Adjusting slope question

[Derath]

I was just curious, as I am a major newbie still with my ibike. I have noticed that on some rides when I play around with the tilt calc everything looks good. But other rides, Ones that don't end where they began, like a one way trip, that the tilt then seems messed up. Am I correct in the assumption that tilt adjustment only really works if you have a ride which starts and ends at the same point? Otherwise I don't see how the software could know if you ultimately ended at a lower or higher elevation.

-D

[iodaniell]

-D,

Change in weather will make the barometric pressures not line up on an out and back. It doesn't have anything to do with tilt, which uses a different device for tracking slope of the road.

[Derath]

-D,

Change in weather will make the barometric pressures not line up on an out and back. It doesn't have anything to do with tilt, which uses a different device for tracking slope of the road.

I understand in the case of an out and back. But I am talking in the case of an out, without the back.

-D

[racerfern]

I do outs (ride to work) completely separate from backs (riding home). As I see it, in the tilt and power advanced screen you have Elevation (blue) as determined by the barometric sensor. As an aside to this, I think you should be able to click a point on the line and tell the software the actual elevation assuming you know what it is.

Then you have the red slope line. This is the graph drawn by the math done on your tilt reading, speed and distance. I know I'm not 100% right on this, but I'm kind of in the ball park.

The green line is the difference. The goal is to get the slope to match the actual elevation. But that is assuming your elevation determined by your barometric sensor is correct. And that is the grey area we are all chasing. The error could be in the tilt/acceleration with the wrong tilt or it could be in the perceived altitude because of barometric changes.

The two attachments show my typical ride to work and too many hours later my ride home.

[racerfern]

BTW, in the example above, my starting elevation is 28' and finishing is 304'. So in the trip to work the correct ending point should be directly in between the red and the blue lines for the ride to work and just about in between the two for the ride home. In reality the post analysis results are off because the software assumes it needs to move the slope line to the elevation line. That is not necessarily the case.

[iodaniell]

BTW, in the example above, my starting elevation is 28' and finishing is 304'. So in the trip to work the correct ending point should be directly in between the red and the blue lines for the ride to work and just about in between the two for the ride home. In reality the post analysis results are off because the software assumes it needs to move the slope line to the elevation line. That is not necessarily the case.

Hmmmm? I don't understand what you are inferring.

When I remember to set my elevation before my commute to work--from USGS, I'm at ~+85ft above sea level--my ending elevation at work is ~+450ft. Since the barometric pressure will change before I leave, I can do three things:

- ignore/forget re-setting elevation back to my ending elevation from the ride to work
- set the elevation to my ending elevation from the ride to work
- post-process the elevation once I get home using iBike2

I usually post-process the data.

Either way, the blue elevation line should be within tens of feet from beginning, not somewhere between the blue and red line. At least that is my experience.

[racerfern]

What I'm saying is that the blue line has not kept up with the actual elevation changes. On the trip to work it falls short and on the trip home it also falls short. After slope analysis the ending point of the combined red/blue lines is really 1/2 way between the two lines in the pre-analysis graph. IOW, the slope line correction assumes the ending elevation is correct. That is not always the case.

This is where being able to input an ending elevation or a known elevation at some point on the trip would be helpful. And also why on known regular trips we get varying climbing totals even though it's the same ride over and over. Subsequently we can get varying wattage readings on rides that are nearly identical because the software is "chasing" an end point that is not pinned down.

[kb1dqh]

HI,

Quick question on your green line. Isn't that a lot of variance in the before analysis? Seems like your power would be thrown off?

Thanks!

[rruff]

What I'm saying is that the blue line has not kept up with the actual elevation changes. On the trip to work it falls short and on the trip home it also falls short.

Actual elevation change is 276ft. On the way to work it says ~250ft so it is 26ft low... corresponding to an increase in barometric pressure. On the way back it says ~280ft so it is real close to reality. Yes, some way to peg the altitude would be nice, but a few feet of error isn't going to effect your power numbers significantly.

[rruff]

Quick question on your green line. Isn't that a lot of variance in the before analysis? Seems like your power would be thrown off?

No... You have to consider the scale. 100ft/ 18mi/ 5280ft/mi = .001... so his tilt is only off ~.1%... and it is +- depending on whether he is riding uphill or down.

[racerfern]

Actual elevation change is 276ft. On the way to work it says ~250ft so it is 26ft low... corresponding to an increase in barometric pressure. On the way back it says ~280ft so it is real close to reality. Yes, some way to peg the altitude would be nice, but a few feet of error isn't going to effect your power numbers significantly.

Yeah but the overall climbing on the way to work is ~650'. At times for the trip to work the climbing total displayed is as low as 510 or as high as 780 depending on the barometric changes. So what is getting distorted either to the low or high side are the camel humps along the way along with the ending number.

The same swings happen for the trip home which has a climbing total of ~450'.

[alienator]

Yeah but the overall climbing on the way to work is ~650'. At times for the trip to work the climbing total displayed is as low as 510 or as high as 780 depending on the barometric changes. So what is getting distorted either to the low or high side are the camel humps along the way along with the ending number.

The same swings happen for the trip home which has a climbing total of ~450'.

But the tilt error is still very small, on the order of 0.15% or so. That's a pretty good calibration in most circumstances. The error in power caused by a 0.15% error in tilt is pretty small. Since there's already error in any definition of altitude at your starting and endpoints, be they GPS or topo derived, it becomes more likely that 0.15% error is even less significant.

[kb1dqh]

So, in this case, would the proccessed post analysis iBike data be less accurate than the original data because of the barometer change? Is the barometer accurate enough to capture small 10, 15 foot up and down changed or would it register as flat and over-write the correct tilt and thus adjust the power inaccuratly?

[rruff]

So, in this case, would the proccessed post analysis iBike data be less accurate than the original data because of the barometer change? Is the barometer accurate enough to capture small 10, 15 foot up and down changed or would it register as flat and over-write the correct tilt and thus adjust the power inaccuratly?

The barometer notices 1-2 ft changes in altitude pretty well and consistently IME... and I wouldn't worry about climbing totals: http://www.ibikeforums.com/viewtopic.php?f=4&t=402