Turbo Vado 2 4.0 or 5.0 or 6.0?

[Jeremy McCreary]

the Pm data has been adjusted with the efficiency curve above to be Pm mechanical.

Spec's SL 2 documentation says that the mechanical Pm and power ratio readings on the TCU and app are just estimates obtained by multiplying measured electrical Pm by a constant assumed efficiency of 80%. Don't recall that they show raw electrical Pm data anywhere.

Just confirming that you backed out the assumed 80% before applying your empirical efficiency vs. cadence curve. If not, the gap between your empirical Pm/Pr ramp slope data and the corresponding Spec MicroTune data is even wider than seen in @mschwett's last plot.

Another reverse-engineering hurdle: The new empirical 40/40 and 40/100 Pm vs. Pr plots once again show that ramp slope S = Pm / Pr varies strongly with both E and M in assist mode E/M.

This is certainly unexpected from Spec's 2024 documentation. The clear message then was that S depends only on E. Nary a hint that S might also vary with M.

So we'll have to map out this mysterious S(E,M) function empirically. Is it the same for all E/M modes, MicroTune and otherwise?

Unfortunately, much more field work ahead.

 

[mcdenny]

Spec's SL 2 documentation says that the mechanical Pm and power ratio readings on the TCU and app are just estimates obtained by multiplying measured electrical Pm by a constant assumed efficiency of 80%. Don't recall that they show raw electrical Pm data anywhere.

Just confirming that you backed out the assumed 80% before applying your empirical efficiency vs. cadence curve. If not, the gap between your empirical Pm/Pr ramp slope data and the corresponding Spec MicroTune data is even wider than seen in @mschwett's last plot.

Another reverse-engineering hurdle: The new empirical 40/40 and 40/100 Pm vs. Pr plots once again show that ramp slope S = Pm / Pr varies strongly with both E and M in assist mode E/M.

This is certainly unexpected from Spec's 2024 documentation. The clear message then was that S depends only on E. Nary a hint that S might also vary with M.

So we'll have to map out this mysterious S(E,M) function empirically. Is it the same for all E/M modes, MicroTune and otherwise?

Unfortunately, much more field work ahead.

The Spec app specifically defines Pm mechanical as one of the data items you can put on the display. In my experience the highest number I've seen while riding is a little over 300 watts.

The .fit file resulting from recording a ride has a field called "motor power (watts)" that maxes out at 380 - 400 watts. That most certainly is the electrical power into the motor. I used the efficiency function posted above to reduce the .fit file number to a Pm mechanical. All the Pm/Pr curves are Pm mechanical. I didn't back out a fixed 80%, I backed out a percentage based on the cadence value of the record. Most of the records had cadence values between 70 and 90 so the efficiency reduction was about 75%. Post #254 has the exact formula.

 

[Jeremy McCreary]

From Spec SL 2 documentation...

Note the last bullet point. The same document said the same thing about the TCU's motor power field. May or may not not apply to motor power data pulled from .FIT files.

 

[mcdenny]

From Spec SL 2 documentation...

View attachment 206733

Note the last bullet point. They said the same thing about the TCU's motor power field. May or may not not apply to motor power pulled from .FIT files.

 

[mcdenny]

Just went for a quick ride to be sure - selecting "Motor Power Mechanic" for the display. In 100/100 tune it's easy to keep motor power maxed out. I rode for about 1/2 mile lightly dragging the brakes to keep speed in check. Motor power mechanic was almost always above 300 watts with max I saw of 316 watts.




-

My 100/100 .fit file has 570 records, Here's a chart showing the frequency of each Pm data point

The most frequent Pm is 384 watts. 66% of the records were over 350 watts. If the .fit file showed Pm mechanical the peak would be around 300 watts. 384 * 0.80 = 307.

The bike display shows motor mechanical watts but the .fit file records motor electrical watts input.

 

[Jeremy McCreary]

I'm encouraged that your field data keeps showing a Pm(Pr) function with a linear ramp-flat structure with (a) the ramp and flat both straight lines, and (b) the flat close enough to horizontal for modeling purposes.

My original Pm(Pr) model from 2024 Spec documents had the same structure, which has the following advantage: If you know the ramp slope S and saturated motor power Pms for a given E/M assust mode, it's easy to calculate the saturation rider power Prs, where the ramp and flat meet.

The bad news: In the original model, I had good reason to believe that S depended only on E, and Pms only in M. But we now know that neither is true — at least on the SL 2. And it may well have been that way all along.

Clearly, S and Pms in the SL 2 both depend on E and M together. We now have a good idea of the S varies with E

The good news: We could still have a fairly simple Pm(Pr) model if we could pin down simple formulas for the functions S(E,M) and Prs(E,M).

 

[mschwett]

Just went for a quick ride to be sure - selecting "Motor Power Mechanic" for the display. In 100/100 tune it's easy to keep motor power maxed out. I rode for about 1/2 mile lightly dragging the brakes to keep speed in check. Motor power mechanic was almost always above 300 watts with max I saw of 316 watts.




- View attachment 206736

My 100/100 .fit file has 570 records, Here's a chart showing the frequency of each Pm data point

View attachment 206737

The most frequent Pm is 384 watts. 66% of the records were over 350 watts. If the .fit file showed Pm elect the peak would be around 300 watts. 384 * 0.80 = 307.

The bike display shows motor mechanical watts but the .fit file records motor electrical watts input.

would be interesting to conpare a few different cadences from both the display and the fit file to see if they’re really just always assuming 80%, or do they take into account their own published curve which is clearly much lower in places…

 

[mcdenny]

would be interesting to conpare a few different cadences from both the display and the fit file to see if they’re really just always assuming 80%, or do they take into account their own published curve which is clearly much lower in places…

A quick test showed Pm max on the display to be about 250 watts at about 40 cadence. Very hard to hold a steady low cadence. The Spec data shows 210 watts for 40 cadence and 255 watts for 50 cadence so the very short test is in the ball park with Spec's chart.