> I'd like to test the Sanyo eneloop Ready To Use (long
> shelf life) Ni-MH AA
> cells to see what their shelf life really is.  Has someone
> already made up a
> small PIC based tester to do this?  Or have any ideas on
> how to proceed?
> http://www.prc68.com/I/RTU-Batt.shtml <- my eneloop page


Postscript - I've added a better way at the end :-)

Post-postscript.
Just reread your page.
You need to sort your tester out first.


______________

You've already done it :-)

Your page nicely describes the only available "quick test"
you are liable to be able to do.

You bought batteries that appeared to be 16 to 17 months old
(rather than the 15 you suggest) and measured the amount of
energy needed to restore them to their charged state.

Your test indicated that you inserted 15% of their nominal
energy to bring them back to a fully charged state. If your
charger is getting it "about right" then -

- NimH cells usually have a lower than nominal capacity for
the first few cycles. This MAY mean that your inserted
energy (hereafter = "charge") was a greater percentage than
nominal ratings suggest.

- Charge energy is not 100% assimilated - a figure of 80% is
used by some. Will vary with technology, maker and wind
direction.

If decay rate was linear (makes sums slightly easier) then
you lost 300/2000 in 16 months or 225 mAh/year ~= 11.3%.
If you allow charge efficiency of 80% that's a loss of only
11.3 x 0.8 =~ 9%.

That's so much lower than their claimed 15% loss per year
that I'd say that it's "extremely likely" that their claims
can be relied on in typical cases.

YMMV.
Check my assumptions.
The initial few cycles low takeup may catch you out.

___________________

BETTER TEST:

DISCHARGE a set of new batteries with known dates (as
above) - measure discharge energy.
Charge - measure charge energy.
Repeat say 10 times and plot the results.

This will allow you to refine most of the above assumptions.

Now put this set aside (preferably several sets) for say 3
months at 20 degrees C and then repeat.
Longer is better.

Even 1 month *may* do but the 1% or so loss over that period
would be low enough that you'd need to know the specific
battery history to sort result from production spread.

        Russell






-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist