>>2999663
>Nucleo.png
Why are they still insisting on the out of date Arduino headers?
They could produce their boards cheaper by making them long bois, remove the Arduino headers and give us a a single row of pins.
Easy to do for the 64 variant, the 144 one would be a challenge since they would exceed standard breadboards / perfboards. Possibly just put the spare ones at the end.

>>2999761
Idk. The fact that Arduino anything has become a standard is nothing but a hindrance.
>bootloaders
>using any programming interface that doesn't also allow for debugging
>paying for programming hardware on each dev-board
>UART printf debugging
>dumbed down and non-performant programming abstractions
>fuses being an afterthought
And somehow they managed to release a dev-board with a socketed microcontroller, while having a native USB microcontroller on that same board explicitly to act as a programmer, while still requiring a bootloader on the socketed microcontroller. What the fuck?

I get an idea for a project, do some research, pats, schematics, software, looks challenging but doable. *checks price for key part* shiiiiit way too expensive. I want a sensor module that detects a narrow range of SWIR, the whole camera is like $5k, but if I look to buy the sensor, connector, board I might be able to get it for less, but then I do not have the case required to attach the lens, and I would need to have the power and POE connectors, are they plug in or need to solder, then loading software. I keep running to this issue where I think of a fun project and then its way too complicated and complex. How to get past these types of roadblocks?

>>3000029
Learn a better intuition for what components of a project are specialty hardware and are likely to be expensive. Optical components usually fall under that category, unless you’re lucky and can find a band-pass filter for your IR wavelength to strap across a common optical sensor. Or want to make rugate filters at home like Applied Science did, man that shit is fascinating.

>>3000034
>like Applied Science
I will look into it.
When I check some websites for the parts and its "call for quote" I know that shit is too expensive. I dont want to pay it in case it does not work as I anticipate, or I am too lame to get it to the finish line, but id I can get it to work I am confident I can make some good money with it.

Is the CH32V003 still the top choice for a dirt cheap microcontroller, with decent hobbyist-tier support?
People still say it's 10 cents, but I'm finding at least double the price in most places (for 2 to 3 digit quantities). I've seen passing mention of a few other chips in the years since but don't know if any have seen wide adoption.

>>3000054
The minimalist we act dev boards are like $2.xx. i think you can grab the chips for about $.20-$.30 in 10 packs direct from wch.

How do I make my cousin's hello kitty toy piano into a functioning MIDI keyboard with the rp2040?

how hard is slapping power, a microcontroller (the chip itself) and routing only the necessary pins for the project? is it achievable without a master's?

>>3000759
Depends on how many board revisions you’re willing to accept. Probably going to take a month or two, either of learning the principles of electronics sufficient for designing an MCU board, or of brute force trial and error. Post your prospective schematics and board layouts here for anyone with spare time to critique if you want, the general is slow enough now that everyone can ask a clanker their embedded programming questions.

>>3000813
>now that everyone can ask a clanker their embedded programming questions.
And just like that, my years of experience are nothing special anymore.

>>3000816
Hey I’ve been googling my way through arduino projects this whole time anyhow. Now it’s just faster.

I think there’s still a lot of room to use your discretion to write efficient code, minimising clock cycles and/or memory usage, where the AIs trained on pajeetcode need to be really coached into doing so. Max this out by going to assembly language, or just specialising in an obscure architecture or subdomain. Perhaps getting the most out of the Parallax Propellor, or learning a specific RTOS, or going for state machine bullshit. End-game is VDHL or whatever.

>>3000824
>state machine bullshit
Come again?
They take a lot less machine instructions than their apparent verbosity makes it look. Reading through one is like a well organized user guide. I have rarely seen one in the "wild" except the ones I write whenever I can if things get complicated and breadthy with if elses.

>>3001034
Never attempted them before, they just seeem like another more obscure type of embedded programming, another way to differentiate your experience from slopcode. If I understand, they’re meant to be robust programs that can never get into an invalid state, though Im unsure how that works alongside interrupts or multiple cores or DMA. Actually seems quite fascinating, probably a lot of pencil and paper to work out a flow chart.

>>3000824
Claude can do verilog without trouble.

>>3001041
-state machine
-interrupts
-using DMA
I assume thoughtful orchestration would be key. At first glance I'd consider putting values of variables inside of a LIFO from within the interrupt. State change code may refer to this queue when thinking of the variable's value.
I too have not done it yet.

What's this talk about state machines? I only did explicit state machines in verilog.

>>3001103
The main difference is that in Verilog you're describing hardware registers and combinational logic instead of software running on a CPU.
The key difference is that the state variable in Verilog is a flip-flop (hardware), not a variable in RAM. Every clock edge physically updates that register.
The biggest mindset shift is that everything in the combinational logic exists simultaneously in hardware, while the state register changes only on the active clock edge. That distinction—combinational logic versus clocked state—is fundamental to FPGA and digital hardware design.

>>3001165
Thanks ChatGPT.
So what do they mean by making a state machine in software? Defining an enum with a lot of states and having a huge switch statement that gets run all the time?

>>3001166
>a lot
>huge
Hopefully not.
But can you ask a better question?
The C Programming Language uses an FSM by the way in one of its lessons on writing C.

>>3001173
>But can you ask a better question?
My question is just what they mean by using a "state machine bullshit" as if it's some novel, unusual concept. State machines are a fundamental concept, but just look at the conversation above and how it's in the context of AI.
>The C Programming Language uses an FSM by the way in one of its lessons on writing C.
Which chapter?

>>3001174
Ritchie makes you write a program where you must check if code has matching opening parenthesises and closing ones. He proposes an FSM for it. The calculator exercises after that may also use one but I don't remember, it's been 10 years since I read it. Maybe chapter 6 or after that.
Computer Design and Digital Architecture is where I saw it first. Little story at the time I wasn't focused and when we went through this chapter I didn't really get the concept and I regret it a lot. I later had an epiphany rediscovering the FSM with Ritchie's book. In the former book the FSM was looking at bits bit-by-bit trying to detect a sequence of bits. States may have been starting, initial detection and continuing to detect. 2 states. If it started to recognize a pattern it stayed in the latter state, if it saw a bit that didn't finish the pattern, it went back into the initial state. Something like that. At the time, the concept of drawing it on the board or on paper was very important, as it should be. There was also no code involved in that class.

>>3001174
>what they mean by using a "state machine bullshit" as if it's some novel, unusual concept.
Probably an anon wanting to get comfy and read opinions without googling about key CS concepts coming from other knowledgeable anons. I've been guilty of that in the past, I'd be the lat one to blame them for it.

>>3001103
>>3001195
Yeah just things I've heard of and inferred were sub-fields of embedded programming. Seems like state machines are useful when you want a very well understood system with no unforseen edge conditions, like traffic lights. Implementing these in direct logic gates is obvious, but it seems reasonable to extrapolate the concept to ASM on a microcontroller, where the program counter contains blocks of your finite states, and advances to different states with gotos or whatever as needed. Or a switch statement I suppose, but I don't know enough about how those compile to know how robust they are against glitches or whatever. It's probably fine, the compiler has never failed me so far.

>>3001174
>in the context of AI
AIs probably have a harder time writing types of code they've seen less examples of. They should also be less trusted in these cases, because there's less examples of them being used in these fields. Both a more elongated sigmoid curve, and a sigmoid curve we're further back on. My conjecture is that you specialise in such a field, you're less likely to be obsoleted by AIs.

>>3001175
(Me)
My bad the textbook was Digital Design and Computer Architecture, not Computer Design and Digital Architecture.

File: Screenshot_20260727_04525(...).jpg (293 KB, 954x786)

293 KB JPG

so since CSG got nuked:
I want a head up display for a bicycle to handle navigation and maybe more. I want the phone to remain protected and save battery and the display to be readable in all weather. I started working on a lilygo display with esp and a companion app that gets turn instructions from notifications, pushes those to esp where a sprite is displayed accordingly.
Now I saw one can get a 5" motorcycle specific android auto / carplay head unit for 30 bucks or so. It may be that a chink company 'Podofo' manufactures those ir they are really only resellers.
I dig the hardware at that price. It's IP rated and should do the job. I dont want their firmware.
So I checked on XDA and I didnt see anything conclusive.
Does anyone know what kind of SoC and MCU is commonly in those things or specifically in this one and how hard or easy it will be to flash your own build or even something entirely different like linux?
I dont want to and probably can't find an exploit that works, root the thing myself, unlock a locked bootloader, build DT and so on.
But maybe there's a sure fire way or someone has been in there already?
>12V
I know its one of the downsides.

File: 65846b00e7ced326e776592fd(...).jpg (225 KB, 1243x1446)

225 KB JPG

when will digikey start stocking the NRF54LM20A
please don't say a year from now because i already started designing something with it

>>3001364
I'm relatively sure btw that I'll find an allwinner with some 16 or 32 MB SPI flash and exposed UART or something. Probably a variant that comes with hw decoding and whatever they need.
Not even sure if I am atill expecting an MCU to be on there, I had the idea because it's typical for car head units but it is because they can be hooked up to conteol various things.
This unit tho: It reads tire pressure sensors, that's about it.

File: membrane-assembly.jpg (184 KB, 983x633)

184 KB JPG

Can I transfer signals if I clamp copper wires on the end of a membrane keyboard?

>>3000689
The hardest part is reverse engineering the toy keyboard, MIDI is just an UART with inverted logic or something like that, the messages themselves are note-on note-off bytes, now what internal signals mean what in the keyboard only god and the Chinese know

>>3001384
Yes. It will likely fuck up the contacts though.

>>3001378
can't you just email and ask them?

sorry if this is the wrong place but what's the cheapest possible sbc that can run linux, even if slowly/without a GUI?
also what happened to /csg/? I'd be posting this there otherwise.

>>3001891
I like your question because lately I've been wondering what benefits Linux brings.

>>3001891
RPI zero perhaps? Very small but still SoC class and with shit tons of memory (compared to real micro controllers).

>>3001891
ARM or RISC-V?
You can get 8 dollar RISC-V SBCs like the pine64 Ox64, and some Milk-V boards are cheap but the documentation and community code examples might be lacking

>>3001913
to be honest it's only useful for networking; anything else & you're better off with an RTOS.
linux is also really heavy (compared to how weak microcontrollers are).

>>3001914
it's embarassing how an rpi zero's cheaper than something like a milk-v duo lol

>>3001916
Anybody here use Yocto? I think it's to burn customized linux OSes onto sbcs.

>>3001924
>burn OS onto SBCs

>>3001939
Thank you Microsoft Word.
Seems like you can have ~100MB image of only the packages that you want and you get to tell it what chip it is going to be run on.

>>3002038
You still haven't convinced anyone you even read the most basic information about, let alone understood, what you eant to be talking about.
I just cross compile and use buildroot on larger projects and slap the product on the hardware.
Guess there's infinite ways to skin a cat and everyone has their preferences and reasons.
But it's sort of advisable to understand things before talking about, let alone working with them. It's all a pretty simple topic. No excuses.

>>3002048
My bad I guess. In a few job descriptions for "embedded Linux engineer" I saw Yocto mentioned, that's why I was wondering if it was niche or not. Personally I'm studying stm32. I have extensive experience with atmel 8-bit and making the switch right now after a hiatus of nearly 10 years.

>>3002191
I probably work a certain way at home because it's what we have at work more often than I would like to admit. Buildroot is such a case.

Hello.
I would like to play with Bluetooth 6, specifically to use its "channel sounding" mode.. What's the best microcontroller to get started with that?

>>2999763
if youre so smart youd just be using 328p chips by yourself, chump
>>2999761
>out of date headers
what are you a jew or something? oh no how dare they maintain compatibility with the biggest and most common ecosystem

this general would be more alive if you guys werent such shitposters

>>3002296
>tarduino enthusiasts

>>3002296
>if youre so smart youd just be using 328p chips by yourself, chump
I am, but dev-boards are still useful, and I can still have an opinion on something even if I use an alternative.

Did Arduino kill off clone support while I was away??
A few years ago, I bought a nano off a secondhand site. Got it working, and built some pretty neat projects with it. Now, 5 years later, I'm trying to use it with the 2.0 IDE and it's completely unrecognizeable.

Did something change? Is there a workaround for this?? I downloaded the CH340 drivers and everything but it's still unable to upload code. What do??

>>3002331
set it to old bootloader maybe

>>3002331
Did you check the things runs a bootloader?

>>3002244
Maybe some of the ESP32 micros? Maybe the C5? Seemed to be one of the most recent ones last time I checked. Disclaimer: I took a quick look at the datasheet and the 1000+ page reference manual, looks like a handful. Maybe some open-source code shows ways to play with bluetooth 6.

>>3002413
(Me)
Looked more, if you mean audio by "sounding", then ESP32 the original one, or ESP32 S3, both would do sound, but the C5 and C6 would not.

>>3002413
The ESP32 framework (esp-idf) is mostly open source. You probably can't use things like bluetooth on ESP32 at all without the couple of closed source blobs the framework pulls in automatically.

File: IMG_20260707_222746939_HDR.jpg (2.3 MB, 4096x2304)

2.3 MB JPG

Finished my envelope generator a week ago. It is based on the Attiny1604, it was a bad choice it just barely works due to doing floating math for the log curve to calculate voltage level for curve parts of the envelope. I did enjoy the challenge in optimization.

File: IMG_20260707_220421620_HDR.jpg (1.27 MB, 3264x1836)

1.27 MB JPG

>>3002436
And here is the back.
(Away from PC I hate typing on my phone)

>>3002436
Floats on an ATtiny? Use fixed-point Accums or use a logarithmic lookup table instead.

It came out nice looking, but that’s the sort of circuit I’d prefer to make analog. Not that I’d know how to make an analog low-drift log-lin converter.

>>3002517
>fixed-point
Damn didn't think of that. I will keep that in mind if need to overhaul the firmware down the line, thx. not enough flash for a LUT though.
I initially planned for it to be completely analog, but I could not figure out how to get the extreme variation in timing constants for the curves for a single RC circuit (I refused to use a selector switch). It could be done with OTAs but it was going to be too complicated.
The advantage of using a micro is you could reprogram it for different wave functions, and control styles (Linear or Log control range).
I have extra PCBs I could also use these as digital LFOs with exponential CV input.

>>3002436
Wait, isnt the appeal behind shit like that doing it analog?

>>3002539
NTA, but analog log-exp converters take a bit of thermal matching/coupling/compensation wizardry to pull off right. It's not super hard, I've made one, but it definitely takes a bit of extra reading

>>3002413
Bluetooth channel sounding is a specific feature they added to the bluetooth 6 spec that allows you to tell the distance between two bluetooth devices.
I want to play with it because it's fun new tech, but can't find any opensource implementations of it.
I guess it's too new. Oh well.

>>3002547
>a bit of thermal matching
A tale as old as digital electronics. The limiting factor for analog electronics is how well manufacturing can match two transistors. Turns out with digital we just have to add bits.
>>3002630
What's the underlying physical principle? Does the standard require a response to a probe with a precise and fixed delay so the round trip can be measured? Attenuation wont work as it depends on everything else besides distance.

>>3002634
AFAIK It's uses a combination of time of flight and constructive/destructive wave interference.

>>3002437
Hey anon, I have the same model of panavise.
In case you're struggling with setting the x axis get some matching nylon washers that way you can continuously adjust it.

>>3002634
Can’t you just buy sufficiently matched transistors in a single package? Maybe even the darlingtons in a ULN2003.

>>2999663
Do car ECUs from the 90s fit in this thread's subject?

>>3002771
So far as the digital design and architecture goes, but the actual programming and configuration is probably outside the scope of this thread.

>>3002771
>>3002812
I will have to CAN this post

>>3002685
Yes you can those are being offered.
I was merely making a general observation regarding the limits of analog design. Many IC designs contain incredibly well matched and thermally couples transistors, like for long tailed pairs. And still the matching puts a limit on the precision we can reach with analog circuitry.
Or to put it the other way round: With dogital design you can at least always add a bit if you need higher precision, there is no technological or manufacturing problem associated with 'one more bit'.
>Darlingtons
Darlingtons suck. There is hardly ever a sensible reason to make use of darlingtons today. I'd suspect ULN2003 aren't even particularly well matched, that's not the design intent and I suspect little to no consideration is given to that. I'd go for something like BCV61 but better avoid it alltogether.

>>3002651
I see, thanks. So I guess they need to be better than 10 cm or so with the ToF and the result will be more like 5cm or below, else the interference thing would be redundant.

>>3002685
You can, but the voltage divider at the input also needs a PTC thermistor (or equivalent NTC resistor network) so the -3300ppm/C of the Vbe junction gets compensated by a +3300ppm/C thermistor network. Then both the matched transistors and thermistor need to be physicaly coupled together so they're the same temp.
Look up the VCO for a classic synth, like the TB-303, for a good example. Synths have to cover 4-5 octaves, which is a pretty big range.

Ordered 2 esp32 C6 dev kits (4MB ram).
Feelsgoodman

>>3003047
>4MB ram
Obscene. Is it just psram?

Can someone help me parse what this datasheet means by:
>To open the relay, the ON pin must be driven to a logic high state. To close the relay, the ON pin must be driven to a logic low state.
Does that mean if you driven ON to high the relay will be "open" in that electricity will flow, or does it mean it wlil be "open" in that the switch is open, electricity will not flow?
I think (and am hoping) it's the first one.

If you could get one and only one free dev board worth any amount less than $50, what would you get?
Imagine you have a coupon or something, but it can only be a single board. No duplicates.

>>3003054
This means, for an ideal relay:
Open: No current flows to the load.
Closed: The relay becomes transparent and the load is directly connected to the source (or the terminals are connected).
In your specific case, assuming TTL:
0V - 0.8V: The internal contacts do not make contact, there is no continuity between the terminals.
2V - 5V: The internal contacts meet and make contact, ideally the terminals will appear to be directly connected.

>>3003051
Nevermind, 4MB flash, 512kB sram

>>3003054
>>3003071
This is unironically the worst part of the hydraulic analogy. That an open switch is the opposite of an open valve.

>>3003076
The worst bit for me is I'm actually using this relay to ultimately open a hydraulic valve.
All the valves in the chain need to stop flow if any of the prior ones fail.
Electricity -> Air -> Water
Normally "Open" (or is it???) -> Normally Closed > Normally Closed.
Reading the datasheet and gleaning what I can from how they use the words "on", "open", and "activated" in relation to the resistance, and some guides online that use this relay, I infer that when they said "open" in the datasheet, they actually meant the switch is closed and electricity is flowing.

Bouta blink an LED on the blue pill. Gonna do it bare metal with Windows command line tools and make a private github repo for it so it's more organized to jump from laptop to laptop and also Windows/Linux. Hopefully i remember my avr fundamentals to jump into the stm32 way of things. Hoping to do spi after that and draw things on my tft color screen through spi. Wanna also set up my toolchain with Linux but scared of the roadblocks doing so.Thoughts?

>>3003796
meds.

>>3003797
Is it that nerve-racking?

>>2999663
Retarded lurker here... I do not have the funds yet to get into at least starting to make stuff with these microcontroller deals... how do I explain to my boomer parent who WAS an Electrical Engineer who worked for SONY way back compact disks released, that getting a proper workstation and work on this will further my cyber security degree route.... I can only think of how these could further my understanding of hardware and software security... I just do not know how to explain it so I am not getting chastized for seeking out knowledge.... other than saying hey the chinese are able to get into your net and finances through your unsecured printed btw

>>3003803
We arw teaversing an era where we must work hard for low pay. Our managers do not care for quality, only speed, accomplishment of task and shotty quality.
Show you can get thongs done to enter the favors of your dad. Suppress your perfectionism. Things will get better maybe, as reason shows its face again after at least 2 years of fast-result-oriented brute work. School would be a crutch right now. AI is a thread but all generations know it so this can be a grounding factor. Also your post was vague.

>>3003806
Yes I am vague because the things I want to make wont sell, they are for me, though I know if I learn the skill to repair smaller and smaller items, I do not see how... even with a "global third party network of repair centers within colpanies" should it nmot just be chinese slave labor or nigerian or wtv the fuck, I don't see how supply side will overdo the eventual repair side.. less and less people know how to repair items and have grown a disdain to it, and I just do not get it...

>>3003803
> I do not have the funds yet to get into at least starting to make stuff with these microcontroller deals
Dev boards are like $5. Microcontroller programming is cheap. Then use a $5 multimeter, maybe a $10 USB logic analyser.

>>3003803
>Getting a proper workstation will...
The /diy/ disease: Collect tools and materials first and perhaps even last. Not skills.
Some of the best EE and CS I know hardly ever touched or touch hardware before they know exactly what they will do. That's a habit from when they learned. Back then they all learned to crunch logic, design systems, do math, even write asm on paper.
And you can do all that. For free. No hardware need it, it's likely you'll become distracted even and wont learn methods but imitation instead. Three blinky sketches on arduino in and kids still don't know anything.

>>3003859
>>3003864
I meant for a whole hot air station, and everything, I only have one old soldering iron. My current idea to start learning about reverse engineering and beyond base level programming knowledge is make a bluetooth keyboard and program it from an old keyboard i have, rather than just go for piecing together a reddit teir cyberdeck, right now the extent of what I know about servers, desktops, laptops, appliances, is simply REPLACE THE WHOLE THING IF BROKE(the board at most). Fuck trying to fix anything... I could say that this sentiment has been crushing me into avoiding learning about how BASE level motherboards works, like yeah theory teaches how to read the motherboard, but I have zero experience in fixing boards, and everyone says doing that is a waste of time..... which I do not understand....

>>3003869
Is your intention to learn reverse engineering and/or hardware hacking? Or is that just a means to the end of designing better embedded hardware? If the latter, I think you're putting the cart before the horse.

>>3003879
both, I intend to get an old keyboard, slap a sneed controller to it, and then use it for pen testing, though before I begin anything I want to see that I can make a board, that can connect to things, and not be as hugee as this 5" thick phablet with a shitty membrane silicone keyboard because I absolutely abhor thar keyboard. So far what I know is, I need to figure out which pin from which key leads to what on a tiny ribbon cable, to then find a way to attach it to a microcontroller, with likely having an expansion board so that it can have a battery, and then put it in the thinnest vs durable 3 printed case, with maybe a display, though again not as big and chunky as this, and without any of the useless extra flavor "cyberpunk" colored plastic and ridges

https://youtu.be/bvV9h5kAPXw?is=gIxHsJauL9ZmSPkn

>>3003883
here is the keyboard they "designed":
https://www.yankodesign.com/2024/08/23/diy-modular-cyberdeck-could-be-the-futuristic-computer-you-can-get-soon/amp/

>>3003883
>make a board
connect boards together***

Getting an Arduino Uno might be a better experience than fantasizing over cyberdecks.

>>3003934
not going for a cyber deck though. I will just make a stupid tiny bluetooth keyboard with a sneed xiao or wtv, if it turns into a cybercrapper cool, but all that can easily be out done with a laptop or a pc, no need for shit gadgets unless somehow there is a special sensor that can do fuck all and fuck about like something out of spy kids lol

>>3003935
Maybe add a rotary encoder to it, that would be neat. I am biased though because I just ordered a few to learn how to use them.

>>3003869
>I meant for a whole hot air station, and everything, I only have one old soldering iron.
I dont think you got what I meant when I wrote >>3003864 .
You have an old soldering iron. You don't need more and you could do with less. You can understand and plan what you want to do before distracting yourself by cobbling things together.

File: 20260811_151249.jpg (3.05 MB, 4744x2192)

3.05 MB JPG

>>3004017
uhhh.... these have seen better days loll

>>3004026
the top one doesn't work but the bottom one does, not sure for how long, though I do expext to only use it limitedly for sure...

>>3004027
Yeah so did you read a book yet, design the system, check the math?

>>3004037
what book talks about finding leads of keyboard input on a wire? other than just taking the voltimeter and tracing from a to b? besides that it would need to be some flavor arduino ide or circuit python

>>3004042
You do you. Just go do that man.

is it a smart idea to make two esps talk over esp now through a vehicle chassis

>>3004162
ESPs have CAN, I’d want to use that. Wireless might work if it’s a reliable protocol or if the vehicle has no spark plugs.

>>3004185
ESP also has ETH phy, so you might want to consider using that as well.
I’d use it, just because I don’t know much about CAN but use ETH all the time.

>>3004208
Fair enough. ETH phy requires isolation transformer and connector hardware, but even then it's probably cheaper than CAN transceivers, and no less robust in an automotive environment. Plus all the connectors and cables are standardised. The main upside to CAN is that it's a bus, you can have multiple devices all on the same wires without worrying about packet collisions or needing proper routing. Maybe it's better for precision timing too.

>>3004185
i dont have spark plugs and im ok with some unreliable connection as long as it fixes itself.
Its pretty much just for sending data to a 2x16 display.
if i have time i might try it out this week

The ESP32 is so unfun. Too complicated to do your own bare metal stuff. Too memory and CPU starved to comfortably do higher level bloatware.

>>3004340
EASY == FUN
Weird take.

>>3004340
Wow man.
I sort of even doubt your high level stuff would be good either.
>too complicated
So let me guess: Arduino is the way to go? I honestly don't see how setting up a uC bothers anyone besides perhaps being a chore.

>>3004379
What's easy about your dumbfuck ESP32 firmware crashing because some peak load condition caused it to run out of memory?

>>3004392
>So let me guess:
Bad guess.

>>3004393
Would be nice to measure that peak load condition (can it be calculated?) to make sure it isn't ever reached.

I want to make an RC car that I would control using an xbox controller. Do I need something other than 2 eps32s, some kind of a H bridge, motors, battery and a car body. Where can I get some cheap car bodies? I want it to look cool but I don't want it to be expensive. I've never worked with esp32 before

>>3004424
Get the system design right first. Why bother with the least of details, that is the car body, first? Worst case you'll be collecting junk and materials for unfinished projects.
>some kind of an H bridge
That is true but unless you have a particular reason just plan with modules and treat them as black boxes that provide a certain function. Like get aj odrive or whatever and ignore how it does what it does.
>batteries
if you want it to be electric this sounds reasonable
>motors
reasonable
>2 esps
2? Idk how xbox controllers work, especially the air interface but perhaps check first if you can with reasonable effort interface the controller (I assume those are wireless) with esp32.
>anything else
well again... why bother with trivialities now... but yoh'll certainly need a chassis and running gear and shit. But for a proof of concept you can just use anything.
You might want to plan with traditional steering and using an rc servo for it unless you want to use differential drive or go straight only

>>3004426
>That is true but unless you have a particular reason
I have one laying around from a project I had to do while in uni, plus I'd potentially go with 4 motors so that I can have one spinning each wheel individually (that's how I'd get torque vectoring eventually)
>2?
I figured I'd need 1 to spin the motors and to handle the steering and the other to tell the first one what to do
>well again... why bother with trivialities now.
because I'm bored and want to see what else I might need before I start ordering stuff, if I'm ordering from China it will take a while for it to arrive

>>3004424
Make the car body from foam, the type used by architects to make their models. Maybe. Cheap and potentially very nice. Very lightweight. A car body made for an RC car will be expensive if nice I assume.
For the xbox controller if you get it to talk to esp that would be insane. I'm assuming you did a search and found people have done it successfully so you're going to try it too.

>>3004430
>I have one
So why imply an unspecific sort of H bridge. You have 4 transistors and a transformer or what? Great. You just explained you might want 4.
>esp32 soviet employment scenario
Have you made plans but not shared them with us and come upon limitations? So far I am confident 1 esp32 is far more than you need to implement this.

>>3004440
>So why imply an unspecific sort of H bridge
because I'm not at home and I forgot the full name of it. I'll get a few eps32s just in case but I'll try it with one. Thanks
>>3004436
>For the xbox controller if you get it to talk to esp that would be insane
It's hard? I got that idea 15 minutes before I wrote my post, as I said I have no experience with this, I only did some basic arduino simulations and basic electronics without microcontrollers. I very well might be in over my head. I'll spend more time researching before starting this. Thanks for the foam recommendation

>>3004442
>because I'm not at home and I forgot the full name of it.
I forgot to mention that it's one of those boards that can drive 4 motors

>>3004442
>It's hard?
There are tutorials but I never tried it myself. Did Arduino Uno + NRF24L01 when I built my wireless servo controller. That was kind of hard to me at the time. This wireless stuff is no joke but very very interesting.

>>3004443
it's the fucking steppy motor half bridge driver chip? on a breakout board? if it's that temu special then you should know it will probably drive 4 motors unidirectionally, 2 bidirectionally or one 4 lead stepper. Just guessing by the suspected internal design.

>>3004436
>xbox controller
wasnt the billionaire crusher operated via xbox controller
do you think or know they have gone to lengths to obfuscate their protocols or what?
so what is the air interface of those things, i mean I dont expect them to use a propertiary physical layer, 2.4 G BT?

File: images.png (17 KB, 549x364)

17 KB PNG

>>3004449
>This wireless stuff is no joke but very very interesting.
I'll try something
>>3004450
It's called L298N, I found this picture on my phone. I used it to drive only two motors so idk if it will work in the way I want it to
>>3004452
they used a logitech controller, I think that it was the f310 model

>>3004422
That's your only chance. I say it's impossible if you want to run a full bluetooth stack. Next time I'll ask the hw people to include a psram.

>>3004460
Nah.
I looked it up, seems to be that if you push bluetooth hard, it can expose poor code easier (bugs). ESP-IDF gives a few testing tools to help catch those bugs.

>>3004459
Well how can you say you don't know. You have just posted the (idealised) internal schematic and it looks to be doing exactly what I expected and unless you don't care about reversing it won't do.
Get another one for 4 wheel operation?

I heard from a trusted source that the F310 will cause submarine implosions when ran at Titanic-level depths.

>>3004465
As I said I found the picture of this diagram on my phone, I had used it over a year go. I'll get another one, thanks

>>3004464
You looked up what? The docs document it quite clearly that malloc failing means you're fucked:
>it is very dangerous to run out of heap memory, as this will cause ESP32-S3 an "undefined behavior". Thus, enough heap memory should be reserved for the application, so that it never runs out of it.
Okay, that was wifi, not bluetooth.

https://documentation.espressif.com/projects/esp-idf/en/latest/esp32s3/api-guides/wifi-driver/wifi-performance-and-power-save.html?title=ESP-IDF+User+Guide+for+ESP32-C3%2C+SDK+v5.4.2.

>>3004474
It should be fine (wifi) if you do your job: "their documentation gives explicit formulas for calculating the maximum Wi-Fi dynamic-buffer heap consumption, and provides configuration profiles ranging from high-performance to memory-saving"

https://docs.espressif.com/projects/esp-idf/en/latest/esp32/api-guides/wifi-driver/wifi-performance-and-power-save.html

>>3004393
Dynamic memory allocation is scary. I’m fine overspeccing my MCUs so there’s no need for that in the first place.

>>3004424
Buy an off the shelf RC car (maybe 2nd hand, especially one with a missing/broken controller or receiver) and modify that. You’d be replacing its receiver unit with a custom one. Unless you actually want to learn how to make hardware too.

>>3004470
Those motor drivers are shit. Bipolar transistors, and no current limiting or overcurrent protection. Not sure about overtemperature protection. They’re low-power and get hot compared to a chip from the last thirty years.

>>3004490
>those motor drivers are shit
Well it depends. As always. I mean this guy insists, mhm fine let him.
Just test the inductance of the motors or test the actual system and define max duty cycle with stalled motors to not OC. Alternatively ad R01 in series to GND and use an analog to get current.
While I'm fully with you and I'm suspicious of several things related to this project I would not ever go as far as deeming any method, component or whatever as 'shit'. It just always depends.
It depends is also the only answer my boss ever gets from me.

>>3004479
>your job
Is that a joke? A shady buffer size calculation in the docs isn't going to help if some other subsystem suddenly allocates too much because of some unexpected situation.
All you can do is freeze the firmware, field test it until exhaustion, and make sure that failure isn't critical (like you can reboot the chip and the user will just accept it).

>>3004490
>Unless you actually want to learn how to make hardware too.
I want to do as much work as I can
>>3004490
>Those motor drivers are shit.
Which ones would you recommend? I was planing on using that only because I already had it on hand

>>3004518
>I want to do as much work as I can
I would strongly advise you to limit the scope of your project at the start. Scope creep is terrible for projects.
>Which ones would you recommend?
I'm using DRV8874s, made some PCBs for them to be used as general purpose motor/solenoid/H-bridge driver modules, but they're only 6.5A max so depending on your motors you may want to choose something else. Brushless would be more work but more payoff. Designing a brushless motor driver is a lot of scope creep.

Also luckily for you, there's an ESP32 library someone's written to allow you to connect to a controller:
https://github.com/tbekas/BLE-Gamepad-Client
So if you want more scope creep, don't use that and write the code from scratch.

>>3004519
>Scope creep is terrible for projects.
This is my first project like this but I really noticed that I have this issue whenever I'm writing code.
Thanks a lot

>>3004495
>shady
Could be. I doubt it though. I will be able to give you my fully-educated opinion of it once I test it myself on my two ESP32 chips I just got, but I ordered them slightly prematurely and doing so is towards the end of my learning roadmap, after fully understanding the stm32 (blue pill for now, maybe 429 after) and all the communication protocols up to CAN and RS285 or 485 or both (no USB for now, i think), as well as RTOS on that platform. I figure the ESP's thousand-page reference manual will be a lot less intimidating after the stm's thousand-page one. No looking up and using libraries for me this time around, a long-term project and probably 3 months at least before I touch the ESPs.. Okay maybe I'll cheat. I don't know.

File: file.png (391 KB, 1200x624)

391 KB PNG

Do I have to be careful in handling dev boards like the pi pico or components because of ESD? I do have a carpet lying around no idea how bad it could get. I wonder if I can just a bag of soil as my ground...

File: 1530336988657.jpg (51 KB, 720x960)

51 KB JPG

>>3005195
>I wonder if I can just a bag of soil as my ground...
had me for a sec
Real answer, it's so incredibly rare to damage modern chips with static that there's no point. Maybe you'd take the precaution on a very expensive chip, but there's so many things that are more likely to destroy dev boards than ESD.

>>3005195
I've never killed anything through ESD and don't take any precautions, but if you're in a dry climate (especially cold outside, heated inside, with no humidifier or ERV), or wearing very static-prone clothing it could be a problem. Also carpet, in case you're not afraid to lose 0603s in your shag pile.
See:
https://www.eesemi.com/tribo_series.htm
Avoid things far apart on this list coming into contact. I've never had issues with silicone rubber, probably because things don't rub on it easily. I think it's more hard vs fluffy situations you should be most wary of, where triboelectric charge transfer events happen more frequently.

>>3005210
In industry it's cheap precaution. A $10 foot strap for every employee and sufficiently conductive flooring was the norm at an electronics assembly place I worked at once.

>>3005210
ESD can kill things you don't even know about and which may cause weird malfunctions later on. But there's no need to be overly paranoid either. I'd just touch some conductive surface which I know is earthed before handling a board when sitting down.

>>3005215
In the lab we wear anti-ESD lab coats (they look cool too) and anti-ESD shoes. Nothing else.

>>3005210
>there's bo point
happened countless times to me at home and at work and it's always a huge PITA because you have no clue and start chasing errors that aren't there.
Worst time at work was
>release date is a week ago
>we got everything done, ready to release
>a multi buck converter chip on one pre series board blows up
>mhmm weird... didn't that happen a few weeks earlier too
>suddenly boards start blowing up
>expand our on-off test rigs because every board must go into on-off testing
>no pattern, some blow up right away, some after countless cycles, some seem to be golden and just hold up fore... ah nah...
To this day no one is sure if it was the cjip suppliers design and them covering up or not, but they eventually admitted to problems after several of their customers connected and they blamed it on ESD or radiation somewhere in the pipeline. New batches came in, we made sure to follow all ESD precautions et voila.

>>2999663
> nucleo
I bought a f401re what now? Did I pick the wrong board if I want to make synths?

>>3005276
Hmmm, I have been wanting to buy a new lab coat.

>>3005302
hardware FPU, very useful for DSP/filter calculations.
512 KB Flash / 96 KB SRAM — plenty for a fairly sophisticated hobby synth.
the F401RE has no built-in DAC. For decent audio output they'll want an external I2S audio DAC/codec (PCM5102A modules are popular and cheap)
At 44.1/48 kHz, there's enough CPU for several oscillators, envelopes, filters, LFOs, effects, etc., provided the DSP is reasonably written.

>>3005310
Can you recommend an ADC or a codec module that has audio input?

File: 1766915048722182.jpg (1013 KB, 2494x3325)

1013 KB JPG

New to embedded programming, not new to programming.
Wanted to get into FreeRTOS using STM32s, usecase is mostly sensor reading (temp, IMU) and simultaneous servo control and similar things.
Is the Nucleo 466RE the best choice for this or would one of the (many) other boards be better? Anything newer? For now: no intention for any of these projects to leave the breadboard stage.

>>3005593
>FreeRTOS using STM32s, usecase is mostly sensor reading (temp, IMU) and simultaneous servo control and similar things.
you can probably find pi picos and other 2040 and 2350 based platforms for much cheaper than anything STM, and ESP can do it if you need integrated networking

why do you want to learn FreeRTOS anyways? its mostly a tool for fatter SBCs like a pi 5 to act less like a PC, a vanilla STM doesnt even need it

>>3005591
It's going to depend on what feeds it.

>>3005593
Yeah get it, nice board. Please send me one too.
For the real time stuff, making your own first is advised, before starting to use an RTOS. Owning and oscilloscope would also be a must.

>>3005593
I read too fast. If you'd said the 446 I would have said great board. The 466 though I don't know what that chip is.

>>3005646
>>3005647
Oops sorry yeah I meant 446re.
>send me one too
I would if I could

>>3005659
It's going to have a good amount of tutorials. The better route is to quickly skim over the datasheet and read only what you need from the reference manual. Coding and uploading the code should be done with the command line until the toolchain is understood.

>>3005747
I'm an emacs fag and generally prefer the command line over fat IDEs
Can I comfortably work with this thing in general this way without making my life fundamentally harder or will I eventually have to switch to something like Cube?
Maybe I should already start looking into the toolchain before it arrives

>>3005747
>>3005786
How does syntax highlighting work for 3rd party editors? Like, how do I get the DDRB to show up in a text editor, such that I can mouse-over it and see what address it is? Or mouse over a bit-mask name and it will tell me what the bitmask is?

>>3005646
>It's going to depend on what feeds it.
line level audio

File: 1780444485902592.png (1.9 MB, 600x1196)

1.9 MB PNG

Nucleo-64 446RE or pay the extra 10 euros for a Nucleo-144 F767ZI?
>why these two
out of the STM32 devboards mentioned or referenced in the sources I checked those seem to be the only ones easily available and not totally overpriced here
>what do you want it for
Just general embedded learning, so versatility is appreciated
>get arduino/pi/esp
No I wanna become a big boy STMer and I prefer bottom up learning anyway

>>2999763
Make your own board then
Its not too bad

>>3005829
okay then this should be fine (with an electret microphone it wouldn't be though)
Digilent Pmod I2S2

>>3005849
I often do, but that doesn’t change the existing standards. People learning to use microcontrollers start at arduino, and it’s just making things harder and more obtuse than they need to be. The Pi Pico would be great if not for encouraging embedded python.

Also just an FYI
Always use Lua instead of MicroPython

>>3005915
>People learning to use microcontrollers start at arduino,
By the way, I hope with LLMs Arduino finally can be killed off. Because non-programmers don't need to rely on libraries made for retards anymore. They can just tell the LLM to write proper code.

>>3005967
20% Embedded Development job market collapse incoming, great

>>3005967
>tell the LLM to write proper code.

>>3005970
More like 100% software dev job collapse.

>>3005967
whats wrong with using libraries ?

>>3005973
In the current stage of AI based software developer, you have to be a good programmer to wrangle the AI in writing good code (like giving it proper instructions and stopping it if it goes the wrong way).
This problem will be fixed in another year or so.

>>3005967
> Because non-programmers don't need to rely on libraries made for retards anymore
Don’t need to rely on libraries at all. If fact, you don’t need any high-level languages like C anymore. LLMs can just emit ultra-efficient machine code directly.
In fact machine code will go away when LLMs just emit netlists and transistor layouts. I almost said VHDL, but that’s another fluffy high-level language for people.

>>3005980
> you have to be a good programmer to wrangle the AI in writing good code
It’s amazing nobody has thought of using AI to write good *prompts* yet, easily solving the wrangling skills problem.

Where can I find some unpaid hands-on experience in electronics and embedded development?

I’m willing to work for free since I have the time to put into it, especially if the project is interesting. I also can’t really find any paid remote work in this field right now, and I live pretty far away from any major city.

So I’d be happy to hear about any projects where I could contribute for free, especially something run by a nonprofit or some other non-commercial organization.

>>3005989
They're trying that with coding models. It's funny how you can have them write slop code, then point out the slop, and then they apologize and clean it up. They only need to hook up better review skills by default.

>>3005954
1-indexed arrays are an affront to God

>>3005977
I believe he was specifically talking about "libraries made for retards," not libraries in general

>>3006001
>1-indexed arrays
This is a really shitty thing. I'm sad they never tried to fix it.

>>3005954
Why?
Micropython is much more fully featured, and despite that amusingly has lower level capabilities. When you need more speed/control, it supports inline compilation, full address space access, and even C module importing.
If you need maximum speed no matter what then C is the way to go, but if you just need something to run the show while peripherals do the heavy lifting, I don't see a situation where Lua makes more sense.

>>3005991
You can get a bit of hands-on work with repair if there’s a local repair cafe. As for design and embedded work and other more professional stuff, I really doubt it. Too much lability for anything other than one-off gizmos. Maybe a robotics club. You’d be better off collaborating on an open source embedded project and learning from your peers.

>>3005991
>I also can’t really find any paid remote work in this field right now,
Because software development is dead now.

>>3006053
Why does everything suck so much

>>3006066
Sad state of affairs indeed. We have to pivot.

File: IMG_3070.png (1.32 MB, 828x1792)

1.32 MB PNG

>>3000054
PY32 is another option, you need the programmer too for ch32
>>3000688
Minimalist is v1772 I got a bunch for .50 but they don't work for bitbang USB, they're $1 now
>>3005991
Cnlohr discord he made ch32fun

>testing arduino circuit with external dac
>finally almost done
>brush scope probe in the wrong area
>power LED goes off
>comes back on dimly
>no longer responds to programmer
it's all so tiresome

How do I start learning to real code with circuit python?

>>3006642
>real code
>circuit python
pick one

>>3006642
Learning to code nowadays has never been easier. Google "learn to code [insert language]" and there will be tutorials with great examples to learn the basics. Then when you have a real project, just start with what you know and ask Claud or ChatGPT to fill in what you don't. Arduino, C++, and Python are easy as fuck, and you can have AI explain how all of the individual lines work so you can learn and understand and tweak it yourself instead of just having it done for you.

>>3006679
I dont want to use AI and I just want to jump from blinking LEDs to the next level like making usb hid or send midi signals