Comparing power values - Tacx Neo vs Assioma pedals

Recently I'd become interested to see how the power reported by my ageing Tacx Neo smart trainer (first generation) compares with my Assioma pedals (i have the dual set). My initial investigations involved fitting the Assiomas to the bike mounted on the turbo and comparing the reported power by eye. This seemed to have a significant different at least at some power levers, so I felt some further investigations were merited.

I decided that my first experiment would record the power recorded by both devices, with the Neo operating in erg mode controlled by TrainerRoad. To do this I set up a custom workout with the following sections:

0-5 minutes at 125W; 5 x 5 minutes at 175W; 5 minutes at 125W, making 7 x 5 minute sections. I planned to keep my cadence at 90rpm as best I could throughout, and varied the gearing from 53x13 in sections 1 and 2, increasing by 1 gear each section to finish in 53x17 for sections 6 and 7.

The trace below uses the DC Rainmaker Analyser to compare the two traces. The Assioma pedals are in purple and the Tacx Neo in a sort of teal blue. The first thing that struck me was that in the first 5 minutes at 125W, the Assiomas read about 10W higher than the Tacx (but that this effect lessened at 175W, and lessened further as the gearing reduced.

I decided to poke into the data for each section of the trace above, and I've tabulated the results below.

Generally speaking I'd say there's a significant difference between the two power meters in the first two sections. This difference appears to be

(a) cadence dependent - see the successive data sets at 175W and

(b) power level dependent - the power reported by the Assioma pedals is closer to that reported by the Tacx when 125W and 175W are compared at 53x13.

At the end of the day, I think the message here is that even after 6 years of heavy usage, my 1st Gen Tacx Neo is still pretty reliable, but that at low power and high gearing there is likely to be a discrepancy. As to which one is the correct value, that would take a bit more experimentation and given that by 175W the two power meters are pretty concordant I don't think it worthwhile.