Low Power MCU vs LiPo Self-Discharge

[colorado.rob]

At what point does the self-discharge rate of a LiPo battery exceed the low power current of an MCU?  Put another way, when using rechargeable LiPo batteries, at what point does the current draw in the circuit become immaterial?  I'm down to a sleep current for my board in the 2-3uA range when powered off.  For my 900mAh battery, that's good for 300000 hours, 12500 days or 35 years.  There's no way the battery will last that long.

What's the prudent cut off point?

BTW, the quoted self-discharge rates for LiPo batteries I've seen are way overstated.  I've got similar boards on my shelf that have been "asleep" for 4 years that still show close to 4V on the 3.7V battery.  Full charge was 4.2V.

[retrolefty]

Well from a quick goggle search:

Lithium Polymer cells experience a self-discharge rate of approximately 5% per month, compared with over 30% per month and 20% per month in NiHM batteries and NiCd batteries respectively.

That's a start so take your battery mAH rating and do the math to get a rough idea. You may need to get a datasheet for your specific battery pack to get more accurate estimate.

Good luck