Over time it is possible that the quiescent current in the amplifiers can drift and become too high or low. If the current is on the high side the amplifier will get warm without any being driven. This does not degrade the performance for THD, but may cause overheating.
The Crimson and BPA designs are almost pure class B, The standing current in the power devices should be in the region of 5mA to15 mA. In the lower power PCBs this will be below 1/2W per PCB. In the higher power amplifiers, this will produce 2W per PCB. A typical heatsink will be 1.5 deg. C per watt. This implies a temperature rise of 1.5 deg. C (2 x PCBs one heatsink) to 3 deg. C for a mono with 120V DC total supply. Up to double this is fine. A good way to check the temperature rise is to use an infra-red thermometer as long as the heatsink is black. The temperature is first taken with the power being off for an hour, then the temperature taken again , power on, 1 hour later.
The only real problem for sound reproduction is if the quiescent current is near zero. This means the amplifier runs virtually stone cold. Crossover distortion will result.
Resetting the bias potentiometer is best done on a test rig and with THD check. The long term stability of the bias can be improved with new values in the bias chain to make the potentiometer action more limited but more fine too.
It is tempting to quantify the quiescent current by measuring the volt drop across one of the current sense resistors (usually OR1, or 0R18 on pre 1998 low power PCBs). This can be done but a slip with the meter probes generally destroys the amplifier ( I have done it more than once).The current needs several minutes to stabilise. Another way is to break one the DC supply lines. Switch on and back the pot to minimum. Advance the pot to get a 5 to 10 mA increase in current. This increase is usually achieved by getting an initial increase of 4 mA and waiting for thermal effects to keep this increasing. The same comment applies to slip with probes making destruction easy and likely.