Ok, so I've been in touch with support. My issue was identical to the video posted earlier in this thread.
The advice provided by Sequential support was quick and helpful, but did not resolve the issue without my own experimentation.
It was only after considerable experimentation that I was able to get the filter 2 cutoff to behave what I presume is correctly.
The initial advice was to run the filter autotune with the F2 cutoff and resonance knobs fully counter clockwise. This initially produced a dead zone between fully counterclockwise and 11 oclock, but addressed the strange bipolar issue I was having.
Then this advice was corrected, and I was told I should instead have run it with those knobs fully clockwise. I did and this reverted the behaviour to the original problematic state.
Running the calibration did not consistently reproduce the above behaviour. It was very hit and miss. Sometimes no effect, sometimes a big effect on the knob behaviour. Nothing consistent. Even when clearing the calibrations in the debug menu first.
I eventually got to a good result, after a lot of experimentation, with the following routine:
I cleared the calibration again.
Then ran the filter autotune / calibration with F2 cutoff / resonance knobs fully counterclockwise
This got rid of the bipolar behaviour in my video, and reproduced the pot’s dead zone from fully counterclockwise to noon before the filter audibly started to respond (ie where I got to the first time).
This time however, each time a calibration was run resulted in successive improvements in the response.
The first run resulted in the bottom 1/2 of the filter pot travel being “dead”
The second run resulted in the bottom 1/3 of the pot travel “dead”
The third run resulted in behaviour “within normal spec” with the far ends of the pot travel being relatively symmetrically dead but the pot working correctly.
Seems to be acceptable in terms of functionality at the moment, but I’m not sure now if it’s “normal” or not.