Login required to started new threads

Login required to post replies

Prev Next
1st (and 2nd and 3rd) experiences with Notio Konnect
Quote | Reply
In Summer of 2017, myself and 4 athletes were one of the first to use the Notio device for aerodynamic testing. In February 2018, I had the opportunity to do a training course with Marc (marcag) in Spain and acquire one of the first available units of Notio Konect *

I am familiar with aerodynamic work (Retül Positioned for Speed ​​by Alphamantis, Manchester 2014) as well as work with Aerolab by Golden Cheetah and estimates via Strava and Best Bike Split, as well as field work with my athletes and with my own person (record Spain time 40-44 with 45.050)

From February to July I have been familiarizing myself with the system and learning to use the Golden Cheetah software and learning about my own methodological errors.

My Thoughts;

Easy to install, easy to use, but requires a training process
Quality of results will depend on meticulousness in the process (it is easy to want to measure something and without realizing introduce other modifications such as a drum not installed or a change of clothes)
Data Obtained easily on the road, without the need to go to a velodrome or wind tunnel
Simple analysis with laptop, but it would be appreciated if the data could be obtained easier in the circuit itself, via mobile app or via Garmin data.
It’s a tool that many things, one of the most important is the efficiency of to overcome the most important force on a flat and rolling terrain


* I have no commercial relationship with Notio or with Argon18, although between 2012 and 2016 Argon18 Spain provided my bicycles to compete. This is how I ended up being an early tester. I have a good personal relationship (I would not dare to call it friendship) with Marc
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [luarca] [ In reply to ]
Quote | Reply
Thank you for your observations

Is it possible to use like this?

Bike 1/2 mile at 200 watts. Watch the Cda in real time. Turn around bike the same

Change helmets repeat above. Know which helmet is more aero

Can that be done with accuracy?

Thank you
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [luarca] [ In reply to ]
Quote | Reply
luarca wrote:
Simple analysis with laptop.

Does Notio supply software for this, or are you using something else?

I'd love to know details on the testing protocol that is needed for good precision, and the spread of results you've seen. For instance if you test the exact same setup 10 times, what would the standard deviation in CdA typically be?

Thanks for posting!
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [RBR] [ In reply to ]
Quote | Reply
Not sure about the new version, but it should be done via app. First at all you will need to do a personal calibration set in a longer course (my course is 3km out and back). You don’t need to fix the power nor speed
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [luarca] [ In reply to ]
Quote | Reply
luarca wrote:
Not sure about the new version, but it should be done via app. First at all you will need to do a personal calibration set in a longer course (my course is 3km out and back). You don’t need to fix the power nor speed

Did you understand my question?

Does it give real time Cda?

Thank you
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [rruff] [ In reply to ]
Quote | Reply



Notio supplies the app and the firmware for the Garmin Unit.

The software is Golden Cheetah

CdA is quite stable in the exact same setup considering that CdA is inherently erratic as power and speed are. For instance, if you start to oscillate in a headwind you CdA will get worst if you go smoothly with tailwind (sorry about my “Yoda style English “ )

For instance, these are 3 runs for a pro triathlete (same position, out and back constant power, out and back progressive power, and out and back in a Shrug position” [inline "Captura de pantalla 2018-07-10 a las 23.30.34.png"]
Last edited by: luarca: Jul 10, 18 14:48
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [RBR] [ In reply to ]
Quote | Reply
Yes to both questions

RBR wrote:

Did you understand my question?

Does it give real time Cda?

Thank you
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [luarca] [ In reply to ]
Quote | Reply
luarca wrote:
Yes to both questions

RBR wrote:

Did you understand my question?

Does it give real time Cda?

Thank you

Wow

To me that is absolutely amazing

Thank you

Weird to me you can’t buy one unless you go to one of their trade shows/exhibitions
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [RBR] [ In reply to ]
Quote | Reply
RBR wrote:
luarca wrote:
Yes to both questions

RBR wrote:


Did you understand my question?

Does it give real time Cda?

Thank you


Wow

To me that is absolutely amazing

Thank you

Weird to me you can’t buy one unless you go to one of their trade shows/exhibitions

May I add that Jaime recently qualified for Kona at IM Lanzarote.
On top of being a great tester, an aero connaisseur, he is a super athlete.
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [RBR] [ In reply to ]
Quote | Reply
RBR wrote:
Weird to me you can’t buy one unless you go to one of their trade shows/exhibitions

I think that's pretty smart, and would have benefited the developers of some power meters, etc (*cough*garminvector*cough*). Get your earliest users to be people to be nerds who understand the technical issues being dealt with and who are able to communicate issues well. Sort of like the OP.

As opposed to just throwing it into the wild and having your product thrown under the bus on Facebook by people who don't appreciate the challenges involved in bringing such a thing to market.
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [luarca] [ In reply to ]
Quote | Reply
Thanks for the info.

What I'm curious about is how much variation you see if you do several out-back laps with the same configuration. So not comparing configurations, rather establishing how much variation is typical when you are trying to be as consistent as possible in your position, with "laps" performed back to back, and also on different days.

Also, why is the wind always higher in the "out" segment than the "back"? Wouldn't you expect them to approximately cancel?
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [rruff] [ In reply to ]
Quote | Reply
rruff wrote:
What I'm curious about is how much variation you see if you do several out-back laps with the same configuration. So not comparing configurations, rather establishing how much variation is typical when you are trying to be as consistent as possible in your position, with "laps" performed back to back, and also on different days.

There are two sources of variability in estimation: random, and systematic. Random variability is the thing undergraduates in stat classes usually learn about, and the thing they focus on in statistical inference. For this method, when you have a good power meter and good speed sensor, the random component of variability is quite small. Systematic variability (either through model misspecification or experimental error or poor quality sensors) can be large. Training in proper usage is intended to minimize systematic variability. So asking about "how much variation do you see on several tests" really is a question about "how well trained do you have to be to eliminate systematic errors?"
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [RChung] [ In reply to ]
Quote | Reply
You need a good powermeter, good speed sensor, good air speed sensor, a good air density measurement, a consistent estimation of Crr and friction that accounts for tire temperature changes. And last but not least, a really good measurement of elevation if you want "real time" or "just riding" data. If not, then you do laps or out-back where your starting and ending points coincide and software that takes advantage of that. Did I miss any of the important ones?

Assuming that you want good precision and do out-back or loops to minimize the importance of the slope measurement, you will still experience variation in measured CdA if any of the measurements I listed above errs in accounting for changing conditions. Even if position is identical, and the protocol is followed exactly. If you know what that variation is, then you have a best case expectation for precision.

But in Notio's case there is another issue, and that is the location of the sensor and its effect on airspeed readings. It is so close to the rider+bike that changes in wind direction and position will effect it, necessitating a calibration while testing. Assuming that wind speed and direction is the same for the "out" as the "back", for instance. And that can be a big error unless you do a lot of laps.

I have no interest in buying one of these unless the manufacturer has enough confidence in the device to prove its ability for field testing. Let's say raising the forearms 15 deg gives me 1% lower CdA than flat forearms. How much time will it take to determine that raised is my best choice? What if it's a 0.5% difference? In less than ideal conditions; wind, temperature changes, pressure changes, etc.
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [rruff] [ In reply to ]
Quote | Reply
rruff wrote:
You need a good powermeter, good speed sensor, good air speed sensor, a good air density measurement, a consistent estimation of Crr and friction that accounts for tire temperature changes. And last but not least, a really good measurement of elevation if you want "real time" or "just riding" data. If not, then you do laps or out-back where your starting and ending points coincide and software that takes advantage of that. Did I miss any of the important ones?
You don't need an a priori estimate of Crr. You can calculate it from the test. And you're still mixing up the random from the systematic sources of error.
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [RChung] [ In reply to ]
Quote | Reply
Is Notio calculating Crr from their test? Don't random and systematic errors both contribute to uncertainty in the results?
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [rruff] [ In reply to ]
Quote | Reply
rruff wrote:
Is Notio calculating Crr from their test? Don't random and systematic errors both contribute to uncertainty in the results?
I don't know what Notio does. I'll be interested to find out.

You were asking previously about variability across tests. Uncertainty is slightly different. For example, we often think of something like MSE when we talk about uncertainty, and we can think of MSE as being comprised of a bias component and a variance component. Some of the things you were talking about a post or two upthread affect bias, mostly, and variance not so much. Other things affect variance, mostly, and bias not so much. Unmeasured but systematic drift in a sensor can be better thought of as having a bias effect over time but not so much as increasing the variance. Proper protocols include doing things like either controlling for drift or at least measuring it if you can't control it so you can apply statistical controls post hoc. This is almost surely more than you wanted to know, and I may have screwed something up by trying to simplify the explanation.
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [rruff] [ In reply to ]
Quote | Reply
rruff wrote:
why is the wind always higher in the "out" segment than the "back"? Wouldn't you expect them to approximately cancel?

Crosswind situation, but only head/tail component displayed?
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [RChung] [ In reply to ]
Quote | Reply
RChung wrote:
This is almost surely more than you wanted to know.

Not at all. I understand bias, drift, and variance. Some error can be mitigated with a good protocol, and some with analysis. The bottom line is the degree of confidence you have in the result and how much time and effort it takes to get there.
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [Andrew Coggan] [ In reply to ]
Quote | Reply
Andrew Coggan wrote:
Crosswind situation, but only head/tail component displayed?

That could give you a small net positive wind, but this test shows a net negative. -1.6 km/hr.
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [rruff] [ In reply to ]
Quote | Reply
rruff wrote:
RChung wrote:
This is almost surely more than you wanted to know.


Not at all. I understand bias, drift, and variance. Some error can be mitigated with a good protocol, and some with analysis. The bottom line is the degree of confidence you have in the result and how much time and effort it takes to get there.
Exactly right. So, it turns out that VE is a way to visualize goodness-of-fit, especially over laps or loops, on a second-by-second basis. You may be more familiar with summary measures of goodness-of-fit, like R^2, or MSE, that condense the entirety of fit (or lack of it) into a single summary statistic. In my day job, creating an estimator is only part of the job; you also want a way to evaluate the estimate to see how robust it is in the presence of different types and magnitudes of error. So in this case, most people want a black box that spits out the CdA, and less often do they want a goodness-of-fit statistic to tell them how "good" the estimated CdA is. I think the VE profile is a much better way to evaluate fit than a single summary statistic. That's why I spent some time explaining how the VE profile responds to different kinds of error.
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [RChung] [ In reply to ]
Quote | Reply
RChung wrote:
rruff wrote:
RChung wrote:
This is almost surely more than you wanted to know.


Not at all. I understand bias, drift, and variance. Some error can be mitigated with a good protocol, and some with analysis. The bottom line is the degree of confidence you have in the result and how much time and effort it takes to get there.
Exactly right. So, it turns out that VE is a way to visualize goodness-of-fit, especially over laps or loops, on a second-by-second basis. You may be more familiar with summary measures of goodness-of-fit, like R^2, or MSE, that condense the entirety of fit (or lack of it) into a single summary statistic. In my day job, creating an estimator is only part of the job; you also want a way to evaluate the estimate to see how robust it is in the presence of different types and magnitudes of error. So in this case, most people want a black box that spits out the CdA, and less often do they want a goodness-of-fit statistic to tell them how "good" the estimated CdA is. I think the VE profile is a much better way to evaluate fit than a single summary statistic. That's why I spent some time explaining how the VE profile responds to different kinds of error.

Have you thought of how these types of sensors could be used to correct for errors in the VE plot? One passing car or bobble can mess up the profile and clipping that out by hand/spreadsheet is a pain. Could you modify the VE equation to accept the sensor data?

I guess it is possible that this is how the post analysis software works already, but I figured I would try doing concurrent VE and Aeropod estimation to pseudo-validate both methods (I am not very good at VE).
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [RChung] [ In reply to ]
Quote | Reply
I think James Webb (skippykitten) said he was primarily focused on designing a device to make VE testing better. With decent air speed and elevation data big improvements are possible. Getting a CdA number while "just riding" might be a nice novelty, but you can get much better data with a good protocol and analysis.

I think the software will be key. Make it easy to use so that anyone can figure it out, with comprehensive analysis and evaluation (confidence level). Put it in a phone app so you know what you have while you are still out testing. Even the Aeropod might be useful for testing with good software.
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [grumpier.mike] [ In reply to ]
Quote | Reply
grumpier.mike wrote:

Have you thought of how these types of sensors could be used to correct for errors in the VE plot? One passing car or bobble can mess up the profile and clipping that out by hand/spreadsheet is a pain. Could you modify the VE equation to accept the sensor data?

I guess it is possible that this is how the post analysis software works already, but I figured I would try doing concurrent VE and Aeropod estimation to pseudo-validate both methods (I am not very good at VE).


Oh yeah, I've thought about it. This is why I've been talking about systematic vs. random error for a long time now. We have different ways to deal with different types and sources of error, and it would be a mistake to treat systematic sources of error as if they were random. For example, most people use the classic regression method as if errors met the Gauss-Markov conditions; but systematic errors obviously don't. The standard way experimentalists try to avoid systematic error is to control the experiment more tightly, like testing on a dead calm zero wind day. But a second way is to measure and then do statistical controls post hoc. This is where air speed and yaw sensors come in, but other sensors do, too.

In the meantime, even if you have a wind speed and direction sensor, there may be other systematic sources of error that you will want to identify. VE appears to work in that context too. Here's something I've long wished GC's Aerolab would do: I wish you could overlay laps, or mirror reflect out-and-backs and then overlay them. That would make it easier to spot a passing car or a bobble. Then I wish it would make cutting-and-pasting to excise obvious passing cars or bobbles. Those are "non-random" (in a statistical sense) so you want to skip over them instead of treating them like a source of random error. That would be a lot handier than tossing the entire trial run. There are other things I think would be helpful, but I'm not a C++ programmer so I've never been able to do them myself.
Last edited by: RChung: Jul 11, 18 23:32
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [rruff] [ In reply to ]
Quote | Reply
rruff wrote:
I think James Webb (skippykitten) said he was primarily focused on designing a device to make VE testing better. With decent air speed and elevation data big improvements are possible. Getting a CdA number while "just riding" might be a nice novelty, but you can get much better data with a good protocol and analysis.

I think the software will be key. Make it easy to use so that anyone can figure it out, with comprehensive analysis and evaluation (confidence level). Put it in a phone app so you know what you have while you are still out testing. Even the Aeropod might be useful for testing with good software.

I think skippykitten has good insight.
Quote Reply
Re: 1st (and 2nd and 3rd) experiences with Notio Konnect [luarca] [ In reply to ]
Quote | Reply
Are you able to describe the step-by-step process you would use to test something simple like a change in helmet or aerobar angle or something like that?

Thanks!

Eric

Dimond Bikes Superfan
Quote Reply

Prev Next