Hi Michael, By the way, didn't there used to be someone by the name of "Gus" who used to post to the list with a signature involving WFT Electronics? Do you know him? Basically you are on the right track with this. The biggest problem I see is that it is unrealistic to think that you will be able to dissipate 200W from one FET in any typical type of package. FET specs can be very misleading. Usually their max dissipation number assumes that you manage to keep the case at 25C, which would pretty much require cooling with liquid nitrogen or lots of water through a copper heatsink. No real application is going to manage to carry away 200W from that package without the case getting very hot. They give junction to case thermal resistance of 0.6 deg C/Watt and typical case to sink of 0.5. Then, you need to figure at least 1 deg C/Watt for thermal spreading resistance through the heatsink material. If you manage to have a huge fan-cooled heatsink, you might manage to keep the rest of the resistance to ambient at 1 deg C/Watt. That's a total from junction to ambient of 0.6+0.5+1+1=3.1 deg C/Watt. If ambient is 25C, and the max junction temperature is 175 C, then this leaves you with a deltaT of 150C max. 150C/3.1 degC/Watt=48W. Note that this leaves no margin for calculation approximation error, accidental overload, part to part variation, or any kind of safety factor (i.e. keeping the Tj max below 175C to increase semiconductor reliability). i think that you are going to end up with multiple devices in parallel. If you end up doing that, note that you cannot just parallel FETs directly when they are operating in the linear (half-on) mode as you are using them here. That means that you would need ballast resistors and at that point it would probably be better to go with several BJTs in parallel. One way to reduce your device power dissipation is to have some power resistors in the circuit. I once built one of these battery constant-current dischargers which used two power resistors (call them R1,R2) in series with several BJTs in parallel across one of the two resistors (R2). When the BJTs were fully off, the circuit resistance was R1+R2. With the BJTs fully on, it was roughly R1. I could then vary the current between those two extremes by varying the drive to the BJT. The maximum BJT dissipation was then roughly 1/4 of what it would be if there were no power resistors (max power for the BJT in this case was at half of max current and half of max voltage). Another potential problem is op-amp instability. You have the FET in the feedback loop of the op-amp. The FET has a high gate capacitance and non-linear behavior, especially as the current gets low. You may have to add a compensation network (some Rs and Cs) in the op-amp circuit to prevent oscillation. You can tell by prototyping it, putting a scope on the op-amp output, and then driving the DAC with step functions (i.e., command a 0.5A to 10A step) and watch for ringing on the scope. As for an op-amp choice, it will be challenging to get near 16 bit accuracy. There will be Rsense inaccuracies, op-amp noise, and op-amp offset error. Analog devices makes some amazing zero-drift op-amps now (which auto-zero their own offset error). A decent general purpose 1mV offset op-amp is the LMC6482. Sean On Tue, Nov 4, 2008 at 1:32 PM, Michael Algernon wrote: > > I have posted a hand drawn schematic at > http://oh-god.com:5080/dir/Electronic-load/ > > The application is a battery tester. I want to load a 1.5 to 5V > battery with a load current of zero to 40A. To the right you see I > plan to drive the op amp with a 16 bit DAC. R3 shuts off the load if > the DAC is disconnected. I put in R1 and R2 as insurance. I am not > sure I need them at all. The EL can be very slow. A response time of > 1 second is okay. Rsense is .01 ohms. Thus I figure that Vsense > will vary from 0.0 to 0.4V. > The FET will be using a heatsink ( and maybe a fan ) and will be rated > at about 80A and 300W. I have calculated the power dissipation of the > FET to be a max of 40A times 5V which is 200W. ( Some of that heat > will appear across Rsense ) > > DAC = DAC8574 > FET = FQP90N10V2 > OP AMP = ??? > > Will this idea work or have I missed the banana boat. What would be a > good op amp to maintain precision ? The op amp need not be fast. Low > power is obviously not a factor. > > MA > > > > > > WFT Electronics > Denver, CO 720 222 1309 > " dent the UNIVERSE " > > All ideas, text, drawings and audio , that are originated by WFT > Electronics ( and it's principals ), that are included with this > signature text are to be deemed to be released to the public domain as > of the date of this communication . > > > WFT Electronics > Denver, CO 720 222 1309 > " dent the UNIVERSE " > > All ideas, text, drawings and audio , that are originated by WFT > Electronics ( and it's principals ), that are included with this > signature text are to be deemed to be released to the public domain as > of the date of this communication . > > > WFT Electronics > Denver, CO 720 222 1309 > " dent the UNIVERSE " > > All ideas, text, drawings and audio , that are originated by WFT > Electronics ( and it's principals ), that are included with this > signature text are to be deemed to be released to the public domain as > of the date of this communication . > > -- > http://www.piclist.com PIC/SX FAQ & list archive > View/change your membership options at > http://mailman.mit.edu/mailman/listinfo/piclist > -- http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist