EmbeddedRelated.com
Forums

Battery backup

Started by Mike van Meeteren January 13, 2005
All,

I would appreciate it if someone at Rabbit Semiconductor could answer this
question:

What voltage is required at the VBAT pin on a RCM3000 or RCM3100 core to
keep the clock functioning and the battery backed ram working? I talked to
someone in tech support who was not very useful in helping me narrow the
answer down. He e-mailed me a code snippet to verify the clock was staying
alive. Well, I already have code to verify this.

So, I would like the following answered:

1) what EXACTLY voltage level is required at VBAT? The battery we're using
is a 3v nominal battery. It has a pretty flat voltage curve as the battery
discharges, but it's not unusual to see these drop to 2.9v with very few
hours on them. I have observed failing real time clocks with battery
voltages as high as 2.95v. If this voltage is too low that pretty much
eliminates using lithium coin cell batteries and forces us to go to a much
more expensive 3.6v cell solution.

2) what external circuitry is suggested? I copied the circuit from the
development kit (3v battery through a 1.4K resistor) and it seems to work OK.

3) what I/O pin connections could cause excessive drain on the backup
circuit while the processor is powered down, or what software loaded could
make the battery consumption excessive?

From what I understand the 3V backup runs the 32 KHz clock, which in turn
executes some stuff on the core module. What is this stuff? Is it
hardware only (not sensitive to software) or is there an ISR running that
could be corrupted?

This stuff should REALLY be part of the manual as well.

-Mike


Message
While I'm not from Z-World I can help answer a couple of your questions. We actually asked Z-World a while ago about the minimum VBAT but didn't get an answer. If you look at the data sheets for the SRAM they are happy with a VBAT right down to about 2.5V from memory. The maximum VBAT voltage is 3.3V. If you go any higher than this you risk discharging your battery through the switch-over circuitry on the core modules. We use 3V 1/2AA Lithium batteries and have a procedure in production and service of rejecting and replacing them if they are below 3.0V (they typically measure ~3.1V new) as the products are often in the field for many years.
 
You can get away with connecting the battery directly to VBAT without any additional circuitry but typically you add a resistor in series to protect the battery against a short to ground in the event that the core module is incorrectly installed (out of alignment on header) and the VBAT accidently connects to ground.
 
Regards,
Nathan
-----Original Message-----
From: Mike van Meeteren [mailto:m...@fastec.com]
Sent: Friday, 14 January 2005 9:46 AM
To: r...@yahoogroups.com
Subject: [rabbit-semi] Battery backup

All,

I would appreciate it if someone at Rabbit Semiconductor could answer this
question:

What voltage is required at the VBAT pin on a RCM3000 or RCM3100 core to
keep the clock functioning and the battery backed ram working?  I talked to
someone in tech support who was not very useful in helping me narrow the
answer down.  He e-mailed me a code snippet to verify the clock was staying
alive.  Well, I already have code to verify this.

So, I would like the following answered:

1) what EXACTLY voltage level is required at VBAT?  The battery we're using
is a 3v nominal battery.  It has a pretty flat voltage curve as the battery
discharges, but it's not unusual to see these drop to 2.9v with very few
hours on them.  I have observed failing real time clocks with battery
voltages as high as 2.95v.  If this voltage is too low that pretty much
eliminates using lithium coin cell batteries and forces us to go to a much
more expensive 3.6v cell solution.

2) what external circuitry is suggested?  I copied the circuit from the
development kit (3v battery through a 1.4K resistor) and it seems to work OK.

3) what I/O pin connections could cause excessive drain on the backup
circuit while the processor is powered down, or what software loaded could
make the battery consumption excessive?

From what I understand the 3V backup runs the 32 KHz clock, which in turn
executes some stuff on the core module.  What is this stuff?  Is it
hardware only (not sensitive to software) or is there an ISR running that
could be corrupted?

This stuff should REALLY be part of the manual as well.

-Mike