Richard Peet's posts

I thought I would post reference to the article before going forth and saying other posts are an outdated opinion.  I suspect it will be years before it is commonly being done to mix battery sizes and ages.

My solar application is for remote monitoring.  In that application I designed the system to reach 100% DOD at 2 full days of overcast using a 33ah agm 12 volt with a 65 watt poly panel.  But I sure prefer to be at 4 days gloom before shut down so last year I carried in a very old 65 ah agm and connected it in parallel. It worked just fine.

This year I updated the charge controller to one with a low voltage disconnect that is current compensated.  I may miss the hardware fuse but we will see.
This is what the battery looks like before jumping other batteries onto it in parallel:
http://home.comcast.net/~soundfiles/sd722bat.JPG

Rich

Best read on the subject I have found:

Abstract – The operation of a 24V VRLA battery
formed by two strings of dissimilar cells and
monoblocs in a 100% d.o.d C3 discharge/charge cycle
service has revealed not unexpectedly that an
asymmetric current flow from the strings occurs but
that its implication on ultimate cycle service life is less
significant than the strict rule of “do-not parallel
dissimilar strings” would indicate

Document found at:
http://neuralfibre.com/paul/wp-content/uploads/2007/05/can-we-now-sin.pdf

Rich