/diy/ - Do It Yourself

>>2553754
I think that guy needs some prescription glasses.

Where else would you keep your spare solder?

>>2553773
Your sister's cunt.

They say shiny solder is a good sign.

Can any analog wizards tell me if having a buffer then an inverting amplifier will give me better or worse results than using an non-inverting amplifier? Resistor values are not final. May need some DC-blocking caps too, and probably a fair bit more voltage gain.

This is my idea of a microphone preamplifier to replace the usual JFET, using a custom capsule with a charged grid on either side, for electric field shielding. This necessitates a bipolar biasing power supply of some sort, so one grid is at +48V and the other at -48V, or something like that. Don't want to go any higher for the electrocution risk. Ultimately the signal will be fed into an ADC, then go via toslink to a S/PDIF-to-USB IC of some sort.

Why do I need an LDO if I can solder resistors to lower the voltage?

>>2553798
>having a buffer then an inverting amplifier will give me better or worse results than using an non-inverting amplifier?

should be the same, as both have six gorillion ohms input impedance.
however, non-inverting amps sometimes need a cap in the feedback loop if the supply is not bipolar, and that has consequences.

>>2553842
>Why do I need an LDO

it adjusts to changing current requirements, while resistors dont.

>>2553842
Works fine for some loads.
People do it all the time.

would someone mind giving me the cliff's notes on instrumentation amplifiers
I want to build a proof-of-concept digital oscilloscope with a microcontroller, with variable (digitally-selectable/auto-ranging) gain. I already have a basic op-amp circuit that lets me do both over-unity gain and sub-unity gain/attenuation
looking at the standard instrumentation amplifier circuit and transfer characteristics, it looks like it's impossible to attenuate which I would need for large signals

>>2553864
wait a second I'm retarded, R2 can be varied for sub-unity gain
but then doesn't that defeat the purpose of the gain-setting resistor?

Okay can you really look at the schematics like this >>2553798 and tell what the device does?

>>2553872
I remember asking the same question in this thread years ago. It was unfathomable to me that anyone could understand the schematics
now I look at that one and it seems simple even, especially compared to the shit I deal with now

the left part is a model of a piezoelectric microphone
the center op-amp is a unity-gain buffer, and the right op-amp is a low-pass filter

>>2553872
The components are labeled, the filename says what it is, it looks similar to other ones I've seen before.
Without the labels, the filename, and the past experience, I'd have a half-assed guess with a good hour of staring.

>>2553872
the resistor on the buffer looks wrong
not how I'd build it anyway

>>2553850
>should be the same, as both have six gorillion ohms input impedance.
More from a perspective of noise. I hear inverting topologies are often better for noise, but cascading two amplifiers together might be worse.

>>2553866
There are two R2s, so you'd have to vary them both matched to one another, which is difficult. There's only one Rgain. Personally I'd set the R2s for the lowest required gain if Rgain is removed, get it all properly matched, then use Rgain as the variable resistor. Though check that for stability, I hear it's often better to use a voltage divider than an amplifier with underunity gain.

>>2553874
>piezoelectric
It's a custom condenser with three plates, I just used the crystal+ground symbol because there was no three-plated capacitor symbol. Though maybe using picrel would be better. But yeah same-difference.
>low-pass
I'd consider it a gain stage foremost, the capacitor was just to lend stability (more important before, when the circuit's gain was set to 100), with a rolloff after ~20kHz. Again, same-diff. Normally I'd go for 100pF or so, but the smallest MKT caps at my local seller are 1nF, and I want to avoid ceramics for this project. Considering that 1nF is forcing R7 to be 10k and so limiting my gain, I think I'll just ditch the inverting bit and go for a single non-inverting amp.

>>2553886
I think I've seen it used to null the input offset current? It's probably incorrect, rather I think it matters when there's high resistances in the feedback path that would cause a series resistor to be placed in series with the non-inverting-input. The opposite of my case.

>>2553928
Is your whole schtick that you do post wrong info and tell everyone the wrong thing? Serious question.

>>2553952
no

>>2553754
I installed the female front USB motherboard connector on the FireWire port. That means I had -12V on the USB, - because polarity was inversed. Didn't know why the USB memory stick wasn't working so I inserted my quick charge phone when the FireWire controller blew up on the motherboard. I don't care about the motherboard as much but it ruined the turbo power charge on my phone, the thing that charges half the battery in 20 mins. Quick charge still works but in 18w mode where the turbo power is 33w. The thing is that the cable was affected also. It can't do data anymore and charging is almost non existent with it. How could the cable be affected by this? I tried other cable on my 33w charger but it didn't switch to turbo mode. Was that cable special in some way? On the classic USB side it still got only 4 pins. The phone and charger are from Xiaomi.

>>2553798
It will be worse from noise perspective. Buffer will add its own noise and then amplifier will amplify it together with its own noise. Depending on amplifier noise performance and microphone sensitivity a varying amount of dynamic range would be sacrificed. If you want to do amplification you need to do it as soon as possible in signal chain to achieve minimum noise (Friis noise formula).

>>2553754
How do you guys avoid feature creep in a project?
I start any PCB project with a clear goal in mind, but whenever I start to design the schematics, I always start going schizo. For instance I am making a wireless temperature logger

>"what if I can only get this IC in X or Y or Z package? better add all package footprints"
>"what if the LDO is not low enough power? better add footprint for a DCDC regulator just in case"
>"what if I can only get this 8 bit MCU instead of the 32 bit ARM one I'm designing for? better add footprint for this!"
>"i can't get a bare XYZ sensor on the local hobby websites (digikey etc is too expensive in my country). welp lets add footprint for breakout board"
>"hehe i can save so much money if I also include footprints for the accelerometer IC in this PCB (another project)"

I always end up with a PCB with tons of through holes, all different footprints, and all features, everything but the kitchen sink. It also makes my projects drag on for months instead of days.

>>2554038
Gotta learn to stop and say "fuck it."
It's not easy. I like to deliberately force myself to half ass projects that should be quick in order to avoid exactly what you're describing.
Every time I stop myself, "oh I should" NO, FUCK IT, we're making this BARELY work with only what I've got laying around, and only once it barely fucking works am I going to even think about doing anything the right way.
Once it's just barely working, I can start on version 2 if I still feel like it, and I often don't.

>>2554038
Order some chinese takeout and roleplay as a chinese engineer

>>2553773
Do you have a picrel?
One of those shelfs should be enough

>>2554109
Has anyone ever tried soldering their fingernails before?

WTF IS THIS SCHEME

>>2554185
It's the secret plan to the hidden Freemason gold under the Capitol

So noob needing a sanity check. I have this diy oscilloscope shit that i bought to learn to solder and starts on electronics, life happened , i moved to another country and its been 5 years in a box.
now i finally back and doing my pending projects, and after soldering everything and folllowing the manufacturer steps, the lcd looks like picrel
-is the lcd fucked?
-is that i didn't solder properly the connection? or that there were a pin missing in the 2x20 header, but i moved it to something that had no connection to the pcb. maybe that pin needed to be unconnected? but there were no indication of which side goes where in that connector
-could it be that it is fucking cold in the warehouse i use as a workshop? (didn't though about that until i came back home)

>>2554246
The LCD is physically damaged.

>>2554185

>>2554248
thanks, that was what i tough

How difficult would it be to design a system like this for my pumps?
>A ~100w pump that needs to run 24hrs
>It's connected to a solar panel and battery
>during the day it runs off of solar power and the battery charges
>at night or when there isn't enough power from the panel it receives power from the battery
>when the battery is nearly depleted and there isn't enough power from the solar panel it begins using mains electricity to run and charge the battery at the same time.

The last one is the one that scares me, I've seen people run small aerators with batteries and solar but messing with mains electricity scares me

>>2554270
Tons of different systems. No system is too complex just by itself (maybe except the inverter) but kinda difficult to get everything made and working. You will need
>BMS for the battery
>battery charger (from both solar and mains)
>inverter to turn whatever your battery outputs to 120V/240V for your pumps
You can buy all of these things by themselves, but I doubt you're on this board of all places to hear that.

>>2554270
An off-the-shelf solar charge controller would be able to power your pump off solar and battery power. You'd just need something to activate a mains battery charger if the battery voltage drops too low, a comparator + relay would do the trick.

>>2553754
hey lads, how would I go about building an LED progress bar like:
https://www.youtube.com/watch?v=kCVzn4byViA
except without using a microcontroller? Input can be either a debounced button or a clock. (I'm a complete beginner)

>0 Ohm resistor

>>2554468

>>2554185
>that ring shaped trace
ah! thats a mixer, forgot what its really called though, thing looks like it runs at about 27ghz by all of that controlled impedance stuff

https://www.youtube.com/watch?v=K-F_T59ZDPw

what are the little stepper-motor-circular-motion-to-linear-motion linkages on this thing called? What would I be searching for?

>>2554556

look at related video to get names for all parts https://www.youtube.com/watch?v=0BDvbljP4Yk

>>2554312
Look at 4000 series logic. How much help do you want.

Atmega328p-pu, Mouser has 14k on order in QFN.
75k in DIP. Is the hobby market the dominant buyer of these or is surface mount not that popular in industry?

What's "better" quality than a lm386 for audio amplification at similar power output and not much more expensive?

>>2554585
I'm pretty sure industry doesn't bother with DIP microcontrollers.

>desoldering an 0201 smd inductor to resolder in a different position
>twice
wish me luck anons

>>2554633
>>2554585
The industry scoffs and recoils at the sight of THT parts. Even Chinese and Indian workers frown and would rather lose their jobs than solder THT parts
But in all seriousness assembly costs for THT parts are ridiculous. My company wanted to get a small PCB assembled, it had only a single THT connector but the per board assembly cost went from 1$ to 2.17$ from a Chinese firm for 1k pieces. We then reached out to a few indian companies and learned that India is somehow more expensive than China, plus india has like 5 public holidays every month, so the workers are going to be sitting on their asses for a lot of the month. Finally were able to get a "reasonable" quote from a Vietnamese firm

what does the dotted line through the heat switch here mean?

>>2554678
>what

it means a ganged switch.
as in, when you flick it, both contacts are engaged.
you cant do one alone.
i usually do one alone, in case the GF's not in the mood,

Any recs for small buck modules? Will be using them to power some hard drives.

>>2554680
is it possible for fan to be switched on without heat being switched on?

>>2554683
>is it possible

of course: just push the fan switch alone.
the ganged switch forces the fan to go on when the heat is on, in case you forget to hit the fan switch.

>>2554683
>is it possible

CORRECTION
i think it's possible in real life, but it's not possible according to that drawing.
in that drawing, the fan switch is completely useless, which is why the drawing must be wrong.

>>2554685
i think they were supposed to put a dot here

>>2554688

yeah, with the dot, everything makes sense.

>>2554689
Were you struggling to figure it out? lmao

>>2554691
>figure it out

i take no joy in figuring things out.
point out other people's faults is what give me kinetic thrills.

>>2554692
Never look at wiring diagrams for home appliances then. It could result in an erection that lasts longer than 4 hours.

>>2554693

the engineers are playing Quake while the interns draw.

>>2554694
It's only because they sit at a drafting table and draw everything by hand, while the engineers play Quake. They don't get extra credit for dots.

Making a gift for a friend, I havent used KiCad in quite some time.

JLCPCB can't draw my board's outline for some reason and I'm worried I've screwed up the layers.

I'm making a sort of con badge where the reverse will have LEDs that shine thru the FR4, I've excluded those areas of soldermask.

Does this look ok to you guys?

>>2554547
HF wizardry?

>>2553754
>Device requires 5V
>It uses three AA batteries that provide 4.5V
Can I power a 5V fan or LED strip with 4.2V then?

i am building a submarine thing. i need motors and propellers for driving the thing. i am a mechnigger and i have no idea what motor i should use for this. in that, i don't know what motors are waterproof. do you think BLDC motors will be ok for up to 10 meters of water? if not, what motor should i use? also drone propellers seem widespread, can i use those underwater?

>>2554754
BLDC would be the simplest implementation.

>>2554754
Playmobil has had underwater motors for decades.

>>2554743
>Can I

you may.
just expect smaller photons, and diminished CFMs.
if the source is a LiPo, over time the photons with get so small you'll stop seeing them except with a magnifying glass.

>>2554759
Heh. But seriously, I am a clueless tinkerer. Resistors, capacitors and the like are currently beyond me.

>>2554768
>clueless tinkerer.

still, these kinda questions can easily be answered after only 10 seconds of experimentation.
over-volting something could be potentially harmful, but under-volting never is.
after all, everything is under-volted when you switch power off.

>>2554779
>but under-volting never is
Are you sure about that, Anon?

>found 5 brand-new arduino nanos at the bottom of a box in the garage
Today is a blessed day

>>2554779
Actually I was asking in preparation for going to the store. I know nothing is useless in the long run, but I have an insecure narcissist stuck on me like a barnacle. Unless anything I do is undeniably flawless in the first try, he will bust my balls for wasting a few euros and mock me relentlessly. His cope is that since he never tries, he has never failed.
I am currently trying to make a lightup sword that I can do tricks with put I'm having trouble with the power source. I tried a cable down my sleeve but it was restrictive. I tried various designs of battery boxes but they ruined the silhouette of the weapon. Now I'm trying to find a way to put the power source inside the thin hilt. Three AA/AAA batteries in a row are a bit too long. I'm going to grab a couple 3.7V batteries to restore a power bank I found and I thought I might be able to put a 4.3V battery in my sword.

>>2554790
What does the circuit do? Drive LEDs and make noise? A single 18650 cell with a boost converter would work fine. Make the 18650 removable so you can charge it with external electronics.

>>2554790
You can grab a random 18650 battery and solder one of these
https://www.aliexpress.com/item/32930640893.html
Wala you get 5V power, and the 18650 becomes rechargeable via usb.

>>2554794
Even a mid-size LiPo pouch would work, but you should measure current draw of the circuit so you can estimate run time before needing to recharge.

>>2554246
So followup question, i am trying to find a replacement for the lcd, i tried looking for the board number in duckduckgo, ali and amazon but couldn't find a new one, then i tried looking for the number on the ribbon, picrel and still cannot find it
this is the lcd boards schematic https://jyetech.com/wp-content/uploads/2018/07/dso138-schematic-lcd.pdf
where else should i check? can anyone provide me with some guidance? is the lcd connection standard and I can just put something equivalent?

>>2554801
There should be a number printed on the metal shield on back of the LCD. You'll have to pull it off the board to see it. It's probably a SPI interface.

https://www.amazon.com/HiLetgo-240X320-Resolution-Display-ILI9341/dp/B073R7BH1B

My problem comes down to pic related. Top is ideal. Bottom is what i'm thinking about. Would it work the same as top or am i misinterpretating kirchhoff? The load on both scenarios being the same of course.

>>2554805
I would check the metal shield, thanks

>>2554807
No, it wouldn't be the same. In bottom when SCL/SDA pulls low you basically get a voltage divider at VDD, so only 80% of 3.3V will get to VDD, before it drops even further when you start drawing some current. You're creating more voltage drop across 2k resistor. In top example only VDD load creates drop, in bottom you have VDD load + SCL/SDA load.

>>2554801
The board that LCD is sitting on is very unusual, I haven't seen it used elsewhere before.
The LCD appears to be a QDtech QD243701 or equivalent.
You can find them on PARALLEL 2.4" TFT modules, not the SPI versions.
The MAR2406 board is incredibly common and easy to find, and uses this LCD, usually sold as "Arduino 2.4-inch TFT Touch Shield" or similar.
They're not always labeled MAR2406, but it's easy to spot the boards, picrel.

You definitely cannot swap the entire board directly, again that one from your kit is unusual, and the Arduino shield boards are obviously intended for Arduinos.
The LCDs connectors on these small/inexpensive boards are almost always soldered rather than using an FFC/FPC.
Desoldering and resoldering the 37-pin flat ribbon cable isn't going to be fun for average joe who doesn't reflow for the fun of it, extremely easy to rape it.

You might be better off replacing the whole thing, given that the entire kit is $20.

>>2554881
Noice. The part number is 109-13700-00A1. Just throwing it out there.

>>2554805
nothing there at all, i even dismounted it completely, very clear that the lcd is broken
>>2554881
>Desoldering and resoldering the 37-pin flat ribbon cable isn't going to be fun for average joe who doesn't reflow for the fun of it, extremely easy to rape it.
well shit, it is soldered see picrel, i do want to learn to do that to fix other shit, so that's not a big deal by the time I learnt. But I needed the fucking oscilloscope as part of my learning.
>>2554885
cannot find shit about that number other than the wiring diagram of the board, already tried to look for it
Ill ask chang nicely if they sell me a replacement lcd board. I know is not a lot of cash to get a new scope, but i am not in the best financial situation right now

>>2554905
https://docs.rs-online.com/784d/0900766b8120b293.pdf

>>2554914
but isn't that one has 45 pin instead of 37 of the board?

>>2554787
>arduino nanos
They belonged in that box of garbage.

guys help me out

I've this cool voice changer toy. Only the mic sucks, wanted to remove it and solder on something simple with a long cable from in ear buds I had lying around. Opened the casing and who would've guessed the mic is fixed to the other half of the casing. Ripped it off with my retard gorilla hands. Now I don't know where to place the red and black wire again for testing, before I solder some other mic on. I tried to simply hold both wires on spots of which I think they could be correct, gave me only scratching noises, my voice wasn't transmitted. Then soldered it on some pins for further testing, but no luck.

I've no clue about electronics

>>2554927
pin 1 - NC
pins 42,43,44,45 = NC (touch)

where are the 3 extra pins?

I just want a cheap thing that I can use to blast data from PC into the device's USB port, and then onto SPI. What should I get?

>>2554938
I can try to desolder the ribbon tomorrow when i go to the workshop, i'll post a pic.
the board doesn't use the touchscreen, i think, so maybe the small ribbon is not conected at all?
i am guessing but maybe the ground, and the backlight are the other 3 pins? there are 2 and 2 pins in the sides of the board, I thouth they were mechnical but maybe not.

>>2554939
usbasp or usbtiny does that exact thing. pic related is 2 for $12. You will need some code to communicate with it in the PC.

>>2554959
Does this use a FTDI serial converter chip? I want full USB bandwidth. Preferably USB 2.0.

>>2554938
>>2554953
it is everything in the diagram, i am an idiot.
https://jyetech.com/wp-content/uploads/2018/07/dso138-schematic-lcd.pdf

>>2554964
It's just an Atmel AVR, and I don't know what USB spec it supports. My old notes say the fast clock rate is 750khz, which is per bit, so that's probably about 75 khz bytes I'd guess, which is pretty pathetic compared with USB 2.0.

Wait for someone else to come by with a better solution.

>>2554964
FT2232H
specs: https://ftdichip.com/products/ft2232hq/

>think about building something
>realise it's too hard.

>>2554979
>think building something is hard
>build it anyway
>fail and fail again
>learn something
>???
>MOSFET

>>2554754
BLDCs can run with water around the windings, but generally you need to ensure that the winding terminations are properly sealed, and that the bearings can handle being surrounded by water while spinning. Even if you go for a cutlass bearing, you'll still likely be unable to prevent water ingress, even one of those spiral shaft bearings that actively purge water would only work until a certain (probably shallow) depth. So I'd look at motors used in those hydroplane electric surfboards, which I think have been designed with water flowing inside the housing with bearings that can handle that.
The other option (as Hyperspace Pirate alluded to in his submarine kayak video) is to have a way of equalising internal and external pressure, if just into the section where the motor is stored. He suggests using something like a balloon with far more inflated volume than the motor, such that as water pressure crushes the balloon it keeps the motor pressure the same as the outside pressure. That way the bearings don't need to hold back the pressure of the water, which is particularly important for him using brushed motors. Potential downsides include instability; you have more buoyancy when at a shallower depth than at a deep depth, giving hysteresis / positive feedback. If you need sufficient ballast to descend when on the surface with full balloons, then you'll need to dump quite a lot of that when you're deep underwater with crushed balloons. Maybe there's an easy way of cancelling this out, but I can't think of one.

>>2554939
CH341 or FT232

>>2554972
>>2554974
>>2554989
Trash.

>>2554992
tell me the right and true answer, oh wise one.

>>2554992
raspee peeko lmao

>>2554994
I don't know, which is why I'm asking.

>>2554999
There is no answer. Go to /mcg/ and ask them and tell them their answers are trash.

>>2554754
if you don't care much about performance, you can do a magnetic transmission so the motor can be inside isolated from water

>>2554905
You can buy just the LCD module: https://www.aliexpress.us/item/2251832657312930.html
With shipping it costs as much as buying the whole kit again, definitely not worth it.

Alternatively, if you really want to save cash and learn a thing, there is a freely available open source firmware for the DSO138: https://github.com/ardyesp/DLO-138
It's Arduino based, which means you could go in and change things up to use a different, more common, inexpensive LCD.
You can get a 2.4" LCD for about $5/shipped from Ali, or step up to a 2.8" or 3.2", whatever.
If it's 320x240 and has an ILI9341, then it'll be easy to work with.
Parallel would be more similar to what's there now and probably be the easier one to adapt, but SPI would very likely be workable too (with refresh rate trade-offs) and those screens are the more common and cheaper ones.

Fewer connections to make, easy to adapt it, all the hard work would be on the software side of things.
You'd have to modify the firmware to utilize the SPI LCD, which wouldn't be a terrible experience.
Right now it uses a version of Adafruit's TFT library for parallel TFTs that's been ported to STM32 platforms.
More importantly, it uses Adafruit's graphics library, so switching to using a hardware-driven SPI display would actually be pretty straightforward, not much to it.

Do it, save the money and learn a thing, do it.

>>2553842
people that call every linear voltage regulator a LDO are fucking retards
try not being one of them

>>2554185
RF black magic
watch TheSignalPath on jewtube if you want to get an insight into this sort of stuff, the guy is a genius (and I don't say this lightly)

>>2554807
are you intentionally trying to get your communication lines to have crosstalk? you better not be

>>2554728
SHF wizardry

>>2554905
honestly just buy another one, if you get raped on the part cost, shipping, and headache of figuring out pinouts and such, it'd better to just buy another $20 kit

>>2554983
you inspired me to buy an arduino and give it a try even though I know I will fail

>>2554992
Ah you're after the full 480Mb/s? You need the Cypress EZ-USB FX2.

>>2555022
i feel like a baby for trying to make a wifi yagi antenna on a pcb

>>2555039
I know and used the FX3 (successor of the FX2), but no suitable Arduino-like boards for software-only retards.

>>2555041
There are FX2s inside cheap logic analysers, flash different firmware onto one of those.

>>2555033
A journey of 1000 miles begins with a single step. Try to fall forward and wear a helmet. Don't copy streetshitter circuits. Simulate stuff in Spice if you're unsure about something. If you get stressed out or stuck, take a break for a day or switch to another project. Watch videos, ask questions and keep notes.

Good luck, Anon.

>>2555033
You should at least read "Getting Started in Electronics" by Forrest Mims. Pretty basic stuff and it's a short book.

>>2555006
i haven't messed around in arduino in like forever, but I am not actually scared to try.
I would report here when i got stuck

>going to try and make a high-power battery bank bms for some yobbo
>200A
>7S20P
wish me luck anons

>and the rocket's red glare
>the bombs bursting in air
GL and Godspeed, banzai Anon.

>>2554827
Thanks! It took me a while, but your explanation took me the right direction.

>>2555033
I was never able to get any further than rgb LED programming and using an LCD segment display library. Which didn't feel right, I don't like using other peoples' code.
Analog > digital

>>2555126
Read through the library, and reference to the datasheets of the LCD controller and the MCU to figure it out. Comment the library as you go. Understanding it shouldn't be out of your ability, unless the library is some 200IQ akashic tome of ASM.

>>2555126
Just remember that most people writing "software" are retards.

>>2555131
I don't like computer code because it's too serial, I have to think in parallel. Can't understand how parts of a program can communicate with each other if they have to run in series.
Analog is better because it just works.

>>2555138
>Analog is better because it just works.
Analog is great I agree, but it hardly "just works", and it's just as serial as any signal path.

>>2555138
There are languages which can express parallelism. This is actually something programming languages for CPUs are behind (and always were). But look at stuff like Verilog. It can even be asynchronous, even though you will burn yourself.

>>2555133
This is the part that bothers me about it. My only cope is that they're probably as bad at making a circuit board as I am at making a program.
There must be a serious difference in the type of mind that can do coding vs circuit design.

>>2555149
I know people who can do both, and people who can do only one of them. If it helps, the people who can do both could write software that was good at what it was supposed to do, but were mostly unable to follow good practices, and in some cases, were unable not to leave a heap of code-shambles behind. Then again, they were much better than lowly code monkeys who could do nothing else.

>>2555154
That makes a lot of sense. Even if I did get decent at coding there's no way I'd ever be able to explain things in comments.
I made an RGB fade/blinky and cannot explain how it works. Red goes 0-255, green 0-254, blue 0-253, and that causes it to de-synchronize and generate random colors, if that doesn't explain it no combination of words can.

>>2555159
It takes time.

>>2555159
>can't explain
How the algorithm works? How it's implemented? Or does your code only work by accident?

>>2555159
>says he can't explain things
>explains thing perfectly

>>2555165
Teensy 3.2 (arduino clone) using analogwrite, is that what you're asking?
3 variables R G and B, add x (which is 1) to each unless they reach the limits, in that case make x = -x
It's the default fade/blinky demo code times three, I call it swirly

I've got an ancient RadioShack era thermometer, but one of the seven-segment displays is faded. Unironically slapping it a couple of times brings it back to normal, but it eventually continues to fade. Sometimes it also lights up all the segments and does the 88.8 thing that happens when you first put batteries in it. I thought it was a battery terminal problem but the other display functions normally. Thoughts?

>>2553754
Would running a wire system like this blow the fuse? I'm going to run an additional line off a pre-existing one, and while I believe that the fuse shouldn't blow, as there should be no increased load on #2; I'm tragically very retarded, and can't afford to swap the fuse at the moment if it does break.

The voltages listed in the picture are only for demonstrating the problem, and aren't the actual ratings I'll be dealing with.

what sort of 12v power supply/adapter does this use?

>>2555263
>what

most newer laptops have switched over from 2-pin (tip-sleeve) type plugs to 3-pin (tip-sleeve-sleeve) ones.

>>2555248
>Would

if it's not possible to replace a fuse, then dont use a fuse, just let whatever happens happen.
pray the source itself is fused.
prayer is the second-most used instrument by electricians.
after heroin.

>>2555280
how about relying on the thermal fuse of the transformer?

>>2555285
>thermal fuse

that's a desperate last-ditch move.
as in: the ship is sinking, but at least we have life-boats.

>>2555285
>thermal fuse

better options would be
- a resetable fuse; dont have to change it, just press a button to reset it
- a self-resetting fuse; it comes back after a few seconds, covers transitory problems
- big-ass lamp, like 150W incandescent; has very low resistance at low current, limits current when load is shorted.

I'm finalizing the install of a Speeduino into my Miata in the next few days. Just need to finish installing the new wideband. The boards can be built for like ~$115 or so and it uses the same tuning software(Tuner Studio) as much more expensive options. It's pretty impressive how full featured this ECU is, and people are running motorcycles, harbor freight predator motors for their go karts, and a pretty wide variety of engines off of these things. After I get the Miata running, I'm going to build one for my Chevy 5.3 LS project. It's basically a megasquirt but for a fraction of the price and you can build it yourself.

>>2555326
sounds really neat, anon. are the chinks selling them assembled by slave labour for $50 yet?

>>2555330
Unfortunately not lol. Best assembled price I can find is $160 for surface mount.

>>2554933
bump somebody

>>2555331
Bet you could beat that ordering one assembled from JLC. What parts are the most expensive anyhow? That pressure sensor? Nice analogue parts or high-spec transistors?

>>2554933
Chances are one of the pins is tied to Vcc or GND. Either trace the circuit board to figure out which one goes back to a positive or negative battery terminal, or just turn it on and measure with a multimeter. I'm pretty sure mic capsules can't be damaged via reverse voltage, so if the failure of sound continues you may have to reverse engineer the whole circuit.

As for audio quality, you may find that it's just an 8-bit ADC and DAC running at like 5kS/s, which will give ass sound quality regardless of mic.

>>2555330
>>2555331
I think part of the reason is, very few people in the automotive world seem to actually be into these yet for some reason. They've had functional designs since 2015 or so but I think the barrier to entry of soldering your own board is beyond a lot of people, so it's sort of a niche within the car tuning world that many aren't going to discover. Also, a lot of car guys are convinced you need the absolute best, top of the line parts to go do burnouts in your shitbox, so if they don't recognize a brand name, they become pussy crybabies about it.

The only reason I know about it is because one of the main devs was a miata guy. So miata guys tend to know about them a little more than other people in the car community.

>>2555334
The pressure sensor is like $20 or so on aliexpress, which was the cheapest I was able to find it. I can steal a map sensor off of a chevy at the junk yard as well, doesn't even have to be mounted to the board, so that's another way to save cost.

I tried uploading the gerber and whatnot to JLC a while back but I was too retarded to get it to work properly or something. I'll probably try again.

There's a lot of cool stuff you can do with this in combination with a Pi as well. Digital Dash, Canbus Touchscreen etc.

>>2555336
>very few people in the automotive world seem to actually be into these yet for some reason
That's not really a reason for it to be expensive. Sounds like you bought all the parts yourself, but did you have to buy the PCB from them?
>I tried uploading the gerber and whatnot to JLC
JLC can be funny sometimes. IIRC OSHpark might be nicer for hobby board designs? Or was it PCBway? What format are the files in anyhow?
Even if you can't get them in a convenient format, it would only take a couple hours to recreate the whole thing in KiCAD.
>a Pi
Yeah that did look like a pi pin header. Neat stuff.

Do you have any idea if this would work for something as recent as a 2009 eurocar? My pops' VW R36 has this "feature" where if you tap the brake, the accelerator does nothing for like a second or two. It's definitely a lot more complex than a 90s car electrically speaking, but I'm not sure how much extra is on the ECU.

>>2555238
i have similar problems with old calculators... i think it's the zebra strip — this carbon/plastic connector linking the lcd panel to the circuit board. i've had limited success by opening it up and re-pressing them down, doing so with heat, trying to add some tape, etc. the most effective one was to stuff some foam inside the case over the zebra strip such that when you close up the case, the foam keeps pressing on the z strip.

>>2555338
>2009 R36
The new engines add a level of complexity because they tend to have variable valve timing, and sometimes it's a matter of decoding a specific engine's vvt system. I know the 1.8t/vr6 are pretty popular speeduino projects, and I've seen several other modernish engines running off of this setup. From what I can tell, that EA390 in the R36 is a VR6, but I'm not current enough to know what they've altered on the newer versions. I bet it could be done, but then you're likely also going to be figuring out discrepancies between the body control module and OEM ECU which might make the car a pain in the ass to deal with. The older cars are definitely a lot easier for a solution like this because there aren't as many codependent electronics that the ECU and dash are expecting to see.
You're probably better off in that car with just a good tune instead. They can typically turn 'features' like that off and tune for fun/responsiveness instead of for emissions standards. Here's one option:
https://www.getunitronic.com/ecu-tuning/Volkswagen-Passat-36l-FSI-2006-2008

>>2555338
I bought the pcb from them. It was only $10. You can find several variants and custom versions on ebay and whatnot. Ordered parts from Digikey
>file formats
Thanks for the recommendations on the other sites and KiCad and whatnot. I'm not super well versed with any of this but I'm using Speeduino as sort of an entry point to figure it all out. Their github and wiki documentation is actually pretty extensive. Here's the git for the gerber/bom and everything. They've got a forum full of user projects as well. https://github.com/speeduino/Hardware/tree/main/v0.4/THT/Latest

Why there's zero stock on STM32 CortexM7 chips for months or years now?

>>2555336
I would love to rig my old carb clasic jap van into one of those. but I may need to add too many shit, i have like two months to do it bc of well registering shit.
god knows i hate that carb, and apparently there is a throtle body that it is mostly compatible.

>>2555341
Thanks for the pointer. This seems to be a common problem with old LCDs of the era.

>>2555354
do you liver under a rock?

>>2555481
There is at least one bot that posts in /ohm/ regularly.

>>2555481
I just started the electronics you moron. And all of you were explaining computer and car chips shortages because they are complicated and there's only two manufacturers of them on the planet. STM like chips are not complicated and have a lot of manufacturers, yet here we are in 2023 and there's still not chips off shelves, only flippers on eBay for 3x the price.

>>2555335
ah thx but I'm too stupid to figure this out, tried for a hour. Instead I gonna simply use these 2 android apps, mic to mobile, mobile via bluetooth to bluetooth speakers. Beats having only 4 switches and the sound is way better.

>>2553754
I want to build frequency counter for my aviation band receiver kit. Basically an MCU driving LED 7-segment display that can measure frequencies between 108 and 137 MHz. I found this: https://hackaday.com/2016/09/09/very-simple-pc-frequency-counter-works-up-to-100mhz/
but unfortunately having a 10ns pulse time it ends about 124MHz.
What's the correct way to do it? Do I need an 280Mhz MCU? Or maybe scratch this plan and measure the voltage on tuning potentiometer?

>>2554468
Have you heard of jumpers, nigger?

>arduino arrived
>need to find a usb-b lead

>>2555591
What is your input signal that you're trying to measure? You need to find local oscillator that is used to mix down what comes from antenna. It's most likely that it will be less than 100 Mhz.

>>2555595
>using the ardunio programmer
>not using an atmel 8 pin programmer
>using arduino studio instead of avr-gcc or even better tcc to program directly in C
ngmi

>>2554468
>-100 ohm resistor

>>2555326
Have you seen the stripped versions?
There's a fella doing a minimalist version for aircooled VWs, you can slap the whole ECU together for about $35. It's not versatile, being intended only for aircooled VW engines, but I'm still impressed at what's possible for so little.

>>2554794
>>2554796
Thanks for the replies. It looks like I can use it without extra components.

>>2555659
Cool. Just make sure you don't drain the battery too much or it'll damage it. 3V is about the lowest it should get before recharging.

>>2555555

>>2555659
Add a basic BMS. ~60 cents/ea on Amazon, or about 10 cents/ea on Aliexpress.
Cheap as fuck, easy as fuck, saves you from accidentally ruining your battery from overdischarge or exploding the bitch with an accidental short.
An 18650 can go full runaway in a split second, nothing you can do then but wait while it pumps toxic smoke into the room and hope that it doesn't catch fire or set light to anything around it.

Pennies to mitigate the risk is more than worth it.

>>2555615
Does tcc have any advantages? I didn't know they even have a backend for this CPU.

>>2555532
car chips were not that complicated either, until they started putting gaming PCs in them.

I don't buy most of the chip shortage excuse bullshit, i can see it for ultra high density new stuff, but car chps are typically > 100nm and those fabs aren't overworked.

a lot of it is the stupid manufacturing idea from toyota where they keep close to zero inventory so any hiccup results in no cars. you can see some of these ideas mistakenly adopted by software now (e.g. kanban boards) but at least most of that just makes no sense.

>>2555768
I mean, Toyota may have pioneered it, but literally everyone picked it up and implemented it.

>>2555768
>>2555771
JIT supply chain got twisted into a pretzel by the lizard people who waged biowarfare on the world.

>>2555775
yes, they didn't learn a damn thing from the 1993 epoxy plant fire. No second source. I'm not sure the military even requires 2nd or 3rd sources any more. Good luck!

a couple of fabs actually caught fire. one fab didn't have circuit breakers or fire supression in 2021 in an earthquake prone region, and in a modern fab factory (where practically nothing burns) makes you go "hmmmm"

all companies eventually get infiltrated with know-nothing entry-level MBAs —who just scheme all day — and hilarity ensues.

not sure if this is more /ohm/ or /sqt/.
if I'm installing a bunch of ethernet cameras around my house, is there a thing out there that allows me to plug all of the cables into itself and condenses it down to one cable that I can run through the wall, and then breaks it back out on the other side to plug it into the dvr? I'm just trying to avoid drilling a whole bunch of tiny holes or one big ass hole in the side of my house to run all of the individual cables though.

>>2555784
>all companies eventually get infiltrated with know-nothing entry-level MBAs
Don't forget about actual spies and domestic enemies in their respective countries who are actively sabotaging all industries that keep civilization functioning. Sprinkle some brainwashed affirmative action hires/bureaucratic chaff/academic windbags with an axe to grind who do exactly as they're told by their superiors and you have a recipe for the shit cake we're all eating right now. Total corruption corrupts totally, or something. lmao

>>2555786
69 gorillion hours in mspaint

>>2555786
>>2555788
that "multiplexer" could simply be an 8 port hub. or, more likely, a switch nowadays. i just picked one up at goodwill for $7. your wires will be cat 5/5e or whatever is cheapest. this is literally what ethernet was invented for.

what worries me is the "demuliplexer" to dvr connection. if it's PC-like, then no worries, the demux could be done away with. if the are video/bnc connectors then life will be hard.

>>2555786
A switch? You can combine them in any topology, just avoid loops. They will need a power supply.

>>2555798
> needs power supply

should mention that, while expensive, PoE switches should be looked into (mainly for the camera side). you can "hack" (or diy) your own non-standard PoE on the unused cat wires though.

>>2555796
>>2555798
of course it would have to be a poe switch tho since the cameras are poe. and the dvr is a standalone dvd player size unit like you get as a kit, not a computer build. but it is all 100% ethernet, no separate video or power just one single ethernet cable per cam. i just dont know if the dvr is like a switch itself where it can see all of the cameras off of 1 port or one per port is absolutely required.

>>2555802
beat me to it, and yeah id rather just spend the few extra bucks either on a poe switch or poe injectors if this idea would even work rather than diy it just for the convenience and ease of weatherproofing.

>>2555802
>unused cat wires
Gigabit ethernet uses all 8 wires.

>>2555806
But all network cameras I've seen use PoE, so chances are they're not using all the pairs. Doubt they actually need to be gigabit in the first place, though once you combine them at the switch they may well need to be. So having the PoE injection at the switch is probably the way to go.
But then the question is if the NVR can handle gigabit through one of its ports, or if it needs it split among them all. And even then, it may not have the routing capabilities to talk properly to the switch(es).

>>2555806
> gig e uses all the wires
huh... totally learned something today.
i'm able to watch netflix over telephone cable/rj11 without problems and never bothered to pull anything beefier, but i was a bit suspicious about "what a waste" the unused wires would have been on futuristic cat cables.... blamed e.u. standards organizations.

>>2555816
>netflix
Streaming uses highly compressed media over internet. 100 megabit would do, maybe even 10 megabits. PoE over gigabit exists, there's also some weird 2-wire gigabit standard, but probably not applicable to normal equipment.

>CAT5 runs to each camera
>PoE network switch located at the center-most point relative to the cameras (star configuration)
>CAT6 run to the DVR, or fiber if the switch and DVR have SFP ports.
Did I miss anything?

>>2555823
>Did I miss anything?
surge protection & battery backup

>>2555823
a router

>>2555810
>But all network cameras I've seen use PoE, so chances are they're not using all the pairs

PoE doesn't necessarily occupy data lines. Mode B does, but Mode A doesn't. Pic related: The magic of isolation transformers.

>>2555831
unless the DVR is a DHCP server.

>>2555796
>what worries me is the "demuliplexer" to dvr connection

They're PoE cameras. They use Ethernet. By definition, the DVR (technically NVR) must speak Ethernet, and will be able to separate out the individual packets from all the cameras just fine. He'll need a PoE switch to plug the cameras into, but the NVR should have no problem working with just one cable.

There are only two potential issues I can think of:

1.) If they're very good cameras and there's a lot of them, they may spit out enough data to overload a single cable. You'd have to have a shitton of really expensive cameras in order to even come close to saturating a gigabit line, though. With consumer-grade stuff, even an old 100Mbps line would probably be fine.

2.) If they're power hungry, you might not be able to easily power them with a single PoE switch. Obvious solution to that is to get a more capable (and more expensive) switch, but it's something to pay attention to.

>>2554682
bumping
Some modules are really small but have similar power ratings to the larger ones.

>>2555906
> tiny screen
holy shit, that's almost usable....
reminds me of my TI calculator with these tiny led digits with big magnifiers on them so you could make the numbers out.

maybe screens will go the same way

>>2555909
Those screens are used in camera EVFs and such.
That one is 0.96" at 1080p IIRC.

>>2554660
de soldering smd is the easiest thing in the world.

Does DMM counts necessarily define its resolution? For example a 20K count DMM may or may not have a 200mV range, right? It might only have a 2V range and higher. The highest resolution would be 100uV which is meh. Or it could have 200mV and even 20mV range with 1uV resolution. And all of these devices would be marketed as 20K?

>>2555928
i think the term "count" came from the fact that a lot of measurements were made by a frequency counter, e.g. a wien brige oscillator, where one of the bridge components (e.g. resistor or capacitor) was the DUT and it would literally count the pulses and scale the result into the right units. higher ranges would be lower frequencies (swapping out other bridge values with larger ones) and less pulses, so inherently less accurate.

>>2555928
>Does DMM counts necessarily define its resolution?

No? It's literally just how many digits it can show at once. You can get an arbitrary resolution by just moving the decimal and changing prefixes.

The thing is, 4.5/5-digit (and very expensive 6-digit+) multimeters generally are more _precise_ than 3.5/4-digit ones. There's no reason to have that many digits in the first place if the least significant ones have to be discarded.

>>2555928
Generally the only reason to have a 60000-count meter is because you have a 16-bit ADC that counts up to 65535. And 4000 count meters would have a 12-bit ADC. Displays are cheap, ADCs are expensive.
But I have a 6000-count meter, so idk. Not sure about ANENG's 9999-count meter either.

How to learn about wiring up micro controller input?
I want to design a custom keyboard that also has a sensor on it, but I dont know how the non-binary stuff works

>>2555976
Just wire it in?
The details depends on the sensor and micro

>>2555768
So why there's chip shortage on STM? That was my question.

>>2556022
in some cases, insanely large resupply times were given by stm... so much so that they might as well be assumed to be dead, so you might want to start looking at alternatives.

I'm not sure they fabbed the stm32f0 series themselves, it's 180nm.

it's a european company owned by some european governments, so you can reasonably assume that a lot of resources have been diverted to military orders (which also tend to be larger dimensions for reliability)

thats my opinion, there's no smoking gun like the only stm32 fab blew up or something.

>>2555976
> non-binary
i assume you mean analog.

well, very generally, if you assume your sensor is a potentiometer, you figure out what the input voltage range for one of the A/D inputs is, and set up your sensor in a voltage divider that yields voltages *within* that range. The sensor circuit is most likely going to use the common ground and the mcu supply voltage.

>>2556043
Alright what Cortex-M7 alternatives I can get today for a proper price not from flippers?

>>2555978
idk which board ill use yet, but i know i gotta write the driver for it
>>2556050
ohh that makes sense. we gonna have rotary knob and proximity sensor

>>2554682
MP1584

if i have a small servo that is supposed to be operating at 5v and i use 3v instead will that damage it?
When i tried it, the servo still works, it's just a little anemic.

>>2556105
No it won't
Don't expect to control it as good

>>2556105
It will run hotter.

>>2556060
knob as in a potentiometer? Because encoders are not analog
Most proximity sensors are also digital.

>>2556112
wouldn't it run colder if voltage is lower?

>>2556120
If it wants 5V and it's only getting 3V it'll draw more current to compensate.

I don't know if this is the right thread, but y'all know electronics more than me.
Would I get more money selling cathode tvs to the junk buyer, or dismantling em and selling the electronics?

>>2556123
Keep the tubes and yoke assemblies, sell the electronics.

>>2556124
keep the degauss coils and ground straps too

>>2556123
Can you actually get money out of this small time trash?

>>2556124
>Keep the tubes and yoke assemblies
why? are they so worthless noone would buy them?
>sell the electronics.
I'm not sure where I'd sell the electronics either, I only know one junkyard
>>2556126
>degauss coils
>ground straps
?
>>2556127
idunno, haven't asked yet

>>2556128
>why? are they so worthless noone would buy them?
People tube-swap arcade cabinet monitors to replace old screens with burn-in or dead tubes more and more often as time goes on. They want CRTs (fewer every day) and will pay decent money for them, but shipping is a pain. Yokes match the tubes and should be kept together. They're married.

Ground straps and degauss coils are sized to the tube. Instead of having to hack something together, just keep them with the tubes and sell it together to bump the price a bit.

The electronics are really not worth much. There are chink replacements for >30" tubes. Search "universal crt board".

>>2556135
<30"

>>2556135
Keep the neck boards and electron guns too.

>>2556144
and purity rings/ferrite rings/convergence rings.

>>2556123
>>2556128
update: went to the junk guy and he told me to take em appart and get all the copper
>>2556135
I don't know if there are people like that in this shithole country, and if there are they're probably far as fuck

>>2556157
>scrap them for copper
Do whatever you want.

https://www.ebay.com/str/crtsolutions

>>2556122
>If it wants 5V and it's only getting 3V it'll draw more current to compensate.

complete nonsense.
can be investigated and disproven in just 30 seconds.

>>2556172
>DEBOONKED
DIsprove it then. Teach me a lesson.

>>2556173
>teach me

no.
people are more convinced by what they see.

>>2556182
Monke see, monke do.

>>2556118
i suppose so, theres some keyboards that come with a volume type knob but I plan to have several on my board.

>>2556060
> need cortex m7
nxp?
what do you need out of the m7 line? DSP?
instead use an M3 and use a capacitor, inductor and resistor to make a filter.

harvard architecture, hahaha... ngmi

>>2556388
>harvard architecture, hahaha... ngmi
What do you mean by this? Practically everything is Harvard architecture, because it can program itself (maybe except FPGA bitstreams, but even those can do firmware update the least).

>>2556400
> harvard

in the olden days, harvard arch was pretty natural because we put the code in ROM.

RAM became so much faster, we had to copy the code out of ROM and execute it in RAM (von neumann) to get acceptable performance.

That, and some of the fastest algorithms ever invented use self-modifying code.

so, it's just hard to see the point of it today. it survived for historical reaons.

t. 8051 programmer

>>2556416
Harvard vs. Neumann is beyond the question whether you store code in ROM or RAM. I'm not even sure whether the question matters in modern design. Most CPUs can execute from ROM or RAM, even if ROM might be better fir some uses. Hell, even Linux is sometimes executed directly from ROM, because of RAM size restrictions.

>>2556418
> harvard vs von neumann
the ROM reference was just an originating historical fact.
Of course, 99.999% of all my work is done on an emulator (as should most people's work) and not actually burning ROMs, and that was true in the 70s as well.
there was some guy that posted about using tcc - he knows whats going on.
very often i need the flexibility to change or translate the code as it's running, i don't need any historical and artificial limitations from a computer built out of relays.

>>2556426
I usually test on real hardware, because most of my work depends on the hardware, and accurately emulating it is more work than what's it worth. Also, rebuilding my code even with gcc is less than a second. It's possible because I'm as nerd autist obsessed with NIH.

>>2556429
> code/run/test cycle on real hardware
i'll be the first to admit that nowadays that is shockingly easy to do. i wouldn't try and emulate something as complex as a GPU or a sig proc board... hardware is pretty cheap.
but i'm sure the hw guys spent a lot of time in sim before cutting silicon.
intel would simulate the transistor level, it took 2 w eeks to get a DOS prompt.

>>2555960
By precision do you mean accuracy, such as 0.05% +3 instead of 0.5% + 3? Because, autistically speaking, precision is something else.

It's just I thought that the more counts the better but it appears that counts are not the whole story.

>>2555973
Some 6000 meters do not have 600mV range.
ANENG 8009 is a 10000 count 0.5%+3 meter and it does have a 100mV range with a 1uV resolution but that sounds fake. It also skips some mA ranges. ANENG 870 is 20000 and 0.05%+3 it has a 200mV range with 1uV resolution. In theory both of them could have a 24-bit ADC. What makes AN870 ten times more accurate? It is probably not. Plus for lower ranges it is the "+3" part that dominates the error. "whatever tiny percentage plus 3 counts" is such bullshit. I bet all of these 4.5 count meters have 10uV accuracy at best.

>>2556437
I sometimes do FPGA stuff. You can reprogram a FPGA in seconds (though the synthesizing step may take dozens of minutes). But often enough I will run a simulation even then. It's all a question of how quickly you can test your shit. Two weeks is still better than taking months to get he fab to produce a new prototype.

>>2556440
> FPGAs
looking forward to fpga/mcu combos. the ultimate in self modifying code. need a 48-bit cross product instruction in one clock cycle and have enough gate capacity? and voila!

I am trying to modify a positive regulator to make a negative regulator out of it, and here is the modification that is working. But it requires a negative reference. Damn, am I back at square one? How do a convert a positive reference to a negative reference or this is a totally wrong approach? I can't find any examples on google.

>>2556442
But they exist. It's pretty boring, and annoying in terms of loading the firmware (need to load FPGA and CPU images).

>>2556444
found something but what the heck is VN that is connected to V+ input?

>>2556439
Again, the range is independant from the counts. Counts only (theoretically) have to do with the ADC resolution. What ranges you use are dependant on the analogue front-end, i.e. what gain settings you have for the programmable gain amplifiers (or just static resistors controlled by the range switch). And even if you have a 16-bit ADC and a high-gain amplifier, doesn't mean you have sufficiently low intrinsic noise or offset to make a meaningful measurement at 1µV. It isn't uncommon to find shitty ADCs with a bit difference between their actual bit count and their effective number of bits.

>>2556444
Swap the non-inverting amplifier for an inverting amplifier. It may be easier if you also swap the PNP for an NPN.

>>2556439
>By precision do you mean accuracy
>autistically speaking

Strictly speaking, they're only more precise. They're almost always also more accurate, though, because the people shopping for them are generally knowledgeable about what they're looking for. If they see a meter that has 5 digits but only 1% accuracy, they're going to go "lolno" and find something else. Conversely, why would a manufacturer spend extra on a high-accuracy 16/24-bit ADC and then be unable to use it because you can only "see" the first ~12 bits? There's an implicit relationship between the display resolution and the quality of the meter, but that's all it is: Implicit.

>>2556453
I don't get it. Neither of this makes sense especially the NPN configuration. It sort of works from 0 to -2V but then it stays at -2V. Should I connect the reference to the inverting input or should I connect it to V+?

I've got an old UPS that's not working as intended and I'm not sure what it could be, I've got no manual or anything as it's something that's always been at a friend's place. I'm looking to get started into this whole DIY electronics thing because it's been fun so far and I could save some money by fixing it myself.
The thing works when wired to the main (battery also connected inside) but it doesn't work without the battery, had it changed I believe was around a year or so ago. Unplugging the UPS from the main makes the backup LED blink until it beeps before shutting down but gives no power in this state as it originally should.
The battery's listed values are 12v 7amp, tested with a multimeter and it was around 11.5v. Left it charging overnight with a car battery charger and went up to 12.5v (read somewhere it should be around 13~14v) but the damn thing still won't do the power backup function when plugged back into the UPS.
Is there anything else I could check to be 100% it's not the battery? Nothing inside seems fried up, though the transformer inside the unit gets a bit hot when I put the UPS into battery charging mode.

>>2556489
This works but requires an extra opamp to invert the Vref. So it is still a non-inverting config with an inverted Vref.

>>2556489
See pic related. And please adapt the convention of "negative voltage is downmost" such that your NPNs and PNPs don't incite confusion.
Also the OP07 can't handle ±24V supply rails.

>>2556501
thanks anon, so your second circuit is a POSITIVE feedback loop?

>>2556497
sounds like the battery is toast. if a lead acid cell discharges too much—even a single time—it's done. Even self-discharge if it sits too long without a charger.

in my youth, i made this mistake washing my new car with the stereo on and the bass turned all the way up.

>>2556507
They're both negative feedback loops, because they approach a stable equilibrium. Positive feedback loops are unstable. I use the non-inverting input for feedback because the feedback line has been inverted by virtue of using an common-emitter transistor. Signal at the base goes high -> signal at the collector goes low.
Note that common-emitter circuits are a lot less stable than common-collector circuits because of the added gain, you may find you need to add a capacitor across R3. In addition to a capacitor from the output to ground/rails. A capacitor across R3 would limit the high-frequency gain of the circuit to stamp out oscillation.

The diode is just there to drop some voltage, in reality I suspect it isn't necessary since I don't think an op-amp will be supplying current when outputting its minimum voltage.

>>2556509
In my opinion, all cars should come with energy harvesting for trickle-charging. All you need is a little solar panel in the roof, maybe also RF or sonic harvesting assuming it's power dense enough.

>>2556509
Damn, sounds about right; a few months back there was a long power outage in the general area and the UPS was probably used for far too long during that time. Thank you, I'll just buy a replacement, no way I'm trying to fix one of those batteries with nasty chemical shit.

>in my youth, i made this mistake washing my new car with the stereo on and the bass turned all the way up.
Had this happen to a neighbor once, but with the headlights on all night long lmao.

>>2556514
> fix lead acid cell
i've found and been given UPSs with dead batteries, and despite many youtube videos claming to be able to bring them back, i've never been successful, so i'm mostly calling bullshit on them, but it's probably possible if the battery was essentially "freeze dried" from a woking state in lab conditions.

i've got one right now with a single bad cell (i took all the caps off and it emits smoke when i try a high current charge) but the battery is heat welded together and it's a nightmare to try and fix.

>>2556512
>inverted signal goes to V+ so it is stable
mind boggling but makes perfect sense.
i always thought that negative feedback goes to V- and positive feedback goes to V+.
wrong assumption.

>common-emitter circuits are a lot less stable than common-collector
Yeah I believe some LDOs use common emitter for a lower dropout. I played with RC in the feedback before to prevent oscillations.

>The diode is just there to drop some voltage
My simulation doesn't work without the diode at all. It outputs -12v (negative rail) instead of -3v as set by by Vref.
And it is funny that even the simulation is oscillating without the cap and a resistor in the feedback path to slow it down. But dropout is super low, so it is worth messing with. In fact in my simulation it goes almost all the way down to -12V rail since there is no p-n voltage drop in the path. No need to bother with FETs! But perhaps it is too good to be true in practice.

>>2556525
Yeah you're better off dumping that shit in the ocean.
Know a few local places buying dead batteries for recycling, but it's probably not gonna worth the gas.

>>2556527
>i always thought that negative feedback goes to V- and positive feedback goes to V+
In the case of an op-amp that's generally true, but there are op-amps with inverted and non-inverted outputs, plus discrete transistor amplifier circuits. The concept of "negative feedback" is much broader and applies to any system with corrective tendencies, like a thermostat controlling a heater.
>Yeah I believe some LDOs use common emitter for a lower dropout
Some even use a P-channel MOSFET.
>My simulation doesn't work without the diode at all
By "in reality" I meant not in a simulator. If you had an RRIO op-amp then it would be different.
>there is no p-n voltage drop in the path
Technically, the collector-emitter saturation voltage of a BJT is nonzero. For a 2N2222, it's 400mV at 150mA, and 1.6V at 500mA. Probably less at lower collector current. These saturation voltages can be lower than a single diode drop, but typically not by orders of magnitude.

>pic
I think that's a wrong cap placement, by adding it from the output of the op-amp to the non-inverting input you've made it more unstable. Place it between the R6/U2 node and the R5/Q2 node. The other option is having a resistor from U2's inverting input and ground, and putting a capacitor from U2's output to U2's inverting input.

>>2556528
Lead prices have really crashed where I live, but apparently they used to be high enough that it really was worth recycling them.

>>2556558
>wrong cap placement
Ah you right, now it is a true positive feedback.

>>2556527
> 1n4007
why don't you have a resistor between the op07 and the base of the 2222? can the op07 sink that much current?

>>2556578
the current is on the order of microamps. when i said that once i removed D1 I got a wrong result it was probably because the current was too high, even though 4mA doesn't seem too high. But if I replace D1 with a resistor, everything works as expected. But I don't know why adding D1 reduces current from 4mA to a few microamps. Probably because of a voltage drop.

>>2556578
>>2556582
It's about sourcing current, and not being able to go below 0.6V to turn the transistor off. Either diode or resistor will drop the voltage enough for it to work. Personally I'd use a 1n4148. Using a resistor instead intrinsically limits the current range, because the value has to be high enough to drop enough voltage to turn the transistor sufficiently off, but low enough that it doesn't impede current. Honestly you probably have a good few orders of magnitude between those two being an issue, but a diode seems to me like a more elegant solution.

Adding an emitter resistor may make it a bit more likely to turn off properly (though not as much as a base resistor) and it would also lower the BJT's effective gain to make it less prone to oscillation.

>>2556583
Can Vref from DAC be connected directly in this configuration or it needs to be buffered?

>>2556598
If you use large enough resistors, chances are you can get away with a direct DAC output. Assuming it's a monolithic voltage-mode DAC, because I think they already have buffers. A discrete R2R DAC won't play nicely with a load resistance, nor will a current-mode DAC. 10k might be fine, 100k almost certainly. Read the DAC datasheet to be sure though.

how stupid of an idea is this ups? 1 charge resistor, 3 balancing resistors, a high-current schottky diode, and that's it. the input is a 12v smps, so backfeeding is no issue. the hdds running on it can handle as low as 10.8v, so i wouldn't get that much out of the cells (maybe 20%), but it just needs to be enough to keep the system running long enough to tell the raspi to shut everything down

if someone can tell me how long an 18650 will last when kept at 4v per cell for years that would be great

>>2556615
i think even the cheapest chinese recharge circuit is more sophisticated than this.

Anyone have a good source on sprockets/chains or will I almost 100% need to 3d print my own? Checked Home Depot and found jack.
We're talking the size of what would go on a chainsaw, so 2-4"
Also do the chain links for sprockets have basically a universal sizing? I found some good model resources for those at least but it would be far more convenient if I were able to purchase them.
I'm planning on powering the setup with a motor that gets controlled by an Arduino so that's why I'm asking here.

>>2556690
https://www.mcmaster.com/sprockets/

>>2554468
You joke but i can make the same circuit board become a NASA satellite computer or a children's toy using 0R resistors

>>2556698
>Sprockets even mention ANSI roller chain
Thanks anon

ANENG has no relative mode? Boo.

I got this lcd panel and I was just about to attach the ribbon to this when the tiny black handle which I can only assume was there to hold the ribbon down, fell off.

I tried to place it back on, but it just falls off again. How do I fix this? I tried binding it with electrical tape but the panel won't turn on with that.

Unfortunately the "diy" part of /diy is missing here. As long as you do what everyone does you are fine, but as soon you try to do something new or not new but really from ground up (hence diy), then there's no suggestions or help available no matte the topic. Shame.

>>2556887

>>2556884
Depends how fixed you need or want it to be.
The "real" fix is to replace the connector entirely, it's some FFC/FPC common shit, cheap and easy to find, but not a fun soldering job for most people.
If you were the type of person who wouldn't shy away from soldering a new one on, I don't think you'd have even asked.

The clip is broken, obviously, depending on the style it might be replaceable without doing the whole connector.
That is, if you've got the whole clip out, it may be possible to snap a new clip in. Depends on the connector, nothing I could tell you from a photo.
I've 3D printed replacement FFC/FPC clips before, you can readily find models for some common varieties.

The most ghetto dirtiest "get it done" solution is sticking the cable where it goes, then inserting a shim behind it to wedge it in and apply pressure to the pins.
Most often you're looking at slips of paper, construction paper, business cards, cereal boxes, whatever does the job.
Once it's wedged in and seated nicely, you test it. If it's working, then while it's working you hotglue the bitch extensively.
It's important to do this while it's on so you know immediately if you've fucked it, can still manipulate things if needed until the glue cools too much.
If done correctly, it absolutely works, and is pretty much permanent. It's a terrible solution in most cases.

>>2556887
>reee why won't random anonymous people on the internet help me with my specific problem
Nobody owes anyone anything. There are 65 posters in this thread, if anyone knows something and answer is short and simple enough you will get responses. If you want elaborate engineering consulting services you're at the wrong place. Nobody will spend hours researching and coming up with suggestions for some stranger for free, unless it's a really interesting problem and knowledge gained is worth the effort.

>>2556969
Right, just buy a working solution on the internet and pretend you DIY like everyone else.

>>2556969
>I DON'T WANT TO DO IT THE RIGHT WAY I WANT TO DO IT A DIFFERENT WAY AND IT HAS TO BE BETTER
>THIS IS DIY I REFUSE TO BUY CALIPERS SOMEONE LINK ME TO A CALIPER MAKING TUTORIAL I HAVE NO BUDGET
>WHY WON'T ANYONE HELP ME REPLACE ALL MY COPPER PIPES WITH AFFORDABLE LEAD ONES
>I KEEP ASKING HOW TO CHECK THE TEMPERATURE OUTSIDE AND EVERYONE KEEPS SAYING "THERMOMETER" YOU FUCKING ANTI-DIY SHEEP CAN'T EVEN COME UP WITH AN ORIGINAL IDEA
>THERE'S NO SUCH THING AS DIY ELECTRONICS BECAUSE YOU HAVE TO BUY YOUR TOOLS AND PARTS

>BUY BATTERIES AND BMS ON ALIEXPRESS
>CONNECT FOUR CLAMPS
>DIY ELECTRONICS

>>2556977
>DIY TOOL BATTERY GUIDE
>STEP 1) BUY KIT FOR GIVEN BATTERY TYPE
>STEP 2) BUY 18650s
>STEP 3) INSERT 18650s
>STEP 4) MAKE 4000 SPOT WELDS
>STEP 5) LOSE THE SCREWS AND GLUE THE CASE TOGETHER
>YOU DID IT YOU NOW HAVE YOUR FROM-SCRATCH TOOL BATTERY CONGRAPULASHINS

>>2556887
The electronics of:
>buying existing transistors and integrated circuits and microcontrollers
>designing your own circuit boards
>programming your own firmware
>assembling a project
is sufficiently cheap, accessible, and useful that it's gained a lot of popularity. It's a maximised utility-per-cost point. Going further up the chain and paying someone to do all that for you is a lot less efficient, but so is going down the chain and making your own silicon and resistors. While there are a couple of people out there doing diy silicon, the extra utility from doing so is minimal, and the extra equipment required serves as a massive barrier to entry, not to mention the knowledge. If you lack the understanding to implement a novel technique because there isn't an instructables article, then it's up to you to acquire that understanding by reading theory and appnotes and patents.

What you describe as doing "what everyone does" is just the first step, it's honing the physical techniques of construction and troubleshooting, and the understanding of how it works. Once you have those under your belt you can start to modify existing circuits, and even develop novel circuits. Look in this thread, we've got a guy understanding how you can build arbitrary feedback loops with an op-amp, a guy building a submarine, a guy making a strange sort of microphone condenser, that sort of thing.

For most people, diy isn't about becoming independant from existing technology or supply-chains, it's about maximising utility-per-cost. If you disagree, fine you do you, but make your own anti-consumer forum for it instead of complaining here.

>I ONLY HAVE A SINGLE INTEREST IN ELECTRONICS AND IT'S TOOL BATTERIES

>>2557017

>live in Afghanistan
>need battery for angle grinder so I can cut lead-acid batteries open to reclaim the lead and smelt it while barefoot and downwind of the charcoal pit
>no gloves, no respirator, no problem
>open /diy/
>I'M THE MASTER OF LEAD WHY AREN'T YOU SMELTING LEADBAREFOOT LIKE ME
>WE HAVE A SPORTSCAR AND $88 BILLION IN U.S. MILTARY EQUIPMENT
>starves to death

Here is real diy. literally diy. I really like this binding post but I don't think it is meant to be mounted on a metal box since it lacks insulation. And the inside rod is way too short.

So I decided to put a heat shrink tube on it.

And another layer for good measure

Adding some washers, a plastic one goes on the inside. And I need to discard one of the two green inserts to make the threaded end longer.

>>2557133
>>2557136
>>2557140
>>2557145
nice diy

>>2557149
thanks, i hope they don't short out some time later lol.

>>2557154
Did you lick it to check it's not shorted?

>>2557157
taste the painbow

>>2557157
i think i am supposed to push 1000v volts through it or whatever high voltage test they are supposed to pass. now i need to build a 1000v pulse generator. or should it be 10kV?

>>2557174
Hook a MOT to it, nothing bad will happen.

i have a seemingly very strange problem that is making me insane. i have exhausted google so i hope that /diy/ autism can solve this mystery.

>teensy 4.1
>BK-SIM7600 modem breakout board, communicating over UART
>logic level converter

the problem is unreliable and corrupted messages coming from the SIM7600. i am sending simple AT commands from the teensy TX, which the modem responds "OK" on the teensy RX.

>around half of all received messages are totally or partial garbled - they are either nonsense, or something close to the correct command, with a few corrupted characters
>tx appears to be working fine, because i can interpret the garbled replies as being exactly as i'd expect, but just somewhat garbled

here's the really weird part. when i put my finger on the RX pin on the teensy - it suddenly works and 100% of messages are uncorrupted. as soon as i lift my finger, the problem returns. attaching my oscilloscope (or touching the RX pin with all kinds of inert but conductive objects like a screwdriver, peice of wire, paperclip etc.) also seems to "fix it". i've tried swapping the modem, swapping the teensy, swapping the board, and the same thing happens. what causes this kind of effect? is the wire/screwdriver/finger acting as some kind of inductor?

>

>>2557284
Shorten the wires. Lower the baud rate. Use twisted pairs.

I am doing something even sillier, the ultimate DIY within DIY. An atomic level of DIY.
I ran out of angled JST connectors, I only have straight ones. But luckily I have right angle headers. So I am going to trim and solder them on the JST contacts and make a right angle JST connector. Hey at least I am not going to bend the pins before soldering them since they are pre-bent! That would be even more diy-fer!

>>2557308
Or wait maybe I could pull them out and replace them with the header pins? Hmm. Trying to be creative. I am afraid I won't be able to sit them properly and they will get loose.

>>2557284
Do they share a common ground?

>>2557133
Those ones are insulated. You just have to drill a hole for this pink diameter, with the panel clamped between the two green washers. Include the D-flat if you want it to be easy to tighten and loosen the binding post without the whole thing rotating and coming loose.

>>2557154
nice fix. i would have just cut a big rectangle out of the front and put in a piece of PVC or other plastic over it. then the binding posts go throuh the pvc area.

Where can I source trackpoint modules besides sprintek?

>>2557133
>>2557351
>pink diameter

to clarify what Mr Pink Diameter is saying:
one or both green washers have a shoulder, whose height should more or less match the width of the metal case.
so if you drill/file a hole that matches the shoulder, the washers will keep the post isolated from the case.
no need for your silly hack job.

What tool do you guys use to choose impedance matching networks for antennas? I'm talking about those softwares with Smith charts.
Also how do I calculate the values manually (just for my info)

>>2554038
Anything worth making is worth making twice, so save it for the next revision.

>>2557360
Old Thinkpads.
Alternatively you go full DIY and recreate the concept with a magnet on a stick embedded in silicone so it can barely move, with 2/3/4 hall-effect sensors.

>>2557382
>magnet on a stick embedded in silicone so it can barely move, with 2/3/4 hall-effect sensors.
Not that anon, but that's how they work? That's really quite easy to diy.

>>2557394
They work with two pressure sensors under the stick.

>>2554181
You need an exhaust fan. You seem to be breathing in too much solder lead fumes.

>>2557293
why? what kind of magical phenomena is occurring?

>>2557318
yes, shared ground

>>2557394
No, that's just the easy-to-implement DIY method.
Real deal uses resistive strain gauges.

They can also be found as replacement C-sticks for the Nintendo New 3DS.
Not the normal analog stick, specifically the C-stick.

>>2557418
Not sure I can complain but noise for the most part. Baud rate too high causes all kinds of weird parasitics to come into effect even a slightly loose connection could cause bit flipping hell even timing issues. I couldnt say for sure but welcome to signaling and communication.

I have a fan i want to control through a transistor from a 6V source, the transistor is a ph2222A.

Pic is datasheet, what's the beta?

>>2557705
By 6V source do you mean for the base or the fan? How many amps is the fan going to draw at the voltage you give it? Why not just use a MOSFET?

Don’t forget to put it common-emitter topology, to use a base resistor, and to use a freewheeling diode.

>>2557760
in the colector, for the fan. The fan just indicates 0.18Amps.

>freewheeling diode
checked

>base resistor
that's why i'm trying to understand what would the beta be,

>>2557351
>>2557372
Yeah I see what you are saying, it probably could be done but I am not sure if you are serious or you are messing with me.
I salvaged these terminals from an old breadboard. They don't meant to be mounted on a panel. This stepped washer is ridiculously thin and the rod is too short. And drilling a large hole like that is quite an effort. Here is the proper terminal for comparison. Notice how long the shoulder is on the washer that goes through the hole and the inner washer goes over it on the other side. And the hole is much smaller, easy to drill in one go. And the rod is much longer so there is enough room for a couple of nuts and a lug between them. I stand by my silly hack lol.

>>2557762
I’d assume minimum current gain of 50, aiming for a base current of 3.6mA.

>>2557798
The thin step is meant for sheet metal, though it isn’t really a bad thing if the step is significantly smaller than material thickness, it’s still enough for solid clamping since most of the force is axial not radial. The large hole is for high electrical isolation.
In my opinion the D-shape that lets you prevent rotation makes that sort superior.

Damn, I bought a bunch of oscillators but didn't realize that the better ones (1ppm) are TINY. Not your typical SMD7050 that are easy to solder manually but a tiny SMD2550. Can I mount them upside down so I could solder wires to them easily? Does orientation matter for oscillators?

>>2557813
Totally fine, no issue. Might affect heat mitigation for some components, I just can't imagine that being a problem for a teeny oscillator that probably wastes less than 10mA.

>>2557813
coil your wires into vibration-dampening springs, anon!

>>2557806
>current gain 50
How did you reach that value?

I tried with a 1k resistir for base and I got something like 450 micro amps and 100mA at emitter.

>>2557293
>Use twisted pairs.
Doesn't that help only with differential signaling?

>>2557999
>How did you reach that value?

not him, but when you design a circuit and share it with the world, you wanna make sure it'll work with any specimen of the specified transistor.
so you gotta take into account min/max values for various parameters, not typical values, or an individual value you've measured.
50 is a prudent number.
one should practice prudence at least once a decade.

>>2558017
but the datasheet >>2557705 says it should be between 100 and 300 for ic = 150mA and Vce = 10V.

>>2558043
>ic = 150mA and Vce = 10V.

but your supply is only 6V, and when the transistor is on, Vce is likely to be around 0.2V.
so your closest condition is one line above, which claims a min gain of 50.

any way to repair LED lights? i don't know for sure if those can be repaired but
>buy shitty taiwanese LED in 2013 when they were the new thing in my country and i wasn't ready to drop big bucks on one of them
>install it outside despite it not being rated for that use
>works
>shell cracks
>works
>install it elsewhere
>still working as of today
meanwhile
>buy any led in current year, brand name ones
>they crap out in like 2-3 years, 4 years at most
this looks sus. granted, the chink LED was much less bright but even so i think they all smelled the flaw in letting customers use products which will last them tens of years.

>>2557873
>>2557943
my main concern was that turning them upside down would result in a 180 phase shift but the phase shouldn't really matter.
and current is only 1.5mA max. should be fine.

>>2558052
>they crap out in like 2-3 years,

if you read the fine print you'll see that the expected lifetimes on the package refers to lamps in free air.
most people have them inside lil canisters facing downwards.
heat flows upwards and degrades the electronics.
worse still, they often cover the electronics in silicone, or similar, thus trapping even more heat around components.
when i take dead ones apart, the LEDs are always good, the electronics always bad.

How does a DJI drone send 18 Mbits/s of video and control data (1080p at 30 fps) over 10 km in straight line using 5.8 ghz and no sticking out antenna while having complete weight of 250g and low power transmitter?

>>2558008
>Doesn't that help only with differential signaling?
If the wires are unshielded, pairing Rx with a ground wire and Tx with a ground wire will reduce external EMI even though it isn't differential mode comms. Adding some resistance to the Tx line at the MCU could help with impedance mismatch, but it has to be calculated.

>>2558052
bigclive breaks down LED lamps all the time, but usually you can't put them together again. Check out some of his videos

>>2558057
>15k hours
>1.7 years total run time
hmmm
btw i don't have any canisters or shit
they're exposed to free air
>when i take dead ones apart, the LEDs are always good, the electronics always bad.
you mean the chips are fine but the rest of the shit is not? it is exactly what i was thinking.
>>2558071
great starting point thansks

>>2558052
>they crap out in like 2-3 years, 4 years at most
The only LED light that died on me caused a fucking mains short circuit.

>>2558088
dropper capacitor?

>>2558059
well for one, radio signals work a hell of a lot better when one antenna is the air and the other on the ground. secondly 5.8ghz is beyond the ism band and also very specific. likely theis something to do with that band that makes it really good for drones/open areas. i suspect also they are using a 1W-3W or higher transmitter. this can still be relatively small size but increase your OTA distance by a significant amount. ive not investigated drones much, but this is my intuition. take it or leave it

>>2558050
>so your closest condition is one line above, which claims a min gain of 50.
Ok, that's what i wanted to know

When dealing with wall power, does it make a difference if the switch goes in the phase line or the neutral line?

>>2558115
>not wanting to have a life wire, that doesn't seem to have power, because connecting a device to it and the neutral wire doesn't do anything
>not wanting to get a critical shock when you touch the life wire and current flows though you to the ground
Answer it yourself.

>>2558115
Never break ground or neutral. Switch the hot side so your circuit breakers and interrupters/arrestors work properly.

>>2558121
>>2558126
Alright, I was just checking.
This is the final result, is everything ok? Its meant to be controllable wall plugs, the last slot on the connector goes to a wall plug, I'll also connect the neutral and earth circuits to a plug, just don't know what to use yet.

>>2558151
Looks okay to me, but triple check it anyway. Put a junction box next to it and secure the Romex, then split off to the wall plugs.

>>2558158
>junction box
shit, good idea.

>Romex
wtf is a romex

Forgot i have a terminal block to connect the wall plug wire to these.

>>2558171
>wtf is a romex
pic

Just realized you're plugging this in to another wall plug, sorry. Make sure you don't overload it. Just run your hot to the Wago connector and splice neutral and ground straight to the plugs. Secure the power input cable with a clamp and strain relief.

>>2558171
You could rip the IEC input off the back of a ATX PSU and mount it to the rear of the plate. Then use a computer power cable to plug it into the wall.

>>2558177
>overload
This shouldn't cross the 300W barrier, the phase wires are rated for 2300W or something, but i might change them later.

>Wago
Almost spent more than double for those.

>Secure the power input cable with a clamp and strain relief.
I'll go buy all of that tomorrow, the clamp goes on the inside of the junction box?

>>2558151
Let us know whether you accidentally killed yourself once you're done.

>>2558184
You don't need the junction box unless you want to use the wiring in your walls to connect it. Search "junction box strain relief" for a better idea.
>>2558185
What would kill him, provided he doesn't stick the wires up his ass?

>>2558184
Mount the relay board on nylon standoffs. Keep clearance in case of arcing.

>>2558183
That'd look cool, but I don't have anything like that laying around. Pic is what I'll use.

>>2558188
I figured it might be cool to have the relays and connectors.

>>2558190
Anything I might find around home that might to the job?

>>2558197
Anything non-conductive that you can drill through and cut to the same length. 5mm should be the minimum clearance in all dimensions.

>>2558197
>Anything I might find around home that might to the job
>>2558201

>>2558197
That terminal block makes my asshole quiver. Use something else.

>>2558201
Was thinking of cardboard, looks janky but whatever

>>2558205
Like what?

it's taken 30 minutes to simulate 7ms for this circuit
I know the switching frequency is high, but what on earth is making such a simple circuit that slow

>>2558206
Use a clamp near the Wago connector.
https://www.essentracomponents.com/en-us/s/cable-clamps/aluminum_metal

Use a strain relief bushing (pic) near the hole where the power cord comes in. Run the hot conductor straight to the Wago connector. Splice the neutral and ground conductors into the plugs.

>>2558207
Your secondary side is floating and simulator doesn't like when things don't have clear connections to ground. You should put ground on secondary side, because mathematically floating voltages don't make sense and it will force simulator to start adding really high resistor values to ground on various nodes in the background.

>>2558211
>splice conductors into the plug
Is this really better than just using a terminal block?

>>2558227
Yes. It's better (safer) than using the terminal strip in your pic.

>>2558227
Never use a terminal block or wago connector or ANY electrical connector where it's not necessary, every connection offers new fun ways the whole thing can fail.
Obviously connectors are necessary for all kinds of shit, but you should only use them when necessary.
Terminal blocks are shit to begin with, twice as shit with stranded wire.
A splice would be safer, offer lower resistance, and would be far less likely to an hero down the line.

>>2558052
IIRC you can buy "dubai style" LEDs in stores these days. Watch clive's channel as the other anon said, he showcases them but also shows how to reduce the brightness of LED bulbs in order to make them last a lot longer.

>>2558207
ground the floating bits

>>2558052
Cost savings shit, fewer LEDs run as hard as they can handle.
Less efficient, won't last as long, but those are problems for the customer.
To the manufacturer, it's just fewer parts, less complexity, making them faster, easier, and cheaper to make.
Early chink bulbs, especially early COBs, they used a SHITLOAD of "meh" LEDs because that was the cheap way to get that amount of light output at the time. Even running those "hard," they'll last considerably longer than a modern high-powered LED.

All of the good stuff follows the same strategy as Phillips with the Dubai lamps.
Lots of LEDs running in their ideal range, so it takes more LEDs to get the same output, but the overall efficiency is way higher.
Higher efficiency means less waste heat being produced, and a longer lifespan for the lamp as a whole.
A cheap 60W "Equivalent" LED bulb puts out 800lm and uses 10-12W, good ones use <7W, best I've seen reach down to 5W.
Those 12W lamps are fucking trash regardless of who made them, it's raping the LEDs to waste that much power, that's comparable to good CFLs (50-70lm/W).

>>2558232
>>2558234
Alright, should I use solder? I'll finish with a thermo retractable sleeve.

>>2558294
https://youtu.be/evrw0ojvzz0?t=50

>>2558294
Use insulated butt splices with a crimp tool.

>>2558207
floating nodes (add a 1meg resistor to ground somewhere) and overdefined diode circuits are usually the problem

>>2558314
that's a yes

>>2558316
crimping tools are out of stock around here

I am about to do some chemical electroplating but I realized I do not know how to safely dispose of Thiourea. I can't seem to find an answer anywhere either. Does anyone happen to know?

Does anyone recognize this thingy? It's connected between a 18650 Li-ion battery and a bt speaker. The speaker works for no more that half an hour before shutting itself off and I believe this here is the culprit, given I've tried all the 18650s I have and none of them seem to drop even a tenth of a volt after the chink voice tells me the battery is low. Am I wrong, though?

>>2558364
You could read the following:
https://pubchem.ncbi.nlm.nih.gov/compound/thiourea
Particularly section 13.5.4
But honestly, considering the low volumes and concentrations in household silver polish, it’s probably fine to just wash the small amounts into the sewage system. At least so long as you keep the volume of it in the same ballpark as what you’d wash off after dipping some silver in it. As I mentioned earlier, if you use a brush then you can really reduce the amount of thiourea you use.

It’s also an option to decompose it with heat or chemicals.

How to make a powerful cathod ray tube which I could use as a laser but with the electron beam? I figured it will need around 60000 V.

>>2558447
Look up how far an electron of a given energy will travel through free air. Also you’ll need one of those funky electron-permeable membranes, Applied Science showed them off a few years ago.

>>2558449
What air? I'll put it in vacuum chamber obviously.

I've watched all videos about cathod ray tubes on YouTube. Yet none explain how it's constructed in the first place. Everyone gives a general overview, but I'm interested in details.

Do you guys actually read books like practical electronics for inventors or just consult on a topic whenever you need?

>>2558453
books are inflexible and overrated

>>2558470
>books are inflexible
just buy paperbacks, broh

>>2558451
By “use as a laser” I assumed you meant “point it at things”.

Actual construction is just a hot cathode with accelerating grid, followed by a few focusing coils or rings, and a current return conductor near/on the target. I’d recommend looking what electron microscopes and e-beam lithography units use, maybe even looking into patents.

>>2558451
https://www.youtube.com/watch?v=lxQS58t39_U

You gotta blow glass cocks.

>>2558432
Looks like a DW01 chip, as paired with dual-MOSFETs like the FS8205, which is probably what the second package is. It's a very common protection chip for a lithium ion cell, though to be honest it lets the cells go down to 2.4V which is really pretty shit. You can buy lithium ion protection boards with those chips on them for very cheap, though they may not be in the same form-factor.

The wire soldered to one side of that resistor is kinda strange though.

>>2558451
A long, boring video about rebuilding tubes:
https://www.youtube.com/watch?v=W3G7b-DcOO4

>>2558447
>I figured it will need around 60000 V
1kV per inch of screen size (diagonal) is normal. The beams are controlled by the yoke. Then it hits the shadow mask or aperture grill which guides electrons to the phosphor coating.

>>2558451
Don't give yourself cancer with x-rays.

>>2558767
I'm not planning to sit inside the vacuum chamber though.

>>2558744
> video about rebuilding tubes:
I've seen it, they still refuse to explain in details how exactly the electron gun on picture is constructed.
>1kV per inch of screen size (diagonal) is normal
I won't need any screens, I want it to weld metals.

>>2558506
> Actual construction is just a hot cathode with accelerating grid, followed by a few focusing coils or rings, and a current return conductor near/on the target. I’d recommend looking what electron microscopes and e-beam lithography units use, maybe even looking into patents.
Thank you.

>>2558786
>constructing a DIY x-ray emitter using a multiple kV electron canon to weld metal
I wonder what could possibly go wrong.

>>2558789
Alright mom, I will sit behind a lead plate. Is it OK now?

>>2558793
Well, I don't actually care if you irradiate yourself.

>>2558795
What else you think might happen? Maybe blow fuses in the house.

>>2558796
It's more likely you'll shock yourself and get some nice high voltage internal burns.
Calling 911 won't do you much good at that point, doctors can't replace the nerves that were vaporized inside of you.
That's just the extremely dangerous high-voltage source, which will be running at current levels high enough to instaban you from a distance.
Once you've actually got the functioning electron beam, forget about it.
There's good reason for the incredibly thick heavy lead-bearing funnel glass of a CRT monitor, outside of well-engineered sealed containment an electron beam is insanely dangerous and will readily kill, plenty of people have died this way working with electron guns.

What you're considering building is dangerous in ways you absolutely aren't grasping.
You won't even be able to judge the danger, you won't be mapping the emission from all points of the case, you probably won't have it in an appropriate case, you won't have interlocks, compliant mounting, warnings, alarms, isolated safety monitoring, or appropriate labels.
You'll be in violation of a handful of federal regulations that are in place to prevent people like you from frying innocent people by accident with their stupidity.

This is a slow, painful, expensive Darwin Award.

>>2558786
Here you go buddy: https://www.youtube.com/watch?v=v7_61DqXchI
This is an example of "pretty safe."
Not safe, but "pretty safe" nonetheless.
Doesn't meet federal regulations, not even close, violates a shitload of FDA guidelines, but hey it's not a commercial product.
Beyond the typical tests and warning labels, it's a well thought-out machine.

>>2558854
Is his goal of using it to weld metal attainable at all, even if dangerous?

>>2558865
Yes absolutely, we've been welding with electron beams since the first viable methods were developed in the 50s.
It's mostly used for high precision tight tolerance shit, putting tiny welds on tiny parts with high precision and repeatability.
While that's the typical use, you can absolutely go bigger with it, but the practicality drops off in a hurry.
You're welding in a vacuum chamber, the difficulty and expense scales rapidly with size and power output.
You're also welding with extreme precision, whether you like it or not, there's no "defocusing" an electron beam, you're whipping it around at high speeds like in a CRT monitor or electron microscope.
Neat stuff.

>>2558872
How much power does it need compared to a classic CRT TV? The light a CRT produces isn't even very bright, even if it's spread over a large surface.