New indoor bike, meet the TrueBike

[jfranci3]

@mrpedro

Here is the test I just did: - TrainerRoad

First sprint is in regular mode, simulating a low inertia trainer. It took me 3 seconds to go from cadence 75 to 127, I reached 1006W and when I backed down it was because I had spun out.

Second sprint was in the “True” mode. Same gear as before, but now it took me 6 seconds to go from cadence 77 to 122 maxed out at 1228W and it felt like I backed down because I was toast. All of this felt very natural, it was much harder to accelerate than the first time.

If it works well, then I’m wrong.

Looking at the first two hills…
(How did you measure this? I see you’ve got a lot of cadence going after you dropped the sprint)

It looks like the ramp rate on the first hill (the low inertia trainer) that the flywheel/resistance you didn’t have enough ummpph to keep up with your fresh legs or you didn’t have your ‘move’ down. It is a lot less steep than the other three. The 1 sec web ‘sample rate’ is hurting us here too, but it takes 2sec to add 800w and 3sec to lower to 200w + 3 sec to get to 122. I’m thinking that +800w was mostly resistance unit since you were able to keep some power on after you spun out. It’s interesting that you’re 122w @ 80rpm after you gave up unless your legs were still flopping about - a bigger flywheel would have you unable to put any power down. Interesting here is that your cadence is decoupled from the power and the flywheel does almost nothing… is this an Elite Qubo or something? It’s an low lower Elite, Id expected it to be pretty rough under 150w or so, and I don’t see that, but it could be the 1sec sample rate.

On the second hill… what’s really interesting is that the power changes but your cadence stays steady. It must be using some pedal-weight sensor or some thing. Even when you get off it, you’re at the same cadence and drop down to 140w in 6seconds again.

EDIT -
This indicates they do have a true L/R power meter (load sensor) reading your pedal load. THis woudld be better than how the Neo infers the pedal position.
https://truekinetix.com/the-science/

[jfranci3]

@anon13702412

I went from a Neo to a Kickr. My experience was that the resistance felt very similar. The only difference was the inertia. Neo is superb in that aspect.

What size ring gear? If you’re in a smaller one, you might not notice the difference.

[mrpedro]

I recorded this with my Assioma Duo pedals attached to the TrueBike.
First two sprints are on flat, while third and fourth are on a 10% and 15% incline.

The first is simulating a lower inertia trainer and you can see that as I back down and then start light pedaling, the cadence is all the way down to 70-80. While in the second sprint, with the True setting that simulates much higher inertia, the cadence is still 110 when I start to soft pedal again.

This is the recording from the bike itself: Follow Pedro on Strava to see this activity. Join for free.
I believe the power it records is a 3s average.

So, you are right that I was using double sided pedal power meter to record the ride. Just not sure I can follow your reasoning? Are you referring to the two consecutive seconds that cadence remains at 124 when power is going down in the second sprint?
I am not sure how accurate the cadence is in these extreme situations. If you look at the Strava link from the TrueBike recording you can also see the cadence has that quick dip right after the second spring. If we compare the seconds before and after that dip:

Assioma:
122 124 124 98 116 113

TrueBike:
118 122 123 101 107 114

[jfranci3]

Oh… it’s the same trainer.

It’s clear looking at the data that the trainer has a pretty good read on your detailed cadence, which is how it keeps the power adjusted while your cadence remains somewhat constant. The Neo, for example, can’t do this. Even if you’re using a separate PM to record, the trainer is adjusting well.
Having said that, the reading on the bike PM is **** or overly smoothed. Looks like your TR/Pedal combo overly smooths the cadence, while Strava/Bike overly smooths the load. I’m thinking the bike might be reading cadence at a very detailed level and just adjusting load accordingly.

Looks like it has a Shimano crank? Does the left arm have a bump for the PM? Can you get a pic of the bikes drivetain under the cover? Are there a series of pedal/crank position sensors? I’m wondering how their are measuring power.

[mrpedro]

The bike does not show instant power on the screen, only 3s, 10s, 30s. It looks like it is recording 3s power, that is why it looks so smooth. I would expect the cadence to have no smoothing. I think as it is recording only once every second the two recordings will always be off when zooming in this much.

I actually asked the manufacturer today where the power meter of the bike is. This because I was reading 10W-15W lower power from the bike compared to the pedals. And if it was in the back then at least the drive train losses could account for a fair bit. And it was indeed the case that the power meter was at the back near the motor.

For cadence, since the bike has a fixed gear and it is attached to a motor, wouldn’t the software have a detailed knowledge of the motor position and therefore crank speed at every split second.

From interview I saw with the CEO, and reading their materials, my understanding is that they measure 1000 times per second what you are doing and an AI determines how to respond with the right resistance.

Edit:
Actually this is from their site:

In order to do this accurately, we adjust the resistance 1000 times per second, and we take measurements of you and your TrueBike close to 10,000 times per second. That is simple math done very quickly by our robot.

Also, since your pedal push is strongest when you move the slowest, we don’t rely on electro-magnetic (also called Eddy current, used in most higher-end indoor trainers), fluid (viscous) or wind forces to guide your pedals. We use a very strong motor in our robot, straight out of a car. Since we labelled this section the science section we will show you in a graph. As you can see, a motor can deliver a force, even at standstill (also known as the holding torque). This is how cranes and elevators work. They can hold your weight like we can hold your pedal force, even at low speeds.

So this looks like a difference with the Neo as far as I can understand it. The Neo uses a resistance force that relies on movement of magnets and coils, while the TrueBike uses a motor that can also stand almost still and apply force.

[jfranci3]

@mrpedro

t, used in most higher-end indoor trainers), fluid (viscous) or wind forces to guide your pedals. We use a very strong motor in our robot, straight out

Seems like the big deal is the direct drive. A motor obviously uses magnets and coils. The direct drive eliminates a lot of math. All the other stuff isn’t significant. Maybe the Neo bike does the same.

[anon13702412]

Big in the front any smallest in the back.

[mrpedro]

@anon13702412

I went from a Neo to a Kickr. My experience was that the resistance felt very similar. The only difference was the inertia. Neo is superb in that aspect.

Do you mean that when you are pedaling the Kickr and Neo feel the same to you, and when you stop pedaling you notice that the Neo keeps on spinning much longer?

@jfranci3

Seems like the big deal is the direct drive. A motor obviously uses magnets and coils. The direct drive eliminates a lot of math. All the other stuff isn’t significant. Maybe the Neo bike does the same.

Sure, it is all coils and magnets in the end. Although I suspect that if they say that the TrueBike has an electric motor taken from a car, it might use coils and coils, generating much higher forces.

This discussion actually got me thinking some more and I think I understand the difference with the Neo. The big deal is the robotic part, that is not just marketing talk. What I make of it is that the TrueBike is actually holding the pedal with the entire static force of the motor in a fixed position. Then as you apply force, it measures the force you apply and computes the next position of the pedal, the motor then spins to that position and allows the pedal to move to that next position, holding it static again for a millisecond.

While with the Neo, the movement of the pedal is more of a direct result of braking force from the magnets, inertia and the force applied through the pedal. Two different worlds.

[anon13702412]

@mrpedro

Do you mean that when you are pedaling the Kickr and Neo feel the same to you, and when you stop pedaling you notice that the Neo keeps on spinning much longer?

They feel the same except for the inertia which is super high with the Neo. In terms of spinning time I don’t know. I only stop pedaling when I get off the bike.

[mrpedro]

@anon13702412

@mrpedro

Do you mean that when you are pedaling the Kickr and Neo feel the same to you, and when you stop pedaling you notice that the Neo keeps on spinning much longer?

I only stop pedaling when I get off the bike.

And what is it that you mean by inertia?

The crank is always accelerating and decelerating throughout the entire pedal stroke, as our legs don’t apply constant force throughout. You can see that in the peanut plot in my first post that shows the pedal force throughout a full revolution of the pedal.

The feel of the trainer is determined by how much resistance there is to each acceleration and deceleration. That resistance is the result of the inertia of the flywheel and the magnetic braking force. Funny detail is even that the Neo has much less inertia than the Kickr, as in the Neo that resistance to acceleration comes mostly from the magnetic brake force that is applied at exactly the right moment.

Then when you stop pedaling the Neo keeps on spinning for a while as it uses a motor to spin the flywheel. This works just like inertia does, except that it is of no use during the pedaling. It only means that when you coast and then start pedaling again, you can do so at a much higher cadence than for the Kickr that would have slowed down more.

[anon13702412]

https://www.saris.com/post/blog-15-cycleops-science-resistance-inerita

That’s what I mean. Both trainers have that. Though the Neo doesn’t do it via the flywheel.

[mrpedro]

Gotcha.

I am quite interested in the differences between the Kickr and the Neo. As I have only experience with the Kickr and the new TrueBike. It sounds then like the Neo is simulating the same feel as the Kickr, but with no flywheel?

Before the Kickr I had a Kinetic wheel-on fluid resistance trainer. My experiences so far is that each trainer does not feel terrible in itself, until you try the next one:

• Kinetic, even with larger flywheel, is just marginally better than riding through mud. Pedals accelerate down extremely fast and then you still need to keep pushing the pedals around until the next power phase.
• Kickr feels much much smoother. Pedals are still accelerating down faster than outside during the power phase, but it then spins nicely through until the next power phase.
• TrueBike has a ton of resistance during the power phase and then pedals turn smoothly until next power phase.

[anon13702412]

The only difference from the Kickr to the Neo to me was the inertia. Especially during longer sustained intervals the Neo was significantly smoother. Though after a couple of weeks that feeling was gone and I was and am now very happy with the Kickr.

[Peter_Verbeke]

Hi there, I just bought a Truebike myself and I was wondering about your experience riding in ERG mode. I tend to overshoot when my target power changes. Also, if my cadans is too high and I lower it a little, resistance (and therefore power) drop too much. Just in general, I have problems maintaining a steady power level.

I do like the feel a lot, but I wonder why the bike doesn’t get me more quickly (and more stably) to the exact power level programmed in the workout.

The bike is still very new for me and I’m just getting my bearings on indoor training. At the moment I’m trying Zwift, but I imagine it would be the same on any program.

Peter

[mrpedro]

Hi Peter,
Do you have experience with ERG mode in other trainers. I found it to be quite similar. For example in this workout: Log In to TrainerRoad
You can get huge power swings when changing cadence fast. Here, when doing form sprints and going quickly from 90 rpm to 140 rpm, power goes up to 500W, although it is fixed at 160W.
Then for the parts where I am trying to keep cadence more constant you see much less fluctuation. The TrueBike transmits a 3 s smoothed power. but still the power fluctuations seem in line with what other trainers would do.

Here is another workout with lots of intervals with changing power: Log In to TrainerRoad
You need to get used a little bit with how ERG mode works, for most intervals power changes quickly and smoothly. However for the interval at 1:00h, I was not paying attention and the cadence dropped when the resistance increased. As I then quickly recovered, it took a little bit of time for the power to normalize.

Is this the level of power fluctuation you are experiencing?

I think that because of how the TrueBike works, it would be theoretically possible to really fix the power in ERG mode. However that would not create a nice or realistic sensation. Any power fluctuation of your legs would then cause a large change in resistance and cadence, and that would be going on the whole ride.

[Peter_Verbeke]

Hi Pedro,
Thanks for getting back to me. I had the old Daum Ergobike 8008 TRS. It always followed the power curve of the workout to a tee (but that might have been the display messing with me;-)

With the Truebike if I stick to the same cadence all the time it’s acceptable, but I still dislike how wobbly the power level is at any level. It seems to me that there is an issue with the feedback loop. It allows me to stray to far from the target even when I keep a relatively stable cadence.

I really dislike the power drop after a sprint. If the program requires me to lower my cadence at the same time I often even drop out or ERG-mode. I try to compensate for this by shifting to a heavier gear, but at the end of a workout I run out of heavier gears (it is a bit silly in ERG-mode to use the cassette setting, better to just increase/decrease resistance).

Good point on the fact that it would result in an uncomfortable ride should the power be really fixed, but I feel it should be possible to improve adherence to the target power without sacrificing the quality of the ride (which I really like a lot).

Anyway, I also took a couple of screenshots to illustrate some of my issues.

[mrpedro]

The fluctuations you are showing in the graphs look normal and this is what you will see in ERG mode with any trainer. A graph that is showing really straight lines in ERG mode is showing a smoothened line by averaging power over e.g. 10 s. But the underlying power is always going up and down like this.

Your example of setting power to 160 W and seeing fluctuations of +/- 12W is as smooth as you going to get. Try riding in free mode at a steady cadence and see how much fluctuation is coming from your own legs.

The overshooting example is the same as what I had shown. If at the beginning of the interval you let the cadence drop and then compensate by increasing cadence again, it will overshoot and take a few seconds to normalize again. You really need to be ready to start cranking out more power as soon as the resistance goes up to keep it as steady as possible.

The power drop after a sprint that you mention is the same as when riding outside. If you sprint with a cadence of 100+ and then start riding with a cadence of 90 rpm, you will be freewheeling until the speed comes down enough. On the TrueBike you can try squeezing the brakes after the sprint until you start feeling resistance at you lower cadence.

[rocourteau]

@jfranci3

The Neo has a “virtual flywheel” function. It guesses what you’re going to do and then adds/lowers resistance mid-stroke with your rhythm , but it’s very aggressive / not responsive to unpredictable or momentary lapses of power application.

There’s no “guessing” needed. The physics of an accelerated rotating object are pretty simple - torque = inertia times angular acceleration. You measure the torque (through the powermeter), you know the inertia (that’s your artificial flywheel’s inertia), you change the motor’s current to get the proper acceleration. No “AI” or “guessing”. If you do the measurement and current metering fast enough (and 1kHz is in the right range for that), the user won’t feel anything abnormal. You can make the inertia as large or as small as you want. And you can add variations for road feel if you want. This is what the Neo products do.

Flight controls in flight simulators are done in a similar (but much more precise, and expensive) way: for non-powered controls such as elevators on smaller planes, you literally feel the inertia of the entire linkage and flight control surfaces, and feel them get loaded as the plane accelerates. There’s no magic (or AI) involved. Simply simulation of physics.

[rocourteau]

Read the info on the web site. I’m not quite impressed - there are some, well, interesting views in there. The whole argument about “less power indoor” hinges on two studies, for which there does not seem to be any info on what sort of trainers were used. It’s hard to state “my trainer is better” when a) you don’t say what you are comparing it to, and b) you don’t repeat the same studies with your trainer to demonstrate the results depend on the trainer. The whole argument about needing a flywheel that can contain the entire kinetic energy of a bike and rider is moot - the only thing that affects riding feel are time derivatives of rotational speed, not the speed itself.

Follows the usual argument about carbon fiber bikes that will turn into small balls of burning plastic if used on trainers. Refer back to the Canyon vs The Trainer saga on that one.

On the spec side of things, it appears that ANT+ is supported, but not BT. Crank length is not adjustable - you can select it when ordering, but that’s it, so that’s one (small) negative in the multi-rider adjustability use case. Virtual gearing appears to be selectable in their app, don’t know how that plays with other apps. Unsure what the display looks like, and whether you can side-load other apps on it.

[mrpedro]

I have the bike connected to TR on an iPad through BT.
The gears are changed on the bike, regardless of what app you use. Just like when you have a regular bike on a trainer.
It is not possible to install any other apps directly on the display of the bike.
I didn’t notice a very big change in power going from using a Wahoo Kickr to the TrueBike. I did notice a much nicer experience of ride feel.