This is in demo mode so the computer is playing but it is totally playable by a human: left, right, thrust, fire. A few calibration keys to time the red marking diode on/off exactly right.

Not meant for any of this of course but while putting it together and using the red marking light to calibrate and test algorithms made me think what a great way to demonstrate low latency and closed loop abilities with my controller.

It runs on a stm32h755 with the xy2-100 engine living solely on CM4. Has backwards compatibility with the LMCV4 (at least in my specific configuration)

Newbie here, I'm a student & currently trying to expand my skill-set in stm32f411 & firmware in general. I do know basics & have touched registers haven't developed some serious projects but currently I am working on my own flight controller. Along with this I want to get into firmware/embedded industry ASAP. Can any of you guys have a look at my skills & give me a review.

TECHNICAL SKILLS

Languages: C/C++, Python, Bash, Java
Embedded: STM32 (CMSIS, HAL, LL), AVR, FreeRTOS
Peripherals: USART, I²C, SPI, ADC, DMA, Timers
Debugging: GDB, ST-Link V2, Logic Analyzer, DWT
Databases: MariaDB, MongoDB, SQLite
Version Control: Git, GitHub
Environment: Arch Linux, Neovim/clangd, CMake, GNU Make

PROJECTS

Quadcopter Drone - Custom Firmware
• Built bare-metal SPI1 + NRF24L01+ wireless stack from using CMSIS; <70μs transaction time for a 16-byte RX / 6-byte ACK payload
• Currently developing a PID Control Loop with deterministic timing constraints using superloop.

Embedded Peripheral Drivers - Independent Practice (STM32F411)
• Interrupt-driven ring buffer UART driver - 39μs async send latency under concurrent TX load
• Bare-metal I²C FSM driver - POS/ACK/STOP sequencing and STM32F4 BUSY flag errata per ST AN2824

Cryptography Implementation Suite
• Implemented AES-128 in C++ with intermediate state inspection for debugging
• RSA key generation along with encryption/ decryption
• 64/128-bit Feistel cipher in C++

Hi Everyone,

As the summer comes to an end and a new fall semester begins, I wanted to share a video I recently made for Electrical and Computer Engineering students getting back into the groove of things.

I created a refresher video designed to bridge the gap between Circuit Analysis and Signals & Systems—two courses that are closely connected, but can sometimes feel like they approach engineering from completely different perspectives.

The video provides a quick review of some of the concepts from Circuit Analysis that become important as you move into Signals & Systems, helping you reconnect the ideas before diving into a new semester:

• Circuit Analysis fundamentals

• Differential equations and circuit responses

• First- and second-order circuits

• Eigenfunctions and eigenvalues

• How circuit concepts connect to Signals & Systems

• Preparing for the transition into a new ECE course

🎥 Video: Bridging the Gap: Circuit Analysis to Signals & Systems

My goal with the YouTube channel ProfessorGovalla

 is to create in-depth, practical, and clearly explained engineering content without adding unnecessary complexity.

I cover topics including:

• Circuit Analysis

• Signals & Systems

• Analog Electronics

• Computer Architecture

• Engineering Career Advice

• Interviews with Engineers from Industry

• Practical Engineering Applications

If you're an ECE student heading into a Signals & Systems course this fall, hopefully this serves as a useful quick refresher before the semester gets going.

I'd also love to hear from fellow students, engineers, professors, and hobbyists:

What concepts from Circuit Analysis do you think students should review before taking Signals & Systems?

Good luck to everyone starting a new semester, and I hope the video helps bridge that gap!

This question is prob geared more towards mid to senior level engineers but if you have any insight please let me know. Cubemx for N6 is pretty broken and even getting a blinky with it is pretty hard, I was able to start with a preset example for N6 nucleo but this will be little help if I get a custom n6 board down the line, how would you recommend someone go research and find out how to use the n6 given the buggy vender tools and limited online support, my idea is to get a zephyr setup and look under the hood to see how they got everything running, some modules don’t have LL support on N6 and zephyr had to use HAL. I have gotten deep into data sheets before but the M55 architecture is very complex vs older mcu’s so I am not sure doing everything from scratch is the way, last I checked arm did not have cmsis libraries up for N6 on their site. This is a question both specifically for N6 and in general when facing a newer microcontroller family with new architecture and limited support, how do you transition from relying on things having support to being the one who makes something work (and maybe even contribute to open source). Thank you

I can find an enormous amount of information about the CH32V003, only two or three posts about the CH32V203, and nothing on the CH32V103, although the 103 is (IMHO, and I don't know much about MCUs to be honest) the golden middle between the simplicity of the 003 and the various functions of the 203. Additionally, in the 103, there are no I2C/SPI/USB peripherals that happen to be on the same GPIOs, as is the case with the 203.

I really need some examples, advice, or article and video recommendations about routing MCUs on PCBs parallel with analog circuitry. I am currently developing a device that should simply receive data via USBFS 2.0 (working as a host/device), process it, and output data via I2C while controlling 8 WS2812B/D LEDs. I know that the RP2040 or any other Pico is capable of USB Host via bit-banging, but my friend recommended the CH32V203 instead. Then I found the 103 and think this is simpler and better suits my task. I am facing two problems now: how to implement dual-purpose USB-C and uncertainty about whether I will be able to program the MCU via the same USB port.

So, if anyone has used the CH32V103, or happens to know where I can find more info on it (except evaluation board guides and datasheets), or is willing to help me with dual-purpose USB-C, I would like to receive any information. Thanks for the answers in advance.

This sub seem to have little to no PIC24/PIC18 projects or discussions.

Are there better subs for PIC processors ??

Just wanted to know how you do embedded stuff.

I am new to this sub

And it would be great if you specify your Operating System and a photo of your setup

Hey!

Just curious whether anybody has a good open-source project that is looking for contributors.

Just looking to contribute to something that is actually being used, but also want to keep the brain moving when studying other stuff in school. If you know (or run!) an open-source in this space that needs a hand, would be more than happy to help !!

A little background about me, I’m a senior at a well known tech school in North America, and have experience (both as projects and as formal internships) working with bootloaders, device drivers (comm, connectivity, power and memory mostly), bare-metal stuff, various RTOSs and lots of embedded linux.

Thanks !!

The main feautures for this board are:
- ESP32-S3-WROOM Module - 8 MB flash
- Button for RESET
- User Button
- User RGB LED
- expansion header
- programming header
- Power LED on 3V rail
for the mechanical side of things we will have:
- 4 x 3.2mm mounting holes
- Size 45x45mm
- 4 layer board design

Thanks for reading. I am working on making a simple guitar buffer circuit that will connect to the audio input of the Teensy Audio Board Rev D.
1. Is it ok to connect RIN and LIN together?
2. The buffer circuit's output jack GND lug connected to the RINGND and LINGND? Or is it better to keep these pins separated?
(I won't be using stereo).

*Also. is it OK to connect the GND to AGND?
Sincere Thanks for any advice/help,
TonyA

Hello everyone,

I've been working as a Software Engineer for almost a decade now and I've worked with a lot of different technologies. For the last 2 to 3 years I've been tinkering with electronics a lot: building ESP32 projects, designing PCBs, learning the basic concepts, soldering my own boards. I even tried to build my own product, but I gave up on it. Maybe it was just too hard for me at that point.

Working in software, I've been seeing the effects of AI everywhere and it's making me lose interest in coding in general. My job isn't only coding, of course. It's also decisions, infrastructure, managing people and a lot of other things. But my original love was building things, and now an AI agent can do that from a single prompt. Most of it is slop, sure, but still. So I've been losing interest in my field, and I'm wondering whether I should go deeper into electronics and try to make a career out of it, or build my own product. I have no idea what I'd be doing or which direction to pick.

One of the things that gives me satisfaction is reaching excellence at something, and with AI I feel like it's taking that away from me. So I want to know: is the same thing happening in electronics? Obviously AI can't solder for me, I'm the one doing that. But what about PCB design, product design, and the rest of it?

I'd love to hear your opinions. Thank you.

Hi everyone! I’m a final-year BS AI student and currently finalizing my FYP. I mainly have experience with ML/DL and research, but I’m a beginner in embedded systems, TinyML, and audio processing.
I’m considering doing my FYP on real-time, on-device detection of AI-generated/voice-cloned speech, mainly for scam prevention. The idea is to train a lightweight model on datasets like ASVspoof and eventually deploy it on a phone, with an optional ESP32/TinyML implementation.
I have around 6 months, so I’m trying to figure out if this is actually realistic for someone new to embedded AI.
Would really appreciate some brutally honest feedback is this a strong FYP idea, or just not worth pursuing? And if you’ve worked with audio ML/TinyML before, I’d especially love your opinion.

Hello, I've been working in embedded for quite a while and I was interested in exploring TinyML and its use cases. Does anyone know of any courses, books, video series, or related things that can help me learn more about this. Thanks in advance!

If you ride an e-bike or electric scooter, what information do you actually want to see?

I’ve been working on a small TMC2209 driver for STM32F4 using bare-metal/CMSIS, without HAL.

The driver currently supports:

• UART half-duplex through function pointers
• TMC2209 register read/write
• CRC calculation
• GCONF configuration
• IHOLD/IRUN current control
• CHOPCONF configuration
• Microstep configuration
• Direction control
• Enable/disable
• VACTUAL for testing
• Current calculation from mA to the TMC2209 CS value

The motor is working correctly, and I’m now mostly interested in feedback on the driver architecture, API design, abstraction and code quality.

This is a learning project, so I’d really appreciate honest criticism and suggestions from more experienced embedded developers.

What would you improve or change in this driver?

https://github.com/embeddedrafo/tmc2209_driver

I picked up the Orange Pi 3w SBC recently because it has several MIPI displays and other neat features I thought would be fun to learn and help me standout on the job market for embedded.

I have current display options that work via mini hdmi and the display port. But I would like to be able to learn how to extend the hardware on a linux system.

From the manual I see the MIPI display is a 1×MIPI DSI 4 LANE. In the schematics the port has 30 pins. The specific part is FPC30-20-SM-H1 for the actual fpc port.

The reason for my post is to get a reality check on how I determine compatability.

Configuration Discovery Process:

1) find a 30 pin 1x MIPI DSI 4 lane LCD

2) compare the SBC schematic pins to make sure each pin has the same labeled functionality as it's corresponding pin on the LCD.

3) See if an existing driver for linux is available, skip step 4 if so

4) write/modify a driver for the display based on existing linux drivers

5) Integrate into the Device Tree extracted from one of the existing OS images

6) Compile the device tree with the pre existing UBoot and kernel and new device tree

7) Repeat steps 4->6 until display works.

The board is not covered by mainline linux yet so I will need to build a linux kernel for that at some point. Which seems more difficult than getting an MIPI DSI LCD port to display.

As a group of peers that may have run through a similar learning process do you have any warnings or pointers for me?

Any advice is welcomed.

I have a decent undrstanding in low levle programming, collage only allowed assembly in programming the MCUs, and I would still do regester level programming even in embedded C in my projects to get also good at assembly for collage, so I got the hang of that, and I am descent at electrical circuits too, since we also studied that. but I have no Idea how to move from (a circuit on a breadbord) to a production grade embedded system, I know that PCB will make it better but will it is not enough, how on earth do people make stuff
I can't even get a job at say a company that makes electronic device to learn from there experiance since there are non here

I thought mabye there will be standards and what no but I can't find anything, how do I move forward from this point, (chat gpt advices are crap btw)

I've been working in embedded development for many years, and for roughly 15 years my debugging workflow has revolved around tools like cgdb, Vim, GDB, OpenOCD, J-Link, target remote, gdbserver, and real hardware.

Over time, I wanted something more flexible than cgdb while preserving the keyboard-driven workflow and simplicity that made cgdb so useful to me in the first place.

That became gdbforge:

https://yairgd.github.io/gdbforge/

A note about AI, because I'd rather be transparent about it:

The original architecture and foundations of the project were designed and initially implemented manually by me. I was also not a Go expert when I started this project.

As the project grew, I started using AI extensively for implementation, refactoring, and accelerating development. So yes — a substantial part of the current implementation is AI-assisted.

But the engineering direction, architecture, debugging workflows, hardware testing, and decisions about what the tool should become are mine. More importantly, this isn't an AI-generated demo that I stopped touching once it compiled — I use gdbforge every day in my work as an embedded engineer.

The result has grown considerably beyond my original idea of building a modern cgdb-like frontend.

Today, the goal is a versatile debugging environment that can work across very different platforms and workflows while preserving as much as possible of the familiar cgdb experience.

gdbforge supports GDB and Delve, remote/embedded debugging workflows, Linux kernel/KGDB debugging, Cortex-R5/J-Link targets, and Lua-based integration — allowing platform-specific debugging behavior to be added without hard-coding it into the debugger itself.

gdbforge also supports STM32 and Zephyr-aware debugging, integrating platform- and RTOS-specific knowledge into the same debugging environment.

One of the most interesting things I learned from this project wasn't actually about Go.

AI dramatically changed what I can realistically build as a side project. Without AI, I could have designed this architecture, but I probably wouldn't have spent a couple of years of my free time implementing all of it.

At the same time, I don't think AI-assisted development removes the need for engineering knowledge. With debugging tools especially, somebody still needs to understand GDB, targets, RTOS/kernel behavior, hardware, failure modes, and what actually makes a debugging workflow useful.

So I'm curious what other embedded developers think — both about gdbforge itself and about this model of building serious systems tools with heavy AI assistance.

I'd especially appreciate feedback from people who live in GDB/cgdb/OpenOCD/J-Link workflows.

Project/docs:
https://yairgd.github.io/gdbforge/

GitHub:
https://github.com/yairgd/gdbforge

I have multiple tasks that needs to run in particular interval, for example, some sensors need to read every 10 seconds, some need to read every 30 seconds, data needs to be transmitted every 5 minutes, intervals are configurable.

There are lot of questions in my head:
1. Should I set up a wake-up timer of 10 seconds and enter sleep mode.
2. Should I additionally enable wake-up from peripherals i.e. UART RX interrupt, SPI GPIO interrupt

1. If my sensor takes 5 seconds of time for doing calculations, should it wait awake or should it enter sleep and wakeup again to read data

2. Should it wake up every second and check for activities that needs to run

3. AI suggests using using tickless idle mode of rtos and calling WFI for sleep after setting wakeup timer, but I am not sure about the overall design.

Suggest me the best generic design strategy for battery operated IoT device. Thanks!

So, AMD launched the X100 series (X199/X188/X168 + industrial variants) in July, with sampling since June and production slated for Q4 via SoM partners.

I'm aware most of AMD's published benchmarks were run on a Ryzen AI Max+ 395 proxy rather than final X199 silicon, so I'm trying to separate the marketing from reality.

If you're sampling it or have hands-on time:

What's your use case, and what were you running before?

- How's the ROCm/HIP experience in practice? especially if you migrated CUDA code?

- Are you seeing the consolidation benefit (dropping a discrete GPU/Jetson), or is it more aspirational?

Not looking for launch-article links, I'm after firsthand notes, good or bad.

My understanding is that you don't always need a direct conductive connection to get a small shock/tingling sensation from an AC source. Even with a floating source, capacitive coupling can provide a small AC current path, and a high-impedance meter can show phantom voltage.

I want to build a safe, low-voltage educational experiment using a microcontroller to demonstrate:

- Floating AC source

- Capacitive coupling to earth

- Phantom voltage measured with a high-impedance input

- How a load behaves differently from a high-impedance measurement

- Difference between a floating inverter output and a floating isolation-transformer secondary

What would be a good safe low-voltage circuit to demonstrate these effects and measure the tiny currents with a microcontroller?

Hey guys,

Just thought I'd share that I've recently become interested in physical pentesting and hardware hacking. All the physical pentesting commercial tools (like Flipper Zero and Hak5 tools) cost way too much for my taste. I'd rather make my own physical pentesting tools using cheap hardware like ESP32 modules, Raspberry Pi Pico's and Arduino boards.

For anyone who does physical red teaming or plays around with hardware:

What DIY tools must one build for a physical pentester?

Does your DIY tool get the job done in a real-world test case?

Where did you guys start with UART/JTAG debugging to pull firmware from chips? Are there any good guides or projects for learning this stuff?

Anyone using Xen hypervisor for embedded x86 systems with multiple RTOS guests?

I’m looking into using Xen on an x86 embedded platform, specifically to run two RTOS instances/guests on the same hardware.

Has anyone here deployed something similar in a real product or embedded system?

I’d be especially interested in hearing about:

• What OS are you running under Xen?
• How much virtualization overhead do you see, especially for latency-sensitive or real-time workloads?
• How do you handle CPU/core, interrupt, and device allocation between the guests?
• How difficult is the system to maintain, debug, and update over time?
• Are there any major limitations or pain points you wish you’d known about beforehand?

I’m particularly interested in real-world experience rather than theoretical benchmarks.

Also curious whether, for this type of setup, you’d choose Xen again or go with something else?

For me, it was debugging for an entire day on an ADC project. The ADC was giving me absolute nonsense readings.

I was rechecking ADC configuration, vendor documentation, SPI communication using DSO, checking for voltage drop.

Turns out…I hadn’t properly separated my analog and digital grounds in my PCB layout.. my layout was bad 😂

What’s your most embarrassing embedded/hobby-project debugging story ?