Control smart trainer's resistance with external power meter - important for Workout Mode

(Lee Mitchell) #161

Michael Henasy - I cancelled my Zwift subscription a while back due to the inability to use Erg mode for workouts using my Power2Max as the power source (I do a lot of steady Tempo/SST/Threshold stuff and find Erg mode more convenient for that). TrainerRoad has the powerMatch stuff sorted.

When I recently found out about the ability to control my Kickr Snap with my external PM (as per the procedures you described above) I was very excited and yesterday reactivated my membership.  I’ve connected to Zwift using Bluetooth only by pairing my Kickr Snap as both controllable trainer and power source (but the Kickr being controlled by the PM) and found that the power being read on Zwift and on my Garmin head unit was very close but not quite the same - but certainly a lot better than the usual delta between what my PM reports vs Kickr Snap.

HOWEVER, when trying to do a workout on Zwift with Erg mode enabled it just doesn’t work.  My head unit (which is picking up the Ant+ signal from the PM) reports a significant difference to what Zwift is displaying and the difference gets larger as the power levels increase.  For example, today a 210w effort in Zwift was being reported at about 240 on my Garmin (and it certainly felt like 240 not 210).  I tried changing cadence up and down and even stopped pedalling a few times to see if that helped but after a couple of goes Zwift simply disabled Erg mode for the workout and it would not engage again.  Not sure what is going on with this…

(Sunil Savkar WBR (C)) #162

Eric, you misunderstood, I was referring to mating the power meter with the Wahoo KICKR Snap, was not referring to anything built into Zwift!

(Eric C. (Zwift HQ)) #163

@Sunil - Gotcha! Was confused for a sec :slight_smile:

(Michael Henasey) #164

@ Lee Mitchell

it was working for me through a Zwift workout and ERG mode. I would suggest starting up the Wahoo Fitness App first and starting a workout there to make sure the KICKR is in fact connecting to your external power meter. The little green lightning bolt with arrows circling icon tells you its connected. maybe the SNAP behaves a little different than the full KICKR.

(Michael Henasey) #165

Based on what Eric C. just said and what Eric M. said to me on Strava, ERG Correct or Power Match is coming to Zwift. We just have to wait a little bit longer. 

(Sunil Savkar WBR (C)) #166

@Michael Henasey @Lee Mitchell

As far as I can tell (as a SNAP owner), no differences between KICKR and KICKR/Snap regarding the power meter sync up.  That noted, I always notice a difference by a few watts between what I see on my Garmin head and in Zwift when I get this set up, because I think the Snap/KICKR use a slightly different smoothing sampling period versus what i use with the Garmin.  Not sure.

That noted, I admit I got a bit tired from the constant playing I had to do each time to ensure things are still working, which is why for now I have moved over to Trainer Road for more serious work outs.  Until I can clearly see Zwift is reading and controlling my trainer directly versus this indirect method.

(Michael Henasey) #167

@ Sunil Savkar

I did the TrainerRoad thing coupled with Zwift for a while but eventually gave up because (1) TR would crash on both Mac and Windows randomly and interrupt my workout and (2) the UI experience was disjointed and not having the glowing arches to go through made a real difference to me. TR on my iPhone was much more stable but then the experience was ever more disconnected while also on Zwift. Ideally I would like to be able to do everything in Zwift and not have to run other apps, etc. Hopefully the new ERGCorrect feature Zwift is currently working on will resolve all of this.

(Sunil Savkar WBR (C)) #168

@Michael Henasey

I am with you on this.  I actually am just using trainer road alone just for training focus, and not at the same time as Zwift, since now when I want “atmosphere” I am really heading more outside as it has warmed up.  Still will sporadically use Zwift here and there for “fun” rides in eve or on a rainy rest day, and certainly will start up more in earnest in the fall I suspect. 

Frankly in a perfect world, I’d love to have the two companies work on a combined effort so that we could use both together - I know that may defeat the efforts Zwift has alone on their own platform, but the combo could be pretty slick, and especially given some people are already trying to do this themselves.

(Eric C. (Zwift HQ)) #169

Please make sure to try out this feature in our latest update (May 3rd, 2017) :slight_smile:

(Michael Henasey) #170

Sweet, it’s here!

I’m going to try it with today’s workout!

(P H) #171

Steady pace for 65 minutes and averaged 1W over target. Looking good!

(Tom DeMay) #172

Yes! With 1" of rain accumulating today and another inch of rain expected tomorrow, this is perfect timing. I’m assuming this is straightforward to configure/set up? Planning on trying it out tonight.

(Arturo G. Artur) #173

First impressions on BKool Classic are fine. Great job! 

(Tom DeMay) #174

OK, I did this last night, and it seems a bit laggy, but does seem to work.

I found Zwift having trouble keeping up as I’d pedal faster / slower, but it would eventually get me around my Target Number in ERG mode. If I accelerated hard, I’d go way over my number, Zwift would compensate, send a signal, and then I’d drop quite a bit below my target number. After a few seconds, it would balance back out and I’d be right around the target number. It definitely did not keep a perfectly solid ERG number like it does when I use the Kickr SNAP as the power instead of my Power2Max meter, there was a bit of lag up/down. I think this is working pretty well for a Beta, but would love to see the lag/adjustment happening quicker.

Question — on my Kickr Snap I have these options via the Wahoo App:

  • ERG Mode Speed Simulation (I have this toggled OFF — not sure what this is even for, I think it should be left OFF)
  • ERG Mode Power Smoothing (I have this toggled ON — is this affecting Zwift’s calculations? Should I turn it off?)

Thanks for any help…


(Brian Dixon) #175

Certainly a step in the right direction but still needs some work. Seems to take about 20-30 seconds to stabilize after a change. Nowhere near as smooth as TrainerRoad yet. It seems the feedback time period is set to high.


(Lee Mitchell) #176

My experience was very similar to Tom DeMay’s.  It’s not as smooth as TR’s PowerMatch but for my purposes works plenty well enough.  Power generally stayed within 5w either side of the target power with occasional “spikes” of up to 10w which were quickly corrected.  Much closer to real world efforts so probably more realistic than a constant power level.

I’m using a Kickr Snap (which I didn’t spin down first because I hate having to do that) with a Power2Max via Ant+.  I disabled Erg Mode Power smoothing and Control with Ant+ Power Meter in the Wahoo App first.

My workout today was long steady efforts so will have to test to see how Erg goes with short hard efforts.  

Pretty happy with this so far.

(Michael Henasey) #177

I feel the new ERG correct is a bit too aggressive. You end up yo-yo-ing over and under the target wattage for some time before it begins to settle. I’d say it can take a good 20+ seconds for this to work itself out.

For reference, i’m using a 2nd generation KICKR and a Stages power meter. The workouts I have tried involved intervals between 1 to 3 minutes long where the target power would change from 200W to 355W. 

I find that if you use the small chainring the yo-yo-ing is worse compared to when using the large ring. For the KICKR, it’s been discussed that ERG mode resistance changes can occur more quickly when using the small ring as opposed to the large ring. I tend to agree with this and this appears to explain why I believe the yo-yo-ing is worse in the small ring vs. the large ring.

When in the small ring, when a new target power occurs, for a good 5-10 seconds, Zwift overcompensates for the amount of resistance needed and my power comes out 50 or more watts than what is targeted. This then causes Zwift to then under compensate and I soon find myself doing 30-50 fewer watts than what is targeted. The yo-yo-ing continues in smaller differences until it finally settles around the target wattage.

When in the large ring, this phenomenon still occurs but not to the same degree as when using the small ring. It’s more tolerable in the large ring as the sudden changes in resistance are not as great.


(Clifford) #178

I experience the same thing on both Zwift and TrainerRoad although TrainerRoad is perhaps a little better at dealing with it (although it isn’t perfect). The issue makes really short intervals almost useless. You need 15 to 30 seconds for things to settle down. It becomes discombobulated with the short 15 to 20 second intervals, particularly if there is a big cadence change.

I wonder if they need to rethink it completely? Instead of trying to compensate on the fly (leading to the yo-yoing) why not have an exercise where they built a profile of your trainer (based on your PM’s readings) that they then use in a pure trainer-based ERG mode? Maybe there is a ramped workout that they use to figure it out, that you could periodically re-do, any then they use that power curve to control things instead of using the raw output of the trainer?

Not sure if that makes sense, but I think it’s ultimately the only way to solve this issue … Of course, the rider would have to be willing to work through the calibration process …



(Michael Henasey) #179

I agree @Clifford Smith that ERG Correct should be more “intelligent” to help alleviate the yo-yo-ing. 

When the next interval hits, Zwift needs to take into account that as the rider adjusts to the new resistance, either their cadence will suddenly increase or decrease and as a result the power reported will suddenly and drastically change as well. When this is happening, Zwift shouldn’t try to re-adjust the resistance until some time has passed to allow the rider to adjust. If the target power goes from 200w to 300w for the next interval, Zwift should set the target power to 300w and then wait 10-15 seconds for the rider to adjust before re-attempting to correct the power. This would help to prevent the yo-yo-ing. If Zwift is calculating an average offset of power between the trainer and external power meter, Zwift could also factor this in when the next interval hits and the power/resistance changes.


(Clifford) #180

@Michael Henasey - that sounds like a good approach! Hopefully someone is listening! :wink:

On TrainerRoad I’ve been trying to work around it by getting my cadence up to my target for the next hard interval in the minute before the next interval hits to get the yo-yoing out of the way, and then trying to hold that cadence into the interval. It’s not how I would like to handle it, but it’s seems to help a little. Of course, that helps get the power up, but then the yo-yoing starts and I struggle to keep the power at the target for the rest of the interval.