In fact most often the PSUs designed for LED lighting are even cheaper and sometimes even more efficient, however they’re less well regulated. But here given that 12V is only used to feed DC-DC converters all over the place, we probably don’t care about that. I’ve bought some small ones for other projects (powering my home lab) but have not not tried feeding the O6 with it yet.
They should work fine. And if they can output egnuth power you can hook them up to a GPU 6 or 8 pin header to power them as well
Jeff geerling made a great point, If they’re gonna create an ARM desktop motherboard why not full send it?
I believe radxa should have spent more time developing this board adding more QoL features to it that make it in line with a regular x86_64 itx desktop motherboard.
Though to be fair there is like… Almost no consumer grade ARM hardware that comes close to this, It’s either SBC or high end stuff.
Perhaps radxa could create a newer revision of this board?
Perhaps with sata ports?
Wake On Lan?
Proper power stuff?
And ideally some case fan headers…
2 Likes
Agree this is really the hybrid of a devkit and a consumer level device.The next iteration has a lot to fix but would make it even more compelling than it already is.
Since they also made a Rock5ITX-Plus I wouldn’t be surprised if things around software get better they consider making an Orion-O6-Plus
3 Likes
Ideally i would prefer if they didn’t rush to make new hardware and instead focus on making this board work properly first.
Most SBC vendors tend to design and sell hardware, Use basic vendor provided software and call it quits.
I’m rooting for a fully mainlined board that can be considered real ARM desktop class hardware, No need for special custom kernels to get features working just like a regular desktop.
I’m really hoping radxa goes above and beyond with this board and probably start leading in this space since as i said… Sadly nobody seems to care about mainline…
1 Like
It’s not nobody. https://libre.computer/ is an example that goes the extra mile beforehand but that upfront investment is basically the opposite of that from the innovator board vendors that enable us to have the hardware at none DevKit Business prices.
1 Like
I’d really like to see a Plus revision with this hardware issues fixed but not before we get at least 90% of software related stuff working under the mainline kernel
2 Likes
So like, It’s been a hot minute since any updates were announced, Anything new?
Any ETA for when most of the major breaking issues will be fixed?
Or ETA for proper mainline?
I know things take time and i do expect roughly a few years until full mainline support because unfortunately that’s just how it is in the ARM world, But this was marketed as an open board?
I’m mostly interested in knowing when the fix for ethernet will come (The one that would fix the occasional lock-ups and the one that would fix the speeds by default)
I’ve also started to randomly experience full blown crashes (Orion would just completely shut down and restart for no apparent reason), This happens like once per i’d say 50 startups or something?
In terms of firmware i’m also waiting for the build of UEFI that would fix the core layout and hopefully enable front IO access, i heard that the 9.0.0 one has that working but i’d rather have 12 cores than a few more bits of IO.
Ideally i would love to have internal GPU/NPU access too but that i expect to be available within 2 years, At least that’s what i waited for on rk3588 with my orange pi 5 plus.
(Sorry for the rant)
1 Like
You can have 12 cores with 9.0.0, you just need to enable them in the UEFI boot menu. Select Device Manager, then Soc Configuration and finally CPU Configuration. Set all the cores to “Enabled”, and press F10 to save. Done!
Now if there was just also a way to get an efifb larger than 1920x1080…
Iirc there were like different versions of 9.0.0, One that did allow you to enable those cores and one that didn’t even have them listed in the UEFI.
I forgot the specifics, But why haven’t we gotten any updates from CIX/Radxa on the firmware situation?
Since nobody cares. Easy as that.
Other possible explanation: maybe everything relevant only reaches us if we’re stupid enough to sign up on Twitter and Discord (which no-one right in his mind would do)
1 Like
I also unfortunately feel like most updates are locked behind their discord group… Not everyone uses discord, Radxa has their own forums, Any updates (Be it small or major) should be here first.
I follow radxa on twitter, Have not heard anything about CIX/Orion since it got certified as systemready.
I hate to be this kind of guy but i feel like all momentum has ceased post systemready status, At least publicly…
I’m still waiting for the proper CPU clock speeds update to arrive, I purchased this pre the reduced clock speed website change i expect more from this board.
Again CIX is a new player but they made a bold claim about trying to be “The world’s first open source ARM v9 board” i thought radxa would also try hyping this up as much as possible putting as much attention to this but nope, Radio silence…
Radxa please… Update us?
3 Likes
Same here, I even planned to reserve a part of my holidays to hack with the released SDK… which never came after the systemsready announce put everything to a halt. So yes, all the momentum that was going by then stopped, that’s sad 
Yeah… Wish someone at Radxa would comment on this.
This board has a lot of potential.
Just got myself a hold of an rx 7600 xt (Pulse sapphire, No i could not find a white version in stock…)
And i can confirm that it works perfectly* in linux!
*Data throughput readings are not available in nvtop, No clue if that’s an nvtop issue or a what issue.
I will say, This card is pretty sick though pretty useless right now…
With the CPU cores clocked at such low frequencies emulated games will be significantly bottle-necked by the CPU.
I did try the 1.0.0-1 firmware buuut that unfortunately still needs more work (It is marked as pre-release just to be clear, It’s clearly not ready)
I wish radxa focused on mainlining this thing faster, This board really doesn’t deserve to be treated like the rest of the ARM boards.
Don’t get me wrong radxa is doing some stuff, But i feel like it’s not happening fast enough…
Once this board has it’s iGPU working and the CPU clock speeds and layout gets fixed (Maybe even some over clocking 
) this board paird with an rx 7600 xt would be a legendary setup.
Finally, A consumer grade ARM desktop capable of every day tasks including proper gaming!
AI is overrated but the very powerful NPU inside this would be amazing for “free” low power local AI tasks.
So uh, I sort of need help debugging something, I’m not 100% sure what the culprit is behind this.
Could be:
The board (Due to CPU scheduling shenanigans which is VERY plausible)
The emulator
Or just the game
Emulated (x86_64) games (Specifically using vulkan) seem to have crazy stutter after an extended amount of playtime.
This seems to be triggered via input device events (Mouse movement or keyboard interaction)
I unfortunately cannot test box64 as it doesn’t work for me or the games that i play but if someone else here can give that a try i would appreciate it.
Games i tested that had this stutter:
WarThunder
Team Fortress 2
Space Engineers
I haven’t noticed this stutter during the time i was using an orange pi 5 plus as a daily driver for a few months hence why i think it’s something to do with the board. (I was gaming on that thing too)
I will be testing native aarch64 games for a bit now just to figure out if it’s caused by emulation or what.
Any help with this would be appreciated. (Can’t send a video as the limit is 4mb… Sorry…)
I’m really hoping the random USB issues would get fixed soon… Or at least hear from someone else about this issue…
I was doing some dev work on something right now when all the sudden it stopped detecting the device, I could have sworn i fried the damn thing, Nope! Just my O6 acting up as usual…
One reboot later fixed the issue…
Not to mention that my O6 crashed this morning when i turned it on and rebooted on it’s own…
I don’t mean to bash radxa for this as i highly doubt it’s their fault but this is awful.
I am on UEFI 1.0.0-1 (Technically latest github commit so even newer than that)
With linux 6.17-rc5.
I understand that “Pre release” software is unstable and i should expect bugs, Unfortunately this is the more stable setup
0.2.0 was completely unstable
0.3.0 was about as stable as this
It’s been almost a year since this board was out for sale i guess (Or rather was supposed to)
Major issues like this should not persist this long.
I’m not sure who to contact or where to post about this but if someone from the radxa team would kindly do some long term testing like i am and get this fixed i would highly appreciate it…
Unfortunately due to the nature of these issues there’s no real steps to reproduce any of this.
Though from my testing it seems like:
Booting up after a full shutdown (Board unplugged from power) seems to have a higher chance of either crashing the board, Crashing the network interfaces, Or flickering the USB ports.
Thank you for reading my rant…
Just a little frustrated with this happening in the middle of a debug session…
Yes, I think that I might have noticed the USB issue a couple of times as well, but it has been infrequent enough, so as to not bother me much. Note that in my case the USB devices are usually plugged into a hub (my display), which is then connected to the Orion O6 board with a single cable.
Otherwise I am now convinced that most of the crashes that I have been experiencing have been due to thermal issues. The stock cooler just seems to be inadequate, especially if the board is installed in a case and some demanding workload such as building software is running, while utilizing all processor cores. I have also switched to a NVIDIA graphics card for a bit, which has eliminated another source of instability.
My mouse and keyboard are plugged directly into the orion, They still sometimes flicker.
The rest of my stuff is plugged into a hub, Fortunately post 0.2.1 firmware it seems like unplugging/replugging might make things work.
Previously doing so did nothing (It still sometimes does nothing and you have to do a full reboot)
That being said, My crashes were not caused by anything thermal.
At least one would hope so, I don’t think a board that has been sitting offline for a few hours should get that hot the moment you decide to boot it up.
The stock cooler is more than enough for me, Highest temp i saw was like idk 51c?
And i do stress my hardware quite a bit.
1 Like
Out of curiosity, what temps are you getting under load? I get hard crashes while running heavy llama.cpp workloads sometimes and am trying to pinpoint the issue. Temp of the CPU never goes above 70C (usually around 65C under heavy load) so assumed the issue is probably somewhere else. But maybe the P1 is quite sensitive to temp?