Hello,
I bought a case set for the RPI. It turned out that the included power supply works perfectly with the RockPi 5. It also has a switch to turn it on. Just as a hint
https://amzn.eu/d/hY8y9mo 13,99€
Greetings Bernd
Inexpensive power supply that works perfectly :-)
1 Like
This is a 5V, 4A PSU. That means:
- A fully kitted out Rock 5B (with fan, NVMe, Wi-Fi and USB devices attached) might draw up to 30W, and overload it. In the best case, it will shut down due to overload protection, but in the worst case, the PSU may overheat and present a fire hazard. Make sure to only draw no more than 20W from it.
- 5V is the bare minimum required for the Rock 5B to boot. If the voltage sags below 5V (e.g. under load), it may cause problems. (Most power supplies intended for the Raspberry Pi 4 output 5.1V or even 5.2V, but this one claims to be 5.0V.)
1 Like
Ich have alredy USB WI-Fi , NVMe ( Bootloader SD + NVMe, Fan with Fan control … plugged. And it works as a charm ( Armibian) .
Also overclocking is not a Problem.
Im do my own “Retro Amiga Commodore” Workstation thing. So compiling an stressing is everywhere
Next step is to integrate a OLED Display …
Picture is old ( without Fan )
But…you´re right… it could happen…
1 Like
That would have to be an extraordinarily useless and probably illegal power supply.
Where did you get that information?
BTW: using sbc-bench -m you can somehow inprecisely measure input voltage (especially drops under load): Realtime power usage or see below…
As for Rock 5B’s power requirements I really don’t know why everybody thinks 30W would be anything realistic/needed. These ‘30W’ are the rated wattage of Radxa’s USB PD PSU at 12V/2.5A which should be suitable for Rock 5B with lots of host powered USB peripherals and all SoC engines under full load.
If you’ve not lots of host powered USB peripherals and a ‘bad’ SSD with insanely high peak current requirements you’ll have a hard time ever exceeding 15W. When running stress-ng --matrix 0 for example consumption increases just by 10W (measured at wall with a 15W RPi USB-C power brick):
root@rock-5b:/home/tk# sbc-bench.sh -m
...
Time big.LITTLE load %cpu %sys %usr %nice %io %irq Temp DC(V)
11:19:13: 408/1800MHz 0.52 0% 0% 0% 0% 0% 0% 33.3°C 5.23
11:19:18: 408/1800MHz 0.48 0% 0% 0% 0% 0% 0% 32.4°C 5.21
11:19:24: 408/1800MHz 0.44 0% 0% 0% 0% 0% 0% 32.4°C 5.21
11:19:29: 2400/1800MHz 1.05 84% 0% 84% 0% 0% 0% 46.2°C 5.19
11:19:34: 2400/1800MHz 1.60 100% 0% 100% 0% 0% 0% 49.0°C 5.14
11:19:39: 2400/1800MHz 2.12 100% 0% 99% 0% 0% 0% 50.8°C 5.21
11:19:44: 2400/1800MHz 2.59 100% 0% 99% 0% 0% 0% 52.7°C 4.99
11:19:49: 2400/1800MHz 3.02 100% 0% 99% 0% 0% 0% 53.6°C 5.18
11:19:54: 2400/1800MHz 3.42 100% 0% 99% 0% 0% 0% 54.5°C 5.04
11:19:59: 408/ 600MHz 3.79 72% 0% 72% 0% 0% 0% 45.3°C 5.40
11:20:04: 408/ 600MHz 3.48 0% 0% 0% 0% 0% 0% 41.6°C 5.21
^C
BTW: every 5V power source that claims more than 15W/3A is nonsense anyway since the voltage drops when exceeding 15W are too much. There’s a reason none of the USB-C and USB PD power modes define anything at 5V that exceeds 3A.
2 Likes
What do you refer to as ‘overclocking’? Setting cpufreq governors to performance all the time?
BTW: when acquiring a random power brick especially with nonsense specs (5V @ 4A) I would better test this thing.
The only thing you now know is this thing being a dumb 5V PSU not suffering from Rock 5B’s USB PD negotiation hassles. As such it’s working (as any other dumb power source) but whether it works ‘perfectly’ is a different story.
You can measure input voltage drops using sbc-bench -m and by installing stress-ng set the CPU cores under full load and see what happens. If the board survives I would then run a random I/O test on an attached SSD in parallel to further increase consumption (stress test the PSU) and maybe even attach something sucking significantly juice to the USB3 ports.
Worst case conditions at least once tested…
