Climbing

Post Reply
travispape
Site Admin
Posts: 392
Joined: Fri Jan 11, 2008 9:52 am
Location: Durham, NC

Climbing

Post by travispape » Mon Jul 14, 2008 1:28 am

For any given ride, there is no single correct answer for how much climbing you did. Another way of saying that is that you should not expect two different cycling devices to agree on their climbing numbers for any given ride. Let me briefly explain:

Climbing is a sum of positive elevation gains during the ride, ignoring the small elevation gains and also ignoring elevation gains that happen over such long stretches of road that the grade is insignificant. The basic problem is that there is no standard definition of what elevation gain is too small or what road grade is too small to be included in the climbing totals. Gaining 200 ft of elevation in a mile is a climb. Only gaining 20 ft of elevation on a straight incline that is a mile long is not a climb. For most riders, the aerodynamic frictional losses are much greater than the work done against gravity for that mile. Now the question is is a 20 ft gain in only 200 ft of riding a climb? I think most people would want to include that effort in their climbing tally. The issue is that you have to draw the line somewhere, and there is no standards body out there that I know of defining what should and should not be included in the climbing sum.

Imagine for a moment that you have 2 different devices that make perfectly accurate elevation measurements. They would produce elevation profile that perfectly agree with each other; however, they would disagree on their reported climbing numbers because the software in each device will be using a different filter to remove the insignificant elevation gains from the climbing totals. The problem is actually worse on roads that are nearly flat or mildly rolling because are large portion of your ride is near those filter thresholds. One device might include lots of little elevation gains that the other device ignores. Over time, those differences add up to a big disagreement. If you are doing serious climbing up a mountain, both devices will have an easier time agreeing with each other; however, you will still have differences when the grade temporarily levels off.

Note that the exact same issue exists for geological survey data. Say that there is a perfect source elevation data for a ride that you do based on survey data accurate to a mm grid. Again, the climbing total calculated based on that perfect elevation data has to make the same judgment call about what elevation gains are too small or take too long to be included in the climbing total.

So even if you have two devices that have perfect elevation accuracy, they will disagree on their reported climbing numbers. The problem is magnified when you consider that neither altimeters nor GPS devices make perfect elevation measurements. They each have some noise in their signals that has to be filtered out. Yet another source of disagreement is that both types have some systematic errors. Altimeters suffer when atmospheric pressure changes and GPS-based systems are even less accurate on the vertical axis than altimeters.

The bottom line is is that you should not expect two different devices to agree on their climbing totals. If two do happen to be close to each other for a given ride, you should not expect them to repeat the agreement on a different ride with a different profile.

Travis

Charles
Posts: 3
Joined: Sat Feb 07, 2009 1:19 pm

Re: Climbing

Post by Charles » Sun Feb 08, 2009 7:45 am

Travis you made me a little uneasy. The reason is that I was convinced that the iBike uses atmospheric pressure as an altimeter, but does not use it to compute the slope. In stead it uses the vertical acceleration as an electronic level to measure slope. The vertical accererometer has a much greater precision than the athmospheric altimeter.
Small systematic errors in the slope measurements are corrected when doing a ride analysis. As I understand it these corrections are made by comparing the computed total elevation gains of both methods, accelerometer and altimeter.
However this auto analysis is also a source of error. Suppose a long ride during which the atmopheric pressure changes because of bad weather, not of altitude. The ride starts with good weather and high athmospheric pressure, and it ends a few hours later after a round trip at exactly the same place with bad weather and a lower athmospheric pressure. The altimeter accuses an overall elevation gain. Now this false elevation gain is used to correct for an offset of the accelerometer, which falsifies the data!
Please tell if I am mistaken
If not, here is a possible solution; Doing a set up at the beginning and at the end of the ride and setting the correct altitudes. The auto analysis software could then do the right job

User avatar
racerfern
Posts: 1356
Joined: Fri Jan 11, 2008 1:47 am
Location: Baldwin, NY
Contact:

Re: Climbing

Post by racerfern » Sun Feb 08, 2009 10:36 am

Suppose a long ride during which the atmopheric pressure changes because of bad weather, not of altitude. The ride starts with good weather and high athmospheric pressure, and it ends a few hours later after a round trip at exactly the same place with bad weather and a lower athmospheric pressure. The altimeter accuses an overall elevation gain. Now this false elevation gain is used to correct for an offset of the accelerometer, which falsifies the data!
This is exactly my situation on many rides. Living at the ocean, I see wild swings in atmospheric pressure. I'll start off at 20' above sealevel and at the end of a long ride returning to the same spot might have me returning at either -50' or +100'. Those are extremes; usually the ending elevation is within 40' or so.

A couple of times, I adjusted the altimeter a few times along the way to "adjust/compensate" for atmospheric changes. However, the change in wattage was so slight as to make it insignificant. After comparing my manually adjusted rides to some that I left alone, the difference over the 40+ mile course I do on Sundays was about 1.5w with the max difference on one ride being 3.2w. That's just not enough to fret over.

I also do intervals on a hill nearby(about 75' of climbing). Less than 2 minutes up, down slowly, then over and over. On short climbs like this especially at the end of the day there are amazing atmospheric and temperature changes that can make the interval humps gradually work their way up or down. However, the actual power produced on the climb portion is spot on. Climbing for each interval is the same. At the end of the downhill/coasting portion my altitude might not show up the same but the power portion that I'm interested in is good.
Attachments
Picture0003.png
Picture0003.png (7.87 KiB) Viewed 9428 times
Fernando

travispape
Site Admin
Posts: 392
Joined: Fri Jan 11, 2008 9:52 am
Location: Durham, NC

Re: Climbing

Post by travispape » Tue Nov 24, 2009 11:06 pm

Sorry for the long delay in responding. The forum-watching has been in other able hands for a while now, but I'm going to be more active on the forums again.
Charles wrote:However this auto analysis is also a source of error. Suppose a long ride during which the atmopheric pressure changes because of bad weather, not of altitude. The ride starts with good weather and high athmospheric pressure, and it ends a few hours later after a round trip at exactly the same place with bad weather and a lower athmospheric pressure. The altimeter accuses an overall elevation gain. Now this false elevation gain is used to correct for an offset of the accelerometer, which falsifies the data!
Only by a small amount compared to the large errors that were possible before we introduced this feature. Believe us that this feature is worthwhile and does much more good than harm.

You shouldn't see altimeter errors of more than a couple dozen feet based on weather changes during the ride; however, it is on our list to add a way of correcting for weather changes.

Travis

Zoltan
Posts: 190
Joined: Mon Jan 23, 2012 10:47 am
Location: HUNGARY

Re: Climbing

Post by Zoltan » Mon Jul 02, 2012 5:42 am

travispape wrote: Now the question is is a 20 ft gain in only 200 ft of riding a climb? I think most people would want to include that effort in their climbing tally. The issue is that you have to draw the line somewhere, and there is no standards body out there that I know of defining what should and should not be included in the climbing sum.
It is why I always add every centimeter to my total ascents. I never apply any minimum level of ascent of a stretch to add the amount of ascent of that stretch to the total ascents
If you cleverly smooth the barometric elevation data in order to eliminate the spikes then it will not distort to add small 'climbs' to the total. Just to be clear I am not speaking about iBike this time, but about some GPSs with barometers-

On the other hand when I segment my rides to flat, climbs, descents i apply a slope of +/-1% as a separator. I had google'd a lot and I analyzed lot of my rides before making this decision, and although I am not a governing body I think 1% is a fair value as a threshold. 1% slope is something which a cyclist can clearly identify in windless conditions.

Post Reply