I'm using 3 comparators for sensorless back EMF commutation of a BLDC motor. This involves comparing the voltage of each motor phase to a neutral point to detect zero crossings. See Microchip AN970 (PDF) for further details on the application.
Unfortunately, I can't find a comparator that I can run off my wide 6-50V VCC input range. The best I can find in that regard is the LM339 that takes 2-30V, but nothing higher.
That leaves me in a situation where I have to divide the back EMF voltages down to fit whatever supply voltage I run the comparators from. If I run the comparator off a fixed voltage supply (call it VDD) and divide the signals down so that the max signal voltage @ VCC = 50V is within VDD, I'd be worried about attenuating the signal too much when VCC is lower. I guess my first question is, would this signal attenuation actually be a problem? I'd be looking at attenuating it 10:1 to run a comparator off my 5V supply.
Assuming the answer to the above is yes, I've come up with a couple of possible solutions so far (alternatives welcome):
- Divide the signals down as described above, keeping the fixed VDD as high as possible to reduce how much I have to divide the voltage signals down. I'd prefer to avoid adding a Boost converter off my 5V rail for cost reasons, but a simple charge pump doubler like an LM2767 would at least get me to VDD = 10V and 5:1 attenuation.
- Generate a VDD supply that is proportional to VCC (probably about VCC/2 for the LM339) and scale the voltage signals by the same amount. This would be preferable, but I'm not sure about is how to best generate a VDD = VCC/2 supply that I can use for the comparators. Can I get away with a resistor divider here? I was considering a step-down charge pump but I didn't have any luck finding a charge pump regulator with a high enough input range. Maybe even just switching VCC through a FET at 50% duty cycle?