Problem with a short power loss running from LIPO battery

38 views
Skip to first unread message

benjg...@gmail.com

unread,
May 31, 2016, 7:58:35 AM5/31/16
to BeagleBoard
Hi all,

We've got a real problem with a BeagleBone Black running Debian Jessie connected to a LIPO battery through pins TP5-8. The system will  be connected to a power supply that can be turned on or off without notice, so we added the battery to enable a soft power down and prevent system damage. 

Normally this works brilliantly, but if the power only goes off for a couple of seconds then the beaglebone is still powering down when the power is re-applied then the "PWR" LED comes on, but none of the user LEDs light up and the system won't work. While it's in this state powering off and on again makes no difference, it just stays like this until the battery runs out or until the reset button is pressed.

Has anybody got any ideas?

Thanks,

Brn.



evilwulfie

unread,
May 31, 2016, 8:27:28 AM5/31/16
to beagl...@googlegroups.com
smart watchdog that cycles the battery via a ssr.
--
For more options, visit http://beagleboard.org/discuss
---
You received this message because you are subscribed to the Google Groups "BeagleBoard" group.
To unsubscribe from this group and stop receiving emails from it, send an email to beagleboard...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/beagleboard/03ff547f-8cf3-4379-8a69-a9fa2e08c6e6%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


William Hermans

unread,
May 31, 2016, 12:16:11 PM5/31/16
to beagl...@googlegroups.com
On the software side of things. The Debian package "acpi" is mostly responsible for this immediate shutdown. How exactly, I'm still not 100% sure. However . . .

william@beaglebone:~$ cat /proc/interrupts
. . .
189:          0      INTC   7 Level     tps65217
. . .

william@beaglebone:~$ cat /proc/irq/189/spurious
count 0
unhandled 0
last_unhandled 0 ms


This seems to be the file from which userspace can watch for pgood signal interrupts from the PMIC. I've briefly tested pushing the button, *and* removing input power, and this file tracks interrupts from both in this file. That is, the PMIC sends this interrupt, through an NMI pin.

So, you could get your hands on the acpi package source code, and rewrite is so a power "blip" does not necessarily power down the board. *Or*, you can uninstall acpi, and write your own code to watch this file, and behave according to how you feel it needs to behave. Personally, I've been seriously considering the latter, but I have not gotten to writing said code for our ow project yet.

Reply all
Reply to author
Forward
0 new messages