Hi all.

I have designed a PCB using the OPA4990IDR op amp in a comparator configuration. I have use the OPA as a unity gain buffer also before the comparators.
The purpose of this PCB is to measure battery voltages of up to 10 Li-Ion cells (3.3V each , 38V MAX).
I am having a problem tho with the temperature of the OPA990IDR.
When attaching the V+ to the last battery cell and V- to GND, the chips are already getting very hot,
furthermore, attaching the batteries to the board on at a time, starting with the first battery, by the 3rd battery (10V) the chip is starting to produce smoke.
Any idea whats going on here would be helpful. Ive attached a photo of the circut design and PCB ( i am aware the traces will need to be enlarged in the next version)

 

That op-amp has a maximum supply voltage rating of 40V. A fully-charged LiIon cell has a voltage of ~4.2V, so ten in series would be ~42V.

 

I agree with @Alec_t - try running on 6 or 7 cells only to validate general operation, ie replace cells 8 -10 with a short circuit. Secondly this is not a comparator circuit; your 2nd row opamps are configured as differential amplifiers which serve to ground reference each indvidual cell voltage.

I don't know what accuracy you expect to measure to, but with discrete 1% resistors your output voltages for a 3.3v cell will be 3.25 - 3.35v for Cell 1 through to 2.83 - 3.76v for Cell 10 excluding amplifier errors due to resistor tolerances and poor common mode rejection. To do this correctly you need proper differential amplifiers with laser trimmed internal resistors to 0.01% or better - I use INA2133 directly onto the cells as has 25k internal resistor network; running the amps on 30v, the inputs are good for 60v common mode. You really don't need the voltage followers.