Originally Posted by
nafun
If the PT's torque has been zeroed and the discrepancy persists, i would be inclined to trust the PT over the CT. The CT does not actually measure power, it just uses a coast down test to estimate a power vs. speed curve for the resistance unit. Kind of like what the iBike does, except much simpler because there is only one variable.
The Computrainer does not just use a power vs speed curve. You can simulate high-power low-speed conditions (like a hill) or vice versa. Yes, the coast down is to measure the rolling resistance, but the unit is electronically braked.
OP what calibration values are you getting? Are you recalibrating after the unit/tire warms up?