you need to teach me so many things please 🙏🏾
you need to teach me so many things please 
Zwow I didnt know v2x was a thing already, this is really amazing.
KIn essence - look at hardware in the loop testing
Consider that a lot of the setup is identical
You could actually drive a system like this from git, people drop a link into some things, when it's their turn the system picks up the code flashes it and runs it.
The difficult part here is if you want people to be able to interact with the target in realtime - now just throwing jobs at it doesnt work, it becomes more like booking an appointment.
We were talking about building something like this on another server for an embedded advent of code, we hashed out some ideas but didnt have the time to commit and make it happen
Consider that a lot of the setup is identical
You could actually drive a system like this from git, people drop a link into some things, when it's their turn the system picks up the code flashes it and runs it.
The difficult part here is if you want people to be able to interact with the target in realtime - now just throwing jobs at it doesnt work, it becomes more like booking an appointment.
We were talking about building something like this on another server for an embedded advent of code, we hashed out some ideas but didnt have the time to commit and make it happen
Zlet me read up a bit, keen for this December maybe on this server let me see what resources I can gather, I'll come back with questions
Rwhen working with Zephyr OS for embedded systems, what do you find most beneficial?
@Dtynin @Boss lady @melta101
I noticed you guys have work on some projects using zephyr can you please share some of your insights
@Dtynin @Boss lady @melta101
I noticed you guys have work on some projects using zephyr can you please share some of your insights
KNot tagged but the HAL and the use of device trees gives more hardware abstraction than other options. The closest thing in terms of completeness is the arduino hal
The general quality of the drivers is also good enough that you don't have to hack it up all the time
The general quality of the drivers is also good enough that you don't have to hack it up all the time
RThat's a great point, Zephyr’s HAL and device tree really do provide that extra level of hardware abstraction, which can simplify development significantly. How would you compare working with Zephyr’s device tree to other approaches you've used in the past? Do you find it as intuitive, or are there challenges you’ve had to overcome to fully leverage its potential?
TYeah looking at it like scheduled HIL testing is a good point. There's a way to do this that could be much less expensive (ie, linode) and not reinventing the wheel if we're willing to do round robin 'appointment' scheduling. It's not what was originally envisioned, but perhaps a necessary next step.
We'd certainly welcome the assist, @ke7c2mi. Thanks for the offer. I have a lot going on this upcoming week but hopefully we'll be able to regroup next weekend.
Also, I'm open to any ideas from anyone that would provide an alternative to the "scheduling/appointment" phenomenon, which would be annoying.
We'd certainly welcome the assist, @ke7c2mi. Thanks for the offer. I have a lot going on this upcoming week but hopefully we'll be able to regroup next weekend.
Also, I'm open to any ideas from anyone that would provide an alternative to the "scheduling/appointment" phenomenon, which would be annoying.
TMaybe it's an extension of the AMA we just did with Ming. You have to submit your question and code ahead of time and then we'll select ones that are of the broadest interest and let that user 'drive' under guidance of the instructor/expert.
KI don't think it's particularly intuitive and I think it's quite a steep learning curve coming from bare metal / reading datasheets and poking memory. It has felt cumbersome for me compared to things like freertos/rtic/baremetal.
Device trees themselves are neat, and I think pretty intuitive. It's just some descriptive templatey IDL. On the other hand they map to your entire system in an abstract way so actually doing things with them can be complicated - the device tree itself isn't the problem there though!
Device trees themselves are neat, and I think pretty intuitive. It's just some descriptive templatey IDL. On the other hand they map to your entire system in an abstract way so actually doing things with them can be complicated - the device tree itself isn't the problem there though!
CHey Guys Cherry here, just joined the group. Excited to be here. Little intro about myself. I am a recent grad with over an year experience as an Embedded Software Developer.
BThanks for the suggestion. @Edison_ngunjiri also suggested a tinyML project that involves hardware and code that runs on a microcontroller. Am I correct to assume that it would be the first thing I need to understand as the step towards AI/ML in Embedded Systems?
Zfelt the same to me it was odd
ZWelcome Cherry, let us know what you are working and any way we can assist.
MIt's pretty neat
Was able to ship few modules with Zephyr on it.
Development is pretty direct especially since I like Linux kernel and the whole dts style of development.
One oof moment was the difference in opensource licence between the Linux kernel and Zephyr
Had to take down my PR due to the incompatibility
.
Was able to ship few modules with Zephyr on it.
Development is pretty direct especially since I like Linux kernel and the whole dts style of development.
One oof moment was the difference in opensource licence between the Linux kernel and Zephyr
Had to take down my PR due to the incompatibility
.TWelcome Cherry 
Where do you work/what do you work on currently?
Where do you work/what do you work on currently?
EIf you are interested in ML/AI in embedded systems, I would suggest you go right into it.
Just repeat the already existing projects on tinyML/picovoice.
With these two platforms, it's more of a process than actual coding.
After several projects ,you may want to look at pandas,numpy and scikit-learn for more understanding.
This has always been my approach.
From known to unknown.
Just repeat the already existing projects on tinyML/picovoice.
With these two platforms, it's more of a process than actual coding.
After several projects ,you may want to look at pandas,numpy and scikit-learn for more understanding.
This has always been my approach.
From known to unknown.
BThanks a ton
CHi, I currently work for locomotive manufacturing company its been year. My current work is mostly on testing and integrating the legacy projects to QNX.
Thank you
.
Thank you
.Tinteresting that its qnx. worked with other RTOS?
CNo I haven’t worked with other RTOS yet.
Zhttps://tweedegolf.nl/en/blog/137/rust-is-rolling-off-the-volvo-assembly-line Rust in in production in Volvo ECUs
In my job I get to speak to lots of people about Rust. Some are just starting out, some have barely ever heard of it, and then some people are running Rust silently in production at a very large c ...
Z@32bitSaviour
UIt seems like the automotive industry is preparing to adopt Rust. However, the author admits that Rust may not be suitable for early prototypes and certain use cases. I wonder what those use cases might be?
ZIn my experience Rust sucks if you have to do discovery or prototyping, this because you almost have to build the whole house of cards before it works. However, if you know exactly where you are going Rust is ok, I noticed that the earliest POC was in C then Rust came along, Also this is the 12 Volt non critical ECU so it seems there still is a bit of work to get to the mission critical stuff.
UI hope you're not referring to issues with Rust's tooling.
ZNo the tooling is great! just that compiler won't let you get by with things not done
Zwhich is usually what we do in protoyping, we hack stuff together, rust won't let you do that.
UI think I got your point now
AHey guys, I could use some career advice!
I'm currently in my second year of university studying Electronics and Communication, and I’ve been working with Arduino and microcontrollers since I was 14. Over the years, I’ve learned quite a bit about Arduino, AVR, STM32, ESP32, custom board design, Embedded C, register-level programming, and some IoT as well.
Right now, I’m trying to figure out what to specialize in so I can land a good job in the next two years. I’d love to work in a big MNC within the electronics industry or possibly join my dad’s instrumentation manufacturing business—or even start my own hardware company down the line.
I feel like I’ve got a good foundation in generalized electronics, but I’m not sure which area I should focus on next, especially considering current trends. What fields do you guys think would be best for me to master, given my background? I also feel like my early start gives me an edge compared to my classmates (most of them haven't even touched an Arduino), so I want to make sure I’m leveraging that advantage.
Thanks in advance for any advice!
I'm currently in my second year of university studying Electronics and Communication, and I’ve been working with Arduino and microcontrollers since I was 14. Over the years, I’ve learned quite a bit about Arduino, AVR, STM32, ESP32, custom board design, Embedded C, register-level programming, and some IoT as well.
Right now, I’m trying to figure out what to specialize in so I can land a good job in the next two years. I’d love to work in a big MNC within the electronics industry or possibly join my dad’s instrumentation manufacturing business—or even start my own hardware company down the line.
I feel like I’ve got a good foundation in generalized electronics, but I’m not sure which area I should focus on next, especially considering current trends. What fields do you guys think would be best for me to master, given my background? I also feel like my early start gives me an edge compared to my classmates (most of them haven't even touched an Arduino), so I want to make sure I’m leveraging that advantage.
Thanks in advance for any advice!
ZBasically In C, you can quickly prototype a motor control application by directly accessing registers and using global variables, even cutting corners for speed. In Rust, strict rules like ownership and borrowing force you to handle memory safety upfront, slowing down development during prototyping.
SI want to buy an oscilloscope can someone suggest me one around $400-500
UHere's detailed post with the list, price, and specs (Low-cost but still excellent oscilloscopes for SOHO) I'm sure you'll find useful. Best wishes!
A@Umesh Lokhande i would like to listern your suggestion towards this , as per the indian job market in hardware field
USure, allow me some time. Will add answer to existing thread (general-dev-chatHey guys, I could use some career advice)
TIsn’t it also in production on the space shuttle?
ZI know we have some cubesats running it but not sure about the ISS
DHi folks, 46 year old software engineer by trade, leveraging some time off to tinker around with some embedded, iot, 3d printing, and other various projects
ZWelcome @detiber Where do you work/what do you work on currently?
Where do you work/what do you work on currently?Sanyone using siglent sds804x oscilloscope here?
DHi @ZacckOsiemo, my last gig was in the Open Source Program Office at Cisco and have spent a lot of time over the years working on Kubernetes and Infrastructure/Platform automation.
I've always been interested in alternative computing architectures, microcontrollers, and fpgas, but never "had the time" to really explode. Now that I have a few months off, I'm planning on taking the time to dive in head first.
I'm currently furiously 3d printing some organizational needs for my office, then I'm planning on blogging, recording, and streaming my misadventures in hardware
I've always been interested in alternative computing architectures, microcontrollers, and fpgas, but never "had the time" to really explode. Now that I have a few months off, I'm planning on taking the time to dive in head first.
I'm currently furiously 3d printing some organizational needs for my office, then I'm planning on blogging, recording, and streaming my misadventures in hardware
CHi @Ming A friend of mine was wondering:
How do you approach a situation where a build doesn't work as expected due to variables changing, skipping, or being overridden by included meta-layers? The build is technically successful, but the outcome isn't what we anticipated. How can we debug this kind of behavior and effectively track variable values and overrides?
For a concrete example, consider an SoC vendor meta-layer with two different versions, such as community and proprietary. We aim to include only specific parts of the vendor's meta-layer, but it's interfering with the development kit BSP layers.
How do you approach a situation where a build doesn't work as expected due to variables changing, skipping, or being overridden by included meta-layers? The build is technically successful, but the outcome isn't what we anticipated. How can we debug this kind of behavior and effectively track variable values and overrides?
For a concrete example, consider an SoC vendor meta-layer with two different versions, such as community and proprietary. We aim to include only specific parts of the vendor's meta-layer, but it's interfering with the development kit BSP layers.
ZAmazing keen to see you get dug into those architectures
MWithout knowing the specific details, the approach I would take is:
First:, run
This will show you all the layers you have defined and more importantly, their priorities.
The next thing would be to run
Where
Now you can start searching that file for the variables you are interested in.
One thing I have found is that the quality of layers from vendors is very.... 'variable'! I have been know to create an intermediate repository from a vendor layer and then manually cherry-pick the bits I actually want. It is a pain, you you now have to manually keep the layers up to date, but it can be less effort than tearing your hair out trying to work out how to work around the vendor.
First:, run
bitbake-layers show-layersThis will show you all the layers you have defined and more importantly, their priorities.
The next thing would be to run
bitbake -e <recipe>Where
<recipe> is the image or recipe you are interested in. This will show you all you variables along with how and where they are set. This can be a large output, so I normally pipe it to filebitbake -e my-production-image > image-env.logNow you can start searching that file for the variables you are interested in.
One thing I have found is that the quality of layers from vendors is very.... 'variable'! I have been know to create an intermediate repository from a vendor layer and then manually cherry-pick the bits I actually want. It is a pain, you you now have to manually keep the layers up to date, but it can be less effort than tearing your hair out trying to work out how to work around the vendor.
A@Umesh Lokhande I am unable to reply in the thread, but I have a question - here python you mean micro python right ,
for embedded systems ?
for embedded systems ?
Dhey @Renuel Roberts , my response came in late but, I really appreciate its modular design and the wide range of drivers available. These features make it easy to create flexible and efficient solutions for various embedded applications, all while ensuring that everything runs smoothly in real time. hope this suffices.
UI mean to say generic Python, but it doesn't matter if you'd consider micro Python. Companies usually expect candidates to have problem-solving abilities irrespective of programming languages
AOk thanks
N
Friendly Reminder about Upcoming Event Join us on Friday, October 11th, at 15:00 GMT for a deep dive into "Embedded or Not, Virtualization is Headed Towards a Device Near You." Explore how virtualization technology is transforming embedded systems, learn about the role of containers, and share your own experiences with the community!
See you there!
https://discord.gg/7YmCCMFe?event=1293245463713288233
T@Middleware & OS and @MCU, MPU & Firmware, do you build devices that leverage embedded virtualization? I'm interested to know how many "IoT devices" utilize virtualization these days.
TAnd if so, are you using OSs with AMP that abstract some of that complexity?
