They all have a degree of tolerance to the accuracy.
The flow is +/- 5%
The Zumo is +/-3%
However in your case I’m going to say that going from a wheel on to a direct drive turbo is accounting for the difference. Wheel on are prone to accuracy issues caused by tyre slippage, tyre pressure, not being calibrated etc…
Direct drive tend to be less problematic.
You’ll not be the first to unfortunately discover that the wattage you thought you were putting out isn’t what it is in reality.
Honestly I always thought it was overestimating it… so, should I consider the zumo more accurate?
That’s a fair point about using a wheel. Depending on the pressure and how close the brake is, it will give different measures I think in the same turbo trainer
Is there any thread where this is explained or discussed?
I’m sure there is but hopefully I’ve explained it sufficiently well. Google “direct drive versus wheel on” and I’m sure you’ll find some useful info.
The zumo is the more accurate yes. If you look at the Tacx Neo that claims to be +/-1%.
I for one remember moving from the Tacx Vortex (wheel on) to the Neo and thinking that I’d aged 20 years overnight as my wattage dropped.
Enjoy direct drive, it’s quieter, less fiddly and makes using a turbo a far happier experience.
With wheel on trainers, part of the issue is that the resistance curve assumes you put the resistance unit into contact with the tire with just the right amount of force. Unfortunately, in practice, people probably don’t do that. So, I have to agree that your FTP is likely much closer to what the smart trainer measures.
That said, do be sure to make sure the smart trainer’s firmware is up to date, and to do the spindown calibration periodically. That would ensure its accuracy.