Temperature of Netgate 4100 becomes quite hot in idle mode
-
Hello,
I have two Netgate 4100 devices and I am astonished how hot these devices become and I wonder if there is a thermal problem. Especially because one is much hotter than the other. They are doing basically nothing. CPU is around 0-2 %.
I also have a Netgate 5100, which was always around 40 °C.
FW1:
[23.01-RELEASE][admin@XXXXXXXXXXXX]/root: sysctl -a | grep temperature hw.acpi.thermal.tz0.temperature: 26.9C dev.cpu.1.temperature: 65.0C dev.cpu.0.temperature: 68.0C
FW2:
[23.01-RELEASE][admin@XXXXXXXXXX]/root: sysctl -a | grep temperature hw.acpi.thermal.tz0.temperature: 26.9C dev.cpu.1.temperature: 54.0C dev.cpu.0.temperature: 56.0C
Do you made similar observations?
-
Power saving mode is set to adaptive on both devices.
-
Staying around 49° C.
Even when the load gets higher, it'll not raise above 53° C.
The device is in a small network cabinet which is not actively cooled.
Very happy with the temperature. It's a non-max 4100. -
What's the ambient temperature those devices are in?
-
How are they mounted? Are they sitting flat on a desktop, stacked or wall mounted?
If wall mounted, what orientation? Are the ports on top?I know we like fanless units because they're quiet, but it's interesting how much temps drop when a fan simply moves some air around them. Not "blowing on them" but creating circulation.
-
@mer I have them in a office rack. I attached these rubber feet to increase the gap and I stacked them on a metal tray with some slits. The firewall directly on the metal tray is the one with higher temperature.
I also noticed that by putting some more spacer between the firewalls to increase the gap by a centimeter, it results in 1-2 K lower temperatures.
With a closed front door and active rack ventilation, the internal rack temperature is around 27 °C which results in the temperatures that I posted before (without spacer). With an opened rack door, the rack temperature goes down to 22 °C and the cpu0 temperature goes down to 62 °C for fw1 and 50 °C for fw2 (with spacer).
Because you mentioned that passively cooled devices profit from a little air stream, I ordered a horizontal rack ventilator. I hope that is improves the situation and leads to a longer hardware life.
-
Any active airflow at all there will likely lower the temps significantly. It will be interesting to see the results.
-
@stephenw10 I expect the ventilator to arrive at Friday and will share the results around Monday. It may help other people with the same problem.
-
My 4100 max doesn't do a lot, except running around in circles.
Since January, VDSL 25 Mbits/sec is gone, entering 800 Mbits/sec up/down fiber. That didn't change much. -
@roesh I think it should. Like others, it will be interesting in seeing the delta in temp.
I have mine mounted vertically, with a typical "office fan" (about 9 inch diameter) blowing up at about a 45 degree angle, low speed and I'm seeing temps of about 22-24C. Keep in mind that I'm in NH and my house temp is about 18C :)
But vertical mounting, with air flow even from say a computer case fan on USB helps a lot.
-
I installed the horizontal fan. With a closed door, the core temperature dropped to 51 and 47 degrees, but I have seem lower temperatures as well. When I open the door, the temperature is around 42, so maybe there is a bit more improvement possible with increasing the power of the rack top ventilator.
-