zPower Accuracy: Rollers vs Cycleops Magneto

The Archives Archived: General Discussion
1

What’s the deal with the zpower calculation on rollers?   I’m using a cadence/speed sensor configured with 700x25c wheels.

I have a set of SportsCrafters cadence rollers.  I also have a Cycleops Magneto.  

When I’m on the rollers, I can put a high RPE effort, but zpower calculates a relatively low power output. When I’m on the magneto, I can be putting myself through similar RPE effort, but the zpower calculation is MUCH higher (like maybe 1.5x at both low and high RPE efforts).    

That leads me to believe that either the rollers or the Magneto power graph is  wrong.   I’ve tested this hypothesis and used the Magneto setting while on rollers (physically using the rollers, but Zwift thinks I’m using the magneto) and it “feels” similar to when I’m using the magneto.

Which is “wronger”?  The magneto zpower calculation, or the Rollers zpower calculation?

2

Most trainers publish a power curve graph that represents the resistance in watts that correlate to various speeds of the trainer.    I’ve concluded that the power curve for the rollers that I have is most similar to the power curve for the Fluid2 trainer.   The default unsupported rollers option is pretty useless.