TLV3501: Comparator shuts off with amplifier output, need a way to solve this

Hello Parker,

Most likely the model was designed to "misbehave" when a parameter, such as input range, is violated. And there is a chance the actual device will do the same...but 300mV over will probably not hurt the real device.

Theoretically, the op-amp output cannot exceed the supply, and in fact, should be slightly less than the supply. So there must be some inductance somewhere if the output truly exceeds 3.3V.

Since you do not want to diminish the sensitivity (gain), the only thing is to clamp the output at the upper end of the swing. Once you pass the 110mV threshold, it really does not matter if you clamp at 1V or 2V (you do want to get as much overdrive as possible to get the fastest response time, so get at least a volt over the threshold).

You could add a series resistor between the output and the comparator input, with two or more diodes in series to ground to get a 1.4 or 2.1V clamp. But the resistor and the capacitance of the diodes may slow the response due to the resistor and diode capacitance time constant. That is also tough on the op-amp as the clamp current spikes.

Since you are also supplying the op-amps with a split supply, you also need to ensure that the comparator input is not driven negative. That would require another diode to clamp negative.

Since the op-amps are single devices, another trick would be to place a diode in series with the positive supply of the driving op-amp. This would reduce the positive output swing by a diode drop without sacrificing driving speed. Of course, the positive supply will need heavier additional bypassing after the diode.

But the concern still is the negative swing. The OPA810 is capable of 75mA, which is much higher than the abs max input current of 10mA, so it could damage the comparator should the op-amp output go below -0.3V. Normally a resistor would be placed in series with the op-amp output to limit the clamp current - but again we are adding a resistor and a diode...

If you do not need absolute DC accuracy, and are just looking at pulses above a baseline, it may be easier to go to a single supply for everything and AC couple the amplifier path and bias the last amps output to 110mV.

Try the circuit by adding a 1uF cap in series with the 1k feedback resistor to ground. DC-wise, this makes the op-amps followers, while maintaining the AC gain at high frequencies. Ground the negative supplies (or set to 0V). Set the source generator to have a DC bias slightly less than the comparator reference. That should give you a similar waveform, but no chance of going negative.