i have no probs with .0.0.34, but want to ask, is Simulation mode used in TR, or is it only relevant for Zwift, Tacx App Riding, is it more or less the road feel thing?
Sim is not used in TR. They use a Resistance step as a percentage of the max trainer resistance , which is not the same as Sim.
nothing negative like positive has changed on my Neo 2, the power inaccuray himself i have under control with new TR Powermatch via Vector 3 in TR Win10 App
I raised a support ticket with Garmin/Tacx regarding this as my 2T was consistently reading 4% lower than my OG Neo as compared with my Vector 3 pedals. I supplied them with stacks of evidence including .fit files of the discrepancy. They offered to replace my perfectly functioning pedals!!! 
NEO 2 Firmware Update 0.0.38 Available
2-5% drive train loss is actually spot and has been demonstrated in several drivetrain efficiency studies.
Garmin is correct, if anything your Tacx Neo 2T is reading a touch high.
I just read up some studies at it looks like 5% drive train loss is typical.
Taking into account +/-1 accuracy of the 4iii and +/-1 accuracy of the Tacx that brings us to a possible range of 3%-7% difference (assuming 5% drive train loss) or a 9-21 watts difference meaning the Tacx Neo 2T may very well be within spec.
Furthermore, I have a powertap hub, 2x favero dual pedals and a Neo 2T of which I have run my own experiments and have similar numbers.
Spoiler alert, it means your original Neo has always been reading high.
You lose energy through the drive train so a hub based power trainer should in theory always read lower than a pedal or crank based power meter.
Point is 20 watts seems like a lot but after accounting for drive train loss it’s likely just slightly off. ( if off at all)
Is your 4iii dual or single sided?
You can’t compare a single sided power meter to the Neo which is measuring both legs.
You likely have a power discrepancy between your legs and +/- > 2% is not uncommon.
It’s more than likely your Neo 2T is spot on.
I never said the 2T was perfect, but I did clearly demonstrate that it’s feasibly within factory spec of +/- 1 % accuracy when you take into account drivetrain loss (5%), possible +/- 2% difference of factory specs and possible L/R leg power imbalances (easily 2+%).
The fact your prior Neo and Neo2 were in 1% agreement with the pedal/crank based power meters means one of them is inaccurate.
The numbers seem to suggest otherwise, at least to any logical person.
Someone here is auditioning for a job doing smart trainer technical support
Also, note how all of DCR’s and GPLama’s reviews show the trainers are consistently 2-5% lower than the comparison power meters. /s
I don’t really see the problem. If you have a separate power meter on your bike you should be using that as a power source anyway so your inside and outside numbers match.
I am a happy camper with my Neo 2T so far. The power readings seem to be super consistent for me, which is what really matters and aIlows me to train properly.
I don’t have an external power meter yet, but I am looking forward to my free 20W FTP bump, should I ever get one.
for what it’s worth, i have a 2t and dual sided assiomas. same problem as you. 20w less at 300w. i’d say it averages around 8-10% less.
you don’t see the problem that a person needs a second power meter to make their $1400 trainer accurate? ok.
I don’t think that the mean max graphs are all that interesting when comparing two different measuring devices, and certainly not the main thing I care about. With the mean max graph, we don’t even know if the ‘10 minute’ max power is even over the same interval or not.
What I care about, and what I think what others care about, is ‘for a given interval’, do my power meters agree. Average power is the way to do this. This matters for things like ‘what was my power up that climb’, or ‘what was my power during my 20 minute FTP test’, or 'what was my power during the last minute of my FTP ramp test.
For the Neo 2t (DC Rainmaker Analyzer) for the entire ride:
Dzero: 246.20 watts
Powertap P2: 243.19 watts
Neo2t: 240.56
The neo2t was 2.3%lower than the dzero over the entire ride, and closer to the P2s. Looking at several sections the error was fairly consistent in the 2-3% range to the Dzero, and less to the P2s.
Results from the original neo were generally slightly better, although the SRM pedals used in that test seem way off. DC Rainmaker Analyzer
If people are seeing 20 watts low at 300 watts (ie a 6.7% error), I don’t think this is ‘expected’, or inline with review results. Yes, if you have a horribly maintained drivetrain, you will see higher losses than if it is in reasonable shape. People who are spending for power meters and high end smart trainers, and are detail oriented enough to compare them likely have at least decently maintained drivetrains.
Having dealt with smart trainer inaccuracy issues and customer support for my trainer and a friends, it is absolutely amazing at what customer support thinks is normal and acceptable. Trainers that are 10-15% off from 3 other powermeters are just fine, and it’s the power meters that are wrong.
I’ve done countless dual analysis with different powermeters and the Neo 1 and 2T. At the end of the day, the same conclusion has been reached - the 2T reports power lower than any of the powermeters used and the Neo 1.
What’s interesting is that flywheel speed seems to really matter. At lower flywheel speeds the Neo 2T begins to read closer to whatever powermeter I’m comparing it to.
Whilst the discrepancy doesn’t matter for many, at the top end of eRacing where you must connect your trainer as the primary powersource to Zwift, a difference of 4% is hugely significant. Whilst probably within the accepted tolerance for the manufacturer, we’re getting into the situation that trainers are being required to be absolutely precise and accurate, and used in ways that they were never designed.
I’ve seen this kind of thing with ERG before, and I think the problem is with Zwift and ERG rather than the Neo or powermeter. I’ve linked a dual recoding I did with my OG Neo and old Powertap P1S pedals. This was a workout I imported to Zwift from Xert. Notice how the power reading on the pedals ramps down during the work interval, but the Neo reports a consistent power.
So I did get a pair of Assiomas recently and so far I am happy to report that they track quite closely to my Neo2T (so no big FTP bump for me…). I have done three TR workouts with dual tracking.
Pallisade (interval):
Neo2T: Avg 282W ( -1,5% )
Duos: Avg 286W
Mills:
Neo2T: Avg 194W (+1,2 %), NP 245W (-1,4%)
Duos: Avg 192W, NP 249W
(about 7W seperation at ~320W, so 2,3%)
Jepson:
Neo2T: Avg 211W (-0,5%), NP 241W (-1,6%)
Duos: Avg 212W, NP 245W)
After a hard effort on a local climb, intervals.icu gave an eFTP of 278W with the Assiomas and my last ramp test with the Neo2T gave me 279W.
Both devices show a L/R power distribution of 50%/50%
In conclusion, I cannot really confirm the bad rep the Neo2T seems to get on forums in this regard and I am happy with both purchases 
Although I guess for people with FTP north of 350W the difference could become substantial.
