Hi Michael,

Have had the same sort of issues as you.

Default gumstix kernel enables charging on the RTC coin cell circuit. It's a 50uA rate though, so exceptionally slow.  I'd definitely either turn off the charging (recompile kernel) or forget about using throw away lithium cells. Risky!

With regards to rechargeables: I found that if I bumped things up to the max charge rate, the Lithium vanadium pentoxide cell lasts a month or so (I have not measured past that) with no power. Mostly my problem was that the conservative charging rate that is the default in the kernel does not feed the cell very fast, and thus you need to leave the damned thing on for days to get a charge.

I have experienced a complete loss of time when unmating a TOBI board from a custom carrier board (board loaded up with 2000-01-01 time, which presumably suggests dodgy time in RTC memory). I suspect something shorts a bit when that occurs.  Be careful of noise, therefore, on any expansion headers. 

The other thing I've noticed, which may help Philip, is that on shutdown (depending on your rootfs/kernel) the system attempts to flush systime to the hwclock. As a result of this, if anything goes even a little pear-shaped on a shutdown (let's say your power supply starts going down a little faster than you expect), I have found that a completely dodgy value can be pushed out to the clock. 

My thought was it might be a bad i2c transaction (the amount of time 'shift' as a result of this power-down dodginess was about 12-24 hours). Which I believe amounted to about 1 bit of rot.  Never got around to debugging properly, however. Instead, I avoid writing to the RTC during any situations where my system power supply might be even a little bit dodgy.

Hope this helps.

Cheers,

Akram



On Tue, Aug 6, 2013 at 2:33 PM, Michael Wynne <michael.wynne1@gmail.com> wrote:

I have seen this with the AirSTORM. I tried using some rechargeable 123A lithium cells thinking they would last for long periods of being powered off, but they didn't. Then I tried a regular lithium 2032 button cell and those were dead in a matter of several days. I thought they might heat up and explode because there does seem to be voltage at the RTC Vbackup, though the current must be extremely small because they didn't even get warm. That might explain why the relatively large lithium cells didn't really charge.

Since then I've moved to the recommended Lithium Vanadium Pentoxide rechargeable battery (http://wiki.gumstix.org/index.php?title=Coin_cell). It's very small, but seems to last as long as the others I've tried. Though I haven't conducted extensive testing yet. I have 10 units with this battery and I think I've only seen one go dead and loose the time after being powered off for about 3-4 days. The bottom line is these batteries should be lasting longer. I can leave a desktop PC unplugged for a month and a small battery will keep the RTC going, so what's the deal with these COM's?

At this point I have other projects to worry about, but I intend to investigate further when I can. If anyone can offer any other suggestions I would also be grateful.

-michael

On Aug 5, 2013 5:31 PM, "Nye, Philip" <Philip.Nye@acuitybrands.com> wrote:
I am getting reports of the system clock on the Gumstix (overo earth storm) losing big jumps of time - hours or days. This is most likely happening during power cycles though I have not been able to catch it happening. The overo is on our own board which has a non-rechargeable 3V lithium coin cell backing up the RTC.

Does anyone else have experience of anything like this? or pointers to what might be happening?

Is there anything in the backup battery circuitry which could cause the backup battery to drain at an abnormal rate? I know the circuit will trickle-charge a rechargeable cell and ours has a series diode to prevent unwanted back current.

Philip


------------------------------------------------------------------------------
Get your SQL database under version control now!
Version control is standard for application code, but databases havent
caught up. So what steps can you take to put your SQL databases under
version control? Why should you start doing it? Read more to find out.
http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
_______________________________________________
gumstix-users mailing list
gumstix-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/gumstix-users

------------------------------------------------------------------------------
Get your SQL database under version control now!
Version control is standard for application code, but databases havent
caught up. So what steps can you take to put your SQL databases under
version control? Why should you start doing it? Read more to find out.
http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
_______________________________________________
gumstix-users mailing list
gumstix-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/gumstix-users