ibike Power meter
- Thread starter BILLYHOLMES
- Start date
With a power meter, you have to do this on a flat, straight road in windless conditions. You can't do it as a roll-down (i.e., roll down a hill with no pedalling required) because you have to supply power to "the system" to measure something on the power meter. Because the iBike inherently works off air flow, you can do this as a roll-down test. Again, you need windless conditions for it to work out.rayhuang said:
Of course, you could always do a roll-down test with no meter, just a stopwatch or a riding buddy to use as a control (old school) but the accuracy, resolution and repeatability would probably be a lot worse than with an iBike or a power meter.
There are a lot more short hills to roll down than flat roads around here, for sure. I know of only one flat, straight 1 km section of paved path in a 30 mile radius of me.
Any time you estimate something rather than directly measure it, you are introducing another set of variables that can produce even more error. People with Power Taps and SRMs have enough issues as it is, why make things more difficult on yourself? And don't try to tell me that you are measuring torque with an iBike because you are not.[/QUOTE]
- All power meters estimate power it cannot be directly measured. It doesn't matter whether you measure material strain and calculate torque and hence power or measure opposing forces it's the same. The iBike will give very similar result to an SRM or Powertap. Only in extreme conditions will the unit be off (severe crosswinds and moving around in a small pack). Most of the time it is bang on.
RChung said:
I think there's something being lost/garbled in translation. When I've talked to both John Hamann and Glen Cunningham, the developers of the iBike, they've both been very upfront and honest about the strengths and limitations of their product. They believe in it and don't feel the need to "spin." Curiously, the website gives a different impression. I don't know why the discrepancy.
RE using the iBike for aero testing: some of us on the iBike list have been trying to establish the precision of the device by doing multiple coastdown tests and comparing the results. We're also trying to see how different procedures and conditions affect the variability. My testing found a standard deviation of about 2% in CdA and 6% in Crr. I also found that the two vary oppositely. This was under non-ideal conditions, with a variable 5-12mph breeze. Since coastdown tests are fairly quick to do, I do multiple ones and average them to increase precision. (In fact, I'd recommend multiple coastdowns to anyone wanting real accuracy out of the device.) After a lunchtime's worth of testing, I believe I got both numbers to within 1%.
RE finding the "sweet spot" in a pack, I suspect that what John or Glen was referring to was the ability to read the actual apparent windspeed while riding in a pack. Functionally, you could find it just as well watching your wattage number, of course.
To the engineer in me: They all are estimates and what matters is how well each model's implimentation solves the individual's needs. - TFRChung said:
Edit: In other words, the part of my job that did pay the bills was, with a reasonable knowledge of the theory, to find the most cost effective solution to each individual problem regardless of the model or technology involved. - TF
I could as easily say - "Ultimately", how fast you go or how soon you get there depends only upon the power generated against the opposing forces.Steve_B said:
I'm not saying that torque (wherever it's measured) is not a valid metric for most of our requirements. I am saying that:
- The measurement of torque is no more "real" nor, necessarily, any less of an "estimate" than any other.
- "..for many of us..." doesn't apply to all of us.
TF
That's exactly what I was about to post about, but I'm glad you said it first, since I'm no engineer.
I guess it all boils down to how much you are willing to have other "stuff" affect your "estimate" and affect the consistency of that "estimate" then, doesn't it?Terry Ferguson said:
But having a reasonable level of accuracy is in everyone's interest. I leave it up to the user to define "reasonable".Terry Ferguson said:
Yup.Steve_B said:
Simpler than that. I used a little tree-lined street that crossed a creek in the middle of the block. It had about a 20 foot drop from the ends of the block to the culvert, so I could do many loops at various speeds in only a few minutes without using my brakes or changing position. Basically, I could roll down, or accelerate down, or accelerate up as long as I let gravity slow me down just enough to make the U-turn at the ends of the block without using my brakes. Then I "bootstrapped" each loop to get an idea of the variability.
I did essentially the same thing with Dede Demet's Montreal World Cup data here, but of course she used her brakes so that was simply an illustration.
RChung said:
SD. To elaborate a bit, the 2%/6% figures came from a two-step test. I did 10 coastdowns, and after each coastdown I did a second test modelled on the procedure in the iBike manual for verifying if a coastdown produced "good" or "bad" values. This is where you watch the unit's unfiltered wattage screen (accessible in one of the setup screens) and verify that it stays within +/-20 as you coast. To make this second test more selective, I chose to do it at a higher speed than the initial coastdown test, from 30 to 20mph vs. the 20 to 8mph of the normal coastdown procedure. Only 4 of the 10 coastdowns passed this second test. For reference, SD for all 10 coastdowns wasn't nearly as good. I don't have my data in front of me, but I believe it was around 7-10%.
Yeah, I was expecting it too. Since the accelerometer is reading the sum of aero and friction during the coastdown, it makes sense that shorting one would lead to overestimating the other.
