Login required to started new threads

Login required to post replies

Prev Next
Re: Aero sensors for dummies thread [MTM] [ In reply to ]
Quote | Reply
Hello, I am one of the co-founders of Velosense, and I saw the posts here and thought I would jump in to answer some questions.

Do we have a consumer cycling device on the market:

Not yet! We are working very hard on this and we hope to provide more news later this year.
On the cycling probe, the Ventos, we are working closely with a few select partners at this time. This allows them to improve their aerodynamic performance while providing valuable testing in preparation for the consumer product release.

Why would someone run a probe in the wind tunnel:

The Velosense cycling probe (https://www.velosense.com/sensors-cycling) has software which can automatically calibrate itself during testing by measuring the effect of the presence of the bicycle and the rider. However, if a bicycle probe customer is also wind tunnel testing, it is a useful correlation exercise to verify this calibration in the wind tunnel. Just as it is extremely useful to verify wind tunnel results with actual road testing.

To be very clear, we plan to launch a consumer aero device in the near future, and users will be able to achieve very accurate and repeatable measurements without a wind tunnel. Our software ensures that the automatic road calibration is precise and robust in a variety of conditions from a flat road on a calm day, to an undulating stretch on a gusty day. So including our device in a wind tunnel test like the one pictured above provides useful correlation for us as well as ensuring trust in our measurements with our partners.

Why is yaw angle important:

The difference in drag between 0 degrees and high yaw (+10 degrees) can be as much as 5 to 10% in drag. The aerodynamic drag from most advanced TT Frames and Wheelsets will be within a few percent at 0 degrees wind angle, but at higher wind yaw angles, the drag differences can be as large as 20% between wheels.

Some rider positions, especially hand height are quite sensitive to yaw angle. With some configurations better in low angles, and others in higher yaw angles.

Trying to analyse data without knowing the yaw angle can become difficult, if not impossible.

If the yaw angle can be measured accurately, the effect on drag can be normalised and allow the CdA values to be presented in a simple and clear manner, whilst also allowing more data intensive users to examine more specific differences.

If there are further questions about our upcoming product or how our technology works, we will try and answer them.

John Buckley
https://streamlines.aero
Quote Reply
Re: Aero sensors for dummies thread [John Buckley] [ In reply to ]
Quote | Reply
Hi John,

Welcome!

Your web site talks about high precision altitude. Can you talk a little on how you do that and if you require/suggest any specific test protocols ?
Quote Reply
Re: Aero sensors for dummies thread [rruff] [ In reply to ]
Quote | Reply
rruff wrote:
oh_the_run wrote:
I saw this tweet of Roglic in the wind tunnel and it looks like an aero sensor on the front above the front tire - https://twitter.com/.../1382658115070476288


Quite a bit more reach and stack, too:




Not really that much more reach when you look at at where he sits on the saddle on both bikes. It is a little mor reach but the looks are bit deceiving ;-)

Jeroen

Owner at TRIPRO, The Netherlands
Quote Reply
Re: Aero sensors for dummies thread [marcag] [ In reply to ]
Quote | Reply
Hello Marcag,

In regards to test protocols or use cases, we can generalise this into three cases:

1) Rolling average CdA over a set time, such as 30s or 45s.
2) CdA for variable length sections. Sections can be started and stopped in three different ways:
a. Predefined GPS gates (like strava segments).
b. Start triggered when exceeding threshold speed and finishing when speed drops below threshold or when braking event is detected.
c. Triggering by a user button (which can be used in conjunction with a or b)
3) Post analysis of a ride in which case 2b can be applied to the data

Use case 1, time based moving average, is the simplest and easiest to use.

For the highest accuracy, we recommend out and back testing along a straight route using GPS sections (case 2a) with a test length of 1km.

We rely primarily on barometric pressure measurement for altitude measurement. There are several small highly accurate barometric sensors available which have an off-the-shelf accuracy of around 10 to 20cm. However, we have put a lot of work acheiving this resolution on a moving bicycle. This is suitable in calm conditions.

In windier conditions, the barometric pressure undulates, almost similar to waves in the ocean. This makes the barometric elevation measurement more difficult. To solve this problem we needed to develop some proprietry technology to re-establish the sub 20cm altitude accuracy.

John Buckley
https://streamlines.aero
Quote Reply
Re: Aero sensors for dummies thread [John Buckley] [ In reply to ]
Quote | Reply
Thanks for the detailed response.

What are the various components to be installed on the bike ? We see what seems to be the "wind sensor", then there is mention of a component on the stem ?
Are there any others ?

Which of these components does altitude ?
Quote Reply
Re: Aero sensors for dummies thread [marcag] [ In reply to ]
Quote | Reply
We have two products, the Ventos wind sensor, and the Zenith body position sensor. Each product can be used separately, but they work very well together.

The Ventos contains sensors to measure wind speed, direction, and elevation (from dynamic pressure, yaw angle and barometric pressure). It also calculates the CdA in real time and transmits this information to the head unit for display and logging.

To measure CdA on a bicycle we require a power meter and a speed sensor. We have a strong preference for speed sensors with a magnetic pickup, but if testing on long stretches of road, other speed measurement such as GPS or compass based sensors can work.

The Ventos is currently supported on Garmin head units via a ConnectIQ app and on the Hammerhead Karoo and Karoo 2 via a sideloaded android app. The Ventos connects directly to the bicycle power meter and speed sensor via pass pairing from the head unit. CdA calculations are performed within the embedded code on the Ventos to allow real time CdA calculation. It also logs dynamic pressure, wind yaw angle, barometric pressure, and air temperature which are then recorded into a FIT file. We also log the energy split between aerodynamic, gravitational, rolling resistance, and acceleration. For higher data rates than 1s FIT files, there is onboard file storage, and the Android app can store higher rate data via other file formats.

The Zenith body position sensor records a rider's head and chest position, and can also determine if a rider is standing or in the saddle. This information is again transmitted to the head unit.

The Zenith body sensor arose to solve a problem we encountered in our early testing - we were getting inconsistent aerodynamic results as riders were not maintaining a consistent body position. Even very experienced riders often changed their position as they fatigued or under situations of heavy exertion. To test the repeatability (precision) of the Ventos in real world conditions, we needed something to keep the riders in the same position. However, when we began using the body position sensor to conduct our own internal testing, our test riders realised that this could be a very useful item not just for aero testing, but also training and racing. We decided to create the Zenith as a standalone sensor which measures the riders head and chest position. This information is not only extremely useful when looking at aero data, but also when a rider is training to 'learn' a new aero position.

The Zenith is supported on Garmin head units again via a ConnectIQ app as well as the Karoo via a sideload. On Garmin head units there is also a ConnectIQ Data Field so head and chest position can be combined with a riders customised screen.

John Buckley
https://streamlines.aero
Last edited by: John Buckley: Apr 18, 21 6:28
Quote Reply
Re: Aero sensors for dummies thread [John Buckley] [ In reply to ]
Quote | Reply
one more and I'll let you enjoy your Sunday :-)

In a previous post, you said "Our software ensures that the automatic road calibration is precise and robust"

Is there a calibration process like an out and back ?
If there is an auto-calibration, what is the trigger for it (for example a PM calibrates on coasting...)
Quote Reply
Re: Aero sensors for dummies thread [marcag] [ In reply to ]
Quote | Reply
They are good questions so keep them coming...

To be precise here when we refer to calibration we are referring to the adjustment of pressure data due to the presence of the rider and bicycle, I should have probably said aero calibration or possibly installation calibration, rather than road calibration. Our pressure sensors themselves do not need calibration.

Our recommended practice, for the most accurate testing, is an out and back run. So we will focus on our calibration method for that.

The device comes with a calibration out of the box which is roughly suitable for most time trial bikes. On each out and back run, we calculate the specific calibration for that run. So the real time CdA data the rider sees during that run will reflect the last good calibration, but in the post analysis, a more precise calibration can be applied. So the exact CdA for the previous run can be seen just after the run is completed.

The only action the rider needs to take is whether to update their calibration from the last run, and this will only affect the realtime display. We internally handle the data in an uncalibrated format, so that if the particular calibration is wrong during testing, it can be corrected aftwerwards without causing any errors to the data.

It is worth pointing out that once the sensor is installed in position on a bike, the calibration for that rider and bike usually does not change markedly. A lot of aero changes a rider makes will not result in a significant change to the calibration, and those that do significantly change the calibration, will usually have a much larger effect on CdA than the effect on calibration. Changes to the aero extensions which result in a significant reposition of the sensor would be an exception, but the post analysis data will be correct.

In other words, the calibration is most important for determining the absolute value for CdA rather than relative differences. If the realtime CdA shows a drag decrease or increase, it will usually be correct, however the exact magnitude may sometimes change in the post-processing.

I hope this makes sense.

John Buckley
https://streamlines.aero
Quote Reply
Re: Aero sensors for dummies thread [John Buckley] [ In reply to ]
Quote | Reply
  
Makes complete sense. I appreciate all the answers.

I think your transparency on how it works is greatly appreciated. I think more vendors should have technical people involved in these threads to answer questions.

Your explanation of calibration makes a lot of sense. People, in the past have made a big deal (unjustifiably IMO) on how sensitive these devices are to position around the cockpit. The challenges can easily be handled during post ride analysis and there are clever ways to not require out and backs.

I must say the mount of your device is intriguing. A buddy of mine, a fellow Canuck, said it looked like the Canadian space arm.

I also believe vendors should be publishing more data, specs, white papers etc. It would be great if vendors backed up their claims with real data.


btw, just for disclosure in case you don't know, I worked (past tense) for an aerometer company so am aware of some of the challenges.
Last edited by: marcag: Apr 18, 21 6:37
Quote Reply
Re: Aero sensors for dummies thread [marcag] [ In reply to ]
Quote | Reply
Happy to answer the questions, I hadn't put the name together and your previous employment, but I now remember seeing your posts on aero sensors in the past.

The mount in the photo is one which is well suited for quick swapping of bar extensions and configurations.

We have other options, and I find our diagonal single bar mount shown on our website a bit more aesthetic. The 'elbow' between the diagonal and vertical parts of the arm is a go-pro type adapter which can allow a variety of other mounting configurations.

John Buckley
https://streamlines.aero
Quote Reply
Re: Aero sensors for dummies thread [John Buckley] [ In reply to ]
Quote | Reply
John,

I concur with Marc. Very informative posts.
Quote Reply
Re: Aero sensors for dummies thread [RChung] [ In reply to ]
Quote | Reply
Hello Robert,

You and others in the ST community, who have been the pioneers of road aero testing for bicycles, have been a great resource for us. Thank you for all your posts.

John Buckley
https://streamlines.aero
Last edited by: John Buckley: Apr 20, 21 15:45
Quote Reply
Re: Aero sensors for dummies thread [RChung] [ In reply to ]
Quote | Reply
Quote Reply
Re: Aero sensors for dummies thread [marcag] [ In reply to ]
Quote | Reply
I'd rather have a standardized communication than a free for all... it's a (small) step in some of these sensors path to putting *useful* information in the hands of some of us closer to the neanderthal spectrum :D

My Blog - http://leegoocrap.blogspot.com
Quote Reply
Re: Aero sensors for dummies thread [marcag] [ In reply to ]
Quote | Reply
marcag wrote:
For those following along

https://www.dcrainmaker.com/...ng-for-standard.html

You had an uncredited appearance that day -- we had to 3-D print a bracket for the Konect.
Quote Reply
Re: Aero sensors for dummies thread [RChung] [ In reply to ]
Quote | Reply
You're right...I forgot about that. My stupidity in packing 18 million things for that day before flying across an ocean...except the darn thing I needed to attach the aero sensor to the bike, which, I left attached to my bike at home. Doh!

That was an amazing save!


-
My tiny little slice of the internets: dcrainmaker.com
Quote Reply
Re: Aero sensors for dummies thread [dcrainmaker] [ In reply to ]
Quote | Reply
dcrainmaker wrote:
You're right...I forgot about that. My stupidity in packing 18 million things for that day before flying across an ocean...except the darn thing I needed to attach the aero sensor to the bike, which, I left attached to my bike at home. Doh!

That was an amazing save!

One of the highlights of my career.

You know you've made it (in what my wife refers to as nerdland) when you get a call from DCRainmaker :-)

I really hope you and the crew come back around to testing these devices.

Between all the current "contenders" and the newbies ( Aerotech, Velosense, Gibli and more), someone needs to set the current state of affairs.
Quote Reply
Re: Aero sensors for dummies thread [ In reply to ]
Quote | Reply
Marcag was a guest on the Endurance Innovation podcast, talking about aerosensors. Very nice insights.

https://podcasts.apple.com/...6491?i=1000522462057
Quote Reply
Re: Aero sensors for dummies thread [John Buckley] [ In reply to ]
Quote | Reply
John Buckley wrote:
Our recommended practice, for the most accurate testing, is an out and back run. So we will focus on our calibration method for that.

IME this is woefully inadequate. If wind is very light and consistent, it can work... but that is never the case when I'm testing. A single out-back is not nearly enough to assume that wind has canceled.

I use the entire test session which is ~30-40 runs on a short out-back. And even that induces more error than I'd like.
Quote Reply
Re: Aero sensors for dummies thread [rruff] [ In reply to ]
Quote | Reply
As Marc alluded to in that podcast, the goal is to be able to trade-off strict testing protocols for very sensitive and accurate data sensors. The better the sensor, the less restrictive your protocols can be. That said, VE was itself a way to introduce a fairly sensitive diagnostic into aero testing -- that is, we've long known how to do estimation; what we've been missing was a way to assess goodness-of-fit for those estimates. (I can talk statistical estimation for a while -- don't make me do that).

So, the goal is get sensors that provide accurate and precise data, and then use those data in ways that will be robust enough so even poor experimentalists (like me) can get good results. In the meantime, do lots of laps and analyze the data in a smart way.
Quote Reply
Re: Aero sensors for dummies thread [rruff] [ In reply to ]
Quote | Reply
rruff wrote:
John Buckley wrote:
Our recommended practice, for the most accurate testing, is an out and back run. So we will focus on our calibration method for that.


IME this is woefully inadequate. If wind is very light and consistent, it can work... but that is never the case when I'm testing. A single out-back is not nearly enough to assume that wind has canceled.

I use the entire test session which is ~30-40 runs on a short out-back. And even that induces more error than I'd like.

I agree that a single out and back is inadequate. If a person wants to convince themselves, do the test 3 times in a row and they will probably get 3 pretty different numbers.

That said, I think John went on to describe some of the things they do and I suspect they do much more than a single out and back.

In the podcast I mention the use of a drone for testing. Calibration was probably one of the biggest challenges we tried to solve with the drone. Not because drones are better than cycling but because we can fly drones in the winter and we can't ride or travel to somewhere we can ride.

Canadian winters suck!
Quote Reply
Re: Aero sensors for dummies thread [RChung] [ In reply to ]
Quote | Reply
RChung wrote:
Marcag was a guest on the Endurance Innovation podcast, talking about aerosensors. Very nice insights.

https://podcasts.apple.com/...6491?i=1000522462057

Agreed. Great pod!

http://bikeblather.blogspot.com/
Quote Reply
Re: Aero sensors for dummies thread [rruff] [ In reply to ]
Quote | Reply
rruff wrote:
John Buckley wrote:
Our recommended practice, for the most accurate testing, is an out and back run. So we will focus on our calibration method for that.


IME this is woefully inadequate. If wind is very light and consistent, it can work... but that is never the case when I'm testing. A single out-back is not nearly enough to assume that wind has canceled.

I use the entire test session which is ~30-40 runs on a short out-back. And even that induces more error than I'd like.

For us mortals, one thing is how much we can suffer at race pace to test. Some things are no good at fart around pace.

I follow the cLEAN SPS method of when you test a phenomenon and it causes a difference, you cycle the change on and off again. So, base, change, base.

Example, helmets.....
base
helmet 1
base
helmet 2
base
helmet 1
base
helmet 2
eliminate worst, 2nd worst becomes base
base
helmet 1
base
helmet 1

NOT......base, helmet 1, helmet 2, done. But also not 30 runs.

If a helmet beats base (my road helmet) both times, make the "least" gain helmet the new base. Repeat until you've only got two items testing as "light switch" test. Keep the best performer.

If you can turn a light switch on and off repeatedly and you get the same "lights on, lights off" response each cycle.........I call it good (which VE line wins out on an overlay). Has to win twice.

I know this isn't the super best way at all, but it's what I've got time for with kids and a job and fitting in training.

This is how I used to test golf clubs. I had to use binoculars at a range, couldn't afford renting a flight monitor. But you can see how high the ball goes and see how it lands spin wise in addition to the distance.

If you can do that with something that involves randomized "skill" to hit the ball the same, if all you care about is finding which is "on and off" and not the actual value, not terrible.
Quote Reply
Re: Aero sensors for dummies thread [burnthesheep] [ In reply to ]
Quote | Reply
burnthesheep wrote:
For us mortals, one thing is how much we can suffer at race pace to test. Some things are no good at fart around pace.

I follow the cLEAN SPS method of when you test a phenomenon and it causes a difference, you cycle the change on and off again. So, base, change, base.
That's ... admirable, I guess. That's certainly one way to gain confidence in your measurements. It's not particularly time efficient, however.
Quote Reply
Re: Aero sensors for dummies thread [burnthesheep] [ In reply to ]
Quote | Reply
burnthesheep wrote:
For us mortals, one thing is how much we can suffer at race pace to test. Some things are no good at fart around pace.

That's why you need a short out-back with hills at both ends!
Quote Reply

Prev Next