I know what you mean, but it still doesn't make sense.
Have hansibull measure the pins that are not populated to see if there is a PWM signal there, and if the duty cycle of the PWM changes with temperature.
If this is the case, I take everything back.