/diy/ - Do It Yourself

if I want to design analog/digital synthesizers, filters, effects circuits, and all that fun jazz

would I be looking at specializing in Signal Processing for my EE degree?

>>1539340

never make /ohm/ again retard


>RULES
0. Electrics ≠ electronics. Mains wiring goes to /qtddtot/ or /sqt/. PC assembly to >>>/g/.
1. Do your own homework. Re-read all documentation/datasheets related to your components/circuits, and do an honest web search, before asking.
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.
3. Read posts fully. Solve more problems than you create.

>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)
Your local independent retail electronics distributors
ladyada.net/library/procure/hobbyist.html

>Related YouTube channels:
mjlorton
paceworldwide
jkgamm041
eevblog
EcProjects
greatscottlab
AfroTechMods
Photonvids
sdgelectronics
TheSignalPathBlog
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.

>>1539349

To be fair to OP, /ohm/ has one or more autistic spergloards who REEEEEEEEEEE if you make the new thread just because the old one has hit the bump limit.

>>1539349
>retard
>he can't even read

allow me to post it again, in triplicate
>no OP form post either because I can't find a recent version
>no OP form post either because I can't find a recent version
>no OP form post either because I can't find a recent version

there was no /ohm/ in the archive, the only ones I found on 3rd party archive sites were from years ago

so suck my nuts

>>1539355
>there was no /ohm/ in the archive,

well that's odd, because THAT'S WHERE I FOUND IT.

>>/diy/?task=page&page=3

why not post some more triplicate wisdom, newfag.

>he thinks the only archive is the lame 4chan one.

>>1539360
It is still in the archive
>>1532176
and close to the end the new one is announced
>>1537317

>>1539349
op fights are the dumbest thing about generals. newfags don't read the op.

What size wire is best for breadboards? Can I use Cat5 if it's solid?

>>1540912

I think AWG 24 or 22 is recommended.

I want to use a 339 comparator to detect voltage and send the state to an arduino. My understanding is that the output is compatible with an arduino GPIO but it always reads low. I've tried four different chips and either they are all bad or maybe I have a mistake in my code.

So, the question is, can a 339 drive an arduino pin or do I need some other comparator?

link to 339 data sheet:

http://www.ti.com/lit/ds/symlink/lm2901.pdf

>>1541086
it needs a pull up resistor to whatever logic level you are using.
inb4
>add a 10k resistor from the output to +5v

>>1541130
tl;dr*
I need to stop drinking

Just finished my first ever board project, it's a TS808 overdrive guitar pedal.
What's the best way to test this without wiring everything and cramming it in the enclosure?
Also what's the best wiring diagram for guitar pedals that includes a 9v battery?

>>1541369
>What's the best way to test this without wiring everything and cramming it in the enclosure?

solder temporary leads? Or make the board larger and add headers that you use for testing, and then use for final wiring.

>Also what's the best wiring diagram for guitar pedals that includes a 9v battery?

what do you mean? best circuit design?

>>1541377
Like this, this seems to be the best diagram for effect pedal wiring or at least what everybody uses,but it doesn't include a clip for a 9v battery.
I think wiring the battery to the same leads as the DC input would be dangerous, and I'm curious what other design is recommended.

>>1541399
You want to used a Switched Jack. It's the textbook use case.

>>1541408
O Damn, I think I ordered the wrong power jacks then.
Is the switch on these closed by default and open when you connect something to it? If it's the other way round how do o wire it up?
I could also use a relay switch and the connectors I have but that would make it too messy

>>1541399
Traditionally a pedal is activated by inserting the input plug and switched off by pulling it. This works because the input socket used has three connections (like a stereo socket, for tip, ring and sleeve). The ring contact is connected to the (-) pole of the 9V battery and is connected to system ground by means of the (long because mono) sleeve of the input plug. The tip contact carries the audio signal. This scheme is simple and requires neither an on/off switch nor an indicator LED that signals the power-on state. (Pic top)

The DC input socket also has three connections where one is part of an NC (normally closed) switch that opens as soon as the plug is inserted. This separates the connection to the (+) pole of the battery and feeds in the (+) of the adapter in its place. (Pic bottom)

Unfortunately the switching pole is always the outer one on the DC plug which means that a standard 9V adapter (center +, outer -) can not be used unless modified. There are special models with center - and outer + made for pedals. There are also adapters that have reversible plugs.

anyone here ever spun a transformer? I have a bunch of xformer iron plates cut in `E` shape. How would I spin them? I was planning on makin ghalf of it, making the coil then putting the other half, but that is structuraly weak and allows for a airgap. I took a look at some xformers here and they have alternating plates. How are the coils spun?

>>1541444
I`m planning to do a 240 - > 3000-ish transformer. Mains side will be 100ish turns of bifilar wire used for phone lines because it`s a x2 speed bonus if I wire it as a single coil with the correct polarity, HV side will be 1000ish turns of magnet wire.

>>1541444
What do you mean? You need not only E shape plates, but also I shaped plates that go next to them like this to form a magnetic field loop: EI, as well as the spools, usually made of cardboard or plastic, that you first wind up, and then you put the prongs of your E plates inside the holes of the spools.

>>1541444
It's called a bobbin and you can diy.

>>1541444
Look here for instance, it's easier to show than to describe.
https:// www.electronics-tutorials. ws/transformer/transformer-construction.html

> fucking antispam

>>1541448
I thought they were made with interlocking Es facing oposing sides. Atleast the ones I have here seems like it.
>>1541449
Nice. Thanks.

>>1539340
based picture
I'm in the middle of designing a transconductance opamp phaser

>>1539341
signals would be perfect
digital design is also good, but more useful for creating controls panels for consumer electronics (synths can fit here too though)

enjoy, EE is the best major

>>1541445
Used Neon Sign Transformers are dirt cheap. They produce 10..20kV and the core is NOT alternately stacked because you want a small air gap for current limiting.

>>1541462
I could only find them here for 50$. (Not anglo)

Ozone guy from last thread. Decided to put my noggins to work and tomorrow I`ll build pic related using 350v electrolitic 1uf caps. It has less stages and a improved regulation than a 16 stage schlong and I can put them side by side with a glass insulator for tasty ozone. Thanks for the help last thread.

Btw, nice to see so many anon wanting to work with high voltage and transformers, just take care guise.

>>1541426
Oh! Thanks for the info on the power adapters I had no idea regular ones where any different than the 9v adapters i have for guitar pedals! I've never tired to use them interchanged but it's good to know!

What's the best way to ground all this? Run a wire to all GND connecting them with eachother or slider them to the box?

>>1541585

we're not on reddit! so there's no need for an exclamation point on every sentence! Just some friendly advice!

>>1541585
Grounding is normally done by the (shielded) audio cables between chained devices and the metal box of a pedal is automatically grounded by the sockets. Internally I would use thin shielded audio lines between board and sockets and ground the enclosures of the potentiometers (if metal), same as in the guitar. No more is needed.

>>1541612
Thanks mate

How pratical would it be to generate ~100Khz sinewave at 1-10V and use transistors to amplify it to 110 or 220v ? That is in comparison to use a buttload of LC filters to get something out of a square train.
tl;dr
how i inverter

>>1541838
Apparently it`s quite viable.
https://web.wpi.edu/Pubs/E-project/Available/E-project-042507-092653/unrestricted/MQP_D_1_2.pdf

>>1541838
You have to ask yourself if you really need pure sine and how much power you want to output. Using transistor as an amplifier will be very inefficient, you're basically creating class A amplifier in simplest possible case (25% max. efficiency). It's best to use FETs in a h-bridge configuration and then PWM it to create sine. BJTs in active region for power electronics applications are a big no, unless you're dealing with very low power (signals).

>>1541543
>tomorrow
How's the kilovolts?

What should i do with my fpga board?

>>1542029
7-segment display

>>1541543
Don't forget you can always add more tips onto the end to decrease the "resistance" for generation. A home "ionizer" can pump out a lot of ozone with a dozen points/thin wire.

>>1542058
>>1542026
It worked! It doesn`t generate a big manly corona, but I can get arcs about 0.3-0.8 cm. I get a tiny purple corona at 2 cm and it grows a bit larger when close.
>>1542058
Thanks anon! I can smell the weird ozone smell now with making tips with the stranded wire. Tomorrow I`ll put it in a pvc pipe and seal it in oil (I tried melting hot glue in a water bath with a jar and failed miserably) and put it in oil, I`ll also buy some syringes to make a ``````vaccum``` chamber in a mason jar and see if I can get a discharge. I`ll post pics.
I`m using 8 capacitor on each side, each capacitor makes a pair right? So a 4 stage*2.

I`m thinking now about how to maximize corona discharges, but can`t find lots of info online, it seems generating it is quite niche, with most applications being made to prevent it.

>>1542132
>0.8 cm long*
>now with tips made of stranded wire*
Phone posting never again.

>>1542058
>pump out a lot of ozone
Ozone is a pollutant and adversely affects the respiratory tract above a certain concentration. Some inner cities have to temporarily ban driving on hot and sunny days because of that or face legal trouble. (google: cars produce ozone) The lab rule is: stop it when you smell it.

What is "sizing factor?"

I'm trying to get some info an a switched mode transformer I salvaged from an old laptop power brick and I keep coming across sizing factor and I don't know what it means or how it relates to specs I actually care about like turns ratio, primary inductance, winding configuration, etc.

>>1542132
Look up water purification, apparently it's used a lot in municipal water treatment. Also, there's specially made UV bulbs that generate ozone. Those might be what's used in ozone generation for water, since they're probably a lot easier and productive in volume..

>>1542303
Sounds like it's a number match between some baseline and an inductor of increased size (and therefore current capacity).

>>1542029
suffer

>>1539341
I'd not advise it ... it will help you to better create an analogue synthesizer circuit, but it will also make you far more sceptical about the merits of doing so.

You only really need amateur EE skills and old magazines/books to steal the circuits from. Without all the "useless" knowledge it's much easier to get into the mindset necessary to sell this trash. As a proper EE you'll feel like a whore doing analogue synth for money.

>>1541963
All that shit only works for non reactive loads, for reactive loads you need a far more complex control loop to still generate a sine wave voltage.

ive got a 1998 Lincoln Town Car in pretty good shape, but this past weekend i began having strange electrical problems. the hazard lights do not work, and the blinkers only work intermittently. not only that but the brake lights dont work.
i read online that a common problem is the hazard light switch can collect shit/corrosion/etc and mashing the button for a few minutes can solve this issue. i did this and now the blinkers work most of the time, but the hazards and brake lights still dont.
even stranger, when i hit the hazard switch, the left blinker comes on, but not the right, but i can flip the right blinker on manually with the stick.
i think the problem is the Multi Function Switch/Turnj Signal Switch
anybody have any advice?

I did a PCB file on KiCad before the schematic. Can I associate the footprints I placed there to the schematic? Or did I fuck up and need to do the schematic first, associate the symbols to footprints and then start the PCB all over again?

I'm trying to build a system system that pours water into a plant using an Arduino. I have a small pump that takes in water through one tube and pours it out through another. It has a ground wire and a wire for 12V. How do I get the Arduino to make the pump work when I issue the command?

>>1542803

12v relay module to power the pump then use Arduino to trigger it.

looks kinda like this

>>1542803
A single MOSFET, a resistor, and a diode.

OK, so I'm trying to make a wireless Bluetooth controller. I intend on making some receivers to connect to my old school consoles as well.
I was exploring the idea of using a Bluetooth Low Energy MCU. However, my concern is with how it communicates. It's asleep until it has to send out a command in order to save energy, but I'm wondering if this would cause a significant delay?
I get that the low power aspect means it doesn't get as much range, but mainly I'm wondering if communication to a receiver would be as fast as a classic Bluetooth MCU? Obviously, communication speed is pretty critical for some games, and I realize that once you go wireless, you introduce delay as opposed to a hard wire connection. If the communication speed difference is negligible, I can live with my BLE MCU being nanoseconds slower than a classic Bluetooth MCU.

>>1543121
>it's
what's asleep and who put it to sleep?
>nanoseconds
some perspective:
in USB 2.0, HID reports won't be sent any more than once per millisecond
according to a rebbit thread /r/engineering/comments/4gdfj2/how_does_the_dualshock_controller_get_rid_of/ 8ms latency is quite performant for BT
old school consoles only present a new video field every 16-20ms
if you insist on BT, you might just have to bite the bullet and try it, but you might be optimizing heavily
if you don't, Nordic's nRF protocols are pretty good and relatively low in latency

>>1543121
>>1543155
>relatively low in latency
more perspective: I have seen the nRF24L01+ chips used in photo flash remote triggers, an application where microseconds certainly do matter

So esp32 has capacitance sensing pins, so i taped a wire to a piece of tin foil and connected it to one of the touch enabled pins.
And it worked! great. it could sense a body part at around 4 cm distance. perfect.
But the problem is that after couple of hours of running it always starts giving out high readings as if a touch is dectected. even if nothing is near the tin foil or the wire
Any idea how to reliably eliminate this problem while retaining at least 4cm sensitivity?

>>1543121
Even 1ms is a big time for most MCUs. I don`t think you`ll have problems. I don`t know what kind of signal you are sending, but if it`s just keystrokes and those sticks things position I don`t think you`ll notice the delay. Also, datasheet usually lists the time it takes to wake up and other things like brown out reset time

>>1543163
charge buildup on the foil? probably make a weak pulldown to ground. Check datasheet for applications.

>>1542865
Ok

>>1542872
Why 1k resistor?

>>1543169
there aren¨t really any datasheets for random wires and pieces of foils, what kind of pulldown are we talking here?

>>1543178
10k resistor from pin to gnd.

>>1543182
10k will suck any capacitance dry so fast you will have to sit on the thing to actually get a high output

>>1543183
Ok, m8 you get the idea. You need something to dissipate charge, be it 10k or 10M

>>1543173
the gate is a C. the gate lead is an L. the resulting LC circuit can ring. that can slow turn-on and turn-off, which will cause the MOSFET to dissipate power. that can also be amplified onto the output and cause undesired EM/RF emission or just flat out oscillation. the resistor dampens the LC circuit to limit those effects
>>1543163
did you try a grounded guard ring around the electrode?

>>1543188
>grounded guard ring
a what? also there is no electrode, just a wire connected to a foil

>>1543193
>a what?
Pic related. something to keep the static charge off of the electrode
>a wire connected to a foil
that's an electrode. here, I put an electrode around the electrodie

>>1543207
Guard rings are used to minimize leakage current for precision equipment where even very small leakage currents can throw off measurements. This is typically accomplished by a buffer driving the guard ring with a copy of the signal going to the device. That has absolutely nothing to do with preventing static buildup.

>>1543185
btw you could also set the pin as output and give it a logic low output for a short time then go back to listen for changes.

say I have a 100v capacitor but only a 12v source. is there a way I can fill up the capacitor to 100v?

>>1543247
Yes.
https://en.wikipedia.org/wiki/Boost_converter
I haven't looked but there are almost certainly ICs that'll do all this for you.

>>1543259
Given the nature of his post I don't think he's at the level where he can design his own switching converter yet. And while off the shelf pre-built SMPS exist most are step-down and it's hard to find anything that'll step up to more than 30V-50V.

How's this youtube guide for learning electronics? https://www.youtube.com/playlist?list=PLah6faXAgguOeMUIxS22ZU4w5nDvCl5gs
Any better recs?

>>1543155
>what's asleep and who put it to sleep?
The MCU. Sleeping in the sense that when it's not talking to the receiver, that it's not constantly trying to keep a connection. At least, that's how BLE technology has been explained to me.
>if you insist on BT, you might just have to bite the bullet and try it
Guess I will have to. I wanna use the ATBTLE1000, but I wanna be sure about what I'm getting into before I blow $200 on a dev kit.

I should note that I want to develop my own antennae and matching circuit for this guy. Main reason I'm doing this project is to get exposure to RF technology.

>>1543166
The scope of this project is just to send keystrokes to the receiving end which will be hooked up to the console. I'd like to try get a way to measure the battery capacity and display it on a small 7-segment display,but only when a button is pressed to request that info.
>Also, datasheet usually lists the time it takes to wake up and other things like brown out reset time
I'll have to take a look and see if there is significant time differences.
Does anyone know what are the most commonly used Bluetooth MCUs?

>>1543292
>RF technology
RF circuit design. That's what I meant.

I need a battery that is 5mm wide AT MOST, 1.5v minimum, 200ma

What are my options? I want to shoot a lit led with an airgun

Yay, it finally arrived!

The future is now.

I bet you kids don't even know what this is without googling.

>>1543173
To stop inrush current. No clue what that LC stuff is, though it's probably somewhat true.

>>1543307
>To stop inrush current

Uh, no. It has nothing to do with inrush current.
A.) Higher current is better when switching FETs to get fast clean transitions. There are dedicated MOSFET driver ICs for this purpose that can source/sink 0.5-2A of current into the gate in order to obtain very fast transitions.
B.) You wouldn't need an inrush limiter with an Arduino anyway because the I/O pins can only source a small amount of current. idk what the limit is but I think it's around 40mA per pin. Also the gate is high impedance so it's not gonna draw much current though its current draw will increase with frequency because of parasitic capacitence which is related to that "LC stuff" the other guy was talking about.

>>1543303
I have one of those tools sitting on my desk, in fact. got it back when such things cost about as much as you would have paid a machinist to make you one. I don't bother with proper wire-wrapping anymore but they work fairly well on "Dupont" headers when I need a jumper that's not 4" long and waving in the air

>>1543292
>most commonly used Bluetooth MCUs
Nordic makes some Cortex-based SoCs with BLE built-in. they're not exactly drowning in peripheral support like STM32 but should be good enough for your purposes, and not ridiculously expensive

>>1543294
https://en.wikipedia.org/wiki/List_of_battery_sizes
too bad

>>1543310
I know it's better to drive a FET with high currents, but trying to pull upwards of 5A from an MCU pin probably isn't good for it, though it is for a very short time. Unless the pins are rated for some momentary or even continuous short-circuit condition.

>>1543294
How about one of those LiPos like the ones in apple airpods/the chinese knockoffs?

>>1543319
>but trying to pull upwards of 5A from an MCU pin probably isn't good for it

But you can't. The MCU pins are current limited which I already mentioned. A 40mA current source is never gonna spit out 5A no matter how hard you try to make it.

>>1543319
You can make a push-pull circuit with a pnp and npn transistors to drive FETs, they`ll make a nice interface for the mcu/

>>1543320
They aren’t current limited. Try drawing heavy current from a microcontroller pin and see what happens.

>>1543320
>he doesn't know about absolute maximum values

>>1543320
I know that, but by attaching a low impedance (read: like zero ohms fucko) to a pin with internal impedance and letting this internal impedance cause the voltage to drop in a device that isn't meant for this you could cause a low-voltage condition to some of the internal logic, not to mention the requirement for a larger decoupling capacitor. I very much doubt that it has even an instantaneous maximum draw current over an amp.

>>1543321
He isn't going to be switching the thing on at a few hundred kHz, a simple resistor is fine.

What's a decent meter (reasonable precision and resolution, fast range switching, fast bar graph) for $100-150 new? Still UT61E or maybe something newer?

>>1542132
The noble Art of High Tension

>>1543261
you just need a PWM and a big inductor
https://www.youtube.com/watch?v=FNUqA0TRyys

>>1539340
>tfw the only efficiently affordable OTA is still the fucking god damned LM13700 cunt shit ass piss
TWENTY-FOUR TRANSISTORS in a package that can house an entire fucking computer with modern chip fab technology. And those fucking useless darlington buffers. My rage is uncontrollable.

>>1543363
I said I was going to test some vacuum shit with syringes and mason jars but I did nothing. I changed some meds and I am drowsy as shit.
Sorry famlamgang, I will not deliver soon. but if you want zaps and painfull shocks the design I posted works. I have the negative end with a 33M ohm resistor and the other end without any security, it kicks like a mule and travels a fair distance over and through most insulation to the other end.

>>1543372
Then just build it out of transistors. A bunch of analogue synth/effect retards just isn't an interesting market.

>>1543395
>just build a circuit that requires intrinsically matched and thermally bonded transistors out of discretes
what are you, fucking retarded?

>>1543399
I'm not familiar with the circuit, but you should be able to buy smaller matched transistor packages and the like. Though I've no idea if it would be any better or cheaper than whatever an LM13700 is.

>>1543399
No point complaining about the size then. The performance characteristics may well just not survive being scaled down and crowded together under the expected loads.

>>1543425
no, it wouldn't
instead of babbling ignorantly, you should do some basic research and learn something

>>1543372
>LM13700
What about the NE5517 or CA3080?

>>1543439
>NE5517
slightly more expensive than LM13700 but, intredastingly, the 5517 has an active load on the Darlington buffer's middle emitter, making it slightly less useless
>CA3080
apparently obsolete

>>1543451
I just found those two by searching for Operational Transconductance Amplifiers on Aliexpress (not that I'd buy precision parts from there), so if you look around some more you may be able to find a better solution. Have you just been scouring Digikey or Octopart until now?

>>1543313
>>1543294
sr416 is 4.8mm wide by 1.6(?)mm, 1.5v, you could totally do it

>>1543516
200mA would be a stretch unless you somehow crammed a bunch of them in parallel. I think the airpod lipo is a better idea for now, but if he decides that 200mA is actually not necessary (hint: it isn't) that tiny coin-cell would be great. Also he may want to put two in series since running an LED off 1.5V is probably pretty tough.

>>1543451
>slightly more expensive than LM13700 but, intredastingly, the 5517 has an active load on the Darlington buffer's middle emitter, making it slightly less useless
You still get a voltage drop and some funky non-linearities. There is no reason why a super high impedance voltage buffer couldn't be used in place of a goddamned two transistor circuit when we can cram tens of thousands in the same amount of chip space. That is what anyone who uses an OTA does anyway with an external op amp. Yes, the final output transistors have to be large, but the circuitry needed to linearize and increase the impedance of a buffer beyond that of a darlington is all dead-simple small-signal stuff that's been around for decades. A fucking lm741 wired up as an inverting buffer has greater input impedance and better linearity than a darlington buffer.
See also http://www.openmusiclabs.com/files/otadist.pdf

>>1543435
It's exactly the opposite. The design would be amazing if it were scaled down and given some modern touches. The issue with building with discretes is that it's near impossible to get transistors with perfectly matched characteristics, and even if you got two transistors that were manufactured DIRECTLY next to each other on the die and put them in separate packages, the small differences in temperature that will inevitably happen when they are not in thermal contact will cause nonlinearities in the overall circuit. OTAs rely on the exponential current/voltage relationship of transistors which means all of the pairs have to be matched, unlike op amps which just have ridiculous gain and rely on feedback for linearization.

>>1543425
The only way to make it remotely cost effective would be to copy the guy who makes those outrageously expensive SMD op-amp boards that are pin-compatible with common op amp sockets, because audiophiles will buy fucking anything that's simple enough for them to do something with if you can smear enough bullshit on it.

What do you want to do with an OTA anyhow?

>>1543548
A transistor buffer's main advantage over an op-amp based buffer is speed. Won't really matter in an audio frequency circuit but if you need an OTA for some other signal processing application running at a few hundred kHz to a few MHz having an op-amp based buffer on chip would be too slow and you'd end up with a lot of distortion. These chips are general purpose solutions designed to work across a wide frequency range and in many different applications. It's not all about you.

>>1543556
I don't think there are any other common applications that require OTAs that haven't been supplanted by purpose-made ICs. It's honestly amazing they're still in production because they're second- and third choice chips for audio as well with the advent of DSP. All of the circuit examples in the datasheet are either audio applications or practically toys. Who is actually buying these things, anyway?

>>1542672
Take it to a mechanic? >>/o/ ?

>>1543535
Infrared is in that range, and he could use an infrared cam since he's planning to shoot it.

>>1543556
Not that guy but
>these op-amps are pretty good
>can only get 20-40 gain at 20KHz
>oh here's new ones that have 10x the bandwidth
>cost is more than $20/unit
Why
Why
Why
At some point I think it's easier to use a A2D, do operations digitally, and then D2A. I'm probably wrong but holy hell.

>>1543558
I'd assume, and I'm just guessing that they are primarily still manufactured for companies that are actively maintaining and using vintage equipment. If shit breaks they need replacement parts, preferably identical or at least pin for pin compatible replacements. Aside from that there's the hobbyist market and diy audio people with their synths and stuff. Not everyone is skilled enough to write their own fancy DSP algorithms. I'd even argue most people couldn't even make very effective use of pre-made code libraries of that either. Analog stuff is on some level simpler and more convenient even if it can be more expensive in low volume and have more annoying quirks.

>>1543558
Aren't they the go-to method of making an amplifier whose gain can be changed electronically? JFETs are a bit shit.

>>1543569
An op-amp with a programmable resistor would work just the same, or even a digital out with resistors in parallel + mosfet switches

>>1543582
Digital pots? Kinda shit but not awful. I think I'd still consider using a JFET if I needed to cover more than a couple of orders of magnitude

>>1543558
The LM13700 works very well as a frequency mixer for low frequency signals of reasonably high amplitudes. I'd argue it's one of the best solutions actually.

Glibert cells don't work well with large signal levels regardless of whether they are made with discrete transistors or integrated like the NE602/SA612. Diode ring mixers are shit at low frequency because it's hard to get suitable transformers for it. Cascode mixers are kind of okay but they're a little less tolerant of large signal levels than OTAs, especially when the supply voltage isn't very high. Single transistor mixers are always awful and should never be used anywhere.

OTA just werked. I was able to use a 5Vpp 455kHz LO mixed with a 1-3V audio signal and with very little tweaking I got clean amplitude modulation out. This would be very hard to do with any of the above topologies. I'll admit it's a very niche use case but it was useful to me.

It's obviously garbage as an RF mixer and there are many many better ways to do it but I'm not working with RF.

Is Hantek DSO5102p good???

Nice thread sticky, following make: electronics.
Any alternatives to 4026b counter and 74hc393 counter? Can't seem to find them in my continent.

>>1543595
What do you mean large signal levels? Can't you just amplify it after using an SA602? It's all linear anyhow, can't see how making heat in one part rather than another would matter.

>>1543604
Look at the "list of 7400 series ICs" and "list of 4000 series ICs" wikipedia pages and search for your keywords. You'll find a dozen on each, and then you just need to filter out tristates and other undesirables until you stumble upon the one that people sell on aliexpress. Then you know you've got the popular one.

>>1543606
I can but the problem with that is I degrade the SNR. I have to attenuate the signal to allow the SA602 to mix it properly which for line level audio is about 25-30dB of attenuation. I think have to amplify it by 25-30dB after the mixer and when you do that you don't just amplify the signal by 25-30dB but you amplify any noise by the same amount. I consider this unacceptable for what I want to do. So no, attenuating the signal for the SA602 is not desirable and an OTA is better.

>>1543608
Those mixers can take a Vcc of 8V and don't say anything about maximum input voltages. 30dB sounds a bit extreme, but I've never messed about with mixers even though I have a reel of 10 next to me as I type. What problems does a higher signal amplitude lead to?

You could build a transistor (or tube) gilbert cell to have a far higher voltage rail to this end, though I can see how it would be suboptimal. Arguably a JFET voltage controlled amplifier circuit could be made to work with a fair bit of tweaking too.

>>1543613
The probably is the SA602 is internally biased ridiculously close to the positive rail. If you use large signal amplitudes you get ridiculous levels of distortion.

I've already built transistor gilbert cells and haven't managed to get much better performance. And why use a JFET VCA which might require a lot of tweaking when the OTA requires almost zero tweaking?

>>1543159
I used that chip for my custom consoles, no lag what so ever.

>>1543601
I have the DSO5072, and its a bit derpy compared to a keysight (only other I've used) but in general I haven't had any problems with it (ps there's a fw hack up to 200MHz i think, down from even my model)

>>1543558
the OTAs have generally been folded into larger solutions. it's pretty easy to find them inside current-mode switching power converter ICs for example

>>1543560
what do you need all that gain for

>>1543582
>programmable resistor
>it's as if he can't even conceive of amplitude modulation
the lolduino thread is anywhere but here

Is there something like a usb-gpio adapter so that I can use my laptop like a raspberry pi?
Just need a way to connect sensors to my laptop.

>>1543636
arduino

>>1543635
20-40 gain is not a lot. A microphone amp needs around 30dB (~31x) to get to line levels, and that's with the built in amplifier working already. Even with 20-30 gain you're often losing some dBs at 15-20kHz.

>>1543635
>gee how can a digital amplifier with adjustable gain
>oh I know how about a pot connected to a motor connected to some stupid analog circuit
>or a single programmable resistor

>>1543636
I think it'd be easier to get a USB->SPI or i2c adapter, and use those modules. I've been looking at it on and off for years, and if there's libraries for those, it'd probably be much simpler than direct devices.

>>1543651
if you need 30dB of gain, switch it in, with switches. fine-tune the gain down to 0.1dB resolution in a second stage if you must. there's more to it than a digipot, anyway, especially when you understand that digipots are kinda shitty pots and that frequency compensation varies across gain
>I am a brain-damaged amerifat so I have to reduce everything to a single axis of superiority
I see that
>he actually thinks "electronically" means "digitally"
>he can't possibly conceive of any application where one might want to modulate one analog signal's gain based on another analog signal much faster/smoother than some digital toy could
why are you even here, bit weenie manqué

What is the differences between Gain and Amplification ?
What I belive is :
For Pi = power in, Po = power out
Amplification = Po/Pi [no units]
Gain = 10*log10(Amplification) [dB]

>>15437099
semantics

>>1543709
>Amplification = Po/Pi [no units]
>Gain = 10*log10(Amplification) [dB]
The difference here is it's logarithmic.

I'm done pretending to be sane.
>160 cells, 4p4s x 10
>120Ah rated capacity
>180A surge
I'm going to bring it up to 144Ah, and then I still have enough cells left to build another pack of the same size.
gimme that sweet lithium fire

>>1543746
A gain of three is also ok, and an amplification of 20dB. Both are dimensionless.

>>1543760
Well done for fireworks mode.

>>1543310
>>1543173
But you get ringing from the gate capacitance and parasitic inductance of the trace. You can't critically damp it because you don't know either of those numbers but the gate resistor is a "better than nothing" solution.

I personally tend not to use gate resistors because I don't care very much, but they do have a purpose.

>>1543372
> The LM13600 has designed by Bill Gross and myself in less than 5 minutes. At the time the Consumer Linear IC design group was training a new mask designer. We needed something with a few transistors for her to learn how to arrange them in a optimum circuit arrangement while using minimum silicon area.

http://www.idea2ic.com/LM13600/LM13700.html

>>1543774
It's simpler than that. Don't connect digital outputs to capacitative loads. I=C.dV/dt.

>>1543840
How would the driver affect it? It's not ringing because of the driver, it's ringing because it's an LC circuit.

I have a circuit with a component that is clearly a resistor but it is labeled 'L'. Is this some kind of special designation or function that it has?

I'm new to this shit.

>>1543709
You have to be careful, voltage/amplitude gain is 10log, but power gain is 20log . You run into it more often in radio/antennae amplifiers.

>>1543882
Is the body color blue-green? It's an inductor.

>>1543886
>voltage/amplitude gain is 10log, but power gain is 20log
Isn't in the other way around?

>>1543882
if reference designator begins with L, it's most likely an inductor in an axial-leaded package
there are blue-green-bodied resistors, too, but inductors don't have the same shape of waist

>>1543888
it sure is

>>1543889
>>1543888
Whoops. That's why you need to be careful.

>>1543846
I'm saying that the potential for ringing is the least of the problems. Without a resistor, you're basically shorting the output to ground.

>>1543760
Some kind of EV build, or just want to watch the world burn?

whats the cheapest source of non housefire 18650s?

>>1543979
I'm using it with a Xantrex 1000 watt inverter. In the winter it'll keep the furnace running for a while during a power outage.
In the summer it'll run the fridge.
EV build is in my future I hope, with the knowledge I can get from this simpler, safer project.

>>1543983
clip the thing across your nipples

>>1543241
Good idea i will try that

>>1544051
it won't work.
if the wire with the thing isn't connected to anything then there isn't any potential and the charge won't flow.

I'm thinking of jumping down the SMT+Reflow Oven rabbit hole

I see there's the T962 reflow oven which sucks balls but apparently has a 3rd party open source firmware replacement that makes it a decent little workhorse for ~$200

has anyone tried it out?

What the fuck is this "D E Q" box?
I get theres opamps, XOR gate and OR gate, but what the heck is that?

>>1543819
>we designed it on the fly as a prank

>>1544178
<--- this box

>>1544179
I miss the 1970s

>>1544181
probably a D-type level-triggered latch. (roughly speaking, flip-flops are edge-operated and have clocks, latches are level-operated and have enables)

>>1544211
>probably a D-type level-triggered latch. (roughly speaking, flip-flops are edge-operated and have clocks, latches are level-operated and have enables)
shit. i dont understand any of that lololol

what function is that part doing there? is it related to the switch at the + in the opamp?

>>1544157
A $20 toaster alone will do reflow.

>>1544232
wtf is this true? i need to solder lots of micro shit for modular and pretty sure my huge ass solder gun wont do the deal.

>>1544236
>wtf is this true?

>>1544248
This is misleading. You don't need to hack the toaster at all. Throw the pcb in, turn it on, wait til the solder reflows, then turn it off. Probably not as good as professional level reflow but gets the job done.

For $20, what is there to lose. Try it out.

>>1544222
sounds like you got some theory to read
a latch passes the value on input D through to output Q as long as it's enabled (i.e. input E is at a high level). when disabled, it stops passing through and holds the last value of Q. does that help?

>>1544236
a good pencil iron will work fine for most SMD
unless you have a whole board full of really small stuff (0402 discretes or smaller, or DFN/QFN/xSON packages with no leads), in which case some sort of hot air type of arrangement might be nice

>>1544255
i dunno, a whole board full of parts maybe
>the job
>done
we don't even know what scale of components OP's trying to solder

>>1544255
the point was more to show him to use a toaster oven - not a toaster.

I have an audio cable that keeps getting bad static from a nearby power cable. In my case, I am unable to relocate things, so is there a cheap way to shield the audio cable with something sitting around the house or something I could get in the store?

>>1544248
>>1544236
Never messed with smd much, but there are some very small electric oven things you can get from ebay or aliexpress that could do it. Most of them only have a timer and some metal strip thermal shutoff, but I guess you could make them work for what do you need.

>>1544259
I ideally you want a line driver or similar balanced driver driving a balanced cable like twisted pair that should feed into a suitable 1:1 audio transformer. This should provide you excellent common mode rejection and will basically eliminate all your power supply noise. You can even whack a common mode choke on the output just be sure. Beyond that you need a differential amplifier to take your balanced signal from the transformer output and make it unbalanced again. You need to duplicate this setup for stereo audio.

This is the best way to definitively prevent noise. You can just use an isolation transformer by itself and maybe get adequate results, maybe not. You can also try just buying a shielded aux cable. It probably won't work well but you can try. Beware sellers online that advertise cables as shielded when they actually aren't.

>>1544255
Depends on the particular oven, lots of them don't let the thermostat go high enough for proper reflow.

>>1544259
>bad static
Indoor lightning?
Place is haunted.

>>1544222
It's digital logic. D-type latch = 1-bit memory. In your circuit diagram, it's being used to control the switch.

A gate is a combinational logic element; the output depends only upon the current state of the inputs. A latch or flip-flop is a sequential logic element; the output can depend upon past states of the inputs (i.e. it has "memory"). A D-type latch copies the D (data) input to the output when the E (enable) input is high, and retains the prior state of the output when the E input is low.

>>1544506
right. but which one of those two topologies, or doesnt matter because logic can be interchangeable and reach the same result?

also, lets say i build this d type latch with OR gates, mayeb a single cmos chip, then how the fuck does this "Q" output controls the switch? what in hell is that switch? could it be a 4066 bilateral type switch?

>>1544603
doesn't matter, usually you only care about the output
a CD4066 or similar seems most likely, sure. I suppose you'd have to show the whole diagram if you want us to understand at the system level

I've decided it's kinda fun grabbing random clip art of cute cats and putting them on my silkscreen layers

steep decline

>>1544620
Agreed

Ultra stupid question
Can I use x20/x200 diff probe with a scope that only have x10/x100 probe ratio????

>>1544651
Yes but the readings on the scope will be off by a factor of 2. The scope setting just multiplies in software the input by some amount to make up for the attenuation in the probe.

Hello /ohm/
Right now the Volume pot takes postive and negative waveforms and every frequency gets canceled resulting in lower volume. The High Filter however acts as a variable high pass on the volume only. The lower the volume, the lower the effect of the high pass. But I would like to have the same (or at least a bigger) range of filter in every volume position and I want to do this with passive components only.
Maybe I am just stupid, but is this possible? E.g. take a stereo pot and change the way the filter responds to the volume when changing volume? Adding another capacitor? Any way?

tl;dr
Is there any way to make pic related more effective on low volumes with passive components only?

>>1544706
goddamn accidentally refreshed page and lost my post but tl;not retyping: consider putting the vol pot in series with input and attaching it's wiper to the hi-pass in

>>1544719
How would I do that? I don't get it. The volume works by canceling frequencies, it does not create a variable voltage devider and shunts to ground like usual volume controls, it cancels + and - amplitudes by mixing them. How can I put it in series in this configuration?

>>1544706
Strange circuit, context missing.

>>1544736
The context does not matter. The circuit is exactly what it is: A signal, ideally the same on both terminal but inverted, + and - gets canceled by turning the volume pot, resulting in a lower signal as the waves cancel each other out. The high filter is just that: A filter that only cancels high frequencies instead of the whole spectrum. But it is volume depended, because it is a quasi high pass, a treble bleed on the volume. If the volume is low (the waves cancel) the high filter is ineffective.

So the question is remains: How can this filter act in the same range independed of the "volume" pot? I need some kind of constant load, but how?

>>1544739
It looks like a differential (antiphase) input signal but it has no source, no source resistance and no load. The circuit fragment is funtionally isolated.

>>1544618
whole schematic is this. its a wavefolder. author didnt put any of the values, but should be easy to guess, the logic part is the one giving me headaches. OTA acts as sort of a limiter, though i still dont know how.

>>1544603
> but which one of those two topologies, or doesnt matter
Doesn't matter; both circuits produce the same result.

> how the fuck does this "Q" output controls the switch?
From the diagram, I'm assuming that the switch is closed (short to ground) when Q is high/true/1.

> what in hell is that switch? could it be a 4066 bilateral type switch?
Quite possibly. If "ground" is the negative rail, a BJT would work. If the input can be either side of ground, it would need to be a FET (without a body diode) or a 4066 or similar.

>>1544760
ground is 0, its a bipolar powered circuit, for synths.

>>1544739
context does matter. why you using such a weird-ass volume arrangement

>>1544757
interesting
the designer probably calls for CD4000 "B"-grade logic which are typically specified to handle up to ±10V supplies. if they actually meant a latch, you could use 1/4 of a CD4042B, but if they meant a plain old D flip-flop, you could use a CD4013B

>>1539349
Why not just keep track of it and keep the thread alive instead of being a huge nigger about it?

>>1544782
No, it does not fucking matter. If a simple voltage devider worked, I would have already implemented it. But that is not the case, it has to work by canceling ONLY.

I totally love it when people on the web assume that the problem that is explained simply does not exist and that there are better ways to not encouter it in the first place when it is explicitly stated that this is not the case.

By the way, the circuit as is workes. It is just that the high filter range suffers on low volumes, but it still works OK. I am just looking for a simple, elegant and better way that I have not found yet.

>>1544788
>it has to work by canceling ONLY.
you're not canceling anything. you're simply loading down the source
if you weren't a pajeet trying to get other people to do your homework, you would already know this because you would be able to derive a transfer function for the network as a whole

>>1544179
>>1544211
More surprisingly
>her

>>1544792
actually, if you were not a niggerjew you would have seen that I have already written that I probably need a constant load but I don't know how to get it. and, yes I do indeed not know how to draw a transfer function, that is why I am asking.

and no, the basic setup cannot be changed. and, you really woulreferencesd have to explain this to me, I am not loading the sources because I do not change voltage references any how, do I? literally two signals, one positive, one negative, from different sources.

>>1544788
>>1544792
Yeah the guy is literally shorting one output into the other and doesn't see this as a problem, dunno wtf he is thinking.

>>1544798
yes even when you connect a 1Teraohm resistor to a voltage source and leave the - lead floating, even that is a load. More realistically, your volume control adjust from Rmax -> basically short (what's in parallel doesn't matter now).
>>assume that the problem that is explained simply does not exist and that there are better ways to not encouter it
the problem was explained simply, and step 1 is to scrap your design or give schematic

>>1544797
women have been doing the unattributed scut work in technology for decades now. back to >>>/r9k/ with you

>>1544798
>I know what I want
>but I don't know what I want
pottery
sigh. alright, open wide: the reactance of a capacitor is according to the formula Xc = 1/(2*pi*f*C) where X is in ohms, C is in farads and f is in hertz. knowing this, you can look at an ordinary RC low-pass filter (as if you took your sample from between the high-filter pot and the high-filter capacitor) as a frequency-dependent voltage divider
if we knew the source impedance, call it Zsource, and assuming output impedance is infinite, we could then see that the "volume" pot is just the bottom leg of a voltage divider, whose output voltage is Rvpot/(Zsource+Rvpot) * Vsource, and likewise with the "high filter", where we replace Rvpot with (Rhighpot + Xhighc), and ultimately combine the two with (Rvpot||(Rhighpot+Xhighc))/((Rvpot||(Rhighpot+Xhighc))+Zsource)
however, we know NOTHING about Zsource (other than that it is high enough that your circuit does something other than heat up the volume pot) or Zoutput. therefore we can draw no quantitative conclusions about anything because we don't have
C O N T E X T
O
N
T
E
X
T

>>1544818
Idk silicon optimization/pcb layout seem related. Wouldn't you want a gf who could route your boards optimally for you?

>>1544821
tbqh that sounds better than blowjobs

>>1544818
>>1544821
>>1544826
Do you guys not like PCB design? It's a lot more fun than making the schematic

>>1544873
>t. regards digital designer

>>1544873
It's pretty fun, but I'm not doing high-speed design. Microwave PCB design sounds like a nightmare.

>>1544873
>more fun than making the schematic

How do you measure the thickness of plated copper?

>>1544944
those are probably easier engineered than reverse-engineered

>>1545006
sectioning and magnification?

>>1545012
Like, cut it and put it under a microscope? How do you get a sense um scale to compare it to?

I have small salvaged speakers and know how to solder
Is it cheaper to make a small amplifier from components or to buy a cheap amplifier/amp kit?

>>1545006
Micrometer? You could also use hundreds of layers of something far thinner (aluminium foil?), count how many layers it takes to get up to the thickness of the plated copper and then count how many layers it takes to get up to 1cm or so.

>>1545023
Well, 10s of um. 1oz copper is about 35um.

The aluminum foil method sounds unreliable unless I could get it ridiculously thin. What if I have access to various labs? There's gotta be specialized equipment for this, but I don't know what's most suitable.

>>1545015
Amplifier kit for sure. An alibay stereo class-D amplifier PCB will run you a few dollars and output maybe 50W, they're quite something. You'll also need some sort of power source, for which you'll want to look for something low-noise. I'd either go with enough (protected) lithium ion cells in series to get to approximately the right voltage, or use a SMPS and a LDO linear regulator that runs off mains. You can also buy USB/5V-powered bluetooth audio receivers, which might be a good thing to add too. Naturally this will require its own power supply, and I think it has to be isolated somewhere (either in the power supply or in the audio path).
There's a "DIY Perks" video about making a concrete bluetooth speaker like I describe which looks pretty decent.

>>1545027
If you know its area (pretty easy if you cut it down to a square inch or so) you could weigh it on some really precise scales (not easy if it's on a PCB), or dissolve it and perform a titration to figure out its total mass. You could also weigh the dissolving solution before and after putting the copper in, provided the reaction doesn't release any gasses.

>>1545014
you could compare it to the nominal thickness of the substrate. a shitty USB microscope might be enough
or, if you can peel and clean off a piece of a known area, say, 1"x1", you can weigh the piece, divide by the density of copper, and solve for the unknown third dimension

>>1545015
depends on the money value of your time
that said, an LM386 and the very few supporting components it requires should be in everyone's parts stock for just these sorts of occasions

>>1545029
>>1545035
I've heard that you should match your speakers and your amp, and these speakers are sort of odd to me because the printing on them says 32Ω and 6W max and all the speakers I've ever encountered were like 8Ω
Will I need to do some math and find appropriate components for this?
I don't want to blow them out accidentally or whatever happens when you do shit wrong

>>1545045
32Ω 6W means an RMS voltage of ~13V, compared to an 8Ω 6W speaker that would run on half the voltage. 13V RMS means 40V peak-to-peak, so you may want an amp and power supply to match, though there's nothing particularly wrong with going for a lower voltage (and hence a lower power). I'm not sure if impedance matching applies to class-D amplifiers.

>>1545050
does this seem like a worthwhile undertaking?
My ambition was to build these into my computer desk so I don't have to take my headphones and perch them turned all the way up on my desk when i listen to music
I haven't take any electronics technology classes since 2015 so I'm getting brain fatigue since things seem familiar but just out of my grasp

if I'm understanding this diagram correctly, I have to supply my own voltage to the LEDs, the IC doesn't actually drive them?

>>1545078
You would burn the shit out of most any LED if you tried to power it from the same voltage the IC needs (12V-20V), but the IC is still "driving" the LEDs. It's just that the switch is between the cathode and ground rather than the voltage rail and the anode.

>>1545067
I'd go for it. Something like:
https://www.aliexpress.com/item/yiz/32777392061.html
Will work, but it's only 34V and has a power rating far higher than you actually need, so you might want to go for a lower voltage (12V) amplifier with a matching transformer or some sort of network. You could probably also just run them off 3.5mm directly and not get awful performance, as they're 32Ω, but you probably do want an amplifier.

At such a low power on each channel, going for a linear amplifier might not be an awful idea, I think you just need a dual op-amp and a few transistors to make something like this image, plus a few resistors on the op-amp for gain reasons. Don't forget the heat sink. There are many more complicated topologies that I'm sure other people here can clue you in to, but this would allow you to throw it onto a ±20V power supply if you had adequate heating.

>>1545078
I made a separate power rail within spec for the LEDs. I can only get one of the LEDs to light up intermittently.
Don't know what I did wrong.

>>1545098
oh alright, I see now.

>>1545102
At maximum waste heat (a half maximum voltage square wave), each transistor would average ~3W or so, which is quite possible with TO-220 packages on a heat-sink. Something like this image, with a volume potentiometer instead of the 160kΩ resistor, would work just fine. FET names are just for testing. BJTs might work without an issue since you're only pulling 600mA maximum, but you'd have to either go for darlingtons or deliberately go for op-amps that can supply a moderate current (possibly up to 100mA).

>>1545102
from experience let me tell you that the design in your picture is GARBAGE for audio (and probably everything else). The op amp will attempt to keep that completely unbiased class B stage happy, but it can only slew so fast through that entire volt or two of dead zone and the result will always be lots of nasty crossover distortion. Even if you strap in a biasing scheme and fancy it up a bit, stabilizing these kind of boosted op amps is a guess-and-check kind of thing with various R and C values in the feedback network. At the same time, I wouldn't trust that a $6 board from aliexpress would sound much better (and if it does, it'll probably burn out or burn the speakers out in short order.)
For a simple project, there are plenty of all-in-one power amplifier ICs in that price range; all you would need to do is solder it to a board with a handful of resistors and capacitors and strap it to a heatsink. If you want to go deep, all-discrete headphone amps can still be very simple and produce completely inaudible levels of distortion at full volume. High impedance speakers makes the job even easier as you don't have to worry nearly as much about loading and subsequent instability effects on the output stage as with 8 or 4 ohms.

Whether you want to rip off a schematic and build it quickly or get lost in some technical reading, I highly recommend a look around rod elliot's site (http://sound.whsites.net/index2.html)) and a copy of doug self's (http://www.douglas-self.com) Power Amplifier Design book.

>>1545114
Not surprised, I've barely worked with audio. Is an all-in-one power amplifier IC rated to dissipate 6W of heat by itself? Are there models meant to be used with external output transistors?

I'm probably going to build one of those totem-pole amplifiers just to see how bad it sounds.

>>1545113
if that circuit actually simulates a nice waveform without really brutal crossover distortion, check out the quiescent current going through the FETs. 40V across two Rds on the order of a few hundred milliohms is quite a lot of juice.

>>1545116
Quiescent current through the FETs? You mean source/drain current? It gets up to 500mA, but the FETs waste far more power when they're dropping a significant voltage across them as opposed to when the current through them is at a maximum. Though I didn't simulate any speaker inductance, which might change that significantly. The thing actually does simulate with pretty low distortion which I found fairly surprising, better than the NPN model.

>>1545115
There are some in some pretty beefy packages with metal tabs, so with a large enough heatsink and high impedance speakers, the limiting factor is more likely to be the max voltage the chip can take rather than the power dissipation. It will be uncomfortably loud before you reach either of those points, though. A legitimately rated 10W amplifier is going to get louder and still sound good than most commercial units rated at 100W.

>>1545114
I'll give a read and see if I can't understand at least enough of it to get my work done
see you in a few days

>>1545124
also I have a breadboard and a bunch of cheap components from that class i mentioned but it's probably all just resistors and basic capacitors so I'll take inventory of that too

>>1545120
This is what I'm getting. Only difference is I'm using power FETs, which is what you'd want to use in a power amp anyway. With the gates tied together and no biasing scheme in place, both of the transistors are "on" and the minor variation in voltage from the op amp provides enough to move the expected current through a theoretical 8 ohm load, but the transistors themselves are basically shorting the power supply rails. BJTs with the bases tied together presents the opposite problem: there's a "dead zone" of two Vbe drops around the 0V mark where neither of the two transistors are conducting any current, resulting in crossover distortion.

>>1545130
That's an 8Ω speaker, not a 32Ω like the guy says he has. If it were 8Ω I wouldn't be using nearly so high a voltage. Also you cannot into totem poles because your PMOS transistor is backwards.

>>1545106
it drives them in the sense that it regulates how much current is pulled through them. the LED supply is separate, and if you just read a little lower down the page, you'll see application advice suggesting that one keep Vled below 7V or add dropper resistors

>>1545130
you sure you got the P-FET right side up?

>>1545130
>>1545113
I thought something was strange; I had the PFET upside down. It doesn't short the supplies but here's that beautiful crossover distortion I promised. Input is only 100mV peak to peak. A faster op amp would alleviate it but only to a point. a mosfet output stage needs MORE bias than a BJT stage; I had it wrong in the last post.

>>1545136
That's due to the high RDS_ON, right?

>>1545133
that doesn't make much difference. Either way the op amp output has to swing the sum of both transistors Vgs's (almost 10V) for the output of the full amplifier to just cross from -1mV to 1mV. The blue trace is the op amp's output terminal. What i'm describing is the nearly vertical section where the input signal crosses over 0V. This change is limited by the op amp's slew rate.

>>1545140
forgot pic

>>1545137
Rds(on) only limits the amount of current that can be driven through the speaker. With the mosfet turned on as hard as it can go, the absolute most current it can sink is the rail voltage divided by the sum of the speaker impedance and the Rds(on) by ohms law. Both mosfets are wired as source followers so it'll never go fully "on" anyway, which kind of negates any reason for using a fet in the first place.

>>1545144
>>1545137
tl;dr the limiting factor is the Vgs(th) for mosfets (a few volts) or the Vbe drop of a BJT (a few hundred mV)

tl;dr use BJTs if you're doing linear amplifiers.

>>1545146
>Vgs(th)
Oh yeah that's what I meant.

>tfw I used a transistor with no feedback as an audio amplifier
>tfw listening to the clipping distortion
I'd do it again too.

I want to make this display smoother, but I don't know how to make an AC signal.

I know how to make a triangle wave using a transistor and a 555 though, would that work for this?

>>1545151
>don't know how to make an AC signal.
that's what the 0.22µF cap does
>would that work for this
probably, if you scale it down a bit

>>1545151
>the signal can be a triangle, sawtooth, or sine wave from 60Hz to 1kHz.

>I know how to make a triangle wave using a transistor and a 555 though, would that work for this?

I reckon that would work yes.

>order operational transconductance amp from aliexpress
>envelope arrives battered and torn, nothing inside
>customs/post person wrote "Arrived without contents"
well shit now what do I do

>>1545178
stop sending money to china for american products that weren't even expensive to begin with.

>>1545178
Write them off as a loss and forget about getting your money back. Probably not gonna happen. Next time don't order electronics from fucking aliexpress. Just pay the $2.08/per unit + shipping + taxes + tariff on digikey or mouser.

>>1543303
Is it a roll of wire?
(didn't even google)

>>1545179
I'm in college, I don't have the luxury of ordering domestic stuff

>>1545180
I've been ordering electronic components from Ali for like 2 years now, never had a problem before. Domestic stuff costs like an order of magnitude more than ordering of AliExpress

>>1545179
joke's on you, they were made in china anyway

>>1545182
I've had 99% good luck with chinkshit. shit happens. try again

>>1545181
yes, but

>fried my 7 segment led display

>>1545189
Ah, that's what it was. I used reverse image search on that there and it's definitely related to the thread and has something to do with the wire.

>>1545136
If you use schotkey diodes and two caps, and a resistor-resistor-resistor path with caps, you can bias the mosfets to just near conductance and have AC out. Or two zeners and a resistor in the middle. Or just go with transistors and diodes and resistors like a sane person trying to do a class B amplfier.

Anyone know if it's possible to remove the color filters from an LCD screen to essentially render it black and white?

>>1545301
The "filters" you speak of is literally the switching crystals. Now, to make the crystals respond to any input voltage, that's an interesting question.

Or just go into color settings and set the screen to monochrome or use an extra software.

>>1545301
the filters are under the glass. doesn't look user-serviceable to me

>>1545356
LCD screens can be taken apart pretty easily. I'd question how well it would work, even if everything could go smoothly. Worse yet, you'd have 3x the pixel width density but still 1x the height, which is noticeable.

>>1545362
post pics where you've broken apart the sealed glass sandwich and gotten to the color filters, instead of bragging about you're too cool to look at the exploded diagram provided with my post, let alone comprehend it

>>1545373
I tried. The glass is stuck together too well.

Well um, all those posts about the layout for an amp, which one is the recommended one?

I don't truly understand the differences though I recognize a lot of the terms.

>>1545409
>layout
I hope you mean design. Class D is best, but if you have a fixed output resistance I think an E or F is slightly easier/more efficient.

>>1545409
>recommended
just buy an LM386 chip and save yourself some angst
but if you really don't want to, class AB is probably a good balance between complexity and performance, for such low-power (in the grand scheme of things)
electronics-tutorials ws/amplifier/class-ab-amplifier.html

>>1544818
Well, thank you. Your post really helped me, at least a little bit. I put your formula in a graphic calculator and played with the variables a little bit. Finally I came up with an idea: Make the volume pot a stereo pot that opens up the path to another capacitor that has another value, better suited for lower volumes. It works a little better in theory.

>>1545421
Yeah, I don't use correct terminology. My apologies.
>>1545444
Righto, I'll get to it then.
I really didn't expect so much response to something relatively trivial for y'all, thanks anons

what is the thing which looks like this
i found one in my pants pocket

>>1545444
>LM386
>12V/18V depending on model
That's not going to work too well, though I suppose it might be sufficient. It's also mono, if you look harder you may find a higher voltage stereo (more modern) all-in-one amplifier IC.

>>1545491
LDR/light dependant resistor. Nice pic.

Anywhere I can read on how to drive LED displays? I want to make a 4x4 led square to display some animations using a eeprom+PIC. But I`m having trouble understanding on how to connect the leds to be able to control individual pixels without using 16 pins.

>>1545495
The word you're looking for is "multiplexing", and "matrix" too. use those keywords you'll find loads of articles/instructables/videos
here's one:
https://appelsiini.net/2011/how-does-led-matrix-work/

>>1545507
Thanks! It`d be my first project using I2C.It`ll be hard to think about memory locations and stuff like that for the first time. How much data to read before scanning each line, how to organize images aaaaaaa

>>1545495
2 74HC595 latching shift registers, assuming you don't want the pixels to flicker if you control them from only 8 wires. The shift registers mean only 3 pins (data, clock, latch) of your PIC taken.

DIY spot welder anon here >>1539340
I put some more money and love to the project this weekend so I traveled the entire city to find these hugeass copper rails, some proper 4AWG cable, screws and blank PCB slates for the project to look better and more tidy.
As the previous attempt was a failure I'm redesigning the circuit for the MOSFETs to have proper heat dissipation now.
Also, today received some little MCP1407s to drive the MOSFETs with, so I have a little question:
I have seen schematics which use one driver for each MOSFET and others which use one driver for all MOSFETs. Can I be stingy and drive all MOSFETs with just one MCP1407? Their model is still IRF540N. The IRFB7430 are still on their way.

>>1545578
Misquoted >>1539236

I have this LED strip and no idea how to get it to light up. It has wires for 5V, ground, and two wires for Din. The other end is the same thing expect two wires are for Dout instead of Din. How do I get the entire thing to light up? I can't find the documentation since there's no serial number

>>1545582
Wires

LM3916 anon back again

I've gotten the thing to light up, finally.
However, the LED output is really... chattery.
You can see here: https://www.youtube.com/watch?v=5tI8b_KqdrM
this demonstration has very smooth LED action.

With mine, the LEDs above the current peak flicker.

The only difference is between mine and his is that his has some diodes and a capacitor between his input and his circuit. Is he rectifying the input?

I've put a large capacitor in series with the input and a large resistor in parallel with the input and ground, but it doesn't do much.

>>1545508
16 bits is only 2 bytes. just read the whole frame into a buffer and set up an interrupt to change the row and write out the next columns. that way you can wait as long as you like for the next frame (perhaps based on data from the EEPROM), or implement effects like scrolling, etc.

>>1545450
de nada. anons will argue about anything

>>1545582
odds are it's WS2812 or something very much like it

>>1545578
the 1407 gets slow with too much capacitance on it. safest would be to run 4 and 4

>>1545495
Use just 8 pins: 4 rows as push-pull outputs and 4 columns as open-drains, and switch them real fast. Replicate pic related but connect a 220 ohm resistor to each row to not burn your LEDs.

>>1545582
These are addressable LEDs. Searching a bit I found out each one of these LEDs are of model WS2812B, which, in effect, has a datasheet.
https://cdn-shop.adafruit.com/datasheets/WS2812B.pdf
You control them by sending 24 bits of serial information followed by a delay. Depending on the order of the information the LED will receive or bypass that information to the next LED. There are some Arduino codes you might be interesting in look at to know more details.

>>1545590
>Is he rectifying the input
he is. the cap charges quickly and discharges slowly, so holds the peak for longer

>>1545591
>>1545592
Ok thanks. I will look into WS2812

>>1545495
>without using 16 pins.
I recently made a board with up to 50 leds driving directly from 50 pins because KISS. The problem wasn't using up the pins, it was soldering 50 0.1" headers and resistors. Pins are cheap, the time it takes to solder is not.

Kicad v5 has ngspice built in, but is it any different from actual ngspice? Should I use both or is it no use?

>>1545660
Is ngspice any good compared to LTSpice?

>>1543303
why are those tools so expensive
when I was first getting started I wanted to go that route but didn't because they were fucking expensive

>>1545421
>Class D is best
>I think an E or F is slightly easier/more efficient.
Give me whatever drugs you're on.

>>1545595
What's the low-down on rectifying RF or audio signals that are at a voltage level similar or lower than the forward voltage of your diode? Like a 1V peak signal into a ~0.6V diode, or the case of a 0.5V peak signal into a ~0.6V diode.

>>1545660
> Kicad v5 has ngspice built in, but is it any different from actual ngspice?
No. KiCad just provides the user interface (ngspice doesn't include a schematic editor or plot viewer; it just reads and writes files).

>>1545698
> Is ngspice any good compared to LTSpice?
ngspice is basically spice3 plus frameworks for simulating complex logic components (Xspice) and more accurate (physically-based) simulation of semiconductors (Cider).

If you want to simulate complex logic in bare spice, you have to build it out of gates which will be simulated as discrete components. That's going to be rather slow for VLSI. Xspice lets you model the internal behaviour using code and just use analogue simulation at the inputs and outputs.

LTSpice is basically a version of spice3 that's been tweaked for better simulation of transients (important for switching converters, which is the core of LT's business) with a basic UI for schematic entry and viewing plots.

>>1545736
- Use a diode with lower Vf, e.g. Schottky or germanium.
- Amplify the signal.
- Build an idealised rectifier (superdiode) with op-amps.

>>1545739
What's a good amplitude to amplify up to? Vf*20?

>>1545703
I requires incredibly high machine precision to make them. They literally need to be atom-precise to work properly.
I got one from chinks for twenty bucks

>>1545592
>Use just 8 pins: 4 rows as push-pull outputs and 4 columns as open-drains, and switch them real fast. Replicate pic related but connect a 220 ohm resistor to each row to not burn your LEDs.

Can you explain a little better? Or draw a schematic for a 2x2 or something. Sorry I`m just starting. Also, what is a common FET IC?

do you have a good alibay seller that has great variety and low cost? Shipping to here is expensive so I`d like to buy all my junk from one place.

>>1539340
I need to buy a sensor which exists in two versions - one that outputs a 4-20 mA signal and one that outputs a 0-10V signal. Which is better for use with an Arduino?

It seems like I need a separate shield to deal with 4-20 mA signals, while with the 0-10V signal I could just divide the voltage. Anything else else to take into consideration?

>>1545828
use a resistor to convert the current to a voltage.

>>1545828
You don't necessarily need a whole shield to read a current based output, you can convert a current to a voltage quite simply with a single resistor using ohms law.
The reason current based sensors are used in industry is because (kirchoffs laws) the current is the same everywhere in the loop meaning you can have the sensor very far away and the current will be the same at the sensor and the receiver while a voltage based device there voltage will be different because of the voltage divider effect from resistance in the wires.

If you are just playing at your desk it doesn't matter, if its a prototype keep everything as close as possible to the finished result.

You should also consider input protection for your microcontroller, a zener diode to clip overvoltage and a resistor to ' burn' the difference is simple but effective.

>>1545750
Depends on the context. What are you doing?

>>1545830
>>1545831
Thanks guys. That doesn't sound too bad.

To elaborate, I have to equip a machine at our lab with one of these PLC terminals:
https://beckhoff.com/english.asp?ethercat/anaout.htm?id=1985119910
Does that change your answer?

I just want to get a machine parameter nicely into an Arduino.

what does it mean for a resistor to pull up or pull down

>>1545838
It doesn't care.

>>1545838
Say you want to use a switch to control an MCU. You put one end of the switch on the MCU pin and the other end to ground, so when the switch is closed the pin is at ground and reads as 0. But when the switch is open, what's the MCU see? The input's floating so the state of the pin depends on complicated silicon stuff inside the microcontroller. It won't always be 1 like you want. So you put a pull-up resistor on that net, connecting the other end to Vdd. Now when the switch is closed there's a current from Vdd through the resistor and switch to ground, and effectively all the voltage is dropped in the resistor, so the MCU reads 0. When the switch is open, there's no current, so no voltage drop across the resistor, so the MCU reads 1.

A pull-down is the same situation but the switch is connected to Vdd instead of ground.

>>1545816
it depends on exactly what you're buying
store 614856 Chinese Super electronic market, 1815642 Victory electronic, and 511081 C&G semiconductor get a lot of repeat business from me. there are other stores that specialize in semiconductors, usually I just order from whichever has everything I need at the time and doesn't rape me on shipping, especially if I'm looking for something only a few stores carry

>>1545703
but there's a lot going on in the ends of those

Picked this up for 5€ at the local scrapyard.
Rate my find.

>>1545865
Glass jar/10.

>>1545865
>old eletromechanical blinky shocky things == good
>old electrolytes == bad

>>1545865
250

Really stupid, uneducated question incoming:

Old 8-bit computers had alot more simpler design, but still loads of different individual chips... If someone hypothetically had a huge store of all of the chips used in say, a commodore 64, Speccy, and MSX, could someone theoretically build or maybe even produce a run of an all-in-one non-emulation computer of all of these systems combined that could run code meant for each individual one? What obstacles would stand in someone's way?

I know my question is stupid, and I know there are other options like FPGA, etc, but I've been curious about this for a long time: If it's possible to combine the definine chipsets from different 8bit computers into a single computer that actually functions.

>>1545887
At the simplest level, you could just build all of the different computers and put them in the same case. But you'd probably want to share the components they had in common. For the chips, that wouldn't be that hard. E.g. you'd only need one 6502 CPU for all of the different 6502-based families (Apple II, Commodore VIC20/C64, BBC micro, etc). The RAM could be shared; you'd use a modern 64K (or greater) RAM chip rather than a load of 1K/2K/4K chips like the originals.

Where it gets a bit more awkward is things like the keyboard, which is probably wired differently on each computer, so it would actually be quite awkward to just have one keyboard which worked with all of them. It would probably be simpler to have a custom (or semi-custom) chip to translate between the various keyboard matrices than to actually wire each key to half a dozen different matrices.

>>1545892
That sounds kind of exciting and plausible so far.

Would there be inconsistencies with having one ram chip over several ram modules, or did the old computers of that era just not care and treat them all as one big pool?

>>1545816
adding, even then, I almost always end up ordering from 3-4 stores in order to fill my entire list
also adding, if you compose a BOM and send it to sellers, you might be able to get one-stop shopping

>>1545887
a lack of shared resources, mainly. also, bus loading. these old systems were not necessarily engineered to have more than say 10 devices loading down one device's output, or to operate coherently beyond their system's designated bus speed
you'd just about end up building a copy of each of those systems (often including keyboards, whose matrices differed by model) and then switching and/or translating the I/O and perhaps bus cycles between them

>>1545892
>you'd only need one 6502 CPU for all of the different 6502-based families (Apple II, Commodore VIC20/C64, BBC micro, etc).
kek, no, the variants were not necessarily compatible. e.g. the 6510 had some onboard GPIO whereas the 65(C)02 did not. also, some revisions of 6502s may have had different instruction timings. and then there are the undocumented opcodes that became documented and different operations in later revisions...

>>1545896
each of these machines also handled their own DRAM control, address multiplexing, and refresh, usually in the video chip (but the Z80 CPU was different). each expecting a particular size and type and width of DRAM, each of which would need to be translated to whatever you're using for system RAM

>tfw you didn't get paid yet so you didn't send those boards to manufacturing so you had enough time to look them over and see that you put two inverters backwards

>>1545887
check out MIST board/computer.

>>1545865

>>1545590

you cant just stick a diode and a cap on your music and expect it to look good. you need some proper circuitry with a precision rectifier and a well-designed and tweaked sample-and-hold. you can either spend an evening designing your own, or copy one of the millions of designs available on the intar-webs.

>>1545956

>sample-and-hold
make that ''peak detector''

>>1543303
Wire-wrapping is so 1980's.
Also that spool of unjacketed cable is not what you use to wire-wrap with, you use 30ga pre-tinned solid wire.
>children these days

>>1543399
>just build a circuit that requires intrinsically matched and thermally bonded transistors out of discretes
'''Matched''' transistors used to be a Thing; don't know if it is anymore or not, but they used to be for that specific reason.

>>1545999
That spool is literally wire made specially for wire wrapping, they even sell it in sets with the tool
sorry for embarrassing you son, maybe you can sound smart the next time around

>>1545491
CdS (Cadmium Sulfide) cell; a light-depended variable resistance.

Interesting story about these: back in the days of wired telegraph, they'd use CdS resistors in the line; every morning the dayshift operator would have to go out and 'adjust' them. Eventually someone realized they were light-dependent and that's why they kept changing value, so they switched to something else.

>>1545421
top quality b8

>>1546006
*Shrug* doesn't look like it to me. Supply a manufacturer and part number for that spool of wire; I've never seen wire-wrap wire that came in multiple colors twisted into a cable like that then spooled up, it's always been single-color per spool.

>>1545421
>'''resistance'''
>not '''impedance'''

>>1546003
yes, still a thing
the LM394 "super matched pair" hasn't been manufactured for a while tho

>>1546049
Yeah, similar, except what I'm talking about is literally two discrete transistors, NPN and PNP, that are a 'matched pair', tested to be complementary to very close tolerances, usually for push-pull amplifier use to reduce distortion as much as possible. Or so I remember. Such things are a *long* way back in my memory. God-tier engineers can and do still design circuits that way (all discrete) but there doesn't seem to be that many around anymore. The last one of such I worked with/around was at an ophthalamic ultrasound manufacturer. All the signal amplification were discrete circuits.

I had $23 left on some e-gift card so I got pic related. Potentiometer of different values and a buck/boost converter. I’ll be seeing you goys this weekend ;^)

>>1546055
oh, right, I haven't seen those for a long time. (neither the matched pairs, nor the discrete-level engineers worthy of the name)

>>1546060
What exactly am I seeing in this picture here? Also, are the wires soldered like I would solder a DIP onto a PCB?

>>1546067
That's a magnified view of a die, presumably a transistor array.

>>1539340
I want to make an fm transmitter and fm receiver. There are a bunch of tutorials online but they just list parts and tell you how a monkey should arrange them. I want to know how/why it works

>>1546078
>I want to know how/why it works
then you need a textbook.

>>1546081
This. It's not a small subject, really.

>>1546078
Overview:
An FM transmitter is essentially a voltage-variable oscillator being controlled by an input signal, varying the oscillator frequency around the tuned frequency.

An FM receiver is more or less like an AM receiver up to the detector. There is more than one type of detector design for FM demodulation. One popular type is a PLL (Phase Locked Loop) tuned to the FM carrier center frequency. The demodulated signal is manifest in changes in the control voltage to the PLLs' VCO.
>now you're going to ask a bunch of questions about how all that works
That's why you need to research it yourself and do lots of reading, since it's not a small subject.

I suspect that "do you own homework" is appropriate here..

>>1546067

>>1546070
>>1546089
Neat. Thanks.

>>1546067
>Also, are the wires soldered like I would solder a DIP onto a PCB?
No, they're wirebonded.

>>1546088
>It's not a small subject, really.
no, it's not, but it's amazing how much more you can get out of reading a textbook when you don't have to worry about passing some exam.

>>1546003
95% positive that AD still sells matched complementary pairs under MAT name, just ordered some mat12 npn samples; might be overpriced though

>>1545178
if you're in college m8 sign up for sample programs (google). Maxim, Linear, Analog, and so many more. should honestly been in the sticky in my opinion

first time i found a useful indian video
https://www.youtube.com/watch?v=uLVJfRu_6tI

>>1546136
>tfw none of them ship to brazil
just kill me. Being born here was a very poor choice. If I ask for anything more than leds and lm741 (anno domini 2000+20-1) they look at me like I`m a alien or/and retarded.

>>1546136
What are sample programs? Do they just send you free shit periodically without needing to ask for it?

>>1546124
That was a fun read on wikipedia.

>>1546160
nah homie (repeat: google ). if you are a student, many companies will send you free sample, ad/linear will send 2pc IC (x4) once a month

>>1546158
yeah excuse my privelege, have you checked everywhere? even hw manufactures send stuff( bourns, vishay, coilcraft-they are very friendly)
I'm starting new thread with it in sticky

>>1546196
daily bread

>>1546200
For fuck's sake we're only at page 3

>>1545902
> kek, no, the variants were not necessarily compatible. e.g. the 6510 had some onboard GPIO whereas the 65(C)02 did not.
I was about to say that the ones mentioned all used the original 6502, but it turns out the C64 used a 6510. So that would be an issue. Later iterations of the BBC micro (B+, Master) used a 6512, but this was a significant factor in killing off the line, as a lot of code used "undocumented" 6502 instructions.

> each of these machines also handled their own DRAM control, address multiplexing, and refresh
If you were building a modern clone, you'd just use static RAM. Unless you actually need to correctly emulate what happens when the RAM doesn't get refreshed.

The main thing with a modern clone is that chips are so much faster these days. Those systems ran at 1-2MHz, with modern logic, you have so much extra headroom. So there wouldn't be any issues with inserting glue logic into the bus.

>>1546202
yep
at least he did the OP correctly, and as such, isn't a complete and total faggot, just needs to get his zizi under control

>>1546191
Thanks man!

>>1546167
Operating a manual wire bonding machine is less fun.

>>1546707
I can only imagine. That seems like something that would need to be precise everytime else damage the entire die.

Gud or crap?