/ohm/ - electronics general, B is found within you edition

[Anonymous No.1669514]

released-to-mfg thread: >>1663702

>RULES
0. Electrics ≠ electronics. Appliances/mains stuff to /qtddtot/ or /sqt/. PC assembly to >>>/g/.
1. Do your own homework. Search web first. Re-read all documentation/datasheets related to your components/circuits. THEN ask.
2. Pics > 1000 words. Post relevant schematic/picture/sketch/9001.5 hours in MS Paint with all part numbers/values/etc. when asking for help. Focus/lighting counts.
2.5. State your skill level if asking an open-ended question.
3. Read posts fully. Solve more problems than you create.
4. /ohm/ is an anonymous, non-smoking general.

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

>Project ideas:
http://adafruit.com
http://instructables.com/tag/type-id/category-technology/
http://makezine.com/category/electronics/

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

>Design/verification tools:
LTSpice
falstad.com/circuit/circuitjs.html
NI Multisim
CircuitLab
iCircuit for Macs
KiCAD (pcb layout software, v5+ recommended)

>Components/equipment:
Mouser, Digi-Key, Arrow, Newark, LCSC (global)
RS Components (Europe)
eBay/AliExpress sellers, especially good for component assortments/sample kits (caveat emptor)
Local independent electronics distributors
ladyada.net/library/procure/hobbyist.html

>Related YouTube channels:
mjlorton
jkgamm041
eevblog
EcProjects
greatscottlab
Photonvids
sdgelectronics
BigClive

>Li+/LiPo batteries
Read this exemplary resource first: https://www.robotshop.com/media/files/pdf/hyperion-g5-50c-3s-1100mah-lipo-battery-User-Guide.pdf
>I have junk, what do?
Take it to the recycler.

[Anonymous No.1669519]

this thread's digits brought to you in part by the AD9514 clock distribution IC

[Anonymous No.1669539]

How feasible is it to build your own arm based motherboard/SOC? Only talking complexity, not time or money. Couldn't one just copy current developer boards for most of it and maybe more importantly drivers?

[Anonymous No.1669542]

>>1669539
that depends entirely on your expectations

[Anonymous No.1669543]

>>1669539
see also /ohm/ RULE 2.5

[Anonymous No.1669546]

>>1669542
Dunno, probably something like a RP4. Maybe add another Lan port, remove the GPIO to make it worthwhile.

>>1669543
Good call. Currently playing with my first FPGAs, previously only working analogue, ergo lights around the house.

[Anonymous No.1669547]

>>1669519
why are there so many of these fucking chips? half the new product email ads i get from digikey are clock ics.

[Anonymous No.1669555]

>>1669539
>How feasible is it to build your own arm based motherboard/SOC?
No problem. BUT...
The faster you go, the more care you will have to take. "low speed" everything -> no problem, medium speed stuff (e.g. faster USB) -> take care, fast stuff (e.g. DDRx) -> calculate and control impedance, use more than two layers, ...
I stop here, as RULE 2.5.

[Anonymous No.1669571]

>>1669343
>1-wire
Why not just use coax or some other noise-proof 2-wire standard, with data being received through a high-pass filter and power being received with a low-pass filter? The data itself would just be transmitted via ampltiude shift keying or something, perhaps full-duplex with SIPO at one frequency and PISO at another. Heck, mix audio down it too.

[Anonymous No.1669584]

I tried to derive some basic AC reactive formulas. Does that look right (see pic)?
I am not sure about the expression for I.
I=CdV/dt now plug in V in the Euler form and get I = C d/dt (Vo e^jwt). After taking the derivative,
I=VojwC x e^jwt. How do I show that the I leads V from here? I would want to show that
I=VowC x e^(jwt+90) . So you can get the phasor form from that. But what the heck is the physical meaning of jw in VojwC x e^jwt? That notation throws me off. The result of taking the derivative looks like a mix of a linear and phasor(euler) form.
Take a look at the pic where I typed them nicely. How do I get 6 from 4?

[Anonymous No.1669587]

>>1669584
>tfw no TeX on /diy/

[Anonymous No.1669590]

>>1669571
you know 1 wire isn't actually 1 wire
its 2.

[Anonymous No.1669595]

>>1669547
Fast micros talking to various bus devices. A good clock divider ensures solid clocking and reliable signals.

[Anonymous No.1669597]

>>1669546
ok. basic digital logic is good. but you'll need to learn how to deal with fast signals, possibly above 1GHz depending on what else you connect up to the SoC. also expect to do a little kernel hacking. not as much as you used to have to do thanks to the device tree system which informs the Linux kernel of resources, how they're connected to other devices and resources, and where their I/O base is
proposed intermediate goal: if you find a system-on-module that fits your needs, you can mate the module to a "motherboard" of your own design, which you can have made at a somewhat lower technology level than the module, but still possibly have things like hardwired USB peripherals, maybe a few PCIe lanes to things that need real speed. sort of working your way from the periphery to the core
in either case, tens of picoseconds can still make or break a design, so you'll need to learn measurement technique. Pic related, just some of the lengths they go to in https://www.edn.com/design/analog/4326623/A-streamlined-method-of-PCI-Express-interconnect-compliance-testing

>>1669547
5G is like their date to the senior prom

>>1669571
you can master the W1 bus with a UART, a weak resistive pullup, and (optionally) a strong pullup made from a single MOSFET. you can string nodes together with doorbell wire for tens of meters and it's slow enough that shielding is almost irrelevant. non-destructive arbitration and fast bus enumeration. you can debug it with a slow logic analyzer. boring old TTL levels! what deficiency in """1"""-Wire™ are you purporting to solve, exactly?

>>1669584
I'm terrible at calculus but dividing by j is equivalent to a rotation by -90° isn't it?

>>1669590
yet nobody calls SPI a five-wire bus

[Anonymous No.1669601]

>>1669597
>dividing by j is equivalent to a rotation by -90° isn't it?
yes and multiplying by j is +90. that's understood in expressions like Zc = 1/jwC which is the same as -j/wC. That form is a complex number in a rectangular form, i.e. 0-j/wC. But I am confused by the notation like jVwC * e^jwt. It is not a rectangular form and it is not an euler notation either but rather something mixed or I miss something there. I understand all of that conceptually but I wanted to understand the rigorous manipulations with proper units etc.

[Anonymous No.1669604]

>>1669601
i think i got it
jVwC * e^jwt
i can think of the two parts as two separate parts and i can, for example, convert e^jwt to rectangular notation, ampl=1 and phase=0 so it is just 1(cos(0) + j sin (0) = 1 so it vanishes and won't bother me anymore and let me sleep at night. so i can concentrate on just jVwC which is essentially 0 + jVwC in the strict rectangular form. Now I can convert it to euler and get Vwc exp ^ (90) cause teta=atan(VwC/0) or atan(inf) = 90.
uggrrrh. correct? it is probably a much more painful way and there is some shortcut but at least there are no assumptions

[Anonymous No.1669612]

>>1669601
>Zc = 1/jwC
in datasheets and other practical documentation, in place of jw I have commonly seen 2*pi*f which is the frequency in radians/sec. but that's about all the Laplace I remember, so I hope that helps with your original question

[Anonymous No.1669625]

>>1669597
>https://www.edn.com/design/analog/4326623/A-streamlined-method-of-PCI-Express-interconnect-compliance-testing
Quite an interesting read, thanks!

[Anonymous No.1669627]

>>1669612
thanks yeah, w is just for brevity. honestly it is just my autism at this point, i can't let go even though i've done pretty much all i could from all angles. but i still have a feeling it is not 100% rigorous. i am still uncertain about this math concept: if i multiply two complex numbers, and one is in the euler form and the other is in the rectanuglar form, whether or not i can convert either to the the other form and then just do the calculations (obviously for multiplication the euler form is preferable).

[Anonymous No.1669698]

>>1669604
> jVwC * e^jwt
e^jwt is a helix centred on the t axis, cos(wt)+j.sin(wt). |e^jwt|=1, only the angle changes.

Multiplication of complex numbers in polar form multiplies the magnitudes and adds the angles. ae^jx * be^jy = a*b*e^jx*e^jy = (a*b)*e^j(x+y). IOW, multiplying by a complex number is equivalent to scale and rotation. Rotating a helix is equivalent to translating it along its axis, i.e. a phase shift. e^jp*e^jwt = e^j(wt+p).

A real sinusoid a*cos(x) can be expressed as the sum of two exponentials: cos(x)=(e^jx+e^-jx)/2. If you put W=(1/2)e^jwt, then its conjugate W*=(1/2)e^-jwt so cos(wt)=W+W* and more generally a*cos(wt+p)=(1/2)ae^j(wt+p)+(1/2)ae^-j(wt+p) = (1/2)(ae^jp)(e^jwt)+(1/2)(ae^-jp)(e^-jwt) = cW+c*W* = cW+(cW)* where c=ae^jp (a is the amplitude, p is the phase angle).

Now, conjugation distributes over practically everything: (ab)*=a*b*, (a+b)*=a*+b*, dW*/dt=(dW/dt)*, etc. Because we start with something of the form cW+(cW)* and every operation we're likely to apply to it preserves that form (a complex value plus its conjugate), we typically 1) ignore the conjugate part, operating on one half of the expression and relying upon the fact that the other half will always be its conjugate, and 2) factor out W then don't bother to write it, just writing e.g. V=1+j as a shorthand for V=(1+j)W (if V really was just a constant, dV/dt would be zero). For impedance R=V/I, the Ws cancel and R really is just a complex constant.

[Anonymous No.1669725]

Any good books that focus solely on semiconductors?

[Anonymous No.1669730]

Pcb test with acrylic spray and a 2w laser. The quality is amazing, like from a toner transfer. Can't wait to try with multiple passes to competely remove the paint. If this works i can make a pcb wirh just
1 run laser
2 etch in fecl3
Done, perfect pcb every time
And the best part is I can use the laser to also make a stencil from paper to apply solder mask

[Anonymous No.1669733]

>>1669730
nice
laser source?

[Anonymous No.1669735]

>>1669733
450nm blue laser diode from chinks, cost me only $60 and its attached to a 3d printer i already have

[Anonymous No.1669736]

>>1669730
>acrylic spray
Is that made specifically for PCBs?

[Anonymous No.1669738]

>>1669735
wew, I have to steal this idea!
I used toner transfer in the past, but the process is ultra tedium and results varied too much.

[Anonymous No.1669740]

>>1669736
No just regular cheap spray paint

[Anonymous No.1669741]

>>1669738
Exactly why I did it, toner transfer can suck my 2inch dick

[Anonymous No.1669772]

>>1669514
B found at >>1669518

[Anonymous No.1669774]

I have all these really old Ni-Cd battery cells. They all are basically at 0V. Is there a way to revive at least some of them? I don't have much knowledge about battery cells.

[Anonymous No.1669780]

>>1669774
If 0V and also close to 0Ω then the cell has developed whiskers which short it. There are means to melt them away (current pulse, capacitor discharge) but they will come back. I'd say recycler.

[Anonymous No.1669790]

>>1669730
that looks reversed
otherwise, fuck yeah

[Anonymous No.1669792]

>>1669780
Thanks! I guess they don't worth the effort. I'll buy some LiPo cells instead.

[Anonymous No.1669793]

>>1669730
>1 run laser
>2 etch in fecl3
apply solder mask
run laser
etch in fecl3
run laser to clean solder pads
done

[Anonymous No.1669794]

Is there a way to create a relatively "simple" circuit that measures the duration of a pulse and emits a trigger signal if the input pulse goes over a specified time? I'm thinking about microseconds.

I need to receive the bipolar return-to-zero output of an aircraft avionic as per the ARINC-429 protocol (pic related). After 32 bits of data, there is a 4 bit duration between each transmission where the output is at 0V.
I've already derived a CLK signal from the differential wire pair, but my microcontroller needs to know when a new message is about to start to be able to sync up properly, so I need a circuit that triggers if its input has been held high for long enough, and reset otherwise.

[Anonymous No.1669796]

>>1669793
>solder mask
any epoxy colored toward the red end of the spectrum should do, might even be able to use a clear epoxy. I wouldn't bother with photoimageable if I were just going to photoablate it

[Anonymous No.1669797]

>>1669514
what's that circuit?

[Anonymous No.1669798]

>>1669797
a math problem. see
>>1669518

[Anonymous No.1669799]

>>1669797
Principle of a self-biasing amplifier.

[Anonymous No.1669803]

>>1669796
I am not sure if 2W laser will even be able to vaporize solder mask tho
Plus there is the problem of aligning the board for second lasering because you cannot laser the solder pads before etching, you have to cure the solder mask then etch and then laser the pad holes from the cured uv mask and aligning that for tiny smd pads to be correctly lasered is incredibly hard
just look at how hard it is for people to align double sided pcbs

[Anonymous No.1669805]

>>1669803
are you talking about etch resist or solder mask?

[Anonymous No.1669806]

>>1669805
The other guy suggested using solder mask as etch resist

[Anonymous No.1669843]

Hey guys, welder repair idiot from the last thread. I found the exact capacitors that are in the HF spark-gap-generator circuit that are in my welder pic related. They are $100 a pc and I need 2 connected in parallel to replace what's on the machine. Can any of you geniuses who understand capacitors recommend a cheaper functional equivalent, (either one big one or 10 small ones I don't care) that would have similar or greater voltage and current ratings?
The ratings on this from the mfg. datasheet are:
pn: F250B202JM, capacitance (pF): 2000, Voltage: 5000, Current @3 MHz: 7.5A, @1 MHz: 6.2A, @0.3 MHz: 3.6A, @0.1 MHz: 1.8A
Thanks for your help

[Anonymous No.1669847]

>>1669514
Does anyone know where I can get a 24dBi, 2.4GHz antenna for less than $50? I need to connect to wifi about 2+ miles through houses/trees/etc.

[Anonymous No.1669870]

>>1669806
of course he did, and he left out the process of getting that shit onto and off of the board which is time-consuming and unpleasant, especially compared to dry film application

>>1669843
https://www.tedss.com/2020089224 isn't much cheaper, but it's from the same series and the value should be close enough

>>1669847
have you considered learning what the fuck you're doing and building one instead of being jewish

[Anonymous No.1669876]

>>1669870
>consider learning what the fuck you're doing
I did, no one would help me so I decided I'd buy a cheap antenna, connect to public wifi's, and work on more important things (such as my degree, farming, and other projects.) So, I take it you don't know the answer?

[Anonymous No.1669877]

>>1669847
Altelix AG24G24 Parabolic Grid Antenna $24.99
Remove houses/trees/etc.

[Anonymous No.1669881]

Electronics novice question
This might be completely retarded but I have to ask

In theory could you power an instrument amp by using just the tiny input signal by charging a capacitor as the power source?

Here we go:
The signal from a quarter inch Jack, is split so that part of it goes into a capacitor, the rest of the signal goes through a simple transistor amp circuit as normal (as the emitter), the capacitor can release it's charge and will act as the base in the transistor.

I know if it ever could work it work be very weak/quiet, and I know there might be some fundamental electronics knowledge I'm missing here, but could it work?

[Anonymous No.1669882]

>>1669730
I am a very happy camper. It turned out great! No shorts or coper residue anywhere it isn't supposed to be and it took just two quick steps to make.
I tested for continuity and everything connects correctly.

Now i am going to order a nail polish UV lamp for $10 and a tub of solder mask and then maybe even that little solder cauldron heater so i can dip tin the pads

[Anonymous No.1669892]

>>1669881
diagram >> prose

[Anonymous No.1669896]

>>1669881
thats some nice electronic gibberish

[Anonymous No.1669898]

>>1669892
>>1669896
I never claimed to be a smart man

Something like this?

[Anonymous No.1669900]

>>1669898
lol
you don't 'power' your input. you feed your signal there to be amplified
so your cap would be from + to ground

[Anonymous No.1669902]

>>1669794
>"simple" circuit
>>1669794
>my microcontroller
just use a timer

[Anonymous No.1669903]

>>1669847
>through houses/trees/etc
hmmmmmmmmmmmmmmmmmmmmmmm
https://en.wikipedia.org/wiki/Fresnel_zone

[Anonymous No.1669909]

>>1669902
>"simple"
Exactly why I used quotation marks and said "relatively" as well.
The microcontroller is needed to process the data and send it to the PC over Serial.

I'll take a look at timers, thanks anonbro. I tried all sorts of wordings in my searches and got nothing so I hope this will put me on track.

[Anonymous No.1669920]

>>1669900
But since the signal.also produces an extremely small voltage, couldn't this also be used to charge a capacitor?
Or is that completely spastic thinking?

[Anonymous No.1669924]

>>1669870
>https://www.tedss.com/2020089224
Thanks. Is there something special about the silver-mica-in-a-box series that is, or could I connect in parallel as many of the pancake ones as I need? Also was recommended a doorknob capacitor but can't find current ratings for any of those?

[Anonymous No.1669925]

>>1669909
>I tried all sorts of wordings in my searches and got nothing so I hope this will put me on track.
try something like mcu timer detect pulse
https://www.nxp.com/docs/en/application-note/AN1067.pdf
https://www.st.com/content/ccc/resource/technical/document/application_note/group0/91/01/84/3f/7c/67/41/3f/DM00236305/files/DM00236305.pdf/jcr:content/translations/en.DM00236305.pdf

[Anonymous No.1669930]

>>1669881
lol no. otherwise you wouldn't need an amplifier and could plug directly into a speaker

>>1669794
Boeing the post

>>1669924
you could parallel doorknob caps of the appropriate voltage rating, probably. after all, they did. note, current rating at 1MHz is probably an important figure of merit in this application, and mounting them to the chassis while retaining kilovolts-level insulation (if in fact required) might be tricky

[Anonymous No.1669931]

>>1669920
>>1669920
>But since the signal.also produces an extremely small voltage, couldn't this also be used to charge a capacitor?
Those are not the signals you are looking for.
https://en.wikipedia.org/wiki/Energy_harvesting

[Anonymous No.1669973]

Working on this interface circuit for detecting push button presses on a game front panel.
The idea was to use CD4532 priority encoders to generate an interrupt signal for a microcontroller when any single button "in_0..7" or "in_8..F" is pressed. The microcontroller then immediately responds by giving a clock pulse to the 74HC166 shift register and driving "in_shift" high. Then it can generate 8 more clock pulses and transfer the encoders' data bits serially via "out_data".
It avoids continuously polling & transferring the shift register, but only one key press can be detected at a time. I think that switch bounce could also be problematic with this approach.

[Anonymous No.1670004]

>>1669794
Monostable timer circuit, be it with a functional equivalent to a 555 or something more along the lines of a schmitt trigger circuit. Both make use of RC circuits, so if you need something more precise then I guess using a crystal oscillator with some sort of counting circuit would also be viable. Yes an MCU counts as a counting circuit.

[Anonymous No.1670010]

>>1669794
is there a reason you need extra circuitry? your micro almost certainly has a spare timer/counter in it which could be configured to fire an interrupt after the derived clock is in the idle state for n clock cycles, which is exactly as precise as your system clock source. you need only route the clock to the timer's input pin and configure the timer at reset (varies by chip, read the datasheet). check the micro's errata list to be sure, we don't need another MCAS

[Anonymous No.1670024]

>>1669882
>those negative traces to the left
kek

[Anonymous No.1670069]

Let's say I want to charge two series connected Li ion cell. Should I go for two tp4056 with isolated ground or is there any better way to do it?

[Anonymous No.1670074]

>>1670069
you'll have to put power through the isolator, I'd go for a cheapy alibay 2S balancing BMS instead.

[Anonymous No.1670089]

>>1670069
bad idea. unbalanced cells in an NiMH pack are dead weight, unbalanced cells in a Li+ pack are fire hazards. the cells are drained as a unit, therefore should be charged as a unit. you'll want a balancing BMS in addition as >>1670074 recommended. the MCP73844-820 is a cheap linear charge controller IC (bring your own MOSFET) that works for a 2S pack, but you don't get temperature limiting in the 8-pin package, which may or may not be a problem for your application

[Anonymous No.1670102]

>>1669514
I am curious if i can replace, or make my own pic related with computer fans would be more efficient.

[Anonymous No.1670108]

>>1670102
no, or they'd use them instead

[Anonymous No.1670109]

>>1670108
i would think chink would use cheap non efficient fans.

[Anonymous No.1670114]

>>1669698
thanks for a detailed reply, appreciate it.

>e^jwt is a helix centred on the t axis,
yes, i get that, it is basically 1∠0. like a unit vector of some kind.

>Multiplication of complex numbers in polar form multiplies the magnitudes and adds the angles
Yes, so here is the key.
> jVwC * e^jwt
Converting jVwC to the polar form: VwC∠90 and to the euler: wCV * e^(jwt+90)
Now we multiply wCV * e^(jwt+90)*e^jwt and so the second e^wt basically disappears?
It is like multiplying VwC∠90 * 1∠0
So the answer is I = wCV * e^(jwt+90) . Where wC is really 1/Z. So it is V/Z. Done.

>A real a*cos(x) can be expressed as the sum of two exponentials: cos(x)=(e^jx+e^-jx)/2
Now this is real deep. I didn't think about cos from that perspective. I find it annoying that the entire AC analysis is based on transforming a real cos of a real signal to a complex form cos (wt) + jsin(wt) for the sake of further converting it to the euler form which I thought was kind of like cheating. You add "a little something that doesnt matter" as an intermediate step to convert to something else and then manipulate that something else and convert back and drop that little something and you get your transformed signal back. Ugh. But hey it is convenient as long as you don't overthink it too much. But now that you mentioned that the "real" cos is a sum of two exps, there may be a deeper connection between this fact and the euler formula and the justification of adding a complex part to the real signal to convert it to the Euler form to do AC analysis with phasors.

>If you put W=(1/2)e^jwt
What is W in this context? Just a random letter or is it a well know convention to represent something standard? That and especially W* reminds of me of linear algebra/vector space specific notation like dual vector space, actually i forgot all of that stuff.

>Now, conjugation distributes over practically everything
That last paragraph makes perfect sense. Thanks again.

[Anonymous No.1670117]

>>1670102
>>1670109
Thing is, those fan have big area and low RPM, which means they would be quiet.
PC fans have tiny area and high RPM, which makes them noisy.
As for motor, it is probably single phase shaded pole motor, which have like 20% efficiency or worse.

[Anonymous No.1670118]

>>1670109
but they can't do the job. which, outside of the crack-addled minds of MBA and econ students, makes them irrelevant

[Anonymous No.1670120]

>>1670109
>>1670118
simple test: put your finger into one of your pic related fans with the grille off, then put the remainder of your finger into a PC fan and compare the results

[Anonymous No.1670121]

>>1670120
Instructions unclear, dick stuck in air conditioner fan

[Anonymous No.1670128]

>>1669930
>current rating at 1MHz is probably an important figure of merit
I was worried about that. I can't find any specs on the doorknob caps.
>mounting them to the chassis while retaining kilovolts-level insulation
I didn't think of that but it sounds important, there is another motor-looking capacitor I think to isolate this HF circuit from the other transformers. Full wiring diagram if anyone is interested.

[Anonymous No.1670133]

>https://nz.mouser.com/datasheet/2/76/WM8224_v4.2-1141927.pdf
>16-bit 40MHz or 12-bit 60MHz ADC
>only has 12 parallel outputs, so 16-bit data must be fed out in serial
How the hell am I supposed to feed this serial data into storage? Some sort of speedy MCU with DMA?

[Anonymous No.1670134]

>>1670121
F

[Anonymous No.1670135]

>>1669539
You'd likely have to roll the rev of the board a few times to get all the bugs worked out of it, buying a whole set of components every time, some of which you'd have to have an assembly house install for you, and for the really high density BGAs there'd be no guarantee that all the balls would adhere correctly. Would be a *very* expensive project.

[Anonymous No.1670137]

>>1669898
You're making your otherwise high-impedance input into a low-impedance input and compromising the overall quality of your input signal in the process, and if I'm not mistaken if it's an AC signal you'd be more-or-less low-pass filtering it that way, too, and the output of your amplifier would be flaky at best.

Shorter answer: NO.

[Anonymous No.1670179]

>>1670133
>The output from the WM8224 can be presented in several different formats under control of the
OPFORM register bit as shown in Figure 21
page 25

[Anonymous No.1670216]

>>1670089
From my understanding, a typical battery pack is generally charged individually through the connector, isn't it?
The other thing is that MCP73844 and various other chips claim to balance charge two cells battery, while only provide two pins connected to the anode and cathode of the battery. How do they balance charge without a sensing pin in the middle of two cells?

[Anonymous No.1670223]

>>1670089
>>1670216
>The other thing is that MCP73844 and various other chips claim to balance charge two cells battery, while only provide two pins connected to the anode and cathode of the battery. How do they balance charge without a sensing pin in the middle of two cells?
why is no one reading the fucking datasheet?
>ctrl+f balance
>no results
>5.1.6 BATTERY VOLTAGE SENSE (VBAT)
>The MCP7384X monitors the battery voltage at the VBAT pin.
>This input is tied directly to the positive terminal of the battery.
>The MCP7384X is offered in four fixed-voltage versions for single or dual series cell battery packs, with either coke or graphite anodes: 4.1V, 4.2V, 8.2V, 8.4V
it just loads "battery-packs" with one or two cells in series. thats it. no balancing.

[Anonymous No.1670237]

>>1670114
> Converting jVwC to the polar form: VwC∠90 and to the euler: wCV * e^(jwt+90)
> Now we multiply wCV * e^(jwt+90)*e^jwt and so the second e^wt basically disappears?
There's only one e^jwt involved. You convert wCV*e^(jwt+90°) -> wCV*e^jwt*e^90° -> wCV*e^jwt*j -> jwCV*e^jwt then just elide the e^jwt part because that's kind of implied in anything that's not a ratio and just write I=jwCV.

To go back to the original question
>>1669584
> So the answer is I = wCV * e^(jwt+90). Where wC is really 1/Z. So it is V/Z. Done.
If V=V0.e^jwt, then dV/dt= jw.V0.e^jwt = jwV and I = C.dV/dt = jwCV. Z=V/I = V/jwCV = 1/jwC = -j/wC.

> But now that you mentioned that the "real" cos is a sum of two exps, there may be a deeper connection between this fact and the euler formula and the justification of adding a complex part to the real signal to convert it to the Euler form to do AC analysis with phasors.
The point of using complex forms is that you can represent both magnitude and phase as a complex amplitude. So you can phase-shift by multiplying by e^jp, and represent the relationship between e.g. V and I by a (complex) constant, which you can't do if V and I are real sinusoids (a*cos(x)+b*sin(x) or r*cos(x+p)) with different phases.

> What is W in this context?
Capital Omega, analogous to using w for lower-case omega. I.e. the too-lazy-to-type-greek version of Ω=(1/2)e^jωt.

> That and especially W*
Asterisk here is just the complex conjugate (in linalg it refers to the conjugate transpose).

[Anonymous No.1670240]

>>1670237
Typo:
> You convert wCV*e^(jwt+90°) -> wCV*e^jwt*e^90° ->
You convert wCV*e^j(wt+90°) -> wCV*e^jwt*e^j90° -> wCV*e^jwt*j -> ...
I.e. the entire exponent is imaginary, not just the wt.

[Anonymous No.1670246]

>>1670179
>it's just two parallel outputs
Oh, so I could feed that into two parallel-8bit-input EEROM or flash or RAM ICs? Neat. I guess I could use the falling edge and the rising edge to do my memory "read" inputs, perhaps with a small delay circuit in there if I get unreliable data.
It's by far the cheapest high-frequency 16-bit ADC on Mouser. Now to look for a 60MHz counter, which shouldn't be too tough. Looks like a 74LV4040 will work, though a CD4040 won't. Perhaps I've reached that point where CMOS ICs aren't as preferred for high-frequency circuits. Never heard of the 74LV series, but the LV is a good thing since the ADC wants 3.3V. I'd chain two of them together to count ~16 million samples. Finding DRAM that works at these frequencies will be a breeze from what I've seen.

Thanks so much for reading the datasheet for me, sorry to be a bother. I had basically given up hope since only one of the output pins (D11) was marked as a serial data output, and I had no luck looking for a 500MHz SIPO shift register.

[Anonymous No.1670247]

>>1670246
you're welcome!

[Anonymous No.1670265]

>>1669882
Time to do the tht holes

[Anonymous No.1670266]

>>1670265
It was fun. I forgot to upload the drilling gcode so instead the drill started to mill the traces sideways since it was using the gcode for laser etching

The holes ended up slightly misaligned and not completely through but that was because i etched the board before drilling it and thus moved and then was too lazy to realign it properly.
So increasing the drill depth by 0.5mm and a aligning the board properly on the xy axis and i will have fully automatic perfectly aligned tht holes

[Anonymous No.1670268]

>>1670266
like a glove

[Anonymous No.1670271]

>>1670246
>500MHz SIPO shift register
stop playing with babby blox and get an FPGA already

[Anonymous No.1670272]

Is this an effective over-voltage protection circuit?

[Anonymous No.1670273]

>>1670272
why don't you ask the components? or a simulator?

[Anonymous No.1670282]

>>1670137
Ok! thanks, but ignoring everthing else is it possible to power a transistor amp just using a capacitor?

[Anonymous No.1670284]

>>1670282
>can I use a lever to make a load lift itself
no, read some theory and get this magical thinking back to >>>/g/

[Anonymous No.1670289]

>>1670268
fuck, i realized i designed the slot holder backwards
and this is exactly why i want to prototype shit first before sending it to chinks to make a fancy version for me, if i ordered 50 of them, they would go straight to the bin

[Anonymous No.1670371]

>>1670237
>dV/dt= jw.V0.e^jwt = jwV
So you dropped e^jwt which essentially means you converted the entire expression to the rectangular coords. Since e^jwt resolves to 1. And jw.V0 was already in the rectangular coords.

What I am trying to say is it seems like taking a derivative of something like A.e^jwt equals to Ajw.e^jwt so we can think of Ajw as a complex number in the rectangular coords, right?

[Anonymous No.1670628]

>>1670240
Also, this threw me off:
>e^j90° = j
I am not used to see an angle such as j90 as a sole argument in the euler notation.
I don't know what to make of it.

The way I think about converting wCV*e^j(wt+90°) is I leave wCV alone and only convert e^j(wt+90°) which is 1L90 in polar coords, so it becomes cos(90) + j sin (90) = j. Maybe directly converting e^j90° -> j is some kind of a short cut that assumes this??

[Anonymous No.1670641]

>100mA output current
>3nv/sqrt(Hz) input noise
>rail to rail
>big supply range
>$1.75/circuit

https://www.digikey.com/product-detail/en/texas-instruments/OPA2156ID/296-53082-ND/9860897

Someone please tell me what the catch is before I stick them in my headphone amp, because the specs to price ratio looks ridiculous to me.

[Anonymous No.1670643]

>>1670628
I think of the e^(jx) function as a function that traces a circle in the complex plane. The circle always has length 1, so if x is 45° = π/4 then you'll get something like 0.707+0.707j. If x is 90° or π/2 you get 1j. If x is 180° or π then you get -1. It isn't that e^(jx) is shortform for cos(x)+jsin(x), but rather the original form of them.

In physics it's quite common to replace a cos(x) function with an e^(jx) because it's easier to work with around differential equations. The j part also naturally causes exponential decay or increase if x is a complex number, which crops up with a wave propagating through a lossy medium. In that particular instance it was the electric permittivity that was complex, as derived from the E and D fields via Maxwell's Equations. A similar helpful use of the exponential function occurs when considering quantum tunnelling if I recall correctly, the negative energy of the particle in the finite well outside the well's boundary causes the wave to turn into an exponential decay; the two imaginary arguments in the function cancel.

[Anonymous No.1670659]

>>1670643
Damn, this brings me back. You're right, there's a square root term of the particle's energy relative to its region in the exponential.

[Anonymous No.1670660]

>>1670643
>In physics it's quite common to replace a cos(x) function with an e^(jx) because it's easier to work with around differential equations.
Yeah I believe it is the case for EE specific DEs such as those used in AC analysis . I think the idea is DE in terms of sin/cos are converted to algebraic equations in terms of e^jx

[Anonymous No.1670663]

>>1670643
Also
>I think of the e^(jx) function as a function that traces a circle in the complex plane. The circle always has length 1,

So after thinking about this, this is no different than the common notation used in EE e^(jwt+θ). Which is just a little more generic since w is frequency so it implies continuous rotation as a function of time and θ is the initial phase. So to solve various EE problems, we just drop wt entirely and only look at θ to get the phase when we convert to a phasor etc. But yeah in a more mathematical sense it is just e^jx or e^jθ or whatever letter denotes the angle.

[Anonymous No.1670669]

>>1670663
Yes. When talking about waves you use the equation f(x,t) = exp(j(wt-kx)), which describes a plane wave (a wave that only depends on x and t, not on z or y) moving both in time and in space. But plane waves are fantasy, so in reality any wave is made of a superposition of plane waves. You can replace k or w with functions derived from the other alongside the speed of the wave. In the case of EM waves, the speed of the wave can be expressed in terms of the permittivity and the permeability of the medium.

[Anonymous No.1670673]

>>1670371
> so we can think of Ajw as a complex number in the rectangular coords, right?
Complex numbers themselves aren't in any particular coordinate system. An expression in "rectangular" form is a+bj where a,b are real, "polar" form is re^jθ where r,θ are real. But expressions don't have to be in either form. In particular, an expression with complex variables or coefficients is neither rectangular nor polar.

[Anonymous No.1670678]

>>1670673
> An expression in "rectangular" form is a+bj where a,b are real,
Which is what Ajw is. It is 0 + Awj. So it is the rectangular form, is it not?

[Anonymous No.1670713]

>>1670678
Provided that A is real. In which case, you could equally say it's in polar form, Awj = (Aw).e^j90°. Really, there isn't much difference for values which are either purely real or purely imaginary. If A is complex, then Awj is neither polar nor rectangular.

[Anonymous No.1670776]

what's going on here

[Anonymous No.1670780]

>>1670776
just some euler algebra, proper applied electronics will resume eventually

[Anonymous No.1670808]

>>1670713
give your head a shake

[Anonymous No.1670814]

>>1670663
> So after thinking about this, this is no different than the common notation used in EE e^(jwt+θ)
e^j(wt+θ)
The exponent needs to be purely imaginary (zero real component) for the expression to have unit magnitude.

[Anonymous No.1670846]

the littlest tcxo

[Anonymous No.1670849]

>>1670846
is that a memes oscillator?

[Anonymous No.1670850]

>>1670849
yes, SiT1552

[Anonymous No.1670859]

>>1670850
kawaii!

[Anonymous No.1670868]

>>1670776
it is about capacitors that we put in DIY. so it is on topic. kinda. btw you will NEVER have such a good conversation on /sci/. there will be 2-3 answers and the thread will die. this place is awesome. theory and practice go together. just like the real part and the imaginary part of jwA.e^jwt+teta.

[Anonymous No.1670872]

>>1670814
yeah my bad, misplaced the parentheses.
but i am not sure why you call it magnitude not phase or angle. Ae^j(wt+θ) : A is magnitude. and wt+θ is phase (or just θ depending on the context)

[Anonymous No.1670875]

>>1670713
i disagree, anon. it cannot be neither.

>If A is complex,
OK let's make it 1+j

>then Awj is neither polar nor rectangular.
No, it doesn't stop being a complex number. And a complex number can not be 'neither polar or rectangular' since you have to pick a notation to express it.

So if A=1+j we get (1+j)(0+jw). This is a product of two complex numbers in the rect notation.

[Anonymous No.1670884]

>>1670875
(1+j)(0+jw) isn't rectangular notation, though. If you simplify it to -w+jw then it's rectangular. But until then it's just a complex expression that's not in either form. Rectangular form is necessarily a+bj with a and b real.

[Anonymous No.1670911]

>>1670884
>(1+j)(0+jw) isn't rectangular notation
Yes it is rectangular. It is a product of two rectangular numbers (1+1j) and (0+jw). Both of them are in a+bj format.
a and b are real. can you see them?

>If you simplify it to -w+jw then it's rectangular.
By simplify you mean you multiply two complex numbers in the rectangular form. Yes you can do that.

[Anonymous No.1670914]

>>1670884
>just a complex expression that's not in either form
this statement makes no sense at all.

[Anonymous No.1670925]

>>1670911
>>1670914
Rectangular and polar forms are just standardized ways of expressing a complex number. (1+j)(0+jw) isn't a number, it's an expression.

https://en.wikipedia.org/wiki/Expression_(mathematics)#Examples
https://en.wikipedia.org/wiki/Number#Complex_numbers_2

[Anonymous No.1670931]

>>1670925
>isn't a number, it's an expression.
ok so you just prefer to be autistic about it.
of course it is an expression. but i am saying that this 'expression' is called a product of two complex numbers x*y. in the rectangular notation. it is like insisting that 2*2 is not an integer number, it is an "expression". and only after you multiply them, then 4 is an integer number. whatever you call them, i am not sure what your original point was.

[Anonymous No.1670935]

>>1670931
My point is that (1+j)(0+jw) isn't rectangular notation. It's the product of two numbers in rectangular form, sure, but that doesn't make it a number in rectangular form. And >>1670713
is right and not every expression involving complex numbers is in polar or rectangular form by default.

Retard.

[Anonymous No.1670947]

>the product of two numbers in rectangular form isn't rectangular notation

[Anonymous No.1670948]

>>1670947
>my retarded argument is correct if I put it in green and ignore its refutation

[Anonymous No.1670952]

>>1670948
>>1670947
>unironically thinking 4chan is for education

[Anonymous No.1670957]

>>1670952
well this particular thread is the exception.
it's been really helpful until he finally sperged out but it was expected considering we were just going in circles pretty much saying the same thing arguing the terminology. but i actually learned/refreshed my memory quite a bit. so it's been great help. even the edgy anon helped me. i love all of you guys.

[Anonymous No.1670960]

I bought a gaming headset for my desktop.
When I use it as a headset for regular phone calls, other people hear the "cellphone close to speaker" sound. When I do VOIP I don't have that issue.
Are there special shielded cables I could buy?

[Anonymous No.1670968]

>>1670960
>>>/g/

Check the settings in your VoIP software.

[Anonymous No.1670969]

>>1670968
what does Hangouts/WhatsApp/Signal settings have to do with the fact that the microphone cable is picking up EM GSM disturbances when making ordinary calls?

[Anonymous No.1670972]

>>1670969
the government is watching you, dude.
F

[Anonymous No.1670975]

>>1670102
fans and their engines are very application-specific. efficiency is key when selecting parts

>>1670121
downvoted

>>1670972
why are you here? you could just have told him to get a shielded cable if you knew anything about electronics

[Anonymous No.1671051]

itt

[Anonymous No.1671108]

>>1670872
> but i am not sure why you call it magnitude not phase or angle. Ae^j(wt+θ) : A is magnitude. and wt+θ is phase (or just θ depending on the context)

>>1670814
> e^j(wt+θ)
> The exponent needs to be purely imaginary (zero real component) for the expression to have unit magnitude
Here, "the expression" is e^j(wt+θ). Assuming that w, t and θ are all real, the exponent is imaginary and |e^j(wt+θ)|=1.

In the original expression e^(jwt+θ), the magnitude is |e^(jwt+θ)|=|e^jwt*e^θ)|=|e^jwt|*|e^θ)|=e^θ.

[Anonymous No.1671117]

>>1671108
>In the original expression e^(jwt+θ)
that was a typo (misplaced parentheses) and so it doesn't make sense. should be e^j(wt+θ) of course. at least in the context of what we are talking about itt.

[Anonymous No.1671286]

Is there a way to find out proper voltage and current for a LED without its datasheet?

[Anonymous No.1671287]

>>1669539
I've seen a project where a guy builds what's basically a DP to eDP converter. That alone already took him several revisions of the PCB.

[Anonymous No.1671288]

>>1671286
Pass 10mA through it to find out the voltage drop. You can't determine the maximum current experimentally (at least not from a single sample; the manufacturer determines the value through destructive testing on hundreds/thousands of samples).

[Anonymous No.1671291]

>>1671286
You don't supply an LED with "proper voltage." The "proper voltage" across the LED is developed as a consequence of getting its quiescent current right. Most regular LEDs will take a current between 1 and 20mA. High efficiency LEDs might give reasonable levels of brightness with just a few mA of current. Older LEDs might need in the range of 10-20mA to be sufficiently bright. That's a good starting point. Some LEDs take up to 30mA to 50mA but that's fairly rare and those will usually get pretty bright at 20mA.

Use an adjustable constant current source if vary it and find where the brightness looks about right. For adjustable constant current source I recommend a two transistor current mirror which is very easy to build and good for low current stuff like this. You could also use a 3-terminal voltage regulator configured as a current regulator though I'm not really a fan of that topology myself.

[Anonymous No.1671292]

>>1671288
>>1671291
So I'd have to find out the proper current by trial and error huh.
Constant current LED drivers adjust their voltage within a range already so I'd only have to find out if the range matches the specs.

[Anonymous No.1671294]

>>1671292
There isn't really a proper current either. LEDs operate over a range of currents. There might be an optimal current for brightness and possibly a different one for optimal efficiency but in general you can operate at any current within a 20-30mA span. You simply don't want to run them beyond their absolute maximum rating. Without a datasheet you'll have to drive one LED up until the point it fails, record that current and then tack a 10 or 20% error margin on that. That's the current you don't want to exceed. Honestly, just use the 20mA maximum rule of thumb for LEDs unless you have something really esoteric. Nearly all LEDs will run between 0 and 20mA just fine and won't exceed any maximum ratings.

[Anonymous No.1671297]

>>1671294
I see.
Was just trying to see if there's a way to safely drive LCD backlight LEDs without its datasheet.
Pic related is for another panel, has everything listed. From what I can tell this means the LED should be driven at ~27mA.

[Anonymous No.1671298]

>>1671297
Forgot pic

[Anonymous No.1671299]

does daylight contain uv light or only direct sunlight?
I want to expose some photosensitive pcbs and wonder if i need like a dark room or some shit

[Anonymous No.1671302]

>>1671299
Also i keep reading about how you can't leave the pcb under the uv light, if you are curing say solder mask, for too long as it will overexpose
that sounds like bullshit to me because if that was true then any pcb board that is left out in sunlight would have to become overexposed

[Anonymous No.1671306]

>>1669514
So I have a bunch of ferrite toroidal cores of several sizes and lots of enameled wire. I'm used to working more with audio and digital shit so please r8 my retard Idea for a welder
>weave a 100-300 turn primary
>secondary made of 1 turn of copper sheet
>switch the fucker at 20khz and about 1A
>weld shit

[Anonymous No.1671318]

>>1671306
you can do this, but there's a difference between welding with AC and DC. also don't saturate your core or it'll explode.

[Anonymous No.1671319]

>>1671299
>>1671302
Windows block almost all UV so you dont need a darkroom. You wash away unexposed mask after the process, if you dont you get weird lumps ans blotches

[Anonymous No.1671321]

>>1671319
>weird lumps ans blotches
Is that a problem thought? I don't mind having lumps as long as the solder mask does its job, it is worse to udnerexpose because then parts of your board will be exposed when you wash away the solder mask

[Anonymous No.1671385]

>>1671292
Dont worry too much, 10mA is perfect. Not too high not too low. Just stick to 10mA and you will be fine.

[Anonymous No.1671387]

>>1671306
>20khz
Do you know the specs of your cores? Maybe they are low freq?

[Anonymous No.1671393]

>disable every peripheral i can think of
>pic still does weird things
there's always another fucking peripheral hiding somewhere

[Anonymous No.1671400]

>>1671385
some smaller LEDs, especially high-efficiency types, have 5mA absolute maximum If

>>1671306
a ferrite core suitable for 20kHz at amps would be very large and probably like $5
if you really want to try this, skip the RF cores and use a power transformer core

>>1671299
you should not leave sensitized boards exposed to blue-end wavelengths unless you're patterning them, and try to avoid exposing them to light at all. nothing heroic, just work with purpose
the development tank should be in the dark so that you aren't exposing fresh photosensitive material right as it reaches the surface, thus slowing development and reducing contrast

>>1671321
but it won't, it might tear or crack when you accidentally drag a soldering iron across it
just follow the process, anon

>>1671393
>starting peripherals in the enabled state
it's almost as if 3 cent micros are just better :^)

[Anonymous No.1671403]

>>1671400
i genuinely love pic10s. this pic16 was just the cheapest thing on digikey with the pin count i needed. it does have a few really well thought out peripherals though. i made a hardware RTC and debounced rotary decoder entirely independent of the system clock.

[Anonymous No.1671405]

>>1671400
>probably like $5
not him. but that sounds dirt cheap. a low freq high power toroidal core can be $100 easily.

[Anonymous No.1671427]

>>1671387
>>1671400
No clue about the cores, they are salvage from a university lab, they range from unmarked CD sized to coin size ones. The ones I plan on using are from some 100W+ supplies I found burnt on the side of the road. Probably will higher than 20khz, just said that because I don't want to hear them hehe

[Anonymous No.1671428]

>>1671393
Check the datasheet, sometimes on a certain peripheral it says
>to use this you need to put this value on this random non-special register before etc
or some shit like that. Also
>mplab

[Anonymous No.1671429]

>>1671428
>mplab
mpasm for life

[Anonymous No.1671440]

>>1671427
>CD sized
Those are probably power cores rated at the mains freq? I don't know what will happen if you feed high freq into them. They will probably have huge losses due to eddy currents etc. Things will get hot

[Anonymous No.1671506]

>>1671440
They aren't iron, they are ferrite. I think they were from a big conversor, the laboratory I found them in was for pwoer electronics, they were dumping out projects.

[Anonymous No.1671585]

Me again, last time I posted I was trying to convert a light variable resistor to a simple twist pot. I ordered a 500k ohm log pot, wired the gnd to the left pin and the power to the middle pin. The pot kept smoking out as I got closer to one end. Did I use the wrong ohm pot? I heard about using it as a voltage regulator but there’s only two leads for the LSR not 3 needed. See schematic

[Anonymous No.1671636]

new board. spot the bodges.

[Anonymous No.1671637]

http://ltwiki.org/index.php?title=Simulating_the_MC34063_in_Inverter_Configuration_with_an_Accurate_TL431A_Model

How do I add the IC to LTspice? I have put the .asy file in the sym folder and the .asc and .circ in the sub folder but I get hierarchical errors

[Anonymous No.1671650]

>>1671403
neat!

>>1671585
you created a variable heater. gj
once you have received a new pot, connect -1 on the swell pedal to the wiper and -2 to one end. no power required or desired

>>1671636
they always were a bodge, weren't they

[Anonymous No.1671653]

whats a searchable term for pic related?

I reverse image searched and it's called a "xpdt" but nothing shows up on wikipedia or ebay.

I understand double/triple pole double throw or w/e but I only ever get rocker switches never buttons like pic related.

[Anonymous No.1671655]

>>1671653
Had a tard moment.

[Anonymous No.1671656]

>>1671637
please if someone could try running this model and explain to me how to include it in spice it'd be GOLDEN.

[Anonymous No.1671664]

>>1671656
pirate multisim instead

[Anonymous No.1671668]

>>1671650
That’s what I did though, -1 to the center and -2 (Gnd) to one end and it just cooked it when I dialed up to that end

[Anonymous No.1671680]

>>1671655
>dat filename
what, Pic related?

>>1671664
anyone in here /microcap/?

>>1671668
cooks with all keys up? if so, check C63 for short
in any case, the LDR in the swell pedal assy resistance probably bottomed out on the order of a few hundred ohms, possibly much more depending on lamp brightness. a series resistor to make up for that wouldn't go amiss

[Anonymous No.1671708]

>>1671668
>just cooked it

that's impossible when it's getting its current thru a 270K resistor. best guess is you wired it to the wrong pins and are putting 27Vdc across it.
in any case, temporarily putting a 5K resistor in series will make sure no smoke comes out even if you wire it wrong (which you can remove once it works more-or-less.)

[Anonymous No.1671722]

Im this guy >>1669843 and I've spent more time reading about capacitors in the past couple days than I ever thought I would in a lifetime. From what I've found the CDM mica capacitors are basically the only ones that give a current rating at different frequencies (6.2A at 1MHz on the one I need) but some radio guys use the tiny high voltage through hole ones for some 2A applications which they are not really rated for. First I was thinking some of these in parallel but they are generally lower voltage and higher capacitance than I need, so I'd need to run them in series as well, and not really saving any money. I still can't find current ratings on any cheap doorknob caps but I am seeing a lot on ebay that have big terminals and higher kV ratings than I need and am thinking of pulling the trigger on these: https://www.ebay.com/i/323883369612
My question is, do you guys think this would hold up in my application? To recap, I am replacing (2) 5kV 2000uF mica caps rated at 6.2A @ 1MHz wired in parallel in a welder high frequency generator circuit that's basically an oscillator with a tungsten spark gap and a coil to generate several? kV pulses at ? MHz.
One of the Chinese knockoffs of this particular doorknob cap says "high frequency machine high voltage ceramic capacitor" which leads me to believe it has been / could be used in similar applications. Also I saw in the welder manual that if one cap fails, you can still use the machine (total capacitance would go from 4000 uF to 2000 uF) but they reccomend tightening up the spark gap and increasing the series resistance on the variable resistor so you dont toast the remaining cap. So I do think I have some leeway on capacitance or at least capacitance change caused by self-heating.
Anyone who bothered to read all that have any thoughts?
PS This is the circuit and C2 is the capacitor in question (there are 2 in parallel, this diagram shows only one). Thanks.

[Anonymous No.1671724]

>>1671708
>>1671680
thanks guys. so is the pot i have now done for since it lit up and smoked at the extreme ends of the slider? i'll go order another one on ebay.here's more of the schematic if it makes any difference. I was so close to getting this project done.

[Anonymous No.1671751]

>>1671724
>is the pot i have now done for

if you burned up only one extremity, then you can use the terminal on the opposite end. the resistive track should be intact at the other end. it could also still be functional at the burned end. depends on the severity of the burn.

>here's more of the schematic

it confirms there's nothing there that could burn the pot.
and sorta confirms you wired it across the terminals for the lamp.

[Anonymous No.1671793]

>>1671393
>there's always another fucking peripheral hiding somewhere
Normally, PICs have ADCs enabled by default. Otherwise, using the Ports for general I/O is the top priority.
Need any help?

>>1671428
>mplab
You mean MPlabX?

>>1671429
>mpasm for life
You know it

[Anonymous No.1671805]

>>1671793
>Need any help?
i'll figure it out by bashing my head against the code eventually, but if you're bored here's a peripheral wiring flowchart and some illegible code with everything but the problem area trimmed out.

https://hastebin.com/raw/juhodazari

the code is for a 16F15356 and the intent is to debounce and decode the input from a rotary encoder with a mom switch, without any software input. i could do this in software but i have all these CLCs and nothing else to do with them.

[Anonymous No.1671858]

Why do analog true rms converter chips rely on analog multipliers instead of a simple precision full wave rectifier plus an integrator on the output? Does that not give you the true RMS value of a waveform, even non-sinusoidal waveforms?

[Anonymous No.1671859]

>>1671858
Pretty sure that gives an average, not rms.

[Anonymous No.1671892]

>>1671680
Yes! Thankyou!
"push button switch" didn't offer anything besides momentary SPDP buttons like arcade buttons.

I then tried "right angle push button switch" and found some results on ebay.

It's not the best way to isolate the part since mometary buttons come up still but at least I have results to sift through now.

[Anonymous No.1671900]

>>1671319
>Windows block almost all UV
window glass
UV-A (315–400nm) pass
UV-B (280–315nm) block
UV-C (100–280nm) block

[Anonymous No.1671903]

>>1669514
>test, post results
Filaments from a failed bulb:
Brightest line tester ever!

[Anonymous No.1671912]

>>1671892
how many poles do you need?
search for the number, 4pdt, 6pdt etc (these things usually have lots of poles) then find the manufacturer datasheet and model that you want. yes its a little bit labor intensive.
you know you can buy switches from electronics/industrial supply places, and thats what people on ebay do then sell them for profit right? what country do you live and someone will help you with a supplier.

[Anonymous No.1671938]

>>1671858
> Does that not give you the true RMS value of a waveform
No, it gives you the mean absolute value.
> even non-sinusoidal waveforms?
Not even for sinusoids. For a sine wave, the mean absolute value is 2/π=~0.6366, the RMS value is 1/√2=~0.7071.

[Anonymous No.1672136]

>>1671637
>performance enhancing parasitic elements
feels good man

>>1671912
and yet, all those little kiosks in Huaqiangbei are there just for tourists? the grey market is a thing and occasionally a very good source

[Anonymous No.1672152]

Where do you guys buy electronics parts now? The last electronics hobbyist store in my area closed down. Any recomended websites?

[Anonymous No.1672155]

>>1672152
of course it depends on the destination country. there's a sources list in the OP to get you started

[Anonymous No.1672163]

>>1671892
oh, just add "latching" or "po-po" (push-on push-off). they're also called "piano key" switches or "radio" buttons (just like the UI widget they inspired)

[Anonymous No.1672182]

I'm taking apart an old GBA and noticed the odd loking capacitor, pic related. Does it look damaged? Other GBA Motherboards dont seem to have this cross in the capacitor. The device boots just fine, havent tried a game with it yet.

[Anonymous No.1672188]

>>1672182
>Does it look damaged? Other GBA Motherboards dont seem to have this cross in the capacitor.
the cap is fine
(predetermined breaking point)

[Anonymous No.1672191]

>>1672182
>>1672188

[Anonymous No.1672197]

>>1672188
>>1672191
Ah, I See! These markings are weakened parts on the material, so the build up pressure can escape without the whole motherboard being ripped apart, right? Thank you.

[Anonymous No.1672199]

where do i find an electric engineer to build free energy devices with and start a breakaway civilization?
and if interested
bobopatch#5615

[Anonymous No.1672206]

>>1672197
>so the build up pressure can escape without the whole motherboard being ripped apart, right?
yep, especially to not rip of solder pads.
scratching off the solder mask from thin traces with a scalpel to recover some copper to desperately solder some wire on it is no fun.

[Anonymous No.1672210]

How the hell are you supposed to desolder components? Solder sucker isn't sucking, wick only pulls <90%. I feel like the best method is to wick, then just roast the lead until it oxidizes, then pushing out the cold solder plug - but that's a great way to rip out the legs
Please I have like 70 diodes I need to reinstall

[Anonymous No.1672214]

>>1672210
solder sucker off as much as you can.
wick up whats left, use wick with flux in it or add your own, wont be worth shit otherwise.
hot air and tweezers.

ideally 3 steps in that order but there is nothing inherently wrong with just jumping to 3.

if you don't have a heat gun diodes might be workable but not fun. if you are replacing diodes (throwing them out) then just snip the leads and do one at a time, heat and pull out with pliers or tweezers. if you want it intact you are going to have to do a real god job getting the hole clean or just work it out very slowly side by side until you are done.

but if you watch some youtube videos you will spot the key ingredient is flux.

[Anonymous No.1672217]

>>1672210
>How the hell are you supposed to desolder components?
- solder over so there is a thick ball
- suck the ball
- remove the rest with wick
- clean the pads/holes from flux

[Anonymous No.1672234]

>>1672199
free energy is EXTREMELY highly censored, so if you ever go public with one of the various devices you WILL be disappeared. us electrical engineers just use them quietly to cut back on power bills.

[Anonymous No.1672239]

>>1672199

[Anonymous No.1672258]

>>1672210
wick, generous amount of liquid rosin flux, hot iron always worked for me.

[Anonymous No.1672272]

>>1672210
9005.5 hours in Doodle related works for me 95+% of the time. very important that the tip of the sucker is relatively flat to create a good seal with the board, and that the tip of the iron is big enough to heat the solder all the way through. if your solder sucker still isn't working, maybe it needs a bit of cleaning

[Anonymous No.1672278]

>>1672272

fuck that noise. only works half the time unless you're a kung-fu master. one of these raises that probability to 95%.

[Anonymous No.1672283]

>>1672278
thats a hot benis bump

[Anonymous No.1672287]

>>1671912
>someone will help you with a supplier
That's very kind but I'm just a entry hobbyist tinkering around.
Just starting micro controllers and I live in the sticks so the closest thing that was even remotely related was radioshack that shut down about 10 years ago so ebay was convenient.
I only really require a handful of a individual part to have some fun.
I got some reed switches, and rotary encoders, and stepper motors which none of them I've ever tried before so I'm excited!
>>1672163
>latching
I noticed this on the spec sheet above and found it to be helpful because apparently these can be momentary non-latching too which is rather strange.
>radio buttons
These are so cool since they work as a mechanical failsafe without requiring code.

Is there a way to buy the type that was used in the radio with those mechanisms?

[Anonymous No.1672290]

>>1672210

Solder succ the bulk of it out, then flux the FUCK out that wick.

I got one of these >>1672278 for when that don't work. Don't know what the hell I'm supposed to do when the bulb fills up/clogs, since none of that assembly actually comes apart, but it works, at least.

[Anonymous No.1672296]

>>1672290
Is that metal ring on the bulb unclasp-able?
If not maybe you can snip it off carefully or rotory tool it without damaging the bulb, clean the interior and then use a small piece of heatshrink for air pressure and a 14awg twisted tightly.

[Anonymous No.1672306]

>>1671751
I dont have access to the unit until next weekend ( not stored at my place) but i swear I didnt wire up the lamp wires by accident. Next weekend I will post in /ohm/ again and show you exacit what I did. my new pot should arrived by then too.

[Anonymous No.1672363]

>>1672306
>i swear I didnt wire up the lamp wires by accident.

You needed help with a pair of speaker wires.
I wouldn't be so sure you know what you didn't do.
>>1669329
>>1669451
>>1670142

[Anonymous No.1672364]

>>1672363
ive never dealt with a speaker that wasnt labeled at all. i did run a multimeter across it to determine it was 8ohm which was very helpful.
do you see >>1672306 ? two brown thin wires, two black heavy wires braided together the thick wire is for the light, the brown wires (pins 1 and 2) are for the resistor.

[Anonymous No.1672385]

>>1672278
for $14 you could have Pic related. joints won't know what hit 'em

>>1672287
>a few of each kind of part
then ali the chinaman's flea market is for you! ten of anything on sale at ten different tables! bring your city boy street smarts and keep your wits about you. or go see honest Li Chuang, he'll set you up with a million of anything if you can afford it, and he's got resistors 4 cents for a hundred
>the type that was used in the radio
probably not. in most cases those weren't actually switches, just a mechanical memory for the tuner, therefore usually part of each product's proprietary mechanical design

[Anonymous No.1672398]

>Finally got my copy of "The Art of Electronics" in the mail today
>It's a counterfeit copy
Fuck you Amazon, I even made sure I wasn't buying from a 3rd party seller.

[Anonymous No.1672423]

>>1672398
G R E E K E D
R
E
E
K
E
D

[Anonymous No.1672475]

where are good places to get lithium batteries? I don't want some physics bending claim of capacity, just modest performance/quality for a decent price, need to rebuild a 54v ebike battery

[Anonymous No.1672502]

>>1672385
>Li has it all for cheap
Thanks for the advice.
There's always room for the next project.

>just a mechanical memory for the tuner
Electronics are so fascinating.
It looks like they actually encoded the station information in magnetically!
Too bad they are so tightly integrated into a implementation like you said and not more modular.

[Anonymous No.1672505]

>>1672364
>two black heavy wires braided together the thick wire is for the light, the brown wires (pins 1 and 2) are for the resistor.

that doesn't sound right. the wires for the LDR carry audio, so they should be using coaxial wire, which is what the black wire looks like. the brown wires are definitely not suitable for audio, coz they can pick up all kinds of ambient electrical noise.

i didnt realize you had a multi-meter. but since you do, it's a super simple matter of reading the voltage across the browns and the blacks: the one that reads approx. 27Vdc is for the lamp. all god's secrets are revealed to a man with a meter.

[Anonymous No.1672507]

So I'm new to this stuff, I'm trying to figure out how to encode all propositional-logic terms in DC circuits on a breadboard. So far I've figured out:
>you can make AND with a transistor (and resistor)
>you can make OR with diodes
but I can't for the life of me figure out how to get NOT or any kind of negation. Someone give me a clue, does it need capacitors or something?

[Anonymous No.1672508]

>>1672507

2-second google search will tell you just need an NPN and a couple of resistors.

[Anonymous No.1672511]

>>1672507
RTL = Resistor Transistor Logic. Google it, it's simple stuff. Diode logic can be used for both OR and AND gates, but no inverting gates as that requires active circuitry. Diode transistor logic is a combination of both, but isn't terribly useful.
Don't forget to always use your transistors in common-emitter configuration with resistors at both their base and collector unless the circumstance specifically needs otherwise (i.e. some shitty NPN pullup transistor or analog circuit).

[Anonymous No.1672516]

>>1672508
wtf I invented this already but it didn't work

[Anonymous No.1672518]

>>1672516
post pics

[Anonymous No.1672519]

>>1672518

[Anonymous No.1672521]

>>1672519
post diagram, stupid pics we have enough

[Anonymous No.1672522]

>>1672519
I only see one resistor aside from the LED’s one. You should have both a collector resistor and base resistor, missing either one could lead to destroying the transistor.

[Anonymous No.1672523]

>>1672519

you gotta take the output from the collector to ground. if you put the LED in series with the collector, you get a NOT NOT gate.

[Anonymous No.1672524]

>>1672521
Diagram without resistors for legibility.

[Anonymous No.1672530]

>>1672524
oh this transistor is probably just fucked

[MrCrummy No.1672532]

Got myself a free plasma cutter, it powers on but doesnt cut.
I suspect the pressure switch on the air supply inlet may be the culprit

[Anonymous No.1672586]

>>1671805
Nothing really jumped out, I didn't check over your config settings though.
What exactly is the problem? Does it not work at all? Does it work, but with issues?
Also, what are you programming with?

[Anonymous No.1672661]

>>1672532
>doesn't cut
explains the price tag

[Anonymous No.1672668]

What's the best way to drive 7-segment displays. Is there any use for BCD-to-7seg driver IC in <CURRENT_YEAR> or should it be done with a chain of serial-to-parallel shift registers with latches (74hc565)?

[Anonymous No.1672678]

>>1672668
what ever way works best for whatever your doing

[Anonymous No.1672681]

>>1672668
>What's the best way to drive 7-segment displays.
Whatever uses the fewest parts. If you're driving from a uC you can serial shift in all the segment bits instead of using decoders. Actually with a uC you can even skip registers and directly drive each digit one at time fast enough to be nearly invisible.
What's your deal? uC or no?

[Anonymous No.1672686]

>>1672668
>BCD-to-7seg
>current year
not really, most useful (i.e. least replaceable) in the case of a 1-of-10 cold cathode driver
>shift regs
this is good. easier to get alphabetic output as in >>1671636 and you also have easier access to the decimal point if you want. multiplexing can deliver better power consumption, at the expense of a bit more time-sensitivity. you can do this yourself or, if you don't mind a smrt peripheral, there are some nicely integrated jellybean options, originally intended for front panels on white goods and consumer electronics, that have internal oscillators and current regulation, software-controlled brightness, no external oscillators or dropper resistors required
MAX7219 manages an 8x8 matrix, can do BCD-decode OR direct-segment mode, easy SPI command set, (iirc) daisy-chainable. about 40 cents on ali in DIP
HT16K33 manages a 16x8 matrix, also scans up to a 3x13 keyboard matrix with interrupt output on key down, standard I2C interface. about 40 cents on ali in wide SOP-20 to SOP-28
there are other keyboard/LED drivers that have buzzers, <blink> mode, other stuff that someone wanted and was willing to buy 100 reels as usual

[Anonymous No.1672707]

>>1672524
How about actually copying a circuit diagram from google instead of this “short to ground” style switch. Meaning, emitter directly to ground, collector directly to the output but also to the V+ rail through a 10k resistor or so, and base connected through a 100k resistor to your input.

[Anonymous No.1672728]

magnetic sensing: so simple, even a monkey can do it

[Anonymous No.1672768]

>>1672532
>namefag
>snapchat
where is your patreon?
I'd like to donate -$1

[Anonymous No.1672778]

>>1672728
adorable

[Anonymous No.1672783]

>>1669627
You can always convert Euler to Cartesian form and vice versa

[Anonymous No.1672794]

Rate my "ideal" diode, /ohm/.

[Anonymous No.1672795]

>>1672794
Stimulate it and see

[Anonymous No.1672796]

>>1672794
take a look inside your OP amp

[Anonymous No.1672797]

>>1672795
didn't simulate it but did build it. works as expected.
if only I could get rid of this 130mOhm Rds. (I used an SSM6J206, not the same fet from the schematic that symbol was just pin-compatible)
>>1672796
???

[Anonymous No.1672836]

>>1672794
adorable/10 but pull up that gate
>get rid of this 130mOhm Rds
AON2405?

[Anonymous No.1672874]

>>1672794
0/6
advanced ignorance
RTFD!

[Anonymous No.1672885]

>>1672836
>pull up that gate
I was planning on having one, but the op-amp drives it high so it didn't seem worth it.
>AON2405?
unfortunately not stocked on digikey. But hot damn, that would get me <30m Rds.
component selection is as much of a science as circuit design, unfortunately nobody really seems to teach that. you just learn through mistakes.
funny, I recently realized got these SSM6J206 kind of by accident, I actually filtered by continuous current rating and chose 1-2A instead of just 1A and up. Found that, thought it was the best I could get. stooopid.
>>1672874
did you?

[Anonymous No.1672924]

>>1672885
>not stocked on digikey
There are 4 such listings on ali, if you dare.
>component selection is as much of a science as circuit design, unfortunately nobody really seems to teach that
Amen to that. I'd kill for a list of commonly used transistors and the like as used by industry leaders, along with RRPs and a guide to valuing them by their specs.

Also wouldn't a comparator work better than an op-amp there? And I think it would only work properly if the load had either a big capacitor in parallel or a battery to be charged, as if your voltage goes low on the LOAD side your op-amp will lose power. That's probably why the guy said to use a pullup.
I assume the all-enclosed ideal diode packages use a N-type FET along with some sort of bootstrap circuit. Some of them can also run without a ground reference if I remember correctly.

[Anonymous No.1672932]

>>1672924
I'm going to try an MCH3383, which will get me <50m Rds.
Low LOAD voltage and high LINE voltage is an illegal condition, it won't happen. The opamp is always powered either from the LOAD side, or through the FET's body diode from the LINE side.
I'm really not sure the value of using a comparator rather than an opamp here. I wanted high gain and rail-to-rail output, and the LMV321 provides. Not shown on this diagram but on the actual circuit I have a cap between + and - input to filter out some differential mode noise but I'm not sure how necessary that is.

[Anonymous No.1672937]

>>1672932
Oh you're right that won't happen. I recommend a capacitor because I think it would have a faster slew rate as it's designed to operate in the non-linear region, as opposed to an op-amp. The cap's likely a fairly good idea. If you could use an NFET instead you'd easily be able to get below 30mΩ, though it depends on the voltage you're going for. Could always put a few PFETs in parallel.

[Anonymous No.1672938]

>>1672924
>I'd kill for a list of commonly used transistors and the like as used by industry leaders
whatever's cheap at the time and works well enough. the oldest of old standbys have largely trickled down into the maker world. they're sold for the lowest prices in quantity but won't guarantee you the best performance or the lowest total BOM. if your company has a take-or-pay contract for 2N2222s in SC-70 packages, you should plan on using some
>RRP
what matters is how much you can get them for tho. shoppu hada

>>1672932
>is an illegal condition
but LINE could exceed LOAD by a diode drop for at least a moment while the op amp starts up. consider increasing R1-R4 and the divider ratios
>comparator
differential input range can be limited to ±~1 diode drop on some opamps (3lazy3see whether this is one of them) so a proper comparator could work better. either would work well if the inputs aren't limited to at most Vdd + a couple hundred mV

[Anonymous No.1672941]

>>1672938
>but LINE could exceed LOAD by a diode drop
that's fine, isn't it? V+ will still be higher than the - input, the opamp output will still drive the gate low and turn on the FET. I wouldn't change the divider ratios unless I had to, these are all being hand assembled so having all 4 resistors the same wins a lot of points.
This opamp is rated for +-Vcc differential inputs so exceeding a diode drop isn't an issue.

[Anonymous No.1672947]

Where do you guys buy electronics to repair? I'm looking at spending some money on just some meh cheap shit to use to improve my troubleshooting and repair.

[Anonymous No.1672948]

>>1672947
2nd hand shops, junk shops, literal dumpsters

[Anonymous No.1672994]

>>1672941
I was a bit concerned about possible latchup in case of a slow cap charging. otherwise, ship it

[Anonymous No.1673025]

Ok I want to have two op-amp circuits A and B with gains controlled by a double-gang potentiometer. The gains should be such that a change in the potentiometer that increases A's gain by a factor of X will result in the gain of B decreasing by a factor of X. If chained together without any circuitry in between, I'd always have a gain of 1 regardless of the potentiometer's position, but the signal amplitude between them would vary.
This circuit would allow me to test out nonlinear audio filters, such as clipping diodes and the like.

I think (after some algebra) that with two inverting amplifiers, I can have RA1 and RB2 be two equal sides of a potentiometer, and as long as RB1 and RA2 are constant resistors of equal value I'll get the characteristics I'm looking for. Pic related. Can anyone confirm? Is this an already existing circuit?

[Anonymous No.1673027]

>>1673025
Oh I'll need a small resistor between the output of that op-amp and the diodes or otherwise I'll be driving the diodes too hard and end up sinking a bunch of current.
Like this image.
I'll likely make those small resistors as comparatively small as possible in order to get a low minimum gain. With ±6V 1k will be fine value, but I could go as low as 150Ω, allowing me to use lower value potentiometers. Not that I'd want to use low value potentiometers when they'd just serve to decrease my input impedance. 100k or 1M should be good. Be nice if there were multiturn dual-gang pots.

Anyone know what other components I can shove in there apart from diodes? Emitter followers? Linear regulators? JFET current diodes? Optocouplers? I want to make my footprints as diverse as possible as I'll just be adding female pin headers instead of components for prototyping purposes, so knowing beforehand what pads to include would help.
It will likely just be a crappy perfboard mockup, though I might take this opportunity to give it a shot on a CNC router.

[Anonymous No.1673028]

>>1672885
>did you?
Standard practice. Found a supply range of 2.5V .. 5.5V for the LMV321.

[Anonymous No.1673046]

I'm trying to make a small mobile DC power source with which I could regulate output from 0 mV to 500 mV. Any ideas where to start? I assume there a kits for this but I have trouble finding something concrete. I'm thinking about adding an LCD display but then I'd need a separate battery for that as well. Any advice?

[Anonymous No.1673055]

>>1673046
>0 mV to 500 mV .. where to start?
Maybe at defining the output current?

[Anonymous No.1673057]

>>1673046
>500mV
That's a really low voltage, usually power supplies have some sort of diode drop in there than prevents them from getting that low, or at least not accurately.
One way I could guarantee you getting there (pic related) would be by starting off with a ~5V power brick, getting one of those 5V-5V isolated power modules and using it (or something similar) to make yourself a negative rail, this will serve to power the op-amp without relying on the sometimes unreliable behaviour of rail-to-rail op-amps. Though you could try with rail to rail op-amps instead. Stabilise the negative and positive rails with LDOs and caps. Then just build dual feedback loops combined with diodes (make sure your impedances around that part of the circuit are all damn high) with one based around a current shunt resistor (10 to 100mΩ) and the other based around the output voltage itself, both feeding into a PMOS FET at the upper rail. An NMOS may work too. The use of a FET means there's no diode drop to worry about, though it's worth trying with a BJT if you buffer the base current. You'd use a pair of potentiometers to set the reference values that the current and voltage measurements are compared to. Can probably do it just fine with a dual op-amp IC without any extra negative feedback (i.e. potentiometer with extra resistors above and below it goes into the inverting input, measurement value goes into the non-inverting input, and the output goes to the diode). Op-amp buffered ideal diode circuits could be used if you're feeling fancy, doing this would remove the slight nonlinearity at the border between CC and CV modes.
if you want a proper circuit I can draw that out for you

[Anonymous No.1673058]

>>1673057
Ok shit it also needs something to get the op-amps' rails above the source of the PFET (or the drain of the NFET I guess) meaning you need a 3-way supply or something equally stupid, so it's really not that good a circuit at all. Unless you happen to have a bunch of those + and - boost converters or isolated DC modules lying about. Still may work with rail to rails.

[Anonymous No.1673060]

>>1673057
Proper circuit is always welcome because diagram >> prose.

[Anonymous No.1673063]

>>1673060
if you say so

[Anonymous No.1673071]

I have an lcd monitor that worked flawlessly for 6 years and today out of nowhere it wont turn on anymore, worked perfectly just last night. Not even the button light.
Any ideas what it could be?

[Anonymous No.1673075]

>>1673071
check the capacitators

[Anonymous No.1673083]

>>1673057
>>1673063

Thank you very much for the information.

For context: I work on power cabinets that have sensor inputs but I only have block diagrams without any values. Alarm state on the cabinet starts when there's either 220 mV or 400 mV on sensor inputs. I manually test if alarm state works by discharging a battery to either value and then connecting it for a short time on inputs where sensors usually go. If the cabinet is ok it will detect the voltage and go into alarm. This is a very primitive way of doing it and I wanna construct something to make this much more smoother and stable.

[Anonymous No.1673086]

>>1672707
>How about actually copying a circuit diagram from google instead of this “short to ground” style switch.
I am teaching myself electronics. Some of that is through google and book-learning, and some is through invention/problem solving---like a homework assignment. I'm not sorry I invented a shitty AND gate, because I had fun doing it, and it has set the stage for learning about better AND gates.

Everyone knows that.

[Anonymous No.1673099]

>>1673083
If you just need to output 0.22V or 0.4V we could likely do something simpler, just a voltage divider and possibly some sort of buffer. And a switch to toggle between the two voltages and 0V. Maybe with simple resistive current limiting.

>>1673086
Well you did sound like you looked at google specifically before drawing that diagram (>>1672516) so I was more criticising your ability to use a search engine as opposed to your problem solving ability. I also happened to give you advice on using a common-emitter configuration and ensuring that you have base and collector resistors twice. Resistors are necessary in a circuit diagram so don't leave them out, otherwise a mostly OK circuit could appear to be just shorting something to ground.
I'd learn the basic electronic symbols and how to use them first of all. I imagine you have an understanding of ohm's law? What about Kirchhoff's laws? Next on the list is getting yourself a basic understanding of input and output impedances, the proper use of which (R_input >> R_output) exclude a lot of potential transistor logic gate designs; each logic gate not only needs to perform its logical task but buffer it's output sufficiently to feed any number of cascaded logic gates afterwards. Naturally knowing how to use a transistor in the linear region and how to use it in the switching region are also pretty important in this particular case.

Electronics is a matter of taking lots of tiny sub-circuits and assembling them such that each performs as one circuit, and then assembling those into larger circuits and so on. I wouldn't put too much value on trying to get something to work at such a high "zoom" (i.e. individual transistor logic) when there's so many levels you can zoom out to (individual logic gates, flipflops, registers, memory arrays, PLAs and further integrations, etc. all the way up to whole CPUs or even the entire internet). Though understanding how transistors themselves work is pretty important. Good luck!

[Anonymous No.1673105]

>>1673071
if your lcd has an external power brick then it's easy to stick a $5 multimeter on the end and see if there's any voltage.

[Anonymous No.1673108]

>>1673083

i once had a need for something similar to test voltage meters, so i just used a 6V battery and wired a 10-turn pot across it to give me a variable 0-6V at very little current. i carried a voltmeter for calibration of the pot. if you wanna avoid carrying a meter around, you can use, say, a 5V power bank, then use an approx 10-1 voltage divider, using a 1K resistor and 100-ohm single-turn pot. put markings on the pot every 100mV (and at any other voltages of interest).

[Anonymous No.1673112]

>>1673083
That's certainly portable. Did you ever measure the input resistance or current of your sensor inputs? Maybe you can replace your carefully discharged batteries with a new AAA cell plus a clever voltage divider?

[Anonymous No.1673114]

>>1673028
>2.5V .. 5.5V for the LMV321.
why is that a problem for you? For me it's perfect.

[Anonymous No.1673131]

>>1673099
>ensuring that you have base and collector resistors
Is it OK if these resistors are very small? How small can they be? I think I see why it's good practice always to do this, but it seems like in a circuit with a lot of transistors these resistors would start eating up all the voltage.

[Anonymous No.1673148]

>>1673131
the resistors are intended to "eat up the voltage". V=IR applies here. since trannies have current limits and very low input resistance, you need external resistors to limit current.

[Anonymous No.1673158]

Got a laptop that has a detachable touchscreen.
When they're together and when the screen is above ~80 degrees open the connection between the keyboard and screen fails.
Should I simply add some more metal (by soldering) to the pins to decrease the gap between them?

[Anonymous No.1673200]

I live close to the arctic circle in Scandinavia.
Winter is coming, and the days will be extremely dark.
I'd like to emulate daylight inside my apartment.
Looking at doing something like this:
>Re-painting the ceiling to increase its reflectivity
>Indirect lighting with COB LED arrays
>Color temperature about 5500K
>Time-dependent dimming controlled by rPi/Arduino

Should I go with some random pre-built array or make my own? It's probably a matter of hundreds of individual diodes.
Seeing what's available online makes me lean towards buying pre-built...
https://www.alibaba.com/showroom/ac-cob-led-module.html

Since I'll run them with mains, should I go with standard DC LEDs or "AC LED" chips?
I first thought I'll simply adjust light output, but I guess I'd like to adjust color temperature too since warm-white is more fitting in the mornings/evenings.
Maybe a small part of the diodes can be warm-white ~2000K, and they're only active for a few hours per day?

Seems like "AC LED" chips are a good option, but have only looked at DC COBs so far.
>A further benefit of the AC LED configuration is its intrinsic full-range dimmability, without resorting to a dimming circuit. One of the core features of AC LED approaches is the compatibility with phase-cut (triac) dimmers.
https://www.manufacturer.lighting/info/64/

>AC LED straight off AC system more energy-efficient than DC LEDs off high efficiency AC to DC SMPS.
>DC LEDs had about 20% higher Lumen, but used about 30% more power.
https://pdfs.semanticscholar.org/1c8d/a2bec27fa80586441f886eb0ae4af22ef64d.pdf

>The result shows that DC LED lights are more efficient and reliable than AC lights.
https://ieeexplore.ieee.org/document/7936068

[Anonymous No.1673210]

>>1673200
>emulating sunlight with LEDs
Are you trying to get the supposed vitamin-D benefits and so on, or just a quality of light that looks like a nice sunny day?

[Anonymous No.1673217]

Just bought a new soldering iron because the heating element of my chink 3 dollar one broke. This time I chose a quality German iron. However, it doesn't fucking work. The solder just doesn't stick to it and it doesn't even always make the solder melt. Why is this? I've had this happen in the past, but that's because the tip was absolutely knackered. This is a brand new iron and a brand new tip! wtf I'm mad

[Anonymous No.1673219]

>>1673200
please observe /ohm/ RULE 2.5. build vs. buy decisions are strongly influenced by skill level, especially when dealing with concentrated energy sources
>make my own
sure. get a bag of 1-3W CRI-rated LEDs, a bag of metal-core "star" pcbs, a tube of "heatsink plaster", and a strip of metal of your choice, and get to soldering. I've used this before with success. you could use COB light engines, but they may be more light and heat than you want in one place. constant-current drivers are easy to find at all price points, but can be a bit harder to hook up external control
>intrinsic full-range dimmability
iff you don't mind 100Hz flicker. for me it'd be worse than no light therapy at all
one very powerful advantage of the dc versions is that you don't need to worry about isolating mains-side from project-side which makes it easier to experiment. you might look into the instructables for seasonal light devices. here's one https://www.instructables.com/id/Blue-LED-light-box/ and an update https://www.instructables.com/id/Arduino-Controlled-Sunrise-Alarm-Clock/ if they don't fit your needs exactly, you might be able to adapt a couple of projects together

>>1673217
>cone point
>screwed into place
>"""quality"""
that's a woodburning tool with pretensions
best option: lösche das Spielzeug and buy a grownup temperature-regulated mini-station like a chink 936D clone on ali https://www.aliexpress.com/item/mein-sorgenbenis-XDD/32809597974.html
if you really cannot do that, get some tip tinner so you can get that tip wetted, which you should have anyway

[Anonymous No.1673225]

>>1673219
Ok thanks. I'll try the tip tinner for now. I don't do that much soldering and a bottom of the barrel chinkshit iron worked just fine for me, so I'm just surprised why I can't do anything with this one. Perhaps the temperature is too low?

[Anonymous No.1673228]

>>1673210
Maintain a healthy level of light input, to avoid SAD.
Vitamin D is generated in the skin with the help of UVB. There's nothing supposed about that.
That's not what I'm after, however.

>>1673219
>and get to soldering
Yeah, I'll probably do that as I'm looking to get a rather diffuse light intensity along the corners of the ceiling.

>100Hz flicker
That'll probably push me towards DC then. Thanks for the tips.

[Anonymous No.1673235]

>>1673219
>more light and heat than you want in one place

he should just buy LED strips then

[Anonymous No.1673236]

>>1673225
it melts SAC305 at the point of the tip, temperature is at least hot enough, maybe too much. consider the following https://www.youtube.com/watch?v=2VlvjdaMARM

>>1673228
idea: if you only need diffuse light, and not a huge amount of it, there is LED ribbon sold by the meter, similar to the RGB color LED ribbon. it uses 12V constant-voltage just like the RGB version, but with cool-white and warm-white LEDs instead. no need to hassle with constant current drivers, any 12V regulated power supply will do. it's easy to add PWM with a couple of n-MOSFETs. it's what I'm using in my lab atm, works pretty well

[Anonymous No.1673241]

>>1673236
I can't re-tin it because the tip acts almost as a repellent to solder. It just won't stay on it. Like water on wax.

[Anonymous No.1673245]

>>1673148
>you need external resistors to limit current
Well yes, but in another post you suggested 100kOhm on one end and 10kOhm on the other. Why so high? Are you working with high voltages? My source is 3.3V, I don't think my LEDs would light up with that much extra resistance, especially if there were a stack of multiple such transistor-resistor groups.

[Anonymous No.1673253]

>>1673241
that's because you've oxidized it through heat. the tip tinning cakes have strong wetting agents to bust through the oxide. if you don't have one of those, try this:
while cold, lightly scrub tip with brass or copper wool (or very lightly with steel, if that's all you have), applying a thick coat of paste flux to the first 5-10mm of tip, wrapping wire solder around the first 5-10mm of tip, and heating up over a non-combustible surface. most of the solder will run off. you should be left with a blob of solder on the tip. between soldering, always leave a surface of tin on the tip, more is better than too much. you can wipe the tip on metal wool or a damp sponge just before working a joint, and feed a bit of extra solder on the tip before putting back in the holder

[Anonymous No.1673266]

Does anyone know any good sensors to pick up infrasound (<20 Hz)?

[Anonymous No.1673269]

>>1673266
BMP280 at max 157Hz sampling rate might be sensitive enough

[Anonymous No.1673276]

Can I use the 16MHz oscillator on my arduino to build a radio receiver at about 16MHz, or is this a bad/dumb idea? Skill level zero.

[Anonymous No.1673281]

>>1673269
cheers

[Anonymous No.1673282]

>>1673276
arduino is an anagram for un-radio. terrible idea, entirely different circuitry
the best way to learn radio is by example. hang around with the ham fags, understand the concepts of carriers and modulation, how LC resonant circuits work, how a diode detector works, look into some simple receiver and transmitter designs. there's probably a beginner's book or page about all that, certainly ARRL or your own country's ham radio org has at least one

[Anonymous No.1673285]

>>1673253
Thank you for your tips, will try.

[Anonymous No.1673326]

>>1673253
>>1673285
Some more tips for you:
The rough side of a scotch-brite sponge can also be used to clean oxidation off the tip. Make it damp, rub lightly, and use it sparingly because it will ruin the plating if you do it too much, just like steel wool will.
You can also scrape off stubborn crap with a knife blade, very lightly.
But it's best to keep the oxidation off in the first place. Having the iron too hot will oxidize the tip faster, so will questionable cheap solder. Wipe the tip on a damp sponge or brass wool before putting it back in the holder, re-tin it frequently, and don't let it sit running for a long time untouched.
Keeping the tip clean will also get you much better results when soldering too. When in doubt, wipe it. Before putting it back, after picking it up, every few joints and whenever you can see crap on it.

[Anonymous No.1673329]

>>1673245
If you're trying to power a load with your logic gate, either use a 1kΩ pullup/collector resistor and a 10kΩ base resistor, or make a totem-pole output, or use a seperate transistor to drive the LED with an appropriately sized resistor (200Ω is usually safe).

[Anonymous No.1673349]

>>1673329
> either use a 1kΩ pullup/collector resistor and a 10kΩ base resistor, or make a totem-pole output, or use a seperate transistor to drive the LED
Makes sense. Thanks

[Anonymous No.1673350]

>>1673326
wiping just exposes more of the solder and tip to oxidation, and reduces tip life through thermal shock. the slag don't hurt anything as long as you're not trying to heat something up. wipe before making joints
also never underestimate the power of Pic related

[Anonymous No.1673359]

>>1673350
You can buy copper wool for cleaning? I bet it's cheaper than the stuff designed for soldering.

[Anonymous No.1673372]

>>1673359
>this anon has never been down the cleaning supplies aisle
the ones intended for iron cleaning are yellow brass and will probably last quite a bit longer than pure copper. either way, just need to take it outside and shake the solder droplets out of it every few months or so

[Anonymous No.1673391]

>>1673359
>You can buy copper wool for cleaning?
Yes. You can also buy stainless steel scouring pads.
they're two for $1 at Dollar Tree

[Anonymous No.1673405]

These driverless COBs have a bit of an issue with flickering since they have no filtering. Soldering into after the bridge rectifier would be a bit of a pain, but I already have another bridge rectifier in my control circuit, so I figure I can just add a beefy filter cap after that (100µF) and feed that filtered DC into the bridge rectifier on the driverless COB itself. My only fears being that the higher RMS voltage could cause the COB to pop, but from what I've seen they have their own constant-current drivers in them. So if I had a normal driverless COB that would be the end of it.

But I have an RGB driverless COB (pic related), and am bypassing the strobe control IC with my own PWM signals. So it's possible that there's a temperature or even current sensor elsewhere on the board that was feeding information to the strobe IC, as opposed to that being done for each color channel on the switching ICs. I'll see if I can see any IC numbers/IDs today, but I'm not terribly hopeful.

[Anonymous No.1673530]

Kind of strange question. Im trying to make a desk that I can drop a PC into then easily lift out, and my power supply location is a little restrictive. Is there a sort of connector that doesnt need pinpoint accuracy to connect and also wont kill someone dead if they start poking around in the cavity when the desk is plugged in?

[Anonymous No.1673531]

>>1673530
What's wrong with a regular IEC C13 connector?

[Anonymous No.1673544]

>>1673531
Honestly it might work, but Im just concerned with how easy it will be to put a decent weight of PC and enclosure in there without bending any connectors and getting a good tight connection. The ones on the back of my current PC and monitors need a lot of wiggling around to actually work, but they might just be shit quality. Ill do some testing to see if the heft of the thing is enough to seat it well enough. The IO on the desk surface connects with just copper pads and a few springs, and that worked pretty good with the little 3D printed test, but thats definitely not safe for wall power.

[Anonymous No.1673565]

>>1673544
>and a few springs
Buy yourself a few dozen pogo pins from ali.
though either way a high-speed USB 3.0 or video port might run into some issues.

Sounds like you want a tapered connection of some sort that guides itself in to contact, a heavy (~3mm) chamfer on an IEC C13 and/or C14 could do this for you, if you don't mind deviating from the IEC standard and possibly interfering with the strength of the socket. 4mm banana plugs (the insulated kind) might be able to work if you can get them at the rated current. I'd have a GFCI in your circuit anyhow in case water or something spills down where it shouldn't.
Making an isolation transformer split into two halves (one in the computer and one in the base) could also work if you're not worried about weight or size or efficiency but care most about isolation and safety.

[Anonymous No.1673579]

>>1673544
Anderson PowerPole. usually used for dc but your secret's safe with me, anon
Hubbell also has some other connector ranges which might be of interest to you

>>1673565
don't do this anon. they're sold cheaper by the hundred
>tfw I paid $4 for this fun sized bag

[MrCrummy No.1673580]

>>1672532
>>1672661
Fixed

[Anonymous No.1673583]

Please forgive, I have no clue what I'm doing.

Trying to attach a fan to raspberry pi that can be turned on-off via script. No pulse modulation or anything fancy.. Simply on-off with a pi script.

Here's the diagram of how I think I can do that with a PNP transistor. Please show me where my diagram is wrong.


The listed resistor is just what I have on-hand. What's a more fitting resistor? Should note there are resistors on the daughter board that look like it's leading to the fan's positive lead labeled with an 0.

[Anonymous No.1673594]

>>1673580
Fuck yeah! What was the problem?

>>1673583
Add a diode in antiparallel with the fan (can be known as a freewheel or snubber diode). Your diode can probably be as high as 1.5kΩ without causing too many problems. But I've no clue what that board there is actually doing or what its + and - inputs are.
Pic related is basically guaranteed to work (perhaps with an added noise suppression capacitor).

[Anonymous No.1673608]

>>1673583
That's not a circuit diagram, use common symbols. 8050 is NPN, 8550 is PNP. Which is it? Your circuit doesn't make sense because the transistor shorts the supply voltage (or whatever 'fan controls' means).
>>1673594
>diode can probably be as high as 1.5kΩ
what

[Anonymous No.1673617]

>>1673608
>what
I meant the base resistor

[Anonymous No.1673620]

>>1673583
>no clue what I'm doing
Here are your options.

[Anonymous No.1673635]

>>1673608
I dont know, I think its a PNP. This is what I have a baggie of:
https://www.ebay.com/itm/50-Pcs-Complementary-S8050-30V-700mA-PNP-Silicon-Transistor-S8550-N6Q2/112619659545

Sorry for calling it a "diagram".

Sorry, not fan control, just "fan" as it's the PCB inside the case that connects a header to the GPIO and puts power through it. The PCB has terminals that you can plug a fan into (but the fan cannot be controlled without a transister of course).

So I don't know where ti place the transistor then. The orange wire leads to a GPIO data pin where it can be controlled, while the red connects to, I assume, 5v power supply. I could connect the fan directly to different pins if I wanted, including a 3.3v pin.

[Anonymous No.1673644]

>>1673620
>>1673594
According to both of these I need to buy some diodes.

The board connects to the GPIO and handles power, and on-off switch and allows you to connect a fan directly to it and its positive supplies 5v power to the fan. You can connect the fan directly to it with its red and black wires but it doesn't have a way to control the fan.

[Anonymous No.1673649]

>>1673635
I'm guessing S8050 was a typo, and that they meant to say S8550, since they say PNP in the title and have S8550 on the picture of the transistor itself. For a PNP see the right-most option: >>1673620 and remember that you'll need to drive the output to 0V in order to turn the thing on. I'd also buy some standard NPNs while you're getting diodes, some BC547s or 2N2222s are pretty handy and usually used more often than PNPs. 1N400X series diodes will suit you fine here, 1N4004s are general purpose enough for small signal and mains voltage use.

>>1673644
Then yes, use that right-most circuit with the daughter-board's + connected to the diagram's +, and the daughter-board's - connected to the diagram's __. An NPN would be slightly preferred here as if for whatever reason your daughter board's + output is disconnected, the control signal won't feed into a floating rail. 0V rails are left connected much more often than Vcc rails.

[Anonymous No.1673660]

>>1673635
>50-Pcs
Congrats, good transistors. You can use either, NPN (8050) is for ative H, PNP (8550) is active L. Look at the options diagram posted above. In the NPN case the emitter is connected to GND, the (+) fan wire goes to +5V and the (-) fan wire is connected to the collector. To fully switch the transistor ON (called saturation) you will need a base current of about 1/10 of the collector current (160mA for your fan), so about 16mA. The base voltage at that current is about 0.8V and the collector voltage would be about 65mV (see pic, red line). How to calculate the base resistor? Rb=(Vs-Vb)/Ib=(5V-0.8V)/16mA=263Ω. Your 300Ω resistor is close enough. BTW: you can use a transistor as a diode. Can you imagine how that is done?

[Anonymous No.1673662]

>>1673660
>use a transistor as a diode
Oho, I didn't think of that, nice save.

[Anonymous No.1673665]

>>1673660
I'm admittedly slow. If I connect two of the legs base and emitter, would it work like a diode? probably outting myself as an idiot

[Anonymous No.1673666]

I want to make a simple circuit that flashes different LEDs, slowly enough that it's perceptible to the naked eye. What is the simplest component I can use to get the required time delay?

[Anonymous No.1673673]

>>1673665
>If I connect two of the legs base and emitter, would it work like a diode
Yes, and in the direction of the arrow. You could also use the collector-base junction as a diode too; because a transistor is made of P, N, P semiconductors, it's basically two back to back PN junctions, each of which can act like a diode.

>>1673666
555 timer, comparator, maybe a relay, reverse bias transistor oscillator, there are a few options, but none of them just use a single component; all use resistors and a capacitor or two to set the timing. For the 555 timer method, look up "555 timer astable multivibrator", for the comparator method, look up "comparator relaxation oscillator", for the relay method look up "relay relaxation oscillator" (note this one may waste significantly more power). The reverse-bias transistor oscillator has a pretty low part count, but is pretty shitty in general and depends highly on what transistor you're using (I could only get it to work with a 2N2222, not with a BC548).
There are also LC oscillators like an emitter-coupled oscillator, but you don't exactly need a perfect sine so you should be fine with the ubiquitous 555 circuit.

[Anonymous No.1673674]

>>1673666
oh and for time delays atop your oscillator, monostable circuits using comparators or 555s are pretty simple
ben eater's video on the uses of a 555 timer is surprisingly interesting

[Anonymous No.1673677]

>>1673673
>>1673674
Thanks anon, I'll give it a whirl. I actually like that "shitty" design, it's already taught me something about (unintended?) properties of transistors and their use

[Anonymous No.1673682]

I found someone else's implementation, and they dont seem to be using any diodes or resistors at all. Are they just not needed for this?

Also, thanks to the anon's for their patience and help I get to learn from.

[Anonymous No.1673686]

>>1673682
>Are they just not needed for this?
They are needed, you can blow your transistor or maybe even raspi without the diode
look up what a freewheel diode actually does, it should be a worthwhile read

[Anonymous No.1673687]

>>1673682
if this doesn't fry the pi, it's purely because the onboard capacitance is enough to absorb the motor flyback

[Anonymous No.1673688]

>>1673665
It's actually a Z-diode but the reverse base-emitter voltage is always higher than 5V, so it can be used. You could use the base-emitter junction alone and leave the collector open or use the base-collector junction which has no such avalanche property, but its max current is rarely specified. In case of the 160mA fan this may be critical and the max base-emitter current also isn't always documented. The circuit in the pic is for 5V but can carry the current specified for the transistor itself, at least 0.7A in case of the 8050/8550. This because the transistor goes into active mode.
>>1673682
The diode is merely a precaution to protect the switching transistor at switch-off. It may not be needed at all with a fan. The base resistor I would not omit.

[Anonymous No.1673689]

>>1673688
the missing pic

[Anonymous No.1673693]

>>1673686
>>1673687
>>1673688
Correct me if I'm wrong, but brushless DC motor fans almost always have flyback diodes across the windings which are switched electronically with transistors in the motor (contrast to mechanical commutators in brushed motors).

I think there shouldn't be any need for a flyback diode, unless there is some significant series inductance eg a EMI filter inside the motor. Still, it doesn't hurt to add one just incase.

[Anonymous No.1673694]

>>1673693
>brushless DC motor fans
Oh of course, I was assuming they'd be dumb brushed fans but the image in >>1673583 definitely looks like a PC-style fan so you'd be right.

[Anonymous No.1673713]

>>1673693
>>1673694
>trusting chinkshit to be to spec

[Anonymous No.1673721]

>>1673565
>>1673579
Thanks, thats some pretty good research material.

[Anonymous No.1673802]

>>1673666
I would use a CD4093 oscillator + transistor driver. In fact I made a miniature lighthouse lamp flasher with a single 4093 and a BD136 driver transistor for the lamp.

[Anonymous No.1673814]

>>1672686
>MAX7219
>interface microprocessors (µPs) to 7-segment numeric LED displays of up to 8 digits, bar-graph displays, or 64 individual LEDs.
This is a genius IC. Thanks for sharing. It uses multiplexing to drive 64 LEDs.

Is there anything similar in TI's lineup? So far I've found only TLC59284 which has 16 channels. It's basically two 74HC595 in series with a constant current function and more expensive than MAX7219.

[Anonymous No.1673828]

Total noob here, but I think I can get into this.
As insignificant as it is it sure felt nice the first time I got an led to light up without popping.

Looking for my first bench power supply, anything I should look for in particular? Good brands?
When I looked it up people recommended used keysights and stuff but I dont wanna be 'that guy' who's more interested in buying equipment for his hobby than actually practicing it.

[Anonymous No.1673842]

Is this a U.FL connector?

[Anonymous No.1673891]

>>1673814
>TI
y tho, the mass market is a bit downscale from their tastes and Holtek and Macroblock and others have given designers what they want when they want a front panel on a separate board
however, it is useful to know about TPIC6x595, which is 'HC595 logic with open-drain output drivers. they can be useful as low-side switches for loads like relays, solenoids, split-flap drivers, small incandescent lamps, etc. that draw too much current for the 'HC595 or are powered from a +12 or +24 rail

>>1673842
maybe. measure it

>>1673828
a laptop power brick and a couple or three chinkshit step-down boards are probably enough investment for now. you can get the boards with a single trim pot (for voltage limiting only) or two trim pots (for voltage and current limiting). not sure why you would need the ones with three pots unless you need a battery charger. you can buy the bench-ready power supply system with all the binding posts and knobs when you can demonstrate to yourself a need for better, or you can build those little boards into a bench box of your own arrangement when you know your components better and feel that you're up to it

[Anonymous No.1673895]

>>1673891
>measure it
2mm. Seems so.

[Anonymous No.1673903]

I'm curious, if it is possible to turn proprietary HP printer into free and open sourse?
I realized they sell new printers with same 2 ml cartridges, cheaper than those 2 ml cartridges. So I ended up just buying a new printer every time previous stops printing and I sell them for cheap.
As for mechanics, printer is really simple, I found only two motors. As for print heads, they are bubble jet (resistive heater that evaportate ink)...
Basically I want to remove HP electronics and replace with my shitty one.

[Anonymous No.1673913]

>>1673814
>the only reason I have not added a meter to my bench supply is because I hate programming 7 segment shit
feels bad man

[Anonymous No.1673927]

>>1673903
Get a cheap logic analyser and probe around the control circuit to see what format the data is in, and reverse engineer the control circuits to help build up a full picture
printers need drivers anyhow though, so nothing you make would ever be truly plug-and-play

[Anonymous No.1673929]

>>1673927
>Get a cheap logic analyser
I have been considering getting USB oscilloscope and logic analyser for a long time now.
> see what format the data is in
I think it would be something proprietary.
>and reverse engineer the control circuits to help build up a full picture
Reverse engineering mechanical part would be rather easy, since it uses optical encoders and brushed motors. They are easy to control.
But printing head is kinda difficult. As far as I know, HP uses resistive heaters in head to create bubbles and splash in on paper. But how it is done - I don't know.

[Anonymous No.1673933]

>>1673929
http://spritesmods.com/?art=magicbrush&page=3
Hm... Ok. So it is technically possible to BTFO HP at their own game.

[Anonymous No.1673977]

>bit banging usb with discrete logic, shift registers, and eeproms
possible? with less than 12 ICs?

[Anonymous No.1673983]

unity gain anon (>>1673027) again, just realised I could convert the op-amps into an integrator followed by a differentiator (or vice versa) and get some interesting behaviour by placing clipping diodes in between those. Though at that point I'd need two double-gang potentiometers and a toggle to switch each set to capacitors instead. Or somehow just a single double-gang pot and some interesting switching in order to have it in place of either of the sets of resistors, but that sounds like a lot of work.
Actually just switching the order of the op-amp circuits would work just fine.

[Anonymous No.1674007]

Is motor pwm simply lowering or raising voltage? As in can i use a buck as a pwm for my cnc spindle

[Anonymous No.1674010]

>>1674007
>Is motor pwm simply lowering or raising voltage
No, but since a motor's inductance will resist a change in current it might as well be (assuming you have a freewheel diode). In general, PWM is turning your load on and off thousands of times every second, and varying the ratio of on-time to off-time to vary power. Naturally these cannot result in a higher average output voltage than you put in. For some loads (LEDs, electrochemical systems, etc.) the high-frequency PWM is noticeable as a flickering if you have a fast enough sensor, and trying to represent a lower voltage with a higher voltage being PWM'd can sometimes even cause breakdown or other issues. But as I said, a motor is an inductive load, so the current going through it will probably just ripple a little but stay mostly constant. So long as your PWM frequency is higher than your motor's RL time constant, I'm guessing.

Actually varying a DC voltage is more of a pain, since it requires some form of feedback to work with any arbitrary load. Hence variable buck converter modules will typically have a lesser maximum current rating compared to similarly priced motor PWM modules. But if you already have such a buck converter that's rated at a valid input voltage and output current, then it will be a simple matter to just hook it up to your motor itself, if anything your output speed will be a bit smoother, as the bit of ripple I mentioned with regards to the PWM circuit would be filtered out.

but make sure you use a freewheel diode either way (>>1673620 except your buck converter replaces the transistor, likely the one on the left).

[Anonymous No.1674011]

>>1673929
>HP uses resistive heaters in head to create bubbles and splash in on paper. But how it is done

they use tiny SMD resistors which they hit with some 30V or so. the resistors get hot very quickly coz they're small. the heat causes a spurt of ink to be ejected from a nearby nozzle. this is just a property of the ink: when you heat it, it jumps.

[Anonymous No.1674016]

>>1674010
Thanks, what i want to exactly do is this
I want to buy this chinks motor which comes with power source and a knob to regulate speed
But i don't want to buy the power brick to save money and instead use old computer PSU (which can pump out like 15A so no problem)
And then change the motor speed by switching its connection between the 3V 5V and 12V branches, plus extra booster (not buck oops) that is hard set to 12V -> 48V in case i GOTTA GO FAST
Bust mostly in like 90% of cases i will be using the 12V rail because i don't want to go deaf with the loud 500W shitter

[Anonymous No.1674021]

>>1674016

[Anonymous No.1674024]

>>1673929
>HP uses resistive heaters in head to create bubbles and splash in on paper. But how it is done

they use tiny SMD resistors which they hit with some 30V or so. the resistors get hot very quickly coz they're small. the heat causes a spurt of ink to be ejected from a nearby nozzle. this is just a property of the ink: when you heat it, it jumps (just like the wife).


>>1674010
>if you already have such a buck converter that's rated at a valid input voltage and output current, then it will be a simple matter to just hook it up to your motor itself

except the motor is gonna get very weak as you decrease the voltage. the big advantage of PWM is that you dont lose torque.

[Anonymous No.1674091]

I'm back with another LED question.
If I drive a LED 5 mA above its rating will I kill it early?

[Anonymous No.1674104]

>>1674091
express in percent

[Anonymous No.1674118]

>>1674091
yes though nobody can give you numbers as there's no good public data that i know of. life decreases exponentially with temperature, and higher overcurrents will drive higher temperatures.

[Anonymous No.1674119]

>>1674118
Guess I'll stick heatsinks on it

[Anonymous No.1674124]

I bought a ten pack of MG995 servos years ago and the last one has finally died. Time to get new servos. What do you guys use in this size? Ideally it'd be digital and use metal gears (the MG995's only redeeming feature seems to have been the gearbox), without being extremely expensive.

[Anonymous No.1674126]

>>1674024
>they use tiny SMD resistors which they hit with some 30V or so. the resistors get hot very quickly coz they're small. the heat causes a spurt of ink to be ejected from a nearby nozzle. this is just a property of the ink: when you heat it, it jumps (just like the wife).
Those are not SMD resistors, but silicon ones. >>1673933

[Anonymous No.1674134]

>>1674024
>the big advantage of PWM is that you dont lose torque.

[Anonymous No.1674177]

Today is the day.

[Anonymous No.1674180]

Is it a bad idea to run 12V lads straight off a 12V battery? Power should be fairly consistent from a car battery, right? Should I put a regulator between?

[Anonymous No.1674188]

>>1674180
if you're talking about those generic 12V led panel indicators or equivalent, they should have a current limiting resistor built in and probably won't mind the ~14.7v that battery could give them.

[Anonymous No.1674207]

What's the minimum amount of transistors I need to make a simple cpu?

[Anonymous No.1674216]

>>1674207
about a fuckton
depends on your gender pronoun

[Anonymous No.1674228]

>>1674216
this is the correct answer

[Anonymous No.1674239]

>>1674207
about 2300

[Anonymous No.1674243]

>>1674207
The Intel 4004 (first commercial microprocessor, 4-bit) had 2300 transistors. I don't doubt that you could build something minimally-functional with fewer transistors. It all depends upon whether you want something that's purely academic or something practical. For the former, you could probably get away with a couple of 1-bit registers along with an ALU capable of implementing AND, OR and NOT on 1-bit operands; everything else would be subroutines. The number of transistors would be dictated largely by the indexing registers needed to access RAM.

[Anonymous No.1674246]

>>1674243
>>1674239
I just want to be able to do some really basic instructions: mov, add, sub, or, xor, and, cmp, jmp, jne, je.

[Anonymous No.1674248]

>>1674246
then don't make it out of fucking transistors, use D-FFs and such instead

[Anonymous No.1674249]

>>1674248
but I wanna do it with transistors

[Anonymous No.1674250]

>>1674249
yeah and I want to build a high-tech solar sea-yacht, ain't happening any time soon though

[Anonymous No.1674251]

>>1674249
do it in an FPGA devboard first. it'll be more fun and just as masochistic, but you'll be able to debug it.

[Anonymous No.1674253]

>>1674246
https://www.youtube.com/watch?v=k1hJoalcK68

[Anonymous No.1674254]

>>1674253
i wonder how much he paid for that big board

[Anonymous No.1674255]

>>1674250
i think you underestimate my autism

[Anonymous No.1674256]

>>1674254
I get the impression from the video that he's selling these as kits.

Just bear in mind that only the processor on this thing is made with relays. All the memory and such is ICs.

[Anonymous No.1674257]

>>1674255
Ok, how about using transistor array packages? The common emitter ones so you can save on a few pins and space. You'll still need like $50 in breadboards alone though.

[Anonymous No.1674265]

this parts list S8050, but would other NPN transistors work? Cause I have 2n2222 and BC547B both on-hand and dunno if I should order the other.

[Anonymous No.1674275]

>>1674255
lern2logic first, then lern2processor, then worry about your physical implementation later. I'm not going to tell you that you absolutely can't design a processor at the discrete component level, not least because you or someone else might be just autistic enough to prove me wrong, but I will tell you that you are a fucking moron to not avail yourself of the many simplifying abstractions and layers that have been developed to assist reasoning AND COMMUNICATION about your design and that masturbation is a private activity

[Anonymous No.1674277]

>>1674275
sorry too busy looking up how to make a cpu with trannies and know what? i think ill make the trannies myself too. all i got from your message was "masturbation"

[Anonymous No.1674279]

>>1674277
it's always fun to see an arrogant sperg come i here and ask a vague question only to get mad when told their half baked project is over their head

[Anonymous No.1674280]

>>1674275
>>1674279
just reread your message. again all i got is "masturbation"

[Anonymous No.1674285]

>>1674279
pretty sure it's the same anon and their same low-effort shitposting, predictably showing up whenever /ohm/ hits bump limit

[Anonymous No.1674294]

>>1674285
nah bro it's John from high school. you rember me right?

[Anonymous No.1674301]

>>1674239
why's that so pretty?

[Anonymous No.1674302]

>>1674294
I member...

[Anonymous No.1674304]

>>1674302
not that one the one one that would dress up as a qt anime dog and scream uwu down the hall ways

[Anonymous No.1674444]

Got my first basic Oscilloscope, Going to have some fun with it tomorrow.

[Anonymous No.1674458]

>>1674444
>he didn't get a free CRT dumpster scope
must suck being a channellet.

[Anonymous No.1674460]

>>1674458
>dumpster scope
I live in a state where shit that hits the dumpster is smashed with a hammer before it goes in.

[Anonymous No.1674487]

>>1674460
Massachusetts is the only one I know. that's worse than babies in incubators. the wanton, deliberate creation of useless e-waste. drone air cover for ground attack when?

[Anonymous No.1674572]

>>1674265
No useful information, post diagram. 8050/8550 are for higher currents than what you have. Compare graph to the one (8550) posted earlier. Same conditions, 500mV vs 65mV.

[Anonymous No.1674593]

>>1674444
Very good. Toy scope is better than no scope. You can run it from a 9V battery, self-contained like a multimeter.

[Anonymous No.1674749]

>>1674180
>12V lads
'avin a giggle m8?
Nah most 12v stuff will be OK up to 14v and down to 11v.
Pretty much everything that takes 12v input is just regulating it down to 5v.
If you want a constant 12v output from the battery, you need a buck-boost converter because even a low-dropout regulator will drop the voltage before the battery is actually close to empty.

[Anonymous No.1674753]

>>1674749
oh, also, protip: If you're in doubt about how some piece of equipment will handle voltage, take it apart and see if you can find the regulator near the power input. If you can't recognize the regulator just google the numbers of all the chips near the connector.
Even if you know very little about electronics you'll be able to learn what you can get away with just by reading the intro to the datasheet.

[Anonymous No.1674861]

>>1674460
>tfw brazilian
>touristic capital city filled with beggars and crack users
>anything with copper thrown away vanishes in seconds
>brazilians don't really have electronics as a hobby
kill me

[Anonymous No.1674862]

> want to buy AD9835 or some other DDS IC
to fuss around
>all of them are surface mount
>no way in hell I can scrap copper clad that thin or use that shit in perf board

[Anonymous No.1674869]

r8 my protoboard. small linear regulator circuit included for lulz. power jack not included because tfw no 3d model gf

[Anonymous No.1674870]

>>1674862
now you understand exactly why there are "for arduino" module boards
>no way in hell I can use that shit in perfboard
git gud https://www.youtube.com/watch?v=edERx4x5eY0

[Anonymous No.1674897]

>>1674869
i don't understand why you wouldn't just order a proper board with all your components routed out

[Anonymous No.1674901]

>>1674862
>look up DDS
>only just figured out how PLLs work
oh man I've got a long way to go

[Anonymous No.1674922]

>>1674870
Sorry but I saw a potentiometer board for arduino, literally just one of those big chunky pots.

>bc548
>max vce : 30v
>build a joule thief, get 70V
>ok lets make a pico second rise time pulse generator like my sempai Jim Willians said in App note 47
>transistor is doing fine with 70V.
Thanks obama.

[Anonymous No.1674923]

>>1674922

[Anonymous No.1674924]

>>1674922
spice doesn't model breakdowns

[Anonymous No.1674938]

>>1674861
>taken for copper
Yea, that's what I figured happens, I was going to see about dropping some money on random boxes or a pallet of E-waste, but fucker want like 1,000 USD for these things with barely anything on them.

[Anonymous No.1674940]

I need to power a 12V 60W LED with some very thin (24 AWG) wires. Looking at a chart and doing some quick math shows me that for 12V DC and 60W, I'll be running 5A through the wires. Looking at a chart, I can see that for 5A of power at 12V I'd need a 0.5mm^2 cable. My 24 AWG cables are 0.205mm^2,

Would this work if I use three cables in parallel? The reason I can't just go buy a thicker cable is because I need to run this power through a slip ring.

[Anonymous No.1674945]

Gentlemen, I just need to know, what's your gut feeling on how these doorknob caps will work to replace 5kV 2000 pF mica caps in a high-frequency spark gap generator application? I do not have current ratings on the doorknobs, but the mica caps were rated Capacitance: 2000 pF, Voltage: 5000, Current @3 MHz: 7.5A, @1 MHz: 6.2A, @0.3 MHz: 3.6A, @0.1 MHz: 1.8A.
Thanks.

[Anonymous No.1674968]

>>1674897
he was in a hurry and there isn't an industry standard for WLCSP footprints yet. it's really not all bad with a delicate touch and a decent station. I'm sure that he would have used the QFN if it were available and it would have taken him half as long to bodge it down

>>1674922
>pot board
and then there is the crap for fashionable suckers who can follow a recipe but can't cook

>>1674940
cross-sections add, not diameters. so, you're about halfway there
I'd step up the voltage enough to not exceed the current rating of a circuit (or two or three) on the slip ring but still pass equal power + some margin. on the rotary side I'd step that back down to 12V (if required) to get the amps back. I'm assuming that current carrying capacity of the slip ring does not drop very quickly as voltage rises

[Anonymous No.1674975]

>>1674940
are you trying to blind the driver with an LED in his steering wheel?

>>1674945
rtfd?

[Anonymous No.1674981]

testing 48V power supply with 40V diode. fuggg :DD

>>1674945
they'll be fine. worst case, they'll last until you can flog this beast to some greater fool

[Anonymous No.1674982]

>>1674975
sometimes air bags aren't fast enough and you need the pure photonic energy boost from a Light Bag™

[Anonymous No.1674994]

>>1674975
>rtfd?
I did find something after I posted and it just says 10 amps. Higher is better right? I bought 2 of them for $20 a piece.
>>1674981
>they'll last until you can flog this beast to some greater fool
That was the plan but i have $400 into it so far and it welds really nice. Unfortunately I am currently on a 20A breaker so I can only weld 1" or so at a time before it kicks off so I will need to decide between spending money on a new service or a more modern welder. Thanks for all your help guys.

[Anonymous No.1675017]

>>1674994
Nice work. I look forward to a photo of drawing the longest starting arc you can.

[Anonymous No.1675027]

>>1674940
>three cables in parallel?
Yes. 10A/mm^2 is a good rule of thumb and you have 0.6mm^2 with three AWG24 in parallel.

[Anonymous No.1675029]

>be me
>buy chink 48V 10A power supply
>set switch to 120V
>apply 120V
>measure output with no load
>hmm, only 40V
>twiddle the adjust pot
>hmm, still only 42V, maybe a little load will wake it up
>apply a couple of small loads
>down a couple hundred mV
>apply a 180mA (~6W) load
>hmm, only 35.4V now
>turn off load
>hmm, back to 42V
did I get chinked?

>>1674994
dam son, that's nice
hope that magnificent ox finds a good home with you or someone

>>1675027
intredasting math

[Anonymous No.1675045]

>>1675029
Those cheapo PSUs have a tendency to arrive with borderline failing parts, but you can find troubleshooting videos on youtube, then you just have to buy the part and replace it.

[Anonymous No.1675049]

I've got an old metal case that was used for an old-style soldering rig and solder sucker. I'd like to make new use of the case, but its current condition is meh--there is solder all over the bottom that is caked/stuck onto the case.
Is it foolish to just take a fluxed iron to the large puddle and slowly ball it all up to be discarded, so that I can properly clean the case, and maybe paint it?
I just need something to allow me to carry around my TS100 and its power supply (an old IBM psu), so I figured I'd repurpose it.

[Anonymous No.1675053]

>>1675049
Got a toaster oven?

[Anonymous No.1675054]

>>1675053
I was thinking about this. What would I do to get it out, though? Toaster/reflow oven, then what? Put a bunch of copper wire/wool in there and hope it wets up?

[Anonymous No.1675055]

>>1675029
>be a fart smeller
>dust off the old 1000W 240Vac step-up transformer
>plug 500W 48V 10.4A supply in to step-up xfmr
>power up
>holds 48.0V perfectly with a ~6W lamp load switched on or off
well dat's fugged. so the voltage doubler circuit must not be working at all

>>1675045
>borderline failing parts
indeed, the voltage across the doubler caps was equal to the full-wave peak voltage. the selector switch was actually bad, not just having a particularly rough detent. jumpered across, works great now

>>1675054
place it on a tray in the oven and out comes the muck at 187°C or so (if leaded)

[Anonymous No.1675058]

>>1675055
I guess it'll work if the case is inverted. Good enough point.

[Anonymous No.1675062]

>>1675049

why bother when you can pick up empty tool cases, or small tool boxes, for $3 at the thrift store.

[Anonymous No.1675092]

>>1674924
I know, I've built the thing, it is doing fine with 70v.

[Anonymous No.1675093]

Language question coming in:
I'm looking for a lcd display that I can fit into a watch, so it has to fit into a 32mm diameter circle and have a maximum depth of ~5mm. Also it needs to be low power (like a watch battery).

Here's my issue, I can't find many that would work as a watch display and that I can buy, does anyone know what terms I'm looking for?

I mainly searched for "7 segment 4 digit lcd display" or similar terms. Is this correct or is there another english term I'm unfamiliar with to look?

The best one I found so far seems to be the Lumex LCD-S401M16KR or LCD-S401M16TF

[Anonymous No.1675096]

>>1675093
If it's low power you want, you could consider using e-ink.
From what I've seen, aside from the bulky single or double digit LED seven-segment packages, seven-segment displays tend to be custom built for each use case, same for LCDs in general. You'd have a much easier time going for an LCD or OLED dot matrix. The nokia-style LCDs are pretty cheap. If you're using a microcontroller in there anyway it won't make much of a difference, especially if you're using one of those monochromatic LCDs with a lustrous back that means you don't normally need a backlight (like a calculator, not like a smartphone) which would make for better battery life than an OLED.
Though it's not impossible to DIY your own LCD (or EL) display if you have vacuum deposition or plasma deposition methods or whatever it is.

[Anonymous No.1675121]

>>1675096
Fist of all, thanks for the answer. I considered e-ink, but a 4 digit display seems rare and they cost a whole lot.

I wanted to not use a dotmatrix and instead use a segmented display because it would need less I/O pins and I wanted to keep the I/O pins to a minimum (I'd charlieplex it as well). Also I wanted to get away with some really low power (as in calculating power, not energy draw) and few pin microcontroller of the ATTINY family. I just want to get used to designing a few more pcbs by myself

[Anonymous No.1675130]

>>1675093
unfortunately there are no good search terms that i know of. aliexpress is your friend here but it's polluted with stupid modules and the occasional iphone repair kit no matter what you search. according to eevblog you can order a custom LCD for about $100 if you have cash to burn.

>>1675121
reflective lcds consume nearly zero power but your micro could still spend a lot of power refreshing the screen if you program it wrong. be sure that whatever micro you get will let you reduce the system clock to a few kHz when idle. or better yet get one with an integrated lcd driver.

[Anonymous No.1675145]

>>1675130
>unfortunately there are no good search terms that i know of.

Thanks for clarifying, guess I'll dig deeper
>according to eevblog you can order a custom LCD for about $100 if you have cash to burn.

I most definitely do not have cash to burn unfortunately.

>integrated lcd driver.

I'll consider it.

Thanks again Anon.

[Anonymous No.1675149]

>>1675017
Will do, gotta stress-test the new caps, make sure they blow up when I expect them to and not in the middle of something important. >>1675029
Thanks man.
I'll be back, thinking of getting another one of these if I can get it cheap enough and I also have the opportunity to buy an old 3 phase CV machine for $100 that I'd try converting to single phase. Running out of room though.

[Anonymous No.1675173]

I got a tip about a "great and cheap" German electronics firm, but I didn't get the name of it.
Could it be Reichelt? Mouser?

[Anonymous No.1675176]

>>1675173
reichelt.com exists, no free shipping though.

[Anonymous No.1675178]

>>1675173
>I got a tip about a "great and cheap" German electronics firm
hmm...
https://www.pollin.de/
is the trash king, besides that:
https://www.mikrocontroller.net/articles/Elektronikversender

[Anonymous No.1675182]

>>1675173
https://www.develektro.com/ can also be decent at times

[Anonymous No.1675202]

>>1675176
>>1675178
>>1675182
thanks guys.

[Anonymous No.1675227]

NEW

>>1675224

THREAD

[Anonymous No.1675238]

Which decent EU store has the lowest shipping?