Autoland approaches were binned, so the annoying MFS compass system directional periodicity wasn't a factor - pilots tended to ignore it. But on RW05 at Scampton you could easily get into a divergent mode if trying to chase the azimuth director pointer with a dose of northerly instability making things worse! The Heading Reference System (basically a Master Reference Gyro pinched from the Lightning) was a much better, well damped system with which the plotter used to steer the beast when away from the circuit - there was a MFS/HRS switch to facilitate this. We got in once on absolute minima with the nice steady headings provided by HRS with the plotter frigging the heading on a GCA. Non-SOP, but it worked very well!
Yes, we used to fly approaches with PFCs deliberately failed. I vaguely remember doing so with 2 out on the same side which made things a bit ponderous!
Thanks for your excellent recollections - I've heard from other sources that, despite its pretty appearance, the Comet had all the traditional vices of control harmony, weird sitck force per g changes at higher TAS and the usual 1950's trim changes with any selection of services!