🚀 go-pugleaf

[I posted something similar to this several weeks ago, and didn't get any
responses... so here I go again... :-) ]

I'm currently making a temperature logger that will eventually float around
somewhere off the cost of Florida.  It contains the analog temperature
sending circuitry, a microcontroller, and a radio transmitter.  Every now
and then (at about 10 minute intervals), the microcontroller is turned on
and transmits a bunch of data back to shore.  So far, so good.

This all works fine sitting on a bench here in Wisconsin powered by a nice
regulated power supply, but once it starts floating, it's only going to have
solar cells to power it.  I'm trying to decide whether to to use NiCad or
GelCell batteries in the thign, but I really don't know that much about them.
Here's what I (think) I know:

NiCads -- Can be overcharged all you want.  For the space, you get less a*H's
	then with Gel Cells.  Major problem:  They form memories.  In an
	application like this, where I charge for awhile, drain (probably)
	a very small amount of the capacity of the cell, and they recharge,
	it seems to me that I'm almost trying to create a memory on 
	purpose.  Since the reason NiCads are there to begin with is to keep
	everything going on cloudy days, if they form memories, I'll be 
	dead in the water (no pun intended :-) ) if there ever IS a cloudly
	day.  Finally, NiCads are simple to charge... Solar Cells to diode
	to resisitor to batteries.  Don't mind being deep cycled.
 
GelCells -- Apparently partially damaged is overcharge too much or deep 
	cycles more than a few times.  However, DON'T form memories.  Hard
	to charge properly -- require a regulated charging.  Considering
	that, I'd have to have a circuit powered by the batteries just to
	recharge the batteries. :-( This means that if the batteries ever
	get so low of a charge that I can no longer power the regulator
	circuit, the batteries will probably stay discharged!  Arghh...
 
That's about it.  I'd love to hear from somebody who's designed this kind
of stuff before... 'cuz this is my first time!  I'd love to hear somebody
tell me that NiCads don't form memories!  (Because besides that, they're
just about perfect, it seems.  Whereas getting GelCells into a workable
setup would seem to be a royal pain!)

Sitting right across the batteries is a CMOS timer that goes to a MOSFET.
this is what determines when power is applied to the microcontroller, etc.
The microcontroller is then responsible for shutting itself off.  With this
setup, if I used NiCad batteries and they were totally drained for some
reason or another, the batteries would still have a good chance of being
recharged given enough sunny days.

For those interested, all of this stuff is fairly low power -- each time
the circuit is powered, is draws about 10maH (the microcontroller 
squashes the data together and then burst transmits it -- so evyerthing is
only one for a matter or seconds).  We're looking about batteries in the
1aH range.

Any help is appreciated.  I really want to design a system that WORKS!  
Please respond on the net, or e-mail me at kolstad@cae.wisc.edu.  Thanks!
 
                                          ---Joel Kolstad
					kolstad@cae.wisc.edu