> Dump a lot of current through the cell and fry the dendrites off.
> Monitor the problem cell with a volt meter.  Charge up a big
> cap (say 100,000 ufd) to 10, 15 or 20 volts. Using a low
> impedance delivery method, dump it into the cell

Now that sounds familiar. The rest of your post is new to me, as
is Mike Kendall's and I'll certainly be trying things out. And as you
say, others will their 2 cents worth so we'll see what develops
over the next couple of days. Thanks

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.