Still dont understand how you can stretch/shorten a the steel belt in the tyre to increase/reduce the circumference (rolling diameter)
However, further research tonight I found an ADR paper on testing speedos (
the link for anyone as boring as me that wants to read it). Interestingly, the steel belts don't shrink or stretch as I would have expected, however, its the distortion in the tread on the tyres foot print that creates the difference in speedo readings with tyre pressure changes. An almost bald tyre will show almost no change, a brand new road tyre in the test showed changing from 22psi to 40psi gave only a 1.4km/hr (1.1%) difference in speedo reading at 120km/hr. However, normal tyre wear on the same road tyre (a Dunlop Monza 205/65R15) from new to ready to replace gave a 2% difference in speedo reading (nearly double the difference a change of 22 to 40 psi does).
So yes, speedo accuracy does change with tyre pressure, but tyre diameter does not. Measuring the tyre as per AEV's instructions will only provide an accurate speedo at one particular pressure. Given the tiny real difference from any pressure change, I still believe they would have been better off calibrating it using the measurement of the actual tyre diameter (across the tyre) as this does not change with tyre pressure.
Anyways, after changing the TPMS down to 27 psi yesterday, the clock on the stereo reset itself! Then after the beach run I pumped tyres back up to 32 (5psi over the new setting) and the low tyre indicator would not go off. Increased them to 34, no go. Reset the TPMS to 27 again and all good. So not sure what happened there! But all good now.