Odroid C4 higher idle temperature when USB-SSD or Stick ist plugged in

General Discussion
1

Hi Folks!
i got a weird problem which i don’t understand at all…

Right now, i run a Odroid C2 which runs Pihole in a docker container and a sambashare with a M.2 Drive in an USB-M.2 Case. The idletemperature of the C2 is 27°C. And i like the “who put me in the freezer” comment of dietpi a lot! The cpu temperature only rises if i copy bigger amounts of date via rsync.

Now i wanted to change to the Odroid C4, because of the ability to run the M.2 via USB 3.0 with higher speeds. When i install dietpi with docker and pihole ( on the C4 ), i got idletemperatures of 31-33°C. No longer perfect, but ok.
When i plug in an USB-Stick ( in the C4 ) or the M.2 SSD, the idltetemperature rises to 36-40°C without mounting the Stick or SSD!
When i mount the SSD or stick, the temperature stays at 36-40°C.

The cpu-load ist very low, pihole says 0 0 0, htop says very low cpu loads too.
I knew that USB 3.0 would consume more energy, but i thought this energy would be consumed by the USB bridge. I see no higher cpu-load when plugging in an usb-stick, but the cpu temperature rises.

Can someone explain why my cputemp rises with a plugged USB-Stick?

Thanks in advance!
Greets, Thomas

2

because the temp of CPU is also influenced by every component around it. Maybe the temp of the USB bridge raises so much, that the emitted infrared radiation influences the CPU next to it?!
Do you have the ability to measure somehow the components, e.g. with a infrared thermometer?

3

Hi Jappe,

thanks for your help! As far as i know, the heatsink covers the cpu, and 2 RAM-Packages.
Heat from other components can affect the cpu-temp by ir-radiation or simply by heat-buildup in the case.
Unfortanetly, i have no possibility to measure the componenttemperatures good enough.
I have an IR-measurement-thing, but i do not know what the angle of the radiationdetector is.
I think the only way would be a heatcamera…
I have made bad experiences with these IR-Measurementthings, the material of the surface, as well as coatings, let you measure very weird things…

Greets , Thomas

4

Another I thing I could think of is, that the clock is raised although you have no higher load. You could check that with cpu command.

5

Hi Jappe,

i looked at the cpu frequency via command:
cat /sys/devices/system/cpu/cpufreq/policy0/stats/time_in_state

The output said that over 90% the clock was reduced to 1 Ghz.
Looks fine to me…
This is why i cannot understand the higher CPU-Temp which means for me a higher power consumption…
I think for the moment, i stay with the odroid C2, this device has none of these problems…

Thanks for your help!
Greets, Thomas

6

Well you try to compare two different devices running different CPU type aso. This never will match.

7

Hi there,

i think i got it…
The Odroid C2 runs at 100 Mhz - 1.5 Gigahertz,
the Odroid C4 runs at 1 Gigahertz - 2.1 Gigahertz.

Both are running most of the time at the lowest frequenca, but the lowest frequency is 10times higher on the C4!
Can i change the lower frequency ?? When i try it using dietpi-config → performance options,
i can choose the lower frequency down to 100Mhz on the C2, but only on 1Ghz on the C4.

Is the C4 not capable of lowering the cpu frequency as deep as the C2?

Thanks in advance, Tom

8

According to this site: https://wiki.odroid.com/odroid-c4/application_note/software/set_cpu_freq#tab__odroid-c4m1m1s
the Odroid C4 should be capable of idling at 100Mhz…

The output of
cat /sys/devices/system/cpu/cpufreq/policy0/scaling_available_frequencies

is "1000000 1200000 1404000 1500000 1608000 1704000 1800000 1908000 2016000 2100000 "

Greets, Tom

9

This topic was automatically closed 180 days after the last reply. New replies are no longer allowed.