Systemic tilt error

[fmullin]

At the start of my rides, I consistently do the tilt calibration. Sometimes, I even do a second calibration later in the ride (especially on hot days).

At the end of the ride, I then do the detailed correction and have an opportunity to see the result of the tilt analysis and correction. It ALWAYS shows a downward error in the tilt so that the adjusted power levels are increased.

Is there some problem with my iBike, my techniques, or something else that leads to this consistent error?

TiltError.jpg (88.72 KiB) Viewed 16952 times

[coachboyd]

Is it consistently right around -.4?
What is your riding tilt at the moment?

You may want to adjust your riding tilt number in your profile if you are consistently getting the same tilt error everyday (ex. everyday a -.4% tilt)

[coachboyd]

I think I just answered my own question by playing around a bit. Is your riding tilt at -.350?
Depending on your handlebar position, that could be kind of low. I have mine at -.417 and my weight with the bike 170. If you are consistently seeing -.4 try changing your riding tilt to .-7 and then change your friction number to keep the crr number the same.

That is how *I* got my tilt to read very close to level everyday on one of my profiles. Of course the real answer would be to redo the profile if you are seeing errors everyday. . .but this might be a quick fix until you have time for a 4 mile O&B

[iodaniell]

i had something similar happen, last Sunday, on a long distance training ride wearing a loaded backpack (why weight is 210 lbs) and going very slow. Up until the 25 mile mark, everything looks ok, but after that, tilt seems to take a nosedive.

Notice that I was able to raise 2 high vibration indicators as well.

I'm attributing this to the rise in riding temperature. I started in the high 60sF/low 70s. By the end of the ride, it was the high 80sF. I did not touch the wind offset or tilt for the entire ride once I made the initial settings before the ride.

[rruff]

I'm attributing this to the rise in riding temperature.

Tilt is effected by temperature? That surprises me, but I'm not sure how it works...

My riding tilt was -.6% when I did the CDs and 4mile yesterday. A bit higher than most I guess. It is very important to get this right for your normal (avg) riding posture, then do the tilt cal unloaded and let the iBike compensate.

[iodaniell]

I'm attributing this to the rise in riding temperature.

Tilt is effected by temperature? That surprises me, but I'm not sure how it works...

My riding tilt was -.6% when I did the CDs and 4mile yesterday. A bit higher than most I guess. It is very important to get this right for your normal (avg) riding posture, then do the tilt cal unloaded and let the iBike compensate.

It has to be temperature (or possibly battery voltage related) since my mount is basically glued to my stem.

All of the batteries were replaced a day or so before this ride.

[rruff]

I've experienced similar temperture changes and not seen such a change... certainly is possible though. It looks like the tilt was good both at the beginning and the end of the ride. Did anything else change... like maybe the sun coming out in the middle of the ride? Just wondering if a temperture *differential* in the case (top hotter than the bottom) might cause it to warp enough to cause this. I have a black iBike and I wonder about this...

Just did a quick calc... from 50 to 100 miles the tilt error was -.23%... not a huge amount. I probably have seen changes that large.

[travispape]

I'm attributing this to the rise in riding temperature.

Tilt is effected by temperature? That surprises me, but I'm not sure how it works...

My riding tilt was -.6% when I did the CDs and 4mile yesterday. A bit higher than most I guess. It is very important to get this right for your normal (avg) riding posture, then do the tilt cal unloaded and let the iBike compensate.

It has to be temperature (or possibly battery voltage related) since my mount is basically glued to my stem.

All of the batteries were replaced a day or so before this ride.

I'm starting to think there is another possible explaination. As I've mentioned before, I usually run with 3 iBikes mounted on my bike and they are literally glued on with a 2 part expoxy paste (PC7). For the large magority of my rides, there is hardly any tilt error; however, every once in a while I'll see a change of tilt that occurs at the same point in the ride for all 3 units. The fact that it happens for all 3 units is a strong indication that the change of tilt cannot be explained by any slips in the mounting or within a given unit. Further, the change in tilt is sudden (kind of like yours, Iodaniell) and I can trace it back to some kind of rough vibration event. Once it was a very rough section of road and another time it was speed bumps that I had bunny-hopped.

What I am starting to suspect is that sometimes the bike can "settle" into a new tilt and that the culprit might be found in the steering bearings. Remember that these bearings bear all of the weight and vibration of the front end of the bike and the forces are pretty big in there. Perhaps if the bearings settle into a new nominal position for when you are going straight down the road, the angle of the front end of your bike can change by the small 0.1 to 0.2 % that your iBike can easily detect.

Just a thought. Whatever the source of the tilt error, the iBike2 analysis removes the error. Again, with most of my rides there is not much tilt error to remove.

Travis

[iodaniell]

I'm attributing this to the rise in riding temperature.

Tilt is effected by temperature? That surprises me, but I'm not sure how it works...

My riding tilt was -.6% when I did the CDs and 4mile yesterday. A bit higher than most I guess. It is very important to get this right for your normal (avg) riding posture, then do the tilt cal unloaded and let the iBike compensate.

I did two more very long rides the past two weekends and I had the same tilt 'sag' occur on both rides--similar to the graph I posted earlier. I know for a fact that on last Saturday's ride I recalibrated the tilt at one point, but when I processed the data, it didn't make a difference. I thought I would see the tilt return back to zero, but it didn't. This completely amazed and baffled me. I don't know what is happening. I'm ruling out temperature. Voltage sag? Dunno.

[iodaniell]

Whatever the source of the tilt error, the iBike2 analysis removes the error. Again, with most of my rides there is not much tilt error to remove.

Travis

It's nice that iBike2 does this, Travis, but when I'm riding for 6-7 hours, much of it in the mountains, and metering my effort based on power, a sag in tilt makes me work harder than I should, because a certain pressure on the pedal no longer equates to the same wattage. I may as well go back to riding with a simple HRM. It has basically the same effect, but with a whole lot less calibration hassle.

[rruff]

I must have a weird heart... because HRMs are useless to me for pacing.

Also, you mention climbing. On a 5% grade a .2% error in your tilt will only be a ~4% error in power... but maybe that is enough to mess you up.

Last two rides I've done, the tilt corrections in iBike2 were .00% and .02%. There is some luck involved there I'm sure, since it isn't likely to get the tilt cal that accurate. I don't know why mine would behave differently, but the tilt is very good. The mount is a basic handlebar unit with sticky tape, and the typical roads are chipseal. I *have* had some issues with a brake-bolt mount on my TT bike where it seemed that the mount created a vibration that was making the tilt read low... by ~2.5% and jumping around. Like others, your issue seems to *not* be due to the unit or bike shifting position, since doing a tilt cal during a ride had no effect. Others have mentioned that their tilt checked out after the ride, so I'm wondering if vibration could be effecting things? Seems like on a climb where the speeds are low this would not be the case though...

[alback]

My tilt is points upwards when I download my ride file. It starts within 2km, almost immediately when I ride, not a break away as I see in your graphs. I ride on chipset and paved roads. When I rode on absolute smooth pavements, the tilt pointed upwards as it does now.

This means the power reading as I ride my ibike and iaero are considerably above the watts that the iBike profile brings it back down to. In fact, until viewing the DLP team numbers, I didn't know what numbers to expect, or which of my profiles would provide the most realistic results.

What I don't understand is: once a ride is downloaded and corrected in the iBike2 software, and adjusted under the tilt and power analysis. Why doesn't the software offer an option to the user, to send an adjustment tlt number back to the iBike/Aero unit? Guessing will be pure trial and error? If the software corrected this after every ride, it would improve the power reading for the next day.

A second option in the software could be where the elevation changes at the end of a loop ride. The user could identify the ride as a loop. Then the end elevation will match the elevation at the start. This would provide an opportunity for ibike software to compensate for barometric changes and therefore a more accurate adjustment to the ibike tilt.

The tilt mechanism seams very sensitive, and appears to be a little different from unit to unit, is it possible that the tilt mechanism has a bias which may start at the beginning of its use, or after some time or after some rough pavement.

Alan

[rruff]

A second option in the software could be where the elevation changes at the end of a loop ride. The user could identify the ride as a loop. Then the end elevation will match the elevation at the start. This would provide an opportunity for ibike software to compensate for barometric changes and therefore a more accurate adjustment to the ibike tilt.

I like that suggestion, and I bet something like that will be implemented someday.

As for your tilt always being high, this is easy enough to change by modifying your profile.

[alienator]

I've another quasi-tilt related idea: it be groovy iffin' there was a way to, with a button push or two, to switch to a second profile, third, and etc. profile. In fact, if there an isolated button on the face for this purpose, I think it'd be well worth the redesign. If nothing else, such a button could at least allow time/distance marks in the data set to use in data analysis, later, for the application of said profile.

And as someone else said in another thread, if the hardware inside could be arranged to accommodate it, having the keypad directly over the mount would be ideal for eliminating pressure induced tilt errors during tilt calibration before a ride.

[iodaniell]

Here's the 'analyze tilt' screen from iBike2 for that 110 mile ride.

[racerfern]

Two days ago I moved the mount from the handlebar to the stem with an epoxy base. I also removed the two sided taped and found some 3M two sided tape that is .030" thin. I put one layer on the bottom of the wireless head and one layer on the clamp half. Luckily my stem is close to round so the epoxy was only needed to level the head.

A new set of coastdowns produced dramatically different results. The biggest was the drop in wind offset (primarily because even though I had it on the bar, it was still tilted upward). My tilt correction from the 4OB dropped to -.327 but I suspected that was because of the different ride style between CDs and the 4OB, so I went with -.444 and adjusted my Fric up slightly to keep Crr at .0056.

Today on a long ride, I thought the tilt was off at 17.4 miles so I re-did the tilt on the fly. I had that same feeling at 67.3 miles and re-did it again. Now that I have downloaded the ride, I can hardly tell where I did a tilt correction. I will make an adjustment in my tlt to -.461 for future rides but all in all I'm thrilled with the results of my efforts.

[iodaniell]

Today on a long ride, I thought the tilt was off at 17.4 miles so I re-did the tilt on the fly. I had that same feeling at 67.3 miles and re-did it again. Now that I have downloaded the ride, I can hardly tell where I did a tilt correction.

But, why isn't it visible? The tilt correction should show up as a return to baseline (zero). Shouldn't it?

Notice that I removed the text from my previous message, because after reviewing the cue sheet, I actually made the tilt adjust earlier--somewhere around mile 40--and, like yours, it's not visible.

[racerfern]

But, why isn't it visible? The tilt correction should show up as a return to baseline (zero).

I'm not so sure about that. Although I have seen that in your ride and in some rides I did when I first got my iBike. If you're at x ft in altitude and you re-do your tilt when you ride off again, you're still at x ft of altitude. The only difference should be that the slope line should stay closer to level.

[alback]

But, why isn't it visible? The tilt correction should show up as a return to baseline (zero).

If you consider the tilt is a highly sensitive level meter that is constantly changing. The elevation (altimeter) determines where you are relative to barometric pressure which is more stable. The tilt is needed to create an instant factor (up/down) to evaluate watts while we ride. The altimeter would react too slowly as it did on my Blackburn 6.

I did a manual correction in Excel to my ride where the IAero went crazy last Sunday (rain), I plugged in my watts factor from the Kinectic trainer power file, and added a value based on the tilt (slope) +/- numbers. I was surprised how I was able to completely match the ibike power numbers where it appeared not to be crazy. In the end, I managed to replace all the strange values resulting in a reasonable ride result for my cyclingpeaks. I did not touch the wind factor, other than remove the 130km readings which never happened.

I believe the tilt goes astray in the graph as soon as it makes its first error on the road (over compensates). This could be a bump, a rough spot, or an overall bias up or down. The error value at that moment becomes the new starting point to evaluate the next slope % which is why the graph separates and continues.

So when you did the reset, it would still have the same end point as before, then start from there.

I suspect the tilt apparatus has a mechanical component, if so then it easily could have a bias, get jolted, or change its mode from a significant temperature change. If you move the ibike in your hand, it takes a few seconds to rest at a value, this dampening may be a variable. How quickly you turn your bike 180 and press the reset may throw off your tilt set-up, because the tilt needs a few moments to settle.

As Fernando, I found once I mounted the iBike on the stem, (also using an agressive double sided tape rather than the foam), my tilt became more consistent although it goes upward every ride. I did my calibration on an extremely smooth and almost flat 3 mile road which gave me the best (closest) profile to use in iBike2. However, the profile doesn't make the ibike/iaero read similar watts when I am riding the next time out. My profile brings down my average power by about 33% from the original download. When I ride the next day I try to recall the raw average and that value becomes my power target during the ride.

Since the IAero, Velocomp appears to be focusing the use of the iAero along side other PM's utilizing its unique feature to determine the riders best aero position. I believe their market is more the fitness and performance rider but they know their business best. Personally I have no plans to purchase another PM, or load up my bike with a series of other electonic meters. I purchased the iBike as a more reasonable alternative which could help me ride better for improved physical fitness and download my results. The fact that it gets me motivated on the bike, as I want to evaluate those results, it helps. I just want the watt values to be more reliable when I am on the road. The stem mount was an improvement.

I suspect from the small hints that are made on this forum by the company, that todays unit has achieved its maximum and needs more CPU horsepower, chipset and ram and they are already working on that.

Alan

[iodaniell]

I just want the watt values to be more reliable when I am on the road.

Alan

This is all I'm after.

As a racer who wants to train using this device, I need reliable "on bike" feedback, not post-ride corrected data.

I'm alright with the fact that device may drift during a ride, but I should also be able to get it back into calibration. It should come back to zero.

[coachboyd]

I believe the big jump in slope at mile 60 was when I went over a very bad set of railroad tracks. I didn't think that I hit them too hard, since I'm running wheels with carbon/kevlar spokes (Topolino Carbon Core) and didn't want to risk damaging them, but apparently I jarred the Aero enough to cause that change in tilt.

But if you look at it, the big jump was in your elevation number. . not your slope number. It looks like where that jump happened all of a sudden there was big change in elevation data. By any chance did you do a store stop on this ride or stop for a period of time where the ibike could have sat in the sun? The slope line actually looks pretty good, and notice how it continues the same trend directly after that jump.

[iodaniell]

I believe the big jump in slope at mile 60 was when I went over a very bad set of railroad tracks. I didn't think that I hit them too hard, since I'm running wheels with carbon/kevlar spokes (Topolino Carbon Core) and didn't want to risk damaging them, but apparently I jarred the Aero enough to cause that change in tilt.

But if you look at it, the big jump was in your elevation number. . not your slope number. It looks like where that jump happened all of a sudden there was big change in elevation data. By any chance did you do a store stop on this ride or stop for a period of time where the ibike could have sat in the sun? The slope line actually looks pretty good, and notice how it continues the same trend directly after that jump.

Yes, I did notice that the slope was consistent after that point.

I did a store stop at the base of the first climb. I'm sure that I started the Aero back up, but I noticed that it turned itself off at one point. That up-jog in elevation may have been when I turned it back on--with a little more thought on my part, I should have realized that the rr-track crossing can't cause a big barometric gain as shown.

But, what I'm wondering is why the tilt didn't go back up to zero when I re-did the tilt at about the 40 mile mark.

Here's the elevation profile of the ride as measured by GPS:

[iodaniell]

But, why isn't it visible? The tilt correction should show up as a return to baseline (zero).

I'm not so sure about that. Although I have seen that in your ride and in some rides I did when I first got my iBike. If you're at x ft in altitude and you re-do your tilt when you ride off again, you're still at x ft of altitude. The only difference should be that the slope line should stay closer to level.

Right, the slope line should come back up to match elevation and the difference line should be zero.

[rruff]

But if you look at it, the big jump was in your elevation number. . not your slope number. It looks like where that jump happened all of a sudden there was big change in elevation data. By any chance did you do a store stop on this ride or stop for a period of time where the ibike could have sat in the sun? The slope line actually looks pretty good, and notice how it continues the same trend directly after that jump.

I'm with coachboyd. Your tilt is ~.31% low... change this in your profile and the data will be fine... provided that this is a consistent issue. The one crazy thing is that sudden spike in your elevation. I've gotten little blips there when I've done a new wind offset in the middle of a ride, but nothing that huge.

The reason why you did not get any change in your graph when you did a tilt cal, is because it wasn't bad. The tilt cal will only effect the slope of the line... it doesn't bring it to zero. If your tilt had been off then the slope of the line would have changed after that point and nothing else.

[iodaniell]

Thanks all, for your responses and thank you Travis for your email.

I know what I need to do and will fix it tonight.

[iodaniell]

I want to thank everyone, again. I think I got this pegged, now. I'm hoping to have the same success on much longer rides than this--to be seen this weekend.

[rruff]

That's very cool... what exactly did you change?

[iodaniell]

That's very cool... what exactly did you change?

I adjusted the riding tilt to get that long ride to level out and then adjusted friction to get Crr back to .0070 ( Crr is a little high compared to normal clincher tires. It's because I run Tufo tubular clinchers at 100 psi. They need to be at 140 psi to be closer to tires like Michelin Pro 2s at 100 psi, but they are harsher at that pressure.)

I like the Tufo tires for more confidence on descents. It won't come off of the rim if it flats during a descent, usually losing pressure more gradually. With the clincher version there's no glue to bother with, but the wheel/tire combo is break-even regarding weight compared to a normal clincher tire/inner tube/rim tape/wheel combo.

[racerfern]

Congratulations! That is an awesome setup. Keep us posted on the upcoming rides.

[alback]

Well done! Looking forward to your next post.

Now, how do you go about changing your tilt value to bring these together? I tried in the profile edit, but it did nothing to my graph.

How do you know what value the Crr?


Alan