Power comparison with Vector 3 Pedals

Currently Zwift power is down, but I have around 7-10 watt difference across all sections in ERG mode. Below the image

Here is a graph directly from zwiftpower.

And this one

I used to get a bit less but dunno, it just seems bad and wrong like others reporting.

-are the Assiomas paired to Zwift via BT or ANT+? (i suppose one of the file is Zwift FIT file /Assiomas/, the another one /Neo/ is from a head unit). 
-what was your exact gear ratio during the ERG session
- is your crank lenght set right in the head unit also (You need to set it in the Assioma app and the head unit also at same value otherwise the head unit overwrites the app setting) how long is your crank? 
- did You make zero offset calibration on the Assiomas before the indoor session? 
- do You have round chainrings or oval?

Hi thanks for answer:

Assioma is connected with Ant+ mode on Garmin 1030

Zwift is connected directly to Neo 2t.

Assioma is calibrated via app and garmin directly before drive (I also tried calibrating several times=

Crank length is set both in assioma app and in the garmin to the correct value.

I have oval chainring, but its the newest AbosluteBlack premium which shouldn't cause the problems?

Thanks for any tips.

-what was your exact gear ratio during the ERG session ? was it on small ring and ~middle rear or on big ring front? 

- the ovals could be a modifier (Even the IAV algorithm should need to handle that but never knows... i assume the Assiomas are on the latest 4.xx firmware) 

- a static weigh calibration (with a 20-40kg dumbbell or known weight) is a good method to validate your Assiomas /static/ accuracy, its precedessor (BePro) was famous about its inaccuracy due to the wrong factory slope calibration, the Assiomas are much better now but if one of pedal slope is wrong that could cause such a different .

- what is your balance during these sessions (unfortunately the ZP layout is not showing the power separately from each pedal like the other analyzer tools) ? 50-50 or different? try to pedal with one leg in ERG @60watt for 1 min, 30sec with 2 legs , 1min with other leg, and repeat it 5x times, and we will see is there any different difference between one pedal and Neo than other pedal than Neo. 

- Can You share the link directly to this ZP comparison? on small power the difference is looking less (~3-4watt what is a good different) only @ 200w is looking much bigger.

https://www.zwiftpower.com/analysis.php?set_id=67957

50/34 compact crank, and 11-30 ultegra cassette. I had 34t infront and for this session used 19t in the back. I also tested 17t and others without any much difference. 

The balance for this session was 50.9/49.1, but usually I get 50 50 to 51 49 with all my other rides. I haven't really noticed any issues with pedals in my outdoor rides the results are consistent.

And yes what you haven noticed, at the lower wattage range from 90-120, the difference become smaller and what it should be, but as soon as I go higher in watts the difference becomes bigger, up to 8-12watts.

Yes assiomas are on 4.25

the results are consistent indoor also here as i see just we would like to guess why do You have a slightly higher difference than the normal . So at outside You wont notice any issue unless You can record with an another (spider based 4example) powermeter same time and compare these 2 bike powermeters too . What is annoying me (i noticed it my friend also and at others sometimes who has Duos ) that the cadence is higher almost with 1% with Duos, and probably this overmeasuring in cadence is resulting the overmeasuring in power too (at my friend he tried it with a Power2Max NG meter in same time and he got 5watts more on Assioma always than with NG which is the most accurate and reliable powermeter in the market now /i have 3 since 3 years so i could test these very deeply), so maybe You have the same problem as he has (i told hime that probably his left side pedal is a little bit high but he couldnt test it with static weigh so he is still using it this way...) , indoor the NG is  always 3-4w higher than a Neo ( i tried it with many Neos and with multiple NGs, this 3-4watts are the drivetrain loss at 200w , so always have this different if You measure the power 2 different places and between these points there is not a 100% efficient power transmission), so if yours is also 5w higher than a NG (i assume it ) than the 8-9w is "good" but it is the result of the Assiomas inaccuracy) 

Maybe the single leg drills help to find out .

Here is my friend's comparisons ( a part of it) from outside

Thanks for your answers, however as you see from my analysis the cadence is fairly close. 

And on some of the tests I did the cadence was pretty much same. The difference in cadence comes in the steep dives, where you stop pedaling, and tacx still spins, while assioma is already showing 0w.

I don't think assiomas have a problem. I have a friend who trained with me on the same drill that I posted you yesterday, with same assioma duo, only difference is he has a Neo 2 on older firmware 0.14.

See on his its pretty much spot on, 1-2watt difference. We are using very similair setup. I honestly think Neo 2t has a firmware bug somewhere, and is not working correctly.

then ask Your friend to try your 2T with his setup and You need to go to him with your setup to try his Neo2, so simple :) 

i checked the cadence at many places (I'm an expert power analyst, it is my "job"  ;) ) , especially where You made the 6mins long intervals and your cadence was steady without stops, at all intervals the CAD with Duos is higher almost with 1rpm (so ~1%), just select these intervals and check with your own eyes.

the real problem is that manufacturing very accurate pedal based powermeter in mass is much heavier (or impossible to reach a very high ratio) than manufacturing a spider based one (where the material of the spider is homogene, responding clearly for temp change, and You can glue the strain gauges onto with 100% precision), and there are too many out of spec pedal based power meters on the market due to this problem , so if You take 10pcs Assiomas from the shelf, there will be weird differences between these, if You take 10pcs NG , You will have totally similar measurings and results (just lots of buyers dont know it...) 

Yes that's exactly what I plan to do, once this Covid-19 situations goes a bit down, as unfortunately we are currently in lock down, so I can't travel to his place yet :( But once possible, I'll take my bike and everything on his Neo 2 and he will test mine, and we will post some results to once and for all see where the issue lies

Maybe it really is the oval thing, because the chain is jumping a bit more in the higher speeds, but hard to say. I'll try some more today to see what I get in simulation mode, however seeing other posts look very close to mine, like below you can find 5 different zwift analysis which consistently show Neo 2t having same problems as mine, even somebody posted Assioma vs Neo 2t which has same results.. do you really think both are pedals are the problem? I am really skeptical about that.