While testing the heater loop on r2 I found some problems with how the integrator sum is handled when changing set points or between idle/power/temp loops. Also, the way the temperature slew rate is limited is definitely buggy, at least when lowering the set point. Fix.