, though it looks like the frequency of this converter can change to achieve maximum efficiency. Aside from that minimising ESR of the toroid will be important. The most important part of the transistor is the gate charge/capacitance. If I saw correctly, there are notes in the datasheet explaining how to choose the right inductance, sense resistor, transistor, diode, and capacitors, so just read through it and you should be on the right track. I've never worked with those converter ICs before, but if I'd known it was this easy I would have done so earlier.

>>1569496
Also that's a lot of IN-9s, your project better have a nice looking wood, aged metal, or at the very least transparent acrylic frame.

>>1569517
I just wanna briefly note that in old designs using the MC34063 and TL494 I've always only used the internal switch to drive an external FET. Typically an NTE2900 or IRF730 but I've also tried clamping the flyback spike with a snubber or using a double ended configuration and using an IRFZ44 to get the benefit of the low Rds(on)

>buy a few usb rs485 adapters
>they just an A and B terminal, no GND or anything else, since it uses differential signaling
>connect both with a few meters of twisted wire extracted from an ethernet cable
>put one in one PC, put one in another
>open up a serial terminal in both
>send message from one to another and it just works

How does differential signaling work without a common ground?

>pin 2 and 4 need to be connected to ground instead of only 2 and 3 or 1 and 4
What the fuck where they smoking this makes wiring it up even more tedious

>>1569575
>rotate 90 degrees
>problem solved

>>1569559
Differentially, voltage between the two wires, no signal reference to ground. USB and LAN need no common ground either. It can be used to shield the cable, but that is not related to the signal itself. Maybe you've heard the term 'ground loop'? That's why studio microphones are differential.

>>1569506
the 3x063 is not a cutting-edge chip, sadly

>>1569459
I think you're gonna want custom or semi-custom magnetics for that. on alibay some vendors stock a select few switching power supply transformers. the wattage isn't listed but I think you want one like this
https://www.aliexpress.com/item/1PCS-EE25-200-6-benis-transformer-high-frequency-switch-benis-supply-transformer-XDDDD/32854219030.html but if you really want to be sure, wind your own magnetics
remember to use a low Rds(on) FET and take good care on your layout especially on keeping traces short, wide and direct

>>1569575
two possibilities:
1. putting ground in the middle to shield potentially noisy supply from potentially sensitive output
2. brake cleaner

>>1569596
>USB need no common ground either
>what is the single-ended zero bus state
nice try though

>>1569605
For that price you can just buy a complete nixie tube driver module on Ali.

>>1569606

>>1569607
I pick good and cheap.

>>1569559
i'm assuming you are american because you don't seem to understand that ground is simply a reference point and that all voltage is just a potential difference. there is nothing special about """ground"""" as a 0v reference point in a circuit.
the problem is when your 0v reference is connected to earth ground and suddenly you seem to think it takes on magical properties or something.
signal to ground is just a voltage difference, same as between voltage between A and B is a voltage difference.
when you were a kid and people told you a circuit needs a ground to operate you were being lied to, a circuit needs a closed loop for current to flow, nothing to do with ground. you just happen to decide to call a point ground because it makes life easier. really you should call it 0v and there would be less confusion.

>>1569626
But I'm talking about having a common reference point between the two. Uart needs one, RS485 doesn't. Where is the trick? Unless... the wires are the common reference.

>>1569630

>>1569660

>>1569663

>>1569665

I’m trying to plan a microcontroller project, probably using an AVR, that runs on batteries (something like 2 AAs is preferable) and basically acts as a digital kitchen timer but with random intervals based on a user specified range. So you could set it to anywhere between 20 and 30 minutes and it would randomly choose an amount of time in that range, vibrate a bit when that time has passed, and repeat.

My main question is how should you keep time in something like this, without eating though the battery? For the most part, the device won’t be doing anything except counting time. Would that necessarily consume too much power to keep it powered for very long? And how should it count time in general? All my MCU projects at uni so far haven’t used time greater than the range of milliseconds which are taken care of by the devices internal timers, and none of them have used battery power, so I’m pretty lost here.

>>1569676
Sleep mode with a slow clock 32kHz clock. Set a counter and a wake up interval. That's how some microcontrollers do it.
I think AVRs don't have the capabilities for slow clocks, so a cheap microcontroller such as an MSP430 or any STM based on ARM Cortex-M0 would work wonders.

>>1569676
Related document: https://www.ti.com/lit/an/slaa792/slaa792.pdf

>>1569686
I think they do, AVRs are fully static and support low frequency clocks

>>1569692
Checked again and Atmegas don't have built-in RTCs, which are needed if you want to don't want your processor to be sluggishly slow while they are awake.
These use two separate clocks: The standard 8/12/16MHz crystal for running code and the 32,768Hz RTC one, keep track of the time. MSP430 and ARMs have built-in RTCs, which allow you to enable RTC interrupts in terms of minutes to wake them up with a minimal amount of peripherics.

>>1569660
>banana
pls write electronics textbooks

>>1569676
wake up, decrement a counter, if not expired yet go back to sleep
the big question is, what are you using for the display? there are AVRs and PICs that will directly drive LCD glass for you, if you desire

>>1569686
>AVRs don't have the capabilities for slow clocks
depends on the AVR, of course
the ATmega8A only draws about 200µA in idle mode at 3.3V Vcc when the internal RC is divided down to 1MHz
the ATtiny87 can divide the internal 8MHz calibrated RC down to 31.25kHz but the datasheet doesn't state anything about power consumption when running thus
none of these are especially accurate, but since OP just wants a random timeout that may not matter very much. if accuracy does matter, all AVRs are happy to accept a watch crystal as their main time base, and consume very little current (handful of µA) while doing so

>>1569694
>RTC
what the literal fuck are you talking about
if you can't count time in an interrupt you need to go back to the lolduino thread

>>1569696
>the big question is, what are you using for the display?

Just a 2x16 character LCD.

>>1569686
Not him but do MCUs with builtin RTCs just use an internal RC oscillator?
How is that accurate enough?

>>1569660
So the wires are just a loop?

>>1569696
Also the accuracy isn’t too important. The same level of accuracy as a cheap digital kitchen timer would suffice.

My main concern about counting time with interrupts is whether or not it will consume too much power to run on some common batteries and not drain them quickly.

>>1569702
I can't explain in a simpler way than those drawings, sorry. You should start with the basics

>>1569701
internal RCs are rarely specified to better than 2% over process/voltage/temperature after calibration
if you care about real time, use watch crystals

>>1569703
>power
well, the display alone is eating (typically) 1.2mA plus whatever for the bias generator plus whatever for the signalling between the display and the MCU, all this ignoring the backlight. it's almost certain that the MCU will be consuming somewhat less than that
it's one of those things where you have to look in the datasheet of the particular MCU you choose and do a bit fo reckoning. in Atmel's datasheets, graphs like Pic related can be found in the "Typical Characteristics" section

>>1569710
I should’ve mentioned, the display would be off when it’s counting down, I only need it on to set the time range and to prompt for input when the timer hits 0.

>>1569694
You don't need an RTC to wake up from sleep, a timer is enough.

>>1569712
got it
http://data.energizer.com/pdfs/e91.pdf
a typical AA alkaline supplies about 3Ah, which in a 3xAA configuration is enough to run the display alone for two months or more
you'll probably want a low-quiescent-current LDO to supply a consistent voltage to the contrast divider, with a shutdown/enable option to supply the contrast divider only when the display is on
if you use a watch crystal as your system time base the MCU core and peripherals should only require a few hundred µA
bottom line: how much battery life is "enough"?

>>1569723
Shit, I hadn’t even thought about that. I’d probably use it 5 hours a day, most days of the week. Just not something that’s laughably short!

>>1569667
gold

>>1569506
PNP/PMOS always has worse performance than a similar NPN/NMOS part. The difference is enough that an n-type switch with a high-side driver (voltage boost and level shifter) is far more common than a p-type switch once the power is measured in W rather than mW.

>>1569526
Doesn't that mean you'd need a totem pole or something to drive the external FET fast enough? I imagine the MAX1771 has the FET driver circuitry already inside it.

>>1569459
Boost converters are a choose-your-poison problem:

1. Discontinuous-conduction mode is simple but requires the peak current to be at least twice the mean current, which increases I^2R losses.

2. Continuous-conduction mode is a pain to control. You can't just use bang-bang control (i.e. turn off the pulses when the output voltage gets too high). If the output current drops, you need to reduce the inductor current, which means dumping some or all of the inductor's stored energy ((1/2)*L*I^2) into the output capacitor. So either the output capacitor needs to be large ((1/2)*C*V^2 >> (1/2)*L*I^2), which also makes feedback less sensitive, or you need additional over-voltage protection on the output. Typically, you control the desired inductor current by using PID control based upon the output voltage, and control the actual inductor current either via hysteresis or another PID controller varying the duty cycle based upon the inductor current.

For a fixed output voltage, large step-up ratio and non-trivial power, your best bet is probably a transformer-based converter (push-pull, half-bridge or full-bridge).

I have a retarded question.

I have this audio sampler circuit that I am following from http://www.technoblogy.com/show?QBB

It works fine for what I need, but I find it way too quiet and I want to make the output louder.

However, what is tripping me up (because i'm retarded) is that both PB1 and PB4 are feeding to the output speaker...so does that mean I need to have two sets of amplifier for both PB1 and PB4 like a H-bridge or am I suppose to short PB1 and PB4 and then send it to an AB amplifier?

>>1570005
PB1 and PB4 are in opposing phase. that doubles the voltage swing to the speaker, in theory. shorting them would be Bad
pick either PB1 OR PB4, and run that to an amp, and hope that they aren't being too clever about trying to use all three possible voltage levels across the speaker. if they did you might just have to modify the code to be a little less clever

>>1570023
>to use all three possible voltage levels across the speaker

By that do you mean the 3.7 VCC, and the PWM from both PB1 and PB4?

>>1570041
I mean relative to the speaker: +3.7V, -3.7V and 0
if they do that you'll need to do one of the following:
>change the code to use only one of the PWM outputs, which you amplify as normal
>add a differential amp and some filtering
>do the H-bridge thing

>>1570005
Some class-d amplifier boards might have differential outputs which will work just fine with this, otherwise you could put a differential amplifier op-amp circuit after those pins and feed that to a class-D/other power amplifier topology. The op-amp circuit might be powerful enough by itself anyhow.

Pic related, though the ground is the middle point for the output voltage, so you'll either need a split-rail or some funky circuitry to get around that.

>try to solder a bunch of wires wire onto a SOIC-8's pads
>delam 1 of the 5 I need
>scratch off some solder mask and resolder elsewhere
>realise one joint is on the wrong pad
>desolder it, resolder it elsewhere
>go to remove extra unneeded resistor
>have to juggle two beads of solder on my chisel/K tip
>one for each pad of the resistor
>it comes off but flies right into one of the wires
>desolder it
>another completely different pad delams
>repeat step 3
Holy hell when i'm done testing I'll cover this shit in epoxy so I never have to touch it again. It's an aluminium-backed PCB too so I have to constantly keep my iron in overdrive, because even if the temperature in the element's thermocouple is 300C the temperature at the tip isn't enough to melt solder. I can only assume this is because there's enough thermal resistance between the element and the tip that the amount of heat flow caused that much temperature drop. I'd like to see which company's tip standards have the least amount of this thermal resistance to the tip. I guess other T-12 shapes (that stubby conical one maybe) could help with this.

Pic related. If any other pads start to delam I'll relocate all the joints to places in the middle of a trace, so I can have copper on two sides of the junction. Also I'm using white for Vcc because red refers to the LED color. No, I don't want to reflow that neutral mains connection, I think it's strong enough.

>>1570093
Just heat the board up on a hotplate when you solder it.

>>1570097
I don't have a temperature controlled hotplate lying about. Nor a clothes iron or anything other than a stove (not temperature controlled so no thanks). Nor any way of stopping it from sliding about on the hot piece of metal so I don't burn myself.

In a mosfet or a transistor, does it mater if you flip the polarity of the DC you put through it, or does it work only one way?

>>1570045
>>1570063

Thanks. That should be enough info for a retard like me to go on.

>>1570139
>does it work only one way?
yes

>>1569169
Not exactly sure what your end-goal is or how much dsp power you need, but I've been using microchips dsPIC33FJ128GP804 for my audio projects. Has a built in 16 bit DAC.

>>1570093
>300°C
have you heard about lead-free solder

how do you get into electronics

>>1570389
y tho

>>1570398
i need to make an automated blood sample transport system for hospital prototype

>>1570360
The thing already has RoHS on it, if I introduced 60/40 into the mix I'd get an oddball alloy with suboptimal properties.

>>1570478
then turn your iron up. proper lead-free iron temps are ~350°C, higher still if you're fighting with an aluminium heat sinking plane

>>1570139
BJTs only conduct one way. MOSFETs are bidirectional, but ... most power MOSFETs have a parasitic diode between source and drain so they'll always conduct in the reverse direction regardless of gate voltage.

>>1570499
You think? Even though the melting point is 227°C? I thought the whole point of having a temperature controlled soldering iron is that you don't need to overshoot the temperature by much. Too far above 300 and I seem to cook off the flux too quickly and the tip turns dark. I don't have any of that "tip tinner" solder+flux paste, but the copper wool and a bit of extra rosin seems to do the job. I'm fine with turning on the boost function for the aluminium backed PCB, but on a normal PCB shouldn't 300 be more than enough? I guess I'll do some testing on that.

>>1569630
It's not that simple. Normally rs485 transceivers do reference to a common ground, but it is possible to buy isolated transceivers. (And isolation may even be required.) If you are signalling between two systems with a large ground potential between them, without isolation, you may exceed the receivers common mode voltage range and your shit won't work.

https://www.analog.com/media/en/reference-design-documentation/design-notes/dn228f.pdf
read page 8 from https://www.analog.com/media/en/technical-documentation/application-notes/AN-960.pdf?doc=an-1177.pdf
read pages 6-8 from http://www.ti.com/lit/an/slla272c/slla272c.pdf

It is actually possible to signal between systems without any ground reference whatsoever, ethernet does so with either magnetic or capacitive coupling.

>>1570535
Hotter tip gets the part up to temperature quicker, I usually have my iron on 360 for SAC305.

>>1570415
then you need to hire someone who knows what they're doing and/or pay better attention in class

>>1570502
>BJTs only conduct one way
not strictly true, but their current gain in reverse (alpha vs. beta) is minuscule

>>1570535
>copper wool
good
>extra rosin
should only need to apply some solder to keep the tip wet
>on a normal PCB shouldn't 300 be more than enough
nope, if you watch IPC soldering instructor John Gammell's videos he normally runs his iron at 370°C and for fairly fine work at that
remember, heat takes *time* to transfer. you're not trying to get the board into thermal equilibrium with the heater. you're just trying to get the flux activated, the solder melted, and the surfaces hot enough to form intermetallic bonds, preferably without melting the fiberglass and delaminating the copper. more heat, applied quicker, keeps the heat localized. for similar reasons, spot-welding is preferable to soldering onto battery electrodes

>>1566896
quick question for an ohmlet
Can I create mono audio by taking stereo wires (L + R) and connecting them in parallel to create a mono signal?

>>1570587
>you're not trying to get the board into thermal equilibrium with the heater
I'll keep that in mind thanks. I see how a highly thermally conductive medium (PCB) means that the heat will spread making you want to decrease soldering time, but what about very thermally insulating media? Like one of those extra thin PCBs with narrow traces on it? I suppose it wouldn't particularly matter which way you do it.

It would be nice if I could put thermal feedback on the PCB itself (an IR thermometer pointing at it?) and use that to control the iron. I wonder if the cheapy arduino PIR sensors can be used to measure temperature, and if I could use one with some sort of lens or aperture to measure the temperature of a small point?

>>1570596
Before an amplifier, probably but I'd add some resistors in a simple divider so the output device channels aren't constantly fighting each other. Also look up a summing amplifier, which you'd want instead of simple resistors if you're driving a medium-impedance load like earbuds. After an amplifier, probably not at all, too much power shit going on there. Also note that some songs may have out of phase audio that cancels when you put the channels together.

>>1570596
the layer of fiberglass is insulating, but not that insulating. else it wouldn't be much use for a heat sink for an LED.
>mediocre oversimplification follows
to a first approximation, thermal circuits can be thought of as a series of resistors, and temperatures can be considered in relation to ambient much as voltages can be considered in relation to some ambient voltage we call "ground" or 0V. when you decrease the bottom resistor of a voltage divider, two things happen: the current from top to bottom increases, and the voltage in the middle drops. thus, you have to turn up the voltage (analogous to temperature) to get the desired output in the middle
>end mediocre oversimplification
if the temperature of the joint were that critical, that sort of IR thermo overcomplication might be worth pursuing. but really, just get in, get melted and get out. if you have to remove some little passive, don't count on recovering it, just get it off of there. preheating and/or hot air are the best ways to work on these sorts of boards. if you're gonna a temperature controller, may as well pick up a hot plate at a garage sale and control it

>>1570655
Wrong quote but whatever. I've already used capacitors, resistors, and current sources to simulate thermal systems in Spice, it's pretty handy if you can get your head around the units. Certainly better than doing the half a dozen differential equations by hand. Modelling a thin aluminium piece like the PCB would be something like a bunch of resistors in series with capacitors and resistors to ground at each resistor-resistor junction, right?

I completely understand why turning the heat up heats up your joint more quickly, as it's basically Newton's law of cooling. I also agree that in most cases running an iron with IR feedback is likely not worth attempting, but if I'm turning my iron up to 400 and dealing with sensitive parts, something to stop those parts from reaching their kill temperature might be valuable for repair jobs. Since soldering all the pins of an IC might lead the temperature to buffer nearby. A thermal camera in general to keep track of heat flow across the PCB might also be handy, so you can take the iron away and try again a little later when the nearby components aren't about to reflow themselves. A thermal overlay on a soldering microscope could be a product people would be willing to get.

>>1570666
yep basically a two-dimensional mesh of R with C at each node, if you really want to get down to that level of detail
>A thermal camera in general to keep track of heat flow across the PCB might also be handy
oh absolutely this. might want to get started https://www.aliexpress.com/item/MLX90640-Infrared-Thermal-Benis-Sensor-Module-32-24P-Benis-Measurement-Dot-Matrixs-Benis-Imaging-Sensor-Camera/32974460552.html

>>1570701
I was thinking one of those ones with the screen built in, could be cheaper but they do have all that housing. They also blur the few pixels they have together, for better or worse.

Hey guys so I want to make a shocking device. I am new with electronics, and figured it *may* be a simple project to get started with. So I have scattered info on what I "know"

>volts don't kill you its the amperage
I did some googling and people are saying 150mA is lethal, now I read up on electrical output for tasers, and it seems that most of them only have 1-2mA output while police stun guns have 3-4mA, how can I achieve keeping mA low to not be lethal while keeping voltage high? I am looking for it to be more of a jolt, probably more powerful than a 9v to your tongue.

Maybe using capacitors to shock the person? Like those cheap disposable cameras they use as tasers and the main component is the capacitor.

>>1570755
To shock someone you basically need high voltage. Your body acts like a fairly large resistor, so you're not going to get more than a few µA with a 5V or 12V source thanks to Ohm's law. You could use something like an automotive ignition coil, which is an autotransformer used to step the voltage up to ~10kV for a very short spike, sort of acting like a flyback transformer. But you don't need a voltage anywhere near that high.

I'd build a simple boost converter that goes up to no more than 600V, should give enough of a shock with a good enough capacitor, plus you can use fairly common MOSFETs and diodes for this kind of voltage. It's a pretty nice beginner project too. Look up how 555 oscillators and boost converters work.
>t. built my own 500V boost converter and shocked myself a few times

>>1570760
Ah, so I will end up making boost converter. When making a boost converter should I be concerned with the mA rating or will it be small enough not have to worry about it?

Basically I want to do something similar to this

https://youtu.be/vXQlkAgEiqs

but smaller, im not going to make a hot potato shocker, just similar concept.

Thanks for the info, I'll take a look at it.

Does anyone know a good tutorial for matrix keyboards?

>>1570773
Don't copy that video, he's using an arduino and one of those taser modules. The taser modules aren't a bad way to go if you actually need that high of a voltage, but an arduino is completely unnecessary if you don't need to time it. In general it's the least amount of effort you could possibly put towards such a project.

I'd personally use this project to learn about feedback circuits, inductors and capacitors, and simple power supply design. A simple circuit to do this would use 1 or 2 ICs, 1 transistor, a diode or two, two capacitors and an inductor, and half a dozen resistors.

>>1570773
A boost converter (without any output capacitor) is a good choice because it supplies a fixed current regardless of the load resistance. Applied to dry skin it can produce a hundred volts or more, on the tongue it will only be a few volts; the current is the same in either case.

The inductor current is given by V=L*dI/dt => I = integral (V/L) dt => Ipeak = t*V/L for a square pulse. When you open the switch, the inductor current has to go through the load. The current will be fixed, the voltage will be V=I*R. You need to place a resistor (typically a few hundred kΩ) across the output so that you don't exceed the transistor's breakdown voltage if there's no load.

Also, add a capacitor in series with the output so that the output current has no DC component. DC current causes electrolysis which may have adverse effects (apart from anything else, it tends to cause a burning sensation).

Safety wise, ensure that the terminals are close enough so that you can't reasonably touch different terminals with different hands, as that causes current to go through the heart, which can cause fibrillation even at quite low currents.

>>1570804
What, without feedback? Isn't that a good way of blowing your FET with overvoltage?

>>1570760
I'v given myself shocks testing spark plugs before while also holding onto the frame
it's mildly unpleasant

I would like to slow down a 60mm computer fan that runs constantly at 12V and is loud as fuck.

I have a bunch of those small blue chink resistors, what value should i use to slow it down to about half?

Do i just use a resistor to drop 6V so the fan will be only getting 6V and thus will run at half speed? Is it that simple?

>>1570823
More or less, but I'm afraid that resistor will get far too hot. Better to put a small 12V dimmer module in there, or just run it off your 5V rail instead. With a transistor or something so it is still controlled by your "fan" cable.

>>1570835
it won't be connected to a computer, the power source will only have 12V
maybe i could use a buck convertor, instead and connect both fans to it

>>1570840
I think a dimmer will be slightly more efficient than a buck converter, but there's barely any difference between the two in this case. Just go with whatever's more convenient, though the dimmer/speed controller is more likely to have a nice knob for adjusting right out of the box. Putting two fans in series would also probably work.

>>1570842
I remembed that i have a noctua fan in my pc and noctua ships their fan with these cables which you can use to change speed of your 12v fan by like 20% and 50% or something like that
it seems that those cables contain only those chinky resistors

>>1570823
I would first test whether the fans run when wired in series.
Tested my 50mm SUNON 12V 1.1W KDE1205PFV2:
12V 85mA, 9V 65mA, 6V 46mA, 5V 40mA, 4.5V 37mA, 3V 0.3mA (no start-up)
Two of them run in series on 12V.

>>1570885
So if i connect two identical fans in series to 12V they will each run at 50% of their normal RPM?

>>1570898
without any resistors and such

>>1570898
no but close enough

I need to step up 5V to around 40V for a VFD.
What would be the simplest way to increase the voltage?

I'm thinking boost converter since I'm already using a microcontroller (that could monitor the voltage and provide the appropriate pulse) for the project anyway.

>>1570862
Just a resistor works alright, but would would give a less predictable % reduction for different fans.

>>1570922
Push pull / forward / flyback with two windings on output to create both an AC filament voltage and the anode voltage https://www.tayloredge.com/reference/Circuits/SimpleVFDPower/index.html