HP Probook Random Shutdown Update

In my last post I talked about a possible fix for the HP Probook 45xx random shut-down issue. Basically, the computer would shut down randomly, and I was unable to determine the cause for some time. I suggest updating the graphics drivers, as Intel now offers open source driver support for Linux users.

Although the driver update did help, it turns out that the issue was not entirely fixed. About a week after that blog, it started happening again, albeit less often. I've probably had about 12-15 random power-off's since then. I realized that it happens when performing resource-intensive tasks, in particular graphics related tasks. So this led to me to back to the graphics drivers. It turns out the Intel driver is tainted:

anon@linuxpc:~$ dmesg | grep taint
[    2.134771] drm: module verification failed: signature and/or  required key missing - tainting kernel
[    2.135831] drm: module has bad taint, not creating trace events
[    2.158791] i915: module has bad taint, not creating trace events


I then realized that my computer has been running a lot hotter than it was a couple months ago when I first acquired it, so I started monitoring the temperature while doing different things. As it turns out, when playing high definition video via Adobe Flash, things started to get really hot, with both processor cores running at 80 degrees Celsius, and the power supply running at about 90C. I figured that the computer was probably simply shutting down due to overheating, as every HP computer I've ever owned had that problem at some point.

Since I don't have any duster cans lying around, I grabbed my trusty little Phillips-head and a vacuum cleaner, then got to work. Following standard procedure, and the service manual , I took the computer apart and simply vacuumed all the dust out of the heat-sink, motherboard, and everywhere else I could see dust. If you attempt this, please be careful and be sure to ground yourself to protect your hardware from static electricity. This I accomplish by wrapping a copper speaker wire around a screw and sticking it into an outlet ground, and then wrapping the other end around my wrist.

After the vacuum treatment, I monitored the temperatures while performing various resource intensive tasks, and the temperatures of either core never got above 60 degrees C, and every other sensor was much lower than it was previously. Here is the current output of sensors:

anon@linuxpc:~$ sensors
acpitz-virtual-0
Adapter: Virtual device
temp1:        +16.0°C  (crit = +108.0°C)
temp2:        +54.0°C  (crit = +105.0°C)
temp3:        +50.0°C  (crit = +108.0°C)
temp4:        +48.0°C  (crit = +105.0°C)
temp5:        +26.2°C  (crit = +108.0°C)
temp6:        +45.0°C  (crit = +110.0°C)

coretemp-isa-0000
Adapter: ISA adapter
Core 0:       +51.0°C  (high = +105.0°C, crit = +105.0°C)
Core 1:       +48.0°C  (high = +105.0°C, crit = +105.0°C)

Definitively in the normal range again! I really hope that this solves the shut-down problem for good. I'll update again in a week or so. By the way, I think it's better to use a vacuum cleaner rather than compressed air, because I once wrecked a computer using duster. All it did was blow the dust further into the internal components, making the problem worse. A vacuum does the opposite. Perhaps this method will become the preferred method for minor dust removal. Happy hacking!