www.coridium.us/forum

> Is there any information or speculation on the expected life-time of
> the ML2020 battery when used with the ARMmite? How long might it last
> running the RTC with no power applied for extended periods? How long
> might it last as a backup with +5V power always applied (as in years)?
>

The ML2020 is a rechargeable Li-ion battery. While the ARMmite is
plugged in, it is being recharged with a small current. I don't know
what the shelf life of the battery is, Panasonic may have some specs
on that.

I remember some back of the envelope calculations that showed the
battery should keep the clock alive for at least a week. The design
was meant for a device that is powered on most of the time, but has to
suffer occasional power outages.

I tested one unit for a day or two, but never tried beyond that. I do
know a battery that is continuous use for months will run down, and
does not recover well (I had a unit with one disappear in my office
for at least that long).

According to the NXP spec the worst case Vbat current in standby is 12
uA. Though they rate it with 1.8V on the core supply, not sure why
they do it that way, as its normally off in the interesting
configuration. I hope there's not too much leakage in the parasitic
diodes to the supplies that are off (you'd have to measure it).

Assume there is some leakage through the charging diode, say 2 uA (the
spec on the BAT54 is 4 uA at 25V)

The ML2020 is a 40 mAH battery.

Assume all is right with the world, that works out to a couple hundred
days of battery life, unless you see all the powerdown current on the
core supply as leakage, in which case it would only be more than a
month.