We are running a CC430 at 433MHz with a CW test. Changing the temperature from 25C to 85C we see large deviations in frequency, around a few hundred kHz. Based on oscillator tolerances and stability, our tolerance should only be 36kHz. Doing a hard reset on the unit fixes the problem and the frequency tracks exactly where it's supposed to be. The same happens when moving from 85C to 25C. The ramp rate of the change is 3C/min and the RH is set to 35%.
I would be interested to know what the CC430 or C1101 does under the hood for frequency calibration and how vulnerable it is to temperature changes when running a CW test.