Calibration best practices?

[MultiRider]

What are the best practices for calibration?

I've been iBiking for almost a year and have not been able to get accurate, reliable results. I'm very frustrated.

I did 10 coast-downs and an 8 mile out-and-back on my original iBike and it was very consistent to speeds up to about 25mph, then it started inflating the numbers. Specifically, pedaling backwards on downhills revealed spurious watts starting about 25 at 25mph, increasing to 80 watts at 30mph, and the spurious watts just went up and up as speed increased. I thought it was from vibration, so I bought and installed the VRK. No improvement. So I upgraded to the iBike Gen II. Now I can't even get consistent and reliable results at low speeds.

I've done two calibration attempts with 10 coast-downs and 12 mile out-and-backs with the Gen II. Though frustrated after the first calibration test and subsequent poor results (300 watts while pedaling lightly up a slight hill), I rationalized it because my first attempt might have been compromised by 4 different school buses passing me and gusty winds. But the 2nd calibration effort went well, so I am extremely frustrated by continuing inflated watts.

So I'm going to do another calibration attempt, probably tomorrow. Should I find the flattest possible road? Or is it better to include some hills in the out-and-back? The roads in my area have hills, so that's what I've done so far (both with the original iBike and the Gen II) because it is the easiest. Should I put my bike in my truck and drive someplace to try to find flat roads?

I've heard "more data is better" on this board a lot, so I did 12 mile out-and-backs. But the longer out-and-backs include somewhat rough sections of road (relatively short sections -- 1/4 mile or less). Should I shorten the rides and avoid bumps and rough roads?

I really want my iBike to work. I was pretty happy with the original since it seemed accurate and was very consistent. I just wanted to eliminate watts inflation at speed and expected the Gen II to do that. The Gen II results are worse than ever.

I'm hoping a new calibration attempt will get me back at least to where I was (consistent and reasonable results below 25mph) or optimally get me to consistent, reasonable, accurate results at all speeds.

What are the best practices for a calibration ride?

Thanks.

[racerfern]

A couple of thoughts:

1) More data is better, but there is a point of diminishing returns. When I do a set of coastdowns, switch from the coastdown screen to the main screen and check the aero and fric numbers I got for that coastdown. Just check them and make mental note. Then I do another coastdown and check again. That way I have a good idea immediately if my coastdowns are all within a somewhat reasonable range. If they all seem to fit I stop the process at about ten.

2) For me the nemesis for coastdowns is wind not hills. So I got up at 6am before the wind kicks up and did a set on a level bike trail. I do one, turn around and do the next in the opposite direction. That way in a matter of minutes my coastdowns are over. Then for the out and back, 4 miles is fine, but you want pretty much the same conditions as the coastdowns.

3) Cars have a huge effect on coastdowns and the out and back so try and avoid them at all costs. I prefer flat roads or ever so slightly uphill. You just need to coastdown to 8mph in a reasonable amount of time.

4) FWIW, my Gen II is far and away better than the Gen 1.

5) The only part of the CD process some people have issues with and can be frustrating is the slope/tilt adjustment. That may need to be tweaked manually. there have been numerous discussions about manually adjusting the numbers. But after that, it's good.

Do you have a wireless or wired mount? Is it stem mounted? When you have errors or numbers that seem out of whack, are they consistently a problem from ride to ride? If so, maybe you could upload a ride or two and also your profile.

[MultiRider]

I have a wireless stem mount. Still have the VRK in the stem.

I do my coast-downs on a "saddle" near my home that doesn't get much traffic. I get up to speed on the downhill one way and then coast up the slight (approx 2% grade)uphill the other way. It makes it easy to get coast-downs going both directions and I can get a bunch of coast-downs done pretty quickly.

It is harder to do a no-cars out-and-back. There's one road from my neighborhood that goes to a highway; very heavily traveled. Then there is the road I have been doing the coast-downs on. It is one lane each way and has about a 8' wide bike lane. I try to stay as far away from the cars as possible (near the curb), but they still go by from time to time.

The out-and-back road has a 3 pretty significant hills. They are downhill on the way out, so I hit 35-40mph sitting upright (45+ in a tuck). Then they are significant uphills on the way back -- 6-7% grade. When I get to the bottom, there is a traffic light. That is about 4 miles out and makes an 8 mile out-and-back. There are some gentle turns, nothing sharp (all less than 45 degrees). Though there is a turn on the third downhill, so I'm leaning my bike at about 30mph.

Does that affect tilt? Would it be better to turn around and cut the ride short rather than do a high speed turn like that? It is a gentle bend, but at speed it might affect tilt, right?

For the Gen II, I went through the traffic light and out another couple miles. There are a couple pretty rough sections after the traffic light. So I'm planning on not doing that extra leg in tomorrow's calibration ride.

Should I not do the third downhill due to the high speed turn? Or apply the brakes and do it with the bike as upright as possible?

T'hanks for your help!

[MultiRider]

Here are a couple screenshots of the results of my last calibration attempt.

Please note that I did 7 rides on asphalt and 2 rides on concrete since I ride on both concrete roads and asphalt roads. I should probably weight the concrete more, though, because all the roads around here are concrete and I only ride asphalt when I drive somewhere like to the foothills or some place in a different suburb/city.

As mentioned previously, the iBike displays significantly inflated watts based on the profile that was created with the attached screenshot (I downloaded the rides and created a profile the day I did it, but can't seem to get the graph to display without going through the steps again).

With the profile created by those files, soft pedaling registers 150 watts, medium pedaling register 300-350, hard pedaling registers 550 or more. Based on my experience with the original iBike (which compared favorably to a friend with his PowerTap), those numbers are way off.

I am going to erase all data from my iBike as soon as I submit this post because the memory is 80% full. It has a couple long rides in it. I wish I could selectively delete the long rides and keep the other data, but apparently it is an all-or-nothing deletion process.

[racerfern]

I don't think the rough sections will have a negative effect on the overall outcome. I have yet to see road so rough that new Gen II can't handle. I'm not really thrilled with the description of your 4-8 mile out and back ride. Hopefully. someone with more knowledge than me will jump in. You may want to consider a different location altogether.

The out and back determines the two most important parameters in my opinion. Wind and tilt. My guess would be that extreme downhill or uphill sections will not provide usable data either for wind in one direction or tilt in the other.

Thinking of it at an extreme, I don't think riding downhill for two miles, turning around and pedaling a 6-7% grade for two miles would be considered a good 4 mile O&B. I realize it's not all like that, but a different location might work better. Again, maybe someone with CD experience in hilly areas could pipe in.

I just saw your screenshot as I hit submit. That is a ridiculously low Crr is my first thought and a crazy high negative tilt number. I realize every person is different and every bike and setup are different but its amazing how 99.5% of us fit in a fairly tight circle of numbers. AND, your tilt should be even more negative based on your screenshot. Something doesn't make sense. Are you sure you are correctly setting your tilt prior to coastdowns? Have you checked your tilt after a series of CDs and the 4O&B to make sure nothing got whacked?

After all the effort you've put in, I'm sure you're being anal about the whole process, but you never know...

[MultiRider]

Here are the coast-downs from my 8-25-2008 calibration attempt:

[MultiRider]

Here are the details of the profile that I created in January with my original iBike that provided what seemed like reasonable power readings up to about 25mph. I also attached the details of the current profile. The file name at the top indicates which is which.

Please note that the same wireless stem mount was used both times for both iBikes. It was not quite straight originally, so when I got the Gen II, I loosened the screws and rotated it just a bit. I'm surprised at the difference in tilt numbers. I wouldn't expect that rotating it a little to the left to get straighter iBike alignment would affect tilt that much. Of course, each iBike has its own tilt-o-meter and they are all likely to be a little different. I'm not an engineer and not sure how much difference is reasonable. Just an avid bike rider who wants his stinking power meter to work.

[coachboyd]

Those last results look pretty consistent. The profile you showed before looks like maybe your tilt calibration was a little off when you went for your out and back ride. My suggestion is to find a road with as little traffic as possible. I try to get some very slightly rolling hills (2-3%) and some sections of flat in my out and back ride. Anything with a traffic light and high speed descents is not going to be good.

Definitely take the extra time to do the o&b ride correct. The wind scaling and riding tilt numbers are two of the most important numbers for getting correct on the bike power numbers.

FYI. When I soft pedal with the profile I have on my bike I always show zero watts. If I get down on the drops and soft pedal going down a hill, the watts will go up, and if I go back to my normal riding position, it drops back to zero.

[coachboyd]

For your ride tomorrow, modify your friction and riding tilt numbers to match the results from 1/06. Keep the aero and wind scaling numbers the same. Try that for a ride or two and then send me a couple ride files at boyd (at) totalcyclist dot com. I'll try to get everything working for you and get you the great results that I am getting for my rides.

[MultiRider]

Okay, here's what I'm riding with tomorrow:

[racerfern]

That's great advice from CoachBoyd. That looks like a usable profile that will be easy to fine tune. Looking forward to your post tomorrow. Where are you west coast?

[rruff]

Here are a couple screenshots of the results of my last calibration attempt.

If you are getting W numbers that are too high while riding, then the iBike thinks that:
a) You have more of a headwind than you really have. This would happen if the wind scaling factor was too high. Or it could happen if you don't keep your wind offset zeroed. You can check your wind scaling easily enough by riding back and forth on a stretch of road and seeing if the avg ground wind adds up to zero, via iBike2.
b) You have more drag than you really have... ie CdA is too high.
c) You are going down hill when you really aren't. This would happen if the riding tilt was too low. Riding tilt errors are easy to spot and correct for via iBike2... many threads address this.
d) You have more rolling resistance than you really have... ie Crr is too high.

You need to do the calibration again. Flat to slightly uphill road, little wind. Be sure to do a good tilt cal and wind offset first! The calibration runs are the most important to get right. The 4 mile ride is for determining your riding tilt and wind scaling factors and these can be tweaked to the correct values later if you screw them up somehow.

[MultiRider]

Fernando -- I'm in Colorado. I live in Castle Rock, which is near the front range of the Rockies. I can't think of any truly flat roads that go fairly straight within 15 miles of my house. Cherry Creek Reservoir has fairly flat roads around it, but you're turning pretty much all the time. The fairly straight roads are all pretty hilly. After reviewing the results of this morning's calibration attempt, I've thought of a different road that I could try a calibration ride on that has very little traffic, though some sections of it are pretty steep. I may try that next.

I tried to do a very careful calibration effort this morning and it was the most frustrating ever.

I left my iBike out in front of the garage for 45 minutes to get acclimated to the 50 degree outside temperature. I pumped my tires up to 100psi front and rear. I zeroed the wind offset. I started to do the tilt on the garage floor, but decided to do it inside on the hardwood floor in the kitchen for greater precision. My wife was not pleased, but I had the bike as straight up-and-down as possible and leaning against a wall to ensure no variation while the tilt function did its thing. I double-checked the tilt after the iBike indicated it was done by switching the bike 180 degrees a few times. It was at 0 degress both directions every time.

Took it back outside, got my helmet/etc on and went down the block. It was 50 degrees and calm when I started the calibration effort. Did 6 coast-down runs. Is it critical that you stop pedaling the instant the iBike says "Coast"? My timing was off this morning and I seemed to look up at the road at the wrong moment each time because I only saw it change from "Fast" to "Coast" once. The other times, when I looked down it had already changed. So I gave it an extra half or full pedal stroke each time. Surely the iBike software is smart enough to do the calculations once you get down to 20, even if you start at 22 or 23mph? I hope so.

After the coast-downs, I rode approximately 1 mile to the start of the out-and-back. I hit Trip Reset and zeroed it out. The wind was pretty calm when I did the coast-downs; unfortunately, it was somewhat gusty in the middle of my out-and-back. The road goes through a pretty deep saddle (see graph) and there was constant wind with some fairly significant gusts during that portion of the ride. More traffic than I expected as well. I was passed by 2 school buses, a dump truck, and a fair number of cars.

When I got back, it was 60 degrees. When I checked the tilt in the garage before putting my bike away, it said 0.4 degrees one direction and 4.7 degrees when I swapped the bike around. I checked it a half dozen times because I was so surprised that the iBike had gone from perfect tilt at 8am to a 4.3 degree uphill angle during the approx 45 minutes I was doing the calibration procedure. That is pretty dramatic tilt drift! And it accounts for my observations -- when I first started riding, the watts seemed pretty reasonable and I was very encouraged. By the end of the out-and-back, the watts were clearly much too high again. Here is the result of my calibration efforts this morning. The iBike 2 software won't even create a profile from it -- says the "Riding Tilt is out of range".

[MultiRider]

The web site software seems to hack off the right side of the images I post, so it cut off the coast-down info. Here's the right side of the iBike 2 software:

[coachboyd]

It looks like something happened on your way to the coastdowns to knock your tilt off a bit. It really is worth it to find a good section of road, not rough, with little to no traffic. You keep mentioning that you are getting passed by school buses during your out and back ride. I would look for someplace different where you won't have those outside influences.

Also, before you do your coastdowns I would recommend doing another tilt and wind offset calibration, just in case it has slightly changed from the ride out there. I use the white line on the side of the road, or even a better situation is to find a crack in the road so your bike is perpendicular to the white line on the side of the road (that way it's not off because of an axis not being lined up). Once you have your tilt and wind offset checked, then do the coastdowns all in the same direction and all on the same stretch.

My advice is to accelerate to 23 or 24 mph, then start coasting. The iBike won't start computing aero and friction numbers until you have hit 18mph, but this is what has worked for me best. During the out and back ride, make sure you don't stand up or hit the brakes until the turnaround. The most important thing is to start and finish in exactly the same spot. Really try to avoid any roads where you'll be getting passed by traffic, especially larger vehicles like a schoolbus.

[MultiRider]

For your ride tomorrow, modify your friction and riding tilt numbers to match the results from 1/06. Keep the aero and wind scaling numbers the same. Try that for a ride or two and then send me a couple ride files at boyd (at) totalcyclist dot com. I'll try to get everything working for you and get you the great results that I am getting for my rides.

I think I missed your point. I was so focused on getting a calibration effort that worked, that I thought you were saying to set my profile and then do another calibration attempt. So that's what I did. But as I re-read your post this morning, I think I missed what you were saying. You wanted me to just go do a few regular rides without more coast-downs or an out-and-back, then send the data. My mistake. I was obsessed with getting good coast-downs and a good out-and-back.

I'll do a few rides over the next few days with the manually adjusted profile and send.

Thanks for your help!