How does Smart trainer internal calculation of speed work

#1

From other conversation: We use a recognized formula which includes rider weight. But below is a different conversation :slight_smile:

I think speed from smart trainer does not come from flywheel speed directly but I am also not sure :slight_smile: but generally itā€™s some device conversion from power to speed/distance and the how it works is probably not that important. We use our own conversion math which factors in riders weight and works with our map instead of reading it from the trainer itā€™s self. Also because the trainer slows can be very off a persons outdoor speed it allows us to semi-accurately apply this drag factor to adjust.

Official word from Garmin on activities
Import TrainingPeaks Workout into TrainerRoad Plan
#2

@bvmackin was saying heā€™s recording on his Edge530 on indoor mode. Iā€™m presuming heā€™s running TD at the same time. Iā€™m fairly sure that TD is not sending the data (simulated data) to his Garmin.

Unless of course youā€™re referring to making TDā€™s data SIMILAR to that of his Garmin. Which I think you are, then this number would/should be more accurate vs Garminā€™s.

But heyā€¦ itā€™s indoor. ĀÆ_(惄)_/ĀÆ

#3

Yes, yes I understand. I was just thinking about your idea about Garmin reading flywheel speed. I think flywheel speed should be based on gearing and cadence and as we know in erg gearing is not related to speed (I might be wrong?) . But yes itā€™s indoor, probably not much distance covered :slight_smile:

#4

The Edge picks up the signal from the trainer and it is simply the flywheel speed.

#5

I just read a Wahoo article that says this.

Shifting gears will cause a change in the flywheel speed

So that means if your theory is correct if I am in hardest gear ratio going 90rpm at 100w (fastest flywheel) it would be much faster than middle gear ranges at 90rpm (moderate flywheel) doing 200w. I would be going much slower in the second case even though the watts are much higher. This might be true but I would think it is not.

#6

( deleted )

#7

I understand how BLE broadcasts speed. We use this speed for SwimERG and RowERG and we show the speed on our UI for these. But you have to look at the manufacturers to see how they estimate this speed. Some might use flywheel directly and wheel-on I assume use wheel speed. But do you see my logic that flywheel speed for Wahoo devices does not make sense based on what they say? Gearing shouldnā€™t affect speed in ERG since it does not affect power. Unless when we say flywheel speed in combination with the resistance applied to the flywheel. Hence this is essentially a conversion from power to speed.

#8

Historically, it has always been like that and I would be very surprised if it changed. If you pick up speed from the trainer, itā€™s basically the same as picking up speed from a speed/cadence sensor on your bike.
For wheel-on trainers, it counts the number of revolutions on the trainer roller and multiplies by the circumference of the roller (which is way smaller then a wheel). For direct drive trainers it counts the number of revolutions of the flywheel and consider that the same as a standard bike wheel circumference because the number of revolutions of the flywheel are what would be those of a wheel.
With ERG, it is perfectly possible to ride 100w at a speed of 45k/hr or 250W at a speed of 10k/hr. The resistance is different.

#9

With wheel-on I agree 100% but I would bet a lot a money (hopefully I donā€™t loose it :slight_smile: ) that changing gears in ERG does not change the speed (other than briefly as it re-adjusts) but does change the flywheel speed. Pretty easy to test with a bike computer even our app in rowing mode. I will try to do it soon.

#10

Youā€™re saying flywheel speed donā€™t change or does change (other than to briefly re-adjust it?)

My testing shows that it donā€™t (really) change which was why it was unrealistic to be used for any SIM/Gradient mode workouts, structured or otherwise.

This is a screenshot of data output for when I was developing my Training App.

Screenshot 2022-11-22 at 6.49.12 AM
Screenshot 2022-11-22 at 6.49.12 AM2022Ɨ1192 354 KB

#11

So I was just going to write some stuff here (why I thought I was right) and I decided to test first to see if my theory was correct and I am completely wrong and you guys are completely correct :slight_smile:

55 watts in small gear in front / big gear in back is significanly slower speed coming from device then big gear in the front and small gear in the back at 55 watts. So for ERG the speed coming from the trainer makes no sense at all for a multi-gear device like a smart trainer :slight_smile: Our method of calculating speed is much more realistic unless you happen to be in the right gear on your bike.

#12

Where is that money :rofl:

Any sensible speed output is the result of a power2speed conversion in an app. The direct output from the trainer is not smart, not on a dumb trainer and not on a smart trainer :wink:

1 Like
#13

Exactly, I owe someone somethingā€¦ Itā€™s unbelievable to me this is how the manufactures did it. I tested Wahoo. I need to test Neo and see if it is the same. Maybe I can win part of my money back :slight_smile:

#14

It has always been the ā€˜safeā€™ choice because otherwise they would be flooded with questions from users with a wheel speed sensor indicating another speed that is ā€œ100% correctā€ when riding outdoors.
Once you start converting Power to Speed, you need grade and weight at a minimum. For a lot of people, things become voodoo when you go that way :wink:
Let alone if you also start using frontal area, wind drag,ā€¦
And I guess you knew I was joking in the former post, but just to be sureā€¦

#15

Yes I knew you were joking. I actually really like admitting when I am wrong. I am an expert in in this admission because it happens so often that I am sure I am right and I am wrong :slight_smile:

Yes I was talking to Tacx a few years ago and they said the Neo uses rider weight for something and since FTMS does not support rider weight they did not want to switch to it. Maybe they use rider weight for speed. I donā€™t remember if they told me why they were using it. I thought it had more to do with feel of resistance or something but donā€™t know.

1 Like
#16

Speak for yourself :stuck_out_tongue:

1 Like
#17

He wants the money :slight_smile: Take the moneyā€¦

#18

Screenshot 2022-11-22 at 10.50.40 PM
Screenshot 2022-11-22 at 10.50.40 PM1408Ɨ752 175 KB

1 Like
#19

Yeahā€¦ I do wonder why. But Rider Weight is included as part of the Sim Mode parameters. IIRC so itā€™s there for speed (so uphill can be simulated in terms of speed of climbing)