Lithium button cell, avoid passivation

Do you have a question? Post it now! No Registration Necessary

Translate This Thread From English to

Threaded View

In an embedded system we have a lithium button cell to provide power
for the real time clock IC and some RAM when external power is off.

Since such cells have a phenomenom called passivation, a chemical
process covers the electrode if the cell is not loaded, you cannot
assume it will just work after several years. This means that the cell
should be "waked up" periodically to assure it will work when needed.

The dummy load, which is controlled by the processor, will give a load
comparable to the max. rating of the battery. There is also a voltage
comparator connected to the processor.

The question: Which pulse/pause pattern should be used? 10 - 20 sec. a
day (one pulse), or 1 second every hour? Will an even shorter pulse
every 5 or 10 min. work? The latter would give an alarm earlier when
the cell fails. - And recognize a replaced cell faster.  ;-)

Does anyone have experience with a such system? It may take some time
to test. ;-))

  Mogens Dybk Christensen
  e-mail mdc at mail dot tele dot dk

Re: Lithium button cell, avoid passivation
On 01 Sep 2003 19:07:54 +0200, (Mogens Dybk

Quoted text here. Click to load it

My company sells temperature controllers that use a Panasonic lithium
coin cell as memory backup. In the field, these units are powered up
most of the time. When units come back for repair or cal, I
occasionally have the battery pulled and connected to a datalogger; we
log voltage under a light load to determine residual battery capacity,
killing the battery in a few days. We've never seen any sign of this
effect. In fact, we see no systematic sign that batteries as old as 7
years have any less capacity than new ones.


Re: Lithium button cell, avoid passivation

Quoted text here. Click to load it

Thanks, interesting. Also your discharge-experiment!

I think it depends on type and manufacturer if
you will see the problem. Are you sure your hardware does not have a
dummu load? - See also the comments from Fred.

Another person has pointed me to

which I already had found using google. But it too has no exact

I think we will go for at pulsed load estimated to discharge the
battery in som 20 years. It will probably fail long before that anyway.
But hopefully not after 1 year, as we have seen on another system.

  Mogens Dybk Christensen
  e-mail mdc at mail dot tele dot dk

Re: Lithium button cell, avoid passivation

Quoted text here. Click to load it

The easiest way is to bleed a small continuous current from the battery.  If
you bleed current at the 10 or 20 year rate then this will not have a
significant effect on battery life.

Otherwise measure the typical current you need for your real time clock and
RAM and make up a dummy load with this characteristic.  Use the micro to
switch on the load and at timely intervals measure volts until they are at a
sensible value.

Site Timeline