> I'm looking to develop a NiMH 24V battery pack power supply (20 cells > in series). There are COTS chargers available that I can go with, but > ... if you have suggestions for how to > approach this in a cost effective, and quick development time manner > I'd appreciate it! Off the cuff BOTE without an envelope (or net) etc. I assume that all 20 cells have tap / connection points electrically accessible. Also useful to assume that a thermistor or simkilar per cell can be added. If cells can be separated when charging it' MUCH easier but I assume below that string is hard connected. What size cells? AA or ? What charge current? Constant or settable or ? I assume below1A - easily changed What's the productin volume. What order of cost envisaged? 20 cells with negative delta V detection per cell could be annoying. A reasonably simple reasonably cheap reasonably quick (choose any 3) method albeit not too elegant would be to implement constant current series string charge, with absolute temperature termination per cell with a current bypas= s per cell once tripped. Maybe 10's of cents to well under $US1 per channel electronics only. eg heat sinking if using shunts below costs whatever you want it to - from cents to $s per cell. eg say you series charge at 1A. Monitor a given cell for temperature rise to say 50C or 60C or xC above ambient. When this condition is achieved trigger a 0.950 (say) Amp shunt around the cell. 1 A enters the module and once tripped 50 mA trickle charges the cel= l until all are charged. This takes a few parts per module but not horribly dear. temperature detect is say thermistor and comparator. Apply gross hysteresis to latch until reset. Current shunt is opamp section and shuntin= g transistor and sense resistor. A little more but not vast. Shunt takes here 1A x ~1.5V / cell or 1.5W per shunt at 1A. A similar but different approach is to isolated smps charge each cell and terminate as before. Could even use negative delta V per cell. Using one std IC per cell would allow each to work in isolation if isolated suplies were provided. I say "smps charge each cell above but this need not be a separate full smps per each by eg N windings off a central supply. This approach has advantage of stopping rather than shunting current at termination so no need to dissipate energy in shunts. Of the above the series string with shunts would probably be easiest to get going as everything is very well defined using standard glue parts. Charge termination can be via absolute voltage per cell but this is less precise. In series string required charge voltage will be approaching 40V at higher currents. Op amps may not have enough voltage range if supplied from whole rail. But op amp current is low enough that you can tap from eg top rail to say mid string for Vcc and be invisible to cotroller. Can even tap N cells up and N cells down to get constant supply on all except top and bottom N-1 cells. Russell McMahon --=20 http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .