ok, sounds like we have a way to resolve your question.
Some points of clarification: our power meters are designed to keep their calibration for a long time. When power2max came to market our power meters offered a new set of features that were previously thought not to be possible:
- Ability to change chain rings without affecting calibration
- Precise cadence without a magnet
- Temperature compensated
- Stable calibration
We regularly test units at the factory and are able to review calibration. The testing procedure has been qualified and provides repeatable results. Home testing the power meter whilst installed in a bottom bracket introduces a bias of unknown magnitude, which makes the comparison unreliable. Change your bottom bracket or change the lateral play on your cranks and you might change the measurement by a percent or two (I am just guessing here at the magnitude). If you take a reading you may think the calibration of your power meter has changed, but it hasn't.
This is at the heart of why we don't recommend it. Testing against a Cyclus2 or doing a steep long climbing test are easier to do well, in our experience.
I have a couple of different SRM's, but the one I've had the longest is 5 years old, a Hollowgram spider SRM. During that time I've checked the calibration 11 times-3 different sets of chainrings, 2 sets of crank arms and 4 battery changes (I swapped the battery myself). The slope always tested within 1.5%. This includes moving the unit between multiple frames with different BB's. The other SRM's I've checked have all exhibited similar behavior except for one unit. So, I'm sorry to belabor this point but the repeatability is well within the precision of a Cyclus2 and certainly more precise than a field test back calculation.
Keep in mind, I think that's a good reality check--it was the W/speed relationship on a known climb where I realized that the one SRM had a badly-drifting slope. But guess what? I static calibration quickly confirmed that it was drifting...