I built the latest Nobel build and ended up with Armbian-unofficial_26.02.0-trunk_Nanopiduo_noble_current_6.12.63.img. Even at idle it runs 70C. If you try to do upgrades it will do thermal shutdown. The temp reporting is real because it's hot enough to burn your finger if you hold the heat sink long enough. I used a PS2 usb fan (I call it the ghetto cooler 5000) long enough to downgrade kernel to 6.1.x. Now it idles around 50C. I know there's been heating issues with H3 (Duo uses H2+). I compared DTBs, voltages and cannot find the smoking gun. I played around controlling the voltages, etc. Anyways, I'm OK with 6.1.x kernel since this board was out in 2017. I still have a few I want to use for projects. I guess as long as I can use a current Ubuntu distro I'm OK.

 

OK, unpacked one of my old breadboard projects on focal Linux nanopiduo 5.10.60-sunxi #21.08.2 SMP Tue Sep 14 16:28:44 UTC 2021 armv7l armv7l armv7l GNU/Linux. CPU idles cool, so I do a:

 

sudo cat /sys/class/regulator/regulator.5/microvolts
1100000

 

It looks like 6.x kernel/nobel is using 1.3v. There it is. 1.1V. That confirms the mismatch perfectly. Your old Focal breadboard is running at 1.1V, while the modern Noble board is forcing 1.3V.

On these chips, that 0.2V difference is massive because power (and heat) scales with the square of the voltage. Moving from 1.1V to 1.3V increases heat output by about 40%, which is why one board is a stable 50°C and the other is a 96°C furnace.

 

So more hacking to do. I'll see if I can do a build hack.

 

Have you also looked at the clocks?

I have several NanoPi-NEO's (since around 2021 I think), H3 SoC, never had any overheating issues. I actually did some measurements recently and surprisingly low power, I had expected more. Clock stays at 480 mostly.

I see the Duo2 is H3, Duo I don't see on FriendlyElec's site. But quite different HW/board.

 

Only thing I could think of maybe is that I removed ( apt purge --autoremove) cpufrequtils after in-place upgrade from Bookworm to Trixie. Have not looked at it any further, it just is not in my vanilla Debian Bookworm -> Trixie, so just blind purge I thought. Still all fine, also with 6.18 kernel:

# uname -a
Linux raspi2 6.18.2-edge-sunxi #1 SMP Thu Dec 18 13:03:43 UTC 2025 armv7l GNU/Linux
# cat /sys/class/regulator/regulator.?/microvolts
3000000
3300000
5000000
1100000
# cat /sys/class/thermal/thermal_zone0/temp
27116

 

Even 6.1 kernel has voltage set at 1.3. I'm going to do a kernel patch and set it the same way as 5.x kernel was set.

Hm I briefly remember there was an optional overlay to raise voltage to 1.3 volts..but I don't know why it should be now standard. Did not do any investigation though

@Werner OK, since I have a few Duos (v1.0 and v1.1) I tried same Jammy with 6.1.63 kernel SD in each and let idle. The 1.1 with copper heatsink runs about 40C, the 1.1 with same aluminum heat sink runs around 45C. The v1.0 runs around 50C. So weird. I'll try latest build, but stress-ng causing thermal shutdowns worries me. I used to run decent loads on these little boards. Good thing is thermal shutdown works or it would burn up the board.

So the workaround is to set cpu freq max to 1 MHz. I still have to use a ghetto cooler 5000 when running a process that maxes out the CPU. Seems Duo running 10C+ at idle with latest kernel and Noble. I can live with that as long as I know. The default is 1.2 MHz max which runs too hot at 1.3V.