Actually after many lines of reading I bumped into this
"""The charge-discharge testing of a 24V battery with two
very dissimilar strings formed by 4 monoblocs
6CP100 (6V/100Ah C10) and 12 cells
2CP550(2V/550Ah C10) showed that the currents
flowing from and to the individual strings are
principally governed by the string impedance and
independent of the installed string capacity.
Such “asymmetrical” batteries can be operated quite
successfully and yield nearly their complete cycle life
potential when deep discharges to 100% d.o.d C3 are
carried out.
If power outages cause only shallow discharges then
the “low” impedance string furnishes an overproportional
share of the Ah discharged and may thus
suffer in terms of achievable cycle life.
However if these shallow discharges are infrequent
(»50 to 100/year) then the ultimate battery life will be
governed mostly by calendar or elapsed service time
considerations and not by exhaustion of its cycle life
capital."""
The background to that is batteries for telephone exchanges/basestations.
http://neuralfibre.com/paul/wp-content/uploads/2007/05/can-we-now-sin.pdfSo I think I m going for a simple parallel addition of a same type battery
Thanks a lot for saving me some hours of work !!!