Looking at my notes when I constructed this years ago...
It appears I managed to detect as low as .02 V.
I've persuaded the simulation to resemble something closer to that.
It is important to get the capacitor waveform adjusted, so it is just above the zero line.
Supposedly you can expect an input of .02 V to yield 0.1 Hz. In reality it will be slightly less, because the capacitor charges more slowly, and becomes more gradual as it reaches a plateau.
When the input goes below 0.02 V, the output is flat.