I am in the early stages of a design for a scientific instrument that requires some form of battery backup. In just looking around I see too many choices and would like some pointers from others who have worked through some of the same experiences.
Since it's early in the design process, the specific current demand is not known. What is known is that the circuit is reasonably low power, a draw of less than 100 mA at 3.3V. Most of the time, the circuit can be powered from a wall pack. However, if power fails, the battery must be able to maintain operation for at least 24 hours. If the additional impact is minimal, the ability to operate for up to a week would be an added plus. (Talk to marketing!) In the event of a prolonged power failure, the system must be able to detect this and go through an orderly power down and recover when external power is restored.
In looking at the technologies available, I see Lithium (including LiPO), NiCad, and NiMH. The last two seem more readily available from places like DigiKey and Mouser. But on the other hand, the chip houses seem to be pushing designs supporting the Lithium technologies. I'm also facing the bad publicity that Lithium has received in the last few years. (Note that I'm writing off Lead Acid because of the typical bulk and the need for long term life.)
Another negative that I have noted in the Lithium arena is the lack of standardized packaging. As a scientific application, the design is expected to have a long life, at least 20 years. As it is also low volume, less than 100 units per year, we need to find a solution that will be available for this lifetime that does not require a redesign of the case just because the battery changed it's form factor.
Any recommendations?
