Interesting... this new jig seems to be working better than I had thought.
I just swept all four of the filters I have. The filter I had chosen for my first test, turns to be the one with the greatest asymmetry in its group-delay curve. The other 180 kHz filter also has a lower peak on the high-frequency side but the asymmetry is less. The two 280 khz filters have even less asymmetry, and the lower group-delay peak is on the low-frequency side.
So, I may actually be seeing actual variations between the filters with less bias than I had thought.
