> > 1) Temperature sensor on PCB? > Since the power resistors are also on the PCB, wouldn't this affect the > temperature of the PCB? I would think a remote temperature sensor embedded > in the module would be more accurate. Yes, I considered that. 1) I assume that the resistors won't be On but in special cases, and then we'll know not to trust the temperature reading 2) If fit snugly against a cell, the sensor will sense that cell's temperature. Now, if that temperature increases because of the shunt resistors, then that still is that cell's temperature 3) The resistors dissipate only 2 W, which won't heat-up the cells that much (I assume) 4) The ease and reliability of having everything on one PCB was for me the overriding factor; everything else is just me rationalizing ;-) > 2) ICSP ? which programmer? > I do not have a in-circuit programmer and plan to get one. Which one > are you using and which one you recommend? The EPIC. It's OK, cheap, decent software. I'm looking to get a USB ICSP, much better. It's what my German partners are using, made in the Check republic. But I have to import it. I'll send you a link later. > 3) 2.048V ref might be better? > I think a 2.048V reference voltage would put us closer to the mid point > of the ADC range. I'll look into it. Not much difference. I'd decide based on accuracy. I'd think that 2.5 V is a bandgap volatge, and therefore inherently better. I'll let you know. > 4) Why ground pin 10? > Why is pin 10 on U1 (PIC) grounded? PCB layout reasons. Otherwise I'd need a bridge. (It's a single sided PCB). Davide