I found a good balance of resistors to do this division now.
The FET idea I think is the best.

I calculated what division was max for the different cellcounts to keep
accuracy at a minimum of 5 mv/cell.
Then I calculated minimum division needed to keep max divided voltage
(2v/cell) below 5v.

Then I found that for 1 and 2 cells I do not need division, and for 3-7
cells I only need a divisor of 3.  For 8-20 cells I can use a divisor of 8.
By combining the two resistors for 3 and 8 I can get a divisor of 10 and
that can be used for 21-25 cells.

Result is a R / R1  or R / R2  or R * (1/R1 + 1/R2) division.
Thus:
R = 2*56k
R1 = 56k
R2 = 16k

I will then be able to do 1-25 cells and always be within 5mv or better
accuracy.  Cool eh ?

Now, the next step is making a DC-DC step-up converter that can create a
whopping 50 volts or so at 5 amps.
That will be fun :)

Kyrre Aalerud

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics