On 20 April 2014 15:03, John Gardner wrote: > A follow-up on Prof. Murphy's solar installation... > > http://physics.ucsd.edu/do-the-math/2012/12/death-of-a-battery/ > > I have an email to him in draft suggesting easy ways to eliminate some aspects of the problems that he reported in "death of a battery". No rocket science and nothing that's not obvious to him, but he may be focused enough on the battery as a chemical/electrical/mechanical thing that the obvious idea of using a ha'porth of electronics may be missed. He had a series parallel battery arrangement. The good battery hogged charge energy and then discharged into the bad battery when charging ceased. While one could target "proper"control schemes that balanced charging and even optimally treated each separate battery (or cell) these get complex and more costly. Relatively simple and low cost use of FETs to act as diodes to prevent discharge would stop the cross charging. Controller is very cheap and simple. At 20A losses of under 0.1% of power at 50V are possible with Rdson ~=3D 1 milliohm, , but something <=3D 5 mill= iohm is probably fine. Suitable MOSFETS are somewhat < $1 in modest volume from Digikey and well under that in Asia. It would be relatively easy to also use series FETs to provide division of charging current in whatever ratio seemed good. Controller reasonably simple. Losses potentially higher. eg 20A at 25V divided between two batteries 10A each requires 1 Watt loss per 0.1V of resistive drop required for balance. I'd expect that 0.5V max should be ample so < 5W or <=3D 2% of total power. "Proper" controllers may allow superior results but power losses would not be vastly less. His wiring looks somewhat on the light side. Not perhaps terribly so but real attention to connectors and wire drop needed. Above losses MAY be able to be made up for by improved wiring and connectors. Or not. _____________________ > The longest-lived secondary batteries in my experience are > in single-cell applications > . > > Considering the high efficiences attainable with SMPS, and > readily available deep-discharge lead-acid cells of 400 AH > capacity (that I'm aware of, perhaps more), one wonders what > might be achieved... Certainly attractive. But at say 500W level, a single cell at 2V requires 200A in and out. To get say <=3D 1% losses needs voltage drops of <=3D 2V/100 =3D 20 mV and = total resistances of R =3D (Ptot x 1%)/I^2 =3D (500 x 1%)/200^2 =3D~ 0.15 milliohm. That's losses in any Rdson (switch and synchronous "diodes"), wiring, connectors, inductors if present, ... . Get's 'reasonably hard' [tm]. That said, car starter motor circuits operate at hundred of amps 'for short periods' so the connectors, cables and wiring used may give some guide. However, 2V car battery can read closer to 6V under extreme cranking, so a volt or few loss elsewhere my not be totally unacceptable. ie needs to be at least as good as starter motor circuitry and probably significantly better. Russell > --=20 http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .