Recently I’ve found a glitch in the LT1720/21 models provided in LTSPICE. When it is connected to a pulsating floating voltage source, false high outputs are triggered at the rising edges. I’ve attached the simulation screenshots and files for reproducing the problem.
This problem doesn’t happen to other comparators, such as LT1018/LTC6752, though they also have very small voltage ripples at their output ports. I guess some sinks in these models are incorrectly using the global ground.
