In the spirit of locating the issue first, can you try an erg workout using the tacx app on your phone. That then isolates the issue to either zwift to trainer, or trainer itself.

The tacx app is painful to use I find, and not all that intuitive, but handy in this case.

Note there are two apps, the utility and the training one. I've a feeling you could use the utility one, but not 100%, as I use the training one in situations like this. Just create a 150w 5 min power based session.

Whilst not taking away from the above, then would also be interested to know what the power target is that you are aiming for, and what gear you are in. I ask only as I know my neo struggles to get below 120w if I'm in big ring.

Also can you confirm on zwift if you are taking the power feed from your power meter or the trainer. It should work with both, but just interested. Finally, check on DC Rainmaker review - one of the tacx trainers is known to be dodgy for erg mode. I can't remember which one though, so check his review and see if it mentions it. It was really really clear in the review, so if it's not mentioned then it's a different trainer.

Thanks for the reply.

I did also try running the Tacx app at 150w, it was better at holding 150w at a steady cadence, but same issue once I shifted or changed cadence - no reaction from the trainer.

On the gearing, mostly 52t + 13-15t when I was trying to troubleshoot it at 150 - 240w. But, I did drop down to the small ring (36t) to try that out, no luck. I bought this for my wife and her rides were in the 36t small ring on the tall side of an 11-28 cassette at 70-130w, no luck there either.

I was hesitant about buying a Tacx trainer after the quality issues DCR and other have written on in the past; if I can find a Kickr Core in stock tonight this probably gets returned.

Most smart trainers have an optimal working zone for flywheel speed. Pushing a big gear with high flywheel RPMs can result in them being unable to hold watts in ERG... and their power accuracy ususally goes to shite too.

Try a 250W steady state ERG in the small ring / mid cassette / 90rpm.

As for the tests your wife has performed - Some trainers have a 'wattage floor' and have an optimal working zone for ERG. I know the NeoBIKE was only good for ERG > 140W or so with the initial firmware. I believe this has been addressed with a recent update.

I can't recall the specifics of both these limitations on the FLUX S. I had issues with the spindown and put it aside (it's still aside....)

---

Shane Miller - GPLama
YouTube | Web | Twitter | Instagram | Facebook | Strava

Thanks. FWIW then I have 2x Neo (mk1). One for me 85-90kg FTP 300+ and max power 1100w+ and my wife (much lighter), ftp 140w. Had them for getting on 3 years now and been fine so far. But the general thing is if they work form the box, they work. But there is way too high a failure out the box IMHO for something this expensive and hard to ship/return.

So does sound as if something is odd, especially as you say it tends to work on the training app. Although I must admit you confused me a bit with the comment about constant cadence. On 'normal' varying terrain riding on zwift do you get the changes in resistance? (ie just checking the brake is adjusting).

Where is the cadence coming from? Powermeter or trainer, or specific sensor?

Have you got any other 'things' connecting to the trainer? IE a garmin headunit (I recently got a 530 that caused some issues whne on at same time as zwift).

And to confirm you are connecting via bluetooth on your phone (and the blue light is connected, not the ANT+ on the trainer).

Would be good if you could stick it on a 175w target then ride, record the .fit data from both powermeter and trainer (on separate devices with cadence from trainer / powermeter). Start off holding that 175 at 85rpm, then up the cadence to 95 for a minute, then down to 75 for a minute, then change to the inner ring, and do those 3 cadences again for a minute each.

Appreciate the time and thoughtful responses from you both (GPLama, I'm a big fan of the channel the Fit File Podcast). I was able to find a Kickr Core instock, so I am going to return this.

I suspected it might be an ERG floor but had no luck finding details on that.

@Duncan74 The "just works" factor is the entire value proposition for a smart trainer in this use case, as my wife doesn't want to fiddle around with this thing. I can confirm it was just connected via bluetooth when I did the phone test, even turned off ANT+ speed and cadence data since that was an option in the Tacx app. Also tried FE-C protocol and the other ANT+ option.

I should have put some additional context in my OP, but I got this trainer for my wife. She is brand new to cycling and picked up zwifting during quarantine; but, does not want to shift gears and likes the simplicity of setting a structured workout and just turning the pedals (+getting stars). I did not have an opportunity to test how it reacted to the Zwift changes in terrain, since her use case will be 99%+ ERG.

For reference: the first 30-ish minutes of the strava workout below were me trying to get ERG to work, it's a slow ramping 1 hr "warm-up" from ~150-200w (back half is on a Kickr Core with ERG smoothing). The first 30 of the Garmin workout is the corresponding data from my Assiomas.
https://strava.app.link/SqanzEn4d6
https://connect.garmin.com/.../activity/4871439710

I 100% get the 'just works' factor. That's why I went for a second Neo for my wife - no calibration.

That data is great. So looking at the strava data, then that looks to me as if the trainer is working perfectly. I say this based on looking at the analysis on strava and looking at hte cadence and power on a time based graph.

To me then it looks like the early 15mins (3-18min) is well within the erg consistency I get / expect. And th eback end is amazingly smooth and accurate. You've even got some significant cadence spikes eg 1h24 and 1h29 that is decoupled from power.

Here's a plot from an erg session i did last week. https://www.strava.com/.../3332792858/analysis You can see there is a fair 'wobble' around the target power about 10w. FYI then this is using zwift to control the trainer, but the power is coming from the quarq.

Here's a plot from my wife, much lower power and you can see the 'wobbles' here are more noticable. This is reading power from the Neo. https://www.strava.com/.../3382411802/analysis

So based on that, I'm not sure what the issue is? On zwift you do get it popping on screen to say speed up / slow down, but ignore that for 5 seconds and it'll catchup. My wife had a few issues early too, especially with the cadence based workouts, some very very bad gear choices by her, and the power floor. So I moved to inner ring and mid block, told her to not touch the gears and just pedal as told and let teh resistance change and she's fine now. But I admit that providing this sort of information / advice / coaching is a very delicate process........You have been warned.