Don't rule out misleading coincidences!
Further testing (details below) has proved this was no coincidence. It's also proved this fault is not a hardware one - so at least I don't have to bother changing all those caps. The RMA process has proved problematic so I guess my only hope now is a system flush (which I've not yet tried).
In the following, to determine whether the fault was present at the time I tried Talking Pictures TV (82) and Sky News (233) - both known to cause the cycle of picture freezes and auto-reboots with very roughly a minute of moving pictures each cycle. At least ten minutes was spent each time checking for the fault.
1. Faulty PVR running Humax FW for 36 hours and exhibiting fault.
2. Custom FW installed and on first boot displayed ENTERING, RMA, MODE (this after the usual custom FW message); fault NOT present.
3. Manually (with remote) Into standby and then on again (after waiting 30s), displayed ENTERING, RMA, MODE; fault NOT present.
4. Into standby again, which caused scrolling display to freeze ( Talkin); clock appeared at HDD off.
5. When turned on this time the PVR again displayed ENTERING, RMA, MODE - then put itself into standby.
6. When turned on this time the PVR displayed == RMA ==; fault NOT present.
7. Into standby and on again, displayed ENTERING, RMA, MODE; fault NOT present.
No web interface activity during the above.
Obviously the next step would be to reinstall the Humax FW, which is what I did a couple of days ago - but I strongly suspect this will ultimately take me back to (1.) above. So I will leave with the custom FW for now and see what happens. I assume that if the RMA process is successful it's not possible to start the system as there would be no FW installed - is that right?
Worth mentioning that before bailing-out to a spare PVR the faulty PVR had been showing signs of occasional poor reception. Accordingly, I took a feed of the 12V going to the loft preamp and fed it down an unused coax to the lounge so that I could set up a 'scope to monitor it next to the TV. However there was no correlation and the 12V supply was solid. As we've not had any "reception problems" with the spare PVR I now suspect these occasions were signs that the MPeg decoder was struggling and are related to the freeze-and-reboot problem.