I have some instruments that use the old Beagle BOARD to control them remotely, they have worked for
several years. Now, two of them will run correctly for a few minutes, then fail with a kernel panic.
The first line in the console log is :
Unable to handle kernel NULL pointer dereference at virtual address 00000000
Internal error: Oops: 5 [#1] SMP ARM
and the last line is :
Kernel panic - not syncing: Fatal exception in interrupt
I do have the full trace if there’s anything specific in there that would be helpful.
The first thing I thought to do was to fsck the SD card, and they now come up clean.
This is running Ubuntu 13.04 (GNU/Linux 3.7.10-x13 armv71)
Thanks for any suggestions,
Impressive, that was released on Jun 25, 2013... Almost 6 years ago..
Sadly have you tried swapping to a different board? Something that
hasn't been running for 6 years? Maybe a new SD card?
This unit has not been running continuously for 6 years, but could well have a year or two of total power on time.
It is coming back to me for repair. I will certainly try creating a whole new SD card and see if that works. I have the entire SD card archived, so that should be easy to do.
I don’t have any Beagle BOARDs here to try replacing with. So, the other alternative is to replace it with a Beagle BONE.
I have made an adapter that allows me to use a Bone in place of the original Board.
But, I’d kind of like to know what quit and why.
Well, I still don’t totally understand it, but I did fix it. So, this device is a remote-controlled signal multiplexer used in nuclear research experiments, so you can bring up signals on a remote scope, while the ion beam is blasting the vault and nobody can be inside.
It runs in a NIM bin that provides regulated voltages, including +6 V. So, I used an L4941 low dropout 5V regulator to power the Beagle Board. This worked fine for a couple years. Some of the units did not heat sink the regulator, but it only got mildly warm, so I thought that should be fine. Now, a couple units would crash after running for a couple minutes.
The fix was to bolt the regulator to the case. So, I don;t know if this means that over time the regulators started sagging due to temperature, or the Beagle got more sensitive, or what.
But, the fix seems to be quite solid.
Well, always learning something new…