I let the TVR idle for a few minutes after driving it one evening.
The MAT sensor is calibrated pretty well, so there is only a small residual error in the airflow calculation for generating injector pulsewidth. The main error comes from the MAT sensor beginning to measure engine temperature rather than air temperature.
Although I've fitted a linear trend line in the picture, I don't believe it is quite correct due to the above comment. The calibration looks fine from 20C to 35C, and after that, there is a gradual increase in average fuel correction - implying good calibration and "poor" resistance to heatsoak. This gives a clue as to the value for the fake airden low limit - ie when the MAT will take the Gego out of range of its allowed correction (in my case that is a 3% richening of mixture)
