Turbo Vado 2 4.0 or 5.0 or 6.0?

[Jeremy McCreary]

Perhaps add Pr to your ACE With the Cadence field as well!

Nope, the ACE is totally disconnected from the bike now and only provides GPS services on the bars. And it's very good at that.

That way I get the superior ride data and analysis I much prefer from the Specialized app running on my phone safe in my pocket. Everything else I normally need to see in the saddle, I get from 3 TCU pages.

 

[Stefan Mikes]

Nope, the ACE is totally disconnected from the bike now and only provides GPS services on the bars.

That way I get the superior ride data and analysis I much prefer from the Specialized app running on my phone safe in my pocket. Everything else I normally need to see in the saddle, I get from 3 TCU pages.

Which reminds me of Alfer "The Ice Warrior", a friend of mine who zooms frozen rivers and lakes, and uses his Wahoo for GPS navigation only but he records his rides with the Strava app on the smartphone

It is his own photo where the ice gave up. The name of the river is, ironically, Cold Water

 

[Jeremy McCreary]

I just think the data for the last data row is wrong (Specialized fault). It looks good for other data rows.

Please elaborate. There's no typo.

 

[Stefan Mikes]

Please elaborate. There's no typo.

Input rider's power 85 W, output motor power 300 W. It is just impossible for the SL 1.2 motor. Someone at Specialized had entered a wrong datum.

 

[mschwett]

it's really not that complicated !! how much does the motor help you, as a function of how hard you pedal... and what's the MOST it will ever help you.

as far as i know every decent torque-sensor system from the big manufacturers works the same :

in the mission control days, specialized called the "how much help do you get as a percentage of your power" parameter "support". top slider of left image. the "what's the most support you get, no matter how hard you hammer it," was "peak power"

for TQ, middle image, "assist" is the same concept as support, and max is the same as peak. for mahle, Pp is "peak power" and "Re" is the same as specialized's support and TQ's assist.

the only meaningful differences are :

1 whether they include an acceleration parameter - specialized does not for road/gravel/hybrid bikes, the others do. that's how quickly the motor ramps from no/low assist to high.
2 whether assist is a direct percentage of your pedaling power, as in the TQ example where it's 200% of your power, no other multipliers or factors. in order to keep the scale from 0 to 100 on all their bikes, specialized added the support factor, which is hidden, and ranges from 1.8 or 2.0 to 3.x plus depending on the bike. that's where the confusion comes in, because people set it to 100% thinking if they put in 100w, the motor also puts in 100w, when in fact it puts in 200w or 360w or whatever.

this is totally just a software thing, they could change it if they felt like it... and keep in mind the SL system was co-designed by specialized and mahle, the full power ones by specialized and brose, so really there isn't anything new here that specialized has done other than what they call the terms and set the scales.

 

[Jeremy McCreary]

the only meaningful differences are :

1 whether they include an acceleration parameter - specialized does not for road/gravel/hybrid bikes, the others do. that's how quickly the motor ramps from no/low assist to high.
2 whether assist is a direct percentage of your pedaling power, as in the TQ example where it's 200% of your power, no other multipliers or factors. in order to keep the scale from 0 to 100 on all their bikes, specialized added the support factor, which is hidden, and ranges from 1.8 or 2.0 to 3.x plus depending on the bike. that's where the confusion comes in, because people set it to 100% thinking if they put in 100w, the motor also puts in 100w, when in fact it puts in 200w or 360w or whatever

Interesting way to look at it. Personally, since I have a Pr (rider power) readout on my SL's top bar — and am getting to know what 100, 150, 200, 350, 300W, and so on feel like — I'd much prefer to see a straight percentage of Pr.

Now, if I wanted to add a TCU page, I could see Pr, Pm, and "power ratio" (Pm/Pr?) at once. Will test that tomorrow!

 

[mschwett]

Interesting way to look at it. Personally, since I have a Pr (rider power) readout on my SL's top bar — and am getting to know what 100, 150, 200, 350, 300W, and so on feel like — I'd much prefer to see a straight percentage of Pr.

Now, if I wanted to add a TCU page, I could see Pr, Pm, and "power ratio" (Pm/Pr?) at once Will test that tomorrow!

i probably would too... if i looked at motor power much during the ride!

in part to harmonize the experience between bikes i don't look at motor power much any more. cadence, rider power, pedal balance, heart rate, speed, ride time, that's really all i want and need. i have the displays set up exactly the same on e-bike and non e-bike. with the creo i did fool around a bit with adding some of the ebike fields but just did not find it meaningful while riding. cool to look at after and initially to understand for sure what the bike is doing, and maybe useful for estimating range?

i have my data screen set like this. there's another page with map and most of the same data, another for climbing showing gradients and so on, one with lap data, and one with GPS only info for rides on the commuter!

 

[Jeremy McCreary]

i probably would too... if i looked at motor power much during the ride!

in part to harmonize the experience between bikes i don't look at motor power much any more. cadence, rider power, pedal balance, heart rate, speed, ride time, that's really all i want and need.

I'd normally want to see much the same in the saddle (except pedal balance).

But now I have some detective work to do. Won't hurt to add a TCU page for now, but I generally want just 3 to flip through.

Spend 90% of my time on this TCU page. The 50/50 is the assist mode.

 

[Stefan Mikes]

Everything was easy and understandable until Specialized published its detailed tables

Jeremy, bear in mind the reported Pm or Pm/Pr would be probably based on the electrical motor power. The mechanical power for the SL 1.1 is 0.79 * electrical power. That is, the rated mechanical power of 240 W is achieved at the electrical power of 303 W (that was measured both by me and mschwett in the past).

 

[Stefan Mikes]

Guys,

I've decided to have a look at the Specialized tables with my own eye and just to do all the calculations in Excel.

Where,
Bmax is the rated maximum boost of 2.3
Pm max is the rated maximum mechanical motor power at cadence of 80 rpm, 300 W
E is the Ease %
M is the Maximum Motor Power %
Pr is the rider's power
B is the calculated Boost, Pm/Pr

The assumptions:
-- Maximum Boost of 2.3 may not be exceeded (this is for the last table row)
-- In the Dynamic mode, the Pm max must be achieved even if the E = 0.

The first data series is for the Dynamic Micro Tune (M=100%) and the second series is for the Static case (E=M). The results are... interesting

• There is a single performance curve despite Static or Dynamic
• The relationship is non-linear.

Thoughts?

 

[Jeremy McCreary]

Everything was easy and understandable until Specialized published its detailed tables

Or so it seemed. Sigh.

Like you, beginning to wonder about those official Prs tables for SL 2 MicroTune modes. The E-dependent fudge factor that corrects model Prs predictions for the dynamic M = 100 case DOES NOT correct predictions for the static M = E case.

Makes me doubt that we'll be able to use the official Prs data to refine the model.

Jeremy, bear in mind the reported Pm or Pm/Pr would be probably based on the electrical motor power.

Fortunately, that's exactly not the case. From official Mastermind documentation regarding the TCU "Power ratio" field:

The last bullet point says that the power ratio Q on the TCU is only estimated from electrical motor power by assuming a fixed electromechanical efficiency of 80% — very close to your 79%.

The mechanical motor power shown on the TCU is also estimated this way. So at least it's consistent with Q. In contrast, Pr is measured in the motor.

But we know that real motor efficiency falls off significantly below ~70 rpm. So as efficiency decreases with slowing cadence, Q will be increasingly overestimated.

Fortunately, nearly always pedal above 80 rpm, so the motor power and Q values I see on the TCU should be good enough for reverse engineering purposes. I'll just assume that Q ≈ Pm / Pr.

If the model's slready correct for my SL 1, the TCU will show pre-saturation Q values near

Q ≈ Pm / Pr = E' B ≈ 2 E',

where E' = E / 100, and B = 1.8 ≈ 2 for my bike. It will be easy enough to test Q first for linearity in E in my head: Just double E in MicroTune and see if Q doubles, too.

If not, the pre-saturation model will need a correction factor F nonlinear in E, as explained in my last few posts. Might even be able to use Q and Pr to estimate F.

BTW, in all my posts on the Spec mid-drive PAS, Pm has always been mechanical motor power at the crank. Fully appreciate the distinction.

 

[Jeremy McCreary]

Guys,

I've decided to have a look at the Specialized tables with my own eye and just to do all the calculations in Excel.
View attachment 205199
Where,
Bmax is the rated maximum boost of 2.3
Pm max is the rated maximum mechanical motor power at cadence of 80 rpm, 300 W
E is the Ease %
M is the Maximum Motor Power %
Pr is the rider's power
B is the calculated Boost, Pm/Pr

The assumptions:
-- Maximum Boost of 2.3 may not be exceeded (this is for the last table row)
-- In the Dynamic mode, the Pm max must be achieved even if the E = 0.

The first data series is for the Dynamic Micro Tune (M=100%) and the second series is for the Static case (E=M). The results are... interesting

View attachment 205200

• There is a single performance curve despite Static or Dynamic
• The relationship is non-linear.

Thoughts?

Good work. I'll have to mull this over.

 

[mcdenny]

Guys,

I've decided to have a look at the Specialized tables with my own eye and just to do all the calculations in Excel.
View attachment 205199
Where,
Bmax is the rated maximum boost of 2.3
Pm max is the rated maximum mechanical motor power at cadence of 80 rpm, 300 W
E is the Ease %
M is the Maximum Motor Power %
Pr is the rider's power
B is the calculated Boost, Pm/Pr

The assumptions:
-- Maximum Boost of 2.3 may not be exceeded (this is for the last table row)
-- In the Dynamic mode, the Pm max must be achieved even if the E = 0.

The first data series is for the Dynamic Micro Tune (M=100%) and the second series is for the Static case (E=M). The results are... interesting

View attachment 205200

• There is a single performance curve despite Static or Dynamic
• The relationship is non-linear.

Thoughts?

The only difference between static (E+M) and dynamic (M=100). is motor power is not artificially capped in the static case so the curves are the same. I bet the slight differences are due to rounding in your calculations.

You are assuming row 10 in the spec data table is an error. We just need someone with an SL 2 to go for a ride with the display set to show both rider and motor power, pedal at 80 cadence, pick a gear to give 85 watts of rider input and see what the motor power says. I plan to do this as soon as I get my new bike and the weather warms up a bit.

 

[Stefan Mikes]

My surprise was the curve was non-linear. We always assumed the assistance response would be linear.

I also set the ECO in my Vado 6.0 (full power e-bike) to 100/20%. Bmax = 3.2, Pm max = 520 W. One might assume I would always get 104 W assistance despite my very low leg power. The outcome was most disappointing: the ride was slow and sluggish. I instantly stopped and restored the 40/40% assistance, which I usually set for the ECO on that e-bike.

 

[Jeremy McCreary]

We have to be careful here. @Stefan Mikes and I are using "boost" in significantly different ways. For now, we'll stick with his boost, which is just the average slope S of the pre-saturation ramp in the Pm vs. Pr plot. Whether or not the ramp's a straight line.

Spec led us to believe that the ramp IS a straight line of constant slope S = M' Px / Prs, where Px is max motor power, and M' = M / 100. That's equivalent to saying that Pm is proportional to Pr along the ramp, or that Pm is linear in Pr there.

And we still have no reason to reject that linearity. Nor do we have any reason to reject Stefan's max boost as a scaling constant independent of E, M, and Pr.

What's new here is evidence, from official SL 2 MicroTune data, that S is somehow nonlinear in E. Even though Pm in all likelihood really IS linear in Pr for any given E. The hidden nonlinear fudge factor F(E) I've been talking about is a straightforward way to accomplish that.

May be a little hard to wrap your head around, but no necessary contradiction here. Mathematically, in a 3D plot of Pm as a function of both Pr and E, both things can be true at once.

Made just such a 3D plot on Desmos a year or so ago but never did much with it. Now I'll update it, see what can be learned from it, and share when it's ready for prime time.

 

[mschwett]

I went back and read this https://support.specialized.com/kb/sp_en-us/content/ECV3-6142/03-Turbo-Vado-SL-2-Turbo-SL-System which has lots of cool info.

it’s clear that the previously described assist ratio (the 2x you etc) was either an approximation of a range of values… or they changed how it works since then.

It is obviously nonlinear now, perhaps for electromechanical reasons or perhaps just because it feels better that way. In the event they didn’t change it, and it was always like this, it might explain why I never got quite the precisely predicted 2x values on my creo 1 when graphing rider power and motor power at various assist settings.

 

[mcdenny]

We have to be careful here. @Stefan Mikes and I are using "boost" in significantly different ways. For now, we'll stick with his boost, which is just the average slope S of the pre-saturation ramp in the Pm vs. Pr plot. Whether or not the ramp's a straight line.

Spec led us to believe that the ramp IS a straight line of constant slope S = M' Px / Prs, where Px is max motor power, and M' = M / 100. That's equivalent to saying that Pm is proportional to Pr along the ramp, or that Pm is linear in Pr there.

And we still have no reason to reject that linearity. Nor do we have any reason to reject Stefan's max boost as a scaling constant independent of E, M, and Pr.

What's new here is evidence, from official SL 2 MicroTune data, that S is somehow nonlinear in E. Even though Pm in all likelihood really IS linear in Pr for any given E. The hidden nonlinear fudge factor F(E) I've been talking is a straightforward way to accomplish that.

May be a little hard to wrap your head around, but no necessary contradiction here. Mathematically, in a 3D plot of Pm as a function of both Pr and E, both things can be true at once.

Made just such a 3D plot on Desmos a year or so ago but never did much with it. Now I'll update it, see what can be learned from it, and share when it's ready for prime time.

Clearly Px is a function of E and Pr. Pedal harder, get more motor power. Increase the ease setting, get more motor power. The Spech data table says nothing about how Px varies with Pr FOR A GIVEN E except we know Px is 0 when Pr is 0 and the table gives Pr and Px at saturation for 10 different E values. Time for gathering some real world data.

 

[Jeremy McCreary]

Clearly Px is a function of E and Pr.

I should have been clearer. My Px is the max motor power available at M = 100. To my knowledge, this Px varies only with cadence and choice of motor. It's supposedly independent of E, M, and Pr.

In the official SL 2 Prs tables in question, Px = 300W at 80 rpm.

Time for gathering some real world data.

Amen to that. When I get a chance, my 1st test will be to use the TCU to see if Pm really is linear in Pr on the SL 1 at any given E. If so, the TCU's power ratio should be constant when E and cadence are constant.

If you do the same on your SL 2, we'll gain some valuable info. For example, if the SL 1 is linear in this way, but the SL 2 is not, we'll know that the SL 2 brought some big changes to the Spec mid-drive PAS.

If Pm is linear in Pr on both SLs, then we'll know to focus on correcting the ramp's slope S.

 

[Jeremy McCreary]

I went back and read this https://support.specialized.com/kb/sp_en-us/content/ECV3-6142/03-Turbo-Vado-SL-2-Turbo-SL-System which has lots of cool info.

it’s clear that the previously described assist ratio (the 2x you etc) was either an approximation of a range of values… or they changed how it works since then.

Distinct possibilities now.

It is obviously nonlinear now, perhaps for electromechanical reasons or perhaps just because it feels better that way. In the event they didn’t change it, and it was always like this, it might explain why I never got quite the precisely predicted 2x values on my creo 1 when graphing rider power and motor power at various assist settings.

Can we see those graphs, pretty please? Might be some good clues there.

 

[mschwett]

Distinct possibilities now.


Can we see those graphs, pretty please? Might be some good clues there.

here’s the one that’s most handy, it’s actually closer than i thought ! i should find the xls. the gold is the ratio between rider power and motor power, at four different assist levels with max set to 100% to avoid hitting the limiter, so to speak. at the time we believed that it was always supposed to be 2x the support factor, and that the measured value was electrical power, and the motor was 80% efficient at these optimal cadences. you can also see how spiky the data is (the thinner gold line) which is pretty typical of any power measurements on a bike, no matter how smoothly and consistently you try and pedal. spec’s algorithms probably have all kinds of smoothing.

if you track the red and blue (rider and motor power) and unsmoothed gold, there are often spikes in the support factor when rider power increases, and then sometimes it drops back down. might be related to the responsiveness of their algorithms and controllers scaling up or down.