Tilt Correction for Gen 3

[racerfern]

After yesterday's ride I downloaded the data as I usually do. This is the ride I do every Sunday with the exception that they are doing construction about 25 - 26 miles into the ride. The road was completely torn up with craters, manhole covers and steel plates that stuck up inches above the pavement that looked like a moonscape. After I got back on regular road I looked at my iBike and it appeared fine, not that I expected anything different.

However when I got home I realized my iBike wasn't sitting in the same spot on its epoxy mount. When I put it on the spot that I usually check on it was off by more than 1 degree. Looking at the tilt correction screen I can now see where the Gen 3 did its job seamlessly. You can clearly see the sudden and massive change in the graph, then the quick self-correction that fixed itself in less than five minutes. Because of advances in the hardware and software I didn't have to get frustrated or end up discarding a ride and having to guess what my efforts would have looked like.

I am one of the pre-production testers and I put in over 1,000 miles of testing and reporting, but I never came across anything that would show off the abilities of the Gen 3 like this. Thanks to John, Travis, Chris, Boyd and all the others that made these advances possible.

[nreimche]

Not to be condescending to Velocomp or anybody, but I noticed even the 'after repair' shot I noticed the tilt line was kind of jagged. With that said, would a straight(er) tilt line be more accurate; does the 'after' analysis salvage what it can, then indicate the best tilt (under the conditions of potholes, or like I do, which is, bump the unit!) possible? In other words why is a jagged 'after repair' tilt line, still jagged, and how can I know if a ride is still good, when the tilt is erratic?

[Velocomp]

I'm not sure there is a simple way to answer this question but I will try.

First: the "green line" MUST be jagged. The jags occur because of micro-accelerations caused by pulses in pedaling. Part of the reason the iBike is so accurate is because the accelerometer can respond to the power pulses, and because accelerometer readings are taken many times per second. The jags are a result of the pedaling and a perfectly smooth green line could only occur if the accelerometer had failed.

Second: if there is a severe disturbance to the iBike's tilt calibration (in Fernando's case, caused by the mount moving after hitting a pothole) this will cause the tilt calibration to be "off". The cool thing about the Gen III is that it will sense the disturbance and then self correct the problem, during the ride. It took a few minutes after the huge bump at mile 27.4 to get things right again, but then data was fine again!

This auto-correction feature is unique to the iBike; here is an amusing comment from CTS about a problem Lance had recently with his SRM (a problem he never would have had with a Gen III iBike):

"Lance discovered he had trouble maintaining his power output. Of course, that was partly due to a calibration problem with his SRM – it was reading about 100 watts low – which prompted Lance to ride at what we estimate to be close to 500 watts for the first five minutes of the race."

Lance's SRM calibration problem persisted throughout the entire ride; with a Gen III, calibration problems correct themselves automatically.

Third: don't obsess about the green line! It's kind of a fun thing to look at, especially to visualize the improvements of the Gen III, but the bottom line is NOT the precise shape of the green line; it's the difference between the pre-and-post analyzed rides. In Fernando's case this was 0.1 watt, meaning that the on-the-bike numbers were excellent! In the brief period at mile 27.4 where Fernando hit a pot hole the on-the-bike numbers were most definitely "off" because of the shift in the mount, but the iBike self healed in a few minutes.

Fourth: make sure your mount is ROCK SOLID TIGHT. Then it won't shift, even after hitting potholes

[Ratman]

John, just so I understand, are you saying that the iBike III makes a certain number of potentially dramatic on-the-bike corrections ("healing," in the case of moved mounts), then does additional fine tuning (more subtle "tweeks") after downloaded to the iBike3 software? In other words, do I have it correct that the "before" screen shows "post" on-the-bike corrections (from a data standpoint, more "major"), and the "after" screen shows post on-the-bike corrections plus additional iBike3 corrections (from a data standpoint, generally of less magnitude as major issues have already been addressed)?

Thanks!

[Velocomp]

The Gen III iBike includes most of the tilt correction algorithms of the iBike 3 firmware (we haven't yet figured how to stuff a PC into the iBike!). So, the Gen III does tilt correction in real time, at each second of your ride, to give you extremely accurate wattage readings right on the iBike screen.

The iBike 3 software still has a few extra mathematical tricks in it to provide the last watt of accuracy; so, you can still perform the "Analyze Power" command to compare things. For the hundreds of Gen III ride files we've seen, the pre and post analysis results are generally within 1W of each other, which means the Gen III fw algorithms are working extremely well.

In the Analyze Power screen the the "before" window is the raw data as captured by the iBike, then analyzed by the iBike fw, and as presented on the iBike screen in real time during your ride. So, in the "before" screen you're seeing exactly what was on the screen of your Gen III.

The "after" screen is the raw data of the ride file, with the iBike fw algorithm corrections removed, and THEN with the iBike 3 algorithm corrections applied.

You're correct that the Gen III algorithms do 99% of the corrections right on the bike, and do them extremely well.

I have found the Gen III results so good that I don't even bother to do the Analyze Power task any longer. At some point in the future I suspect we will remove the Analyze Power feature; it is an artifact of earlier generation iBikes.

[iodaniell]

I have found the Gen III results so good that I don't even bother to do the Analyze Power task any longer. At some point in the future I suspect we will remove the Analyze Power feature; it is an artifact of earlier generation iBikes.

+1, I have stopped doing the Analyze Power task as well.

I'm still using an old profile, as I haven't had the right conditions to do a calibration ride, but I can see by the data that I've collected in WKO+ that there are no significant changes in my daily rides to and from work--rides that are similar from day to day and on the same route from day to day.

I likely would benefit from a good calibration ride--when I did do the Analyze Power task, I'm a little less than 2 watts lower then iBike 3's processing--likely from tweaking by hand the ride tilt. But, those <2 watts really aren't bothering me.