Beagleboard RTC

I’ve just got around to fitting (not a nice experience) a battery to my Beagleboard -xM to get the RTC working.

Can someone explain or try this…

I have “Set Manually” ticked in the Boot. configuration – Time and date window and set the current time and date. The next day I power-up the BB the default date and time (2nd January 1970, time 01:01) is shown. Surely with the RTC working the OS should remember the time/date automatically. I’m thoroughly confused – can anyone tell me what is going on? Incidentally I have checked the RTC battery volage across R65 is about 2.5-2.6V.

RISC OS 5.20

Does it work under Linux? And how about what Dave said? Also, does BeagleBoard Battery test offer anything?

I’ve no experience of Linux so am unable to do that.

I have installed Andrew Conway’s Battery configure application and the battery votage is reported as 2.877V. I have the charge set at 3.1V/25uA for my MS621FE-FL11E which was recommended by Jeffrey Lee.

I’ll charge it and see whether it makes any difference.

Thanks for the very useful links – I followed Dave’s advice when installing the battery.

Richard: your battery’s voltage is a bit low. You may want to charge it at the next higher current for a few hours before setting it back to 25 µA for normal operation. 25 µA is a bit marginal for something that is only on for a couple or so hours a day, but only you know your usage pattern.

I have installed Andrew Conway’s Battery configure application

Who’s he then? :-)

As Dave says, your battery voltage is a bit low, so it would benefit from being left on charge for 24hrs (or bump the charge level up for a few hours), and then see if it still gives problems when it has reached 3V.

Who’s he then? :-)

Where’s he then? :-)

Nice to see you couldn’t leave us alone, even on holiday, Andrew!

Sorry about the miss-spelling of your surname Andrew ;-( Incidentally a nice Boot.configure app. The battery voltage is creeping up – will do as you and Dave suggests.

Charged for a few hours to 3.08V before switching off last night. Switched on this morning and the RTC remembers the correct time-hooray. Battery doesn’t appear to hold charge for long as its charging voltage this morning is 2.80V but perhaps not surprising with such a tiny button cell. Do others find this?

Incidentally my thanks to Trevor Johnson for pointing me in the right direction.

but perhaps not surprising with such a tiny button cell

A tiny cell can run a watch with moving hands for over a year. You shouldn’t really be noticing a problem even if you only turn your computer on for one evening a month……

Do others find this?

I did. A few months later, the button cell was completely dead.

I regard that combination of size of cell and choice of charging currents as a problem. The charging currents are very widely spaced. 25 µA is not sufficient to keep anything charged for the job of maintaining the RTC unless it’s on at least a couple of hours a day (don’t forget to include the cell’s self discharge current in any calculations you do). The next higher current is too much if you keep the BB on 24/365.

When I bought the original BB, I bought the largest size of cell I could find with that chemistry, and set the charging current above 25 µA. It wasn’t a problem.

In your experience would it be worth considering a Seiko Instruments MS920SE-FL27E? Even this is limited to a charge current of 50uA but using the 150uA setting I might get away without causing battery degradation since the machine is used intermittently. I don’t really want to keep resetting the date and time settings. Charging currents are fixed in hardware I presume so we can’t change them.

MS920SE-FL27E?

Let me have a more detailed look – sorry, SAUG meeting tonight, I won’t have time until tomorrow.

Charging currents are fixed in hardware I presume so we can’t change them.

Correct.