Question

2

Dell Precision 3660 shuting down for overheating

rated 0 times [ 2] [ 0] / answers: 1 / hits: 719 / 3 Years ago, sun, november 21, 2021, 10:40:51

When monitoring the temperature values of my machine, I see that one of them is slowly increasing under heavy load, ultimately triggering an emergency shutdown from thermald.

I read the types of thermal sensors from /sys/class/thermal/thermal_zone*/type and their corresponding temperatures from /sys/class/thermal/thermal_zone*/temp

However, I didn't find anywhere the definitions of the values returned by /sys/class/thermal/thermal_zone*/type. Some look pretty self-explaining, like TCPU for example. But the offending one for me is AMBF and I have no idea what it refers to.

Does anyone know what this AMBF stands for and/or where to find some documentation about the various types of thermal zones?

EDIT:

I did much searches since the issue first occurred and here are my findings:

I monitored the thermal sensors and made the following chart

One can see that the CPU and memory temperatures are well dealt with (the CPU TCase is 100°C, so never overtaken). However, the AMBF one keeps on growing until it reaches it's critical value of 80°C and triggers the emergency shutdown

Browsing the ACPI table, I finally managed to find that AMBF stands for AMB and Near Fan Temperature (QE3)

I've installed a Windows 10 OS for the unique purpose of testing there. And:
1. The AMBF entry isn't showing up in the ACPI table there (as far as I can see)
2. The issue doesn't materialize on Windows. Running the exact same test doesn't trigger any emergency shutdown

Dell published recently a critical customer advisory

(Dell Technologies Internal Reference ID – Dell Technologies ET004676)

This Customer Advisory is to inform you of a recent discovery involving specific Optiplex and Precision systems which may encounter a thermal shutdown from Modern Standby. Dell Technologies is recommending immediate upgrade to the latest BIOS to maintain optimal system performance and to avoid experiencing any issue.

Needless to say I already was on the latest BIOS version. But I suspect this Customer Advisory might explain why the issue doesn't materialize on Windows... So I tried to boot my Linux system while pretending to be a Windows machine from an ACPI standpoint by using acpi_osi=! acpi_osi='Windows 2019' on the boot command line. Unfortunately it didn't make any difference.

Anyway, I'm still in the same situation, incapable of using the machine for the purpose I bought it. Any further idea very much welcome.

Answers

Only authorized users can answer the question. Please sign in first, or register a free account.

ranquctive

Add To Favorites

Follow

Total Points: 391

Total Questions: 103

Total Answers: 104

Location: South Sudan

Member since Thu, Feb 4, 2021

3 Years ago

ranquctive questions

1 Setting up the $DISPLAY environment variable on WSL2 to use Qt software

Wed, Jul 6, 22, 19:44, 2 Years ago

1 Run shell script as root user while being a regular user

Sun, Apr 10, 22, 23:19, 2 Years ago

1 Gnome shell restart is not working in Ubuntu 22.04 using ALT+F2

Thu, Oct 14, 21, 08:23, 3 Years ago

1 How do I check if command failed... with a pipe?

Sun, Apr 3, 22, 17:43, 2 Years ago

1 Battery life limitation with Ubuntu 20.04 and AMD

Thu, Mar 3, 22, 03:08, 2 Years ago

View All

answered 3 Years ago tionavocad · Accepted Answer

I finally had an answer from Dell support telling me that the AMBF sensor could be safely ignored and the critical shutdown at 80°C disabled.

So I've created a /etc/thermald/thermal-conf.xml file like this:

<?xml version="1.0"?>

<ThermalConfiguration>

  <Platform>

    <Name>Override AMBF critical trip point</Name>

    <ProductName>*</ProductName>

    <Preference>QUIET</Preference>

    <ThermalZones>

      <ThermalZone>

        <Type>AMBF</Type>

        <TripPoints>

          <TripPoint>

            <Temperature>95050</Temperature>

            <type>critical</type>

          </TripPoint>

          <TripPoint>

            <Temperature>80050</Temperature>

            <type>hot</type>

          </TripPoint>

        </TripPoints>

      </ThermalZone>

    </ThermalZones>

  </Platform>

</ThermalConfiguration>

Here, I've decided to set a hot level at 80°C where it initially was critical, and keep a safety net with a critical level at 95°C.

I've also installed the latest thermald from Intel's github repo

From there, I've re-run my test and here is the graph of the execution:

As you can see, all temperatures are still well managed, and AMBF can now freely go over 80°C. It stabilizes at about 92°C at full load, and returns slowly to its normal level when the run is over.

The issue is now fixed and my machine fully operational.