Monday, June 04, 2012

Equipment - Rotational Inertia

I'm revisiting the wheel inertia thing, this time with a suggestion. At some point some time someone told me that I shouldn't complain about something unless either I could do something about it or if the person I was talking to could do something about it.

In other words if your boss, say, doesn't give you a bonus, you shouldn't go complain about it to the counter person at McDonalds. You should instead either talk to your boss or talk to someone that your boss may go to for information or advice. (This is assuming your boss or boss's confidant doesn't work as a counter person at McDonald's.)

I put up a post a little while back, a stream of conscious kind (aren't they all?), and I didn't have any suggestions for a way of proving what I think is true. In the article I basically say that I think wheel inertia makes a difference but I can't prove it.

The post I put up garnered a lot of response, mainly theoretical. No one said anything like, "Dude, I am totally with you on the weight, but being a physicist I have to say that the numbers just don't prove it. I went out with two very different weight wheelsets, did some testing with my SRM, and I found that I couldn't find any difference in power requirements between the two wheelsets greater than the margin of error."

Instead there were those that felt inertial wheel weight made a difference (but like me they couldn't prove it) or those that "proved" that inertial wheel weight didn't make a difference (but no one said that using an 800 gram tire in a 8 turn, 1/2 mile course crit didn't hurt them in the least).

I also received some private responses via email. One person, a racer and a mathematician (the best combo in this case) suggested a blind test. He suggested using lead tape (used to weight tennis rackets and golf clubs to improve swing power because apparently inertial weight does matter there) I should have someone weight a wheel (or not). To keep the system weight (i.e. overall bike weight) the same, I'd need to carry the equivalent amount of weights I don't use when I ride the unadulterated wheel.

I decided that it would be more practical to use two identical wheels (which I happen to have). I'll weigh one down with lead tape, not the other, and have two identical bottles, one with the same amount of weight in it as the lead tape, the other with nothing. (I have to work on the bottles - I don't want to have the weight rattling around). This set up assures me that the bike weighs the same. The only difference will be the inertial wheel weight.

With this set up I'll do repeated acceleration tests. I'll have a helper switch out the front wheel and bottle (and you'll see who this helper will be in just a moment).

I'd weigh and swap both wheels if possible but I want to make the helper's job as easy as possible. The rear wheel is slightly more difficult to put in and it also has the cassette, and I may be able to tell the two cassettes apart. I'll just load up the front wheel with a lot of weight - I'm hoping at least one pound, and I hope to be able to get two or three pounds of lead weight under the clincher base tape.

Ironically I don't care about climbing - I know that even very light wheels don't make me climb faster in a group. When I blow up I'm just as slow, and I can't go fast enough to keep up with everyone without blowing up. This is ironic because the math proves that there is a small but substantial gain when using lighter wheels in an extended climb.

(As a rider that loses 30 minutes over a 2 hour climb on a leisurely ride, reducing that loss to just 29:30 over that same climb doesn't interest me. I'm already taking a 33% hit to finish the climb - I need to chop 20 or 30% off my climbing time, not a half percent here and there.)

I'm only concerned with how wheel weight affects me on flatter roads with a group (i.e. a crit) which is where all this debate occurs. I do crits all the time and find that heavy wheels really affect me, even if they're more aero than my lighter wheels.

Zinn theorizes (and I do too) that part of the light rim preference in racers comes from the idea that lighter wheels allow you to get shelter quicker. I think this is a big part of the inertial thing - getting to shelter quicker and more efficiently. Drafting someone will save a lot more wattage than any aero wheel out there.

In fact when I used a much higher inertia bike (a tandem, with another rider on it), I found that I couldn't maintain a close gap to the riders in front of us. I learned the hard way that I tend to stamp the pedals once or twice, soft pedal, and repeat over and over. With 350+ lbs bike/riders unit (the bike alone is about 40 lbs), my normal "pedal stamp" wasn't enough to close a foot gap quickly. Instead of snapping shut the gap to the rider in front, I found that we struggled for 50 or 100 meters to close that gap, and after a few miles of that we were off the back.

Other theories that I have include my somewhat rough pedal stroke (apparently it's common even with top riders), where I basically have two power strokes per pedal revolution. I can even out my pedal stroke when I'm riding easy, but under pressure there's a distinct "on/off" of power, two surges per revolution. Like the pedal stamping in the draft, the surges need to translate to forward movement immediately. This allows me to stay in the draft and coast when I don't need to surge.

"Coasting?" you ask. "What about the flywheel effect? Wouldn't a bit more weight keep you going better?"

That's true in the case of a time trial or some other very steady, very smooth, very solo ride. In a crit, though, I tend to have to brake going into turns, even when I try and tailgun (to avoid braking, i.e. coasting up to the turns). At my most aggressive tailgunning race, I ended up soft pedaling behind the group for a while before the best tailgunning turn. I found that I still had to use my brakes here and there, and if the group accelerated unexpectedly, I had to push hard to make up ground.

If the flywheel effect was useful in a criterium then everyone would weight their wheels. Mavic's first disk wheel, the "Comete +/-" had removable weights around the wheel (the yellow circle decals on the wheel covered the weight openings). You could add weights to increase the flywheel effect. Apparently this wheel wasn't a big hit although it would make a great wheel for the rotational inertia experiment. A friend had one and used it devoid of weights - I think you could bump the wheel weight up to 2500 grams (about 5 lbs) by filling every circle with a weight.

Let me summarize my thoughts so far:
1. Wheel inertia exists. There are wheels with less inertia than others. Generally speaking I think less inertia is better.
2. Inertia matters in acceleration. Less inertia accelerates more quickly.
3. Acceleration happens in different, unexpected places, like drafting another rider, or accelerating to find shelter from the wind.
4. Although wheel inertia and aerodynamics matter, drafting matters more. The more a rider can shelter in the draft of another, the better off that rider will be later in the ride/race. The most aero wheels will not save as much energy as the energy a rider saves by drafting others.
5. If lower wheel inertia allows you to draft quicker and more efficiently, it will save you exponentially more energy than what wheel aerodynamics can ever save you.

My idea to test all this is to have a Keirin type start - a moto goes by at a set speed, the rider starts accelerating at a given time, and the rider tries to get to shelter as soon as possible (i.e. the moto's wheel). I haven't worked out the logistics but if the moto goes by at about 30 mph, it'll take a good 100 meters or so to get into shelter. The "sprint" is over when the rider reaches the moto.

Using a power meter that records the power, speed, cadence, etc, using the same bike, gearing, rider, and just swapping the different front wheels (identical appearance but with one being significantly weighted) and bottle (ditto on appearance and weight), it should be pretty straight forward to see if there's an acceleration curve difference, if there's a time difference with each effort, etc.

By using a weighted wheel and a non-weighted bottle for the "high inertia" set up and a non-weighted wheel and weighted bottle for the "low inertia" rig, the overall weight of the bike should remain constant. The bottle weight is reasonably low on the bike too, so it shouldn't affect the rocking of the bike too much (attaching the weight at the saddle would make the bike feel sluggish when rocking the bike back and forth).

I believe the inertia math. It makes sense. I just think it's incomplete. I think that it doesn't take into account a rider's pedal stroke (typically not smooth, especially under hard efforts), the energy savings while drafting (especially when the rider is already at their limit), and the energy expenditure when fighting to get the draft.

I hope to find a suitable set up for this experiment, with enough in my legs that I can do a good dozen hard accelerations (I figure groups of 3, two with weights, two without). I'd need a partner in crime to drive either a moto (scooter) or a car (I happen to have a hatchback that has a perfect drafting rear hatch window), someone that can swap a front wheel and bottle without any difficulty (and, ideally, add the weight to one of each independently).

I figure I'll need two days of testing. Not two whole days, just two separate sessions. The first dozen sprints (Session One) will teach us flaws in our basic logistics, like if 30 mph is too slow or too fast, what gear to use to start, how to do a standing start (or if the rider should roll at 15 mph until the car draws even), stuff like that.

The second dozen (Session Two) should be better, with the details worked out. I hope to get some decent data. I probably need more data but this will be a start.

I'm really curious to see what I find and to see if it's a valid experiment. If nothing else I'll get a couple dozen hard jumps in (if I'm the rider in the equation). Now I just need takers, at least one (a driver or a sprinter, and if the former then someone that can hide a few pounds of weights on a rim or in a bottle) or two (a holder or someone that has a holder rig idea... actually I have a rig idea so maybe then we need a camera person). Anyone? Anyone? Ferris?


DJ said...

It's a good experiment, except that the number of trials proposed is not enough to be statistically significant. Maybe enough to observe a numerical "trend" but not to prove it either way. Also experimental bias: you want to believe that the hypothesis is true and could subconsciously effect the outcome by "placebo effect" or "willing it to be true". So to be "valid" in the scientific sense you need:

(1) More trials, to reach statistical significance

(2) Your friends have to do the riding, not yourself (to remove subjective bias)

(3) Blind, randomized trials, the rider cannot be allowed to know whether they are using a weighted or non-weighted wheel

Maybe first do it your way, and then more scientifically if it seems worthwhile.

Aki said...

I realize the number of trials would be limited; if I could see a trend then that would encourage me to do more tests/trials.

1. More trials - yes.

2. Friends do riding - yes, I thought about this. It would dilute the data for myself (since my pedal stroke, my power curve, etc, are mine). I was thinking that whoever would be patient/good enough to drive the "moto" would have to be a rider, and such a rider could then do a bunch of sprints while I rested.

3. Blind randomized tests. I considered this too, but I'd need a third person. The front wheels I have are identical (one will have the lead tape), and I have two identical bottles (and one will have the same amount of lead tape in it, or just extra water). The weighted wheel and lighter bottle will be one set, the unweighted wheel and heavier bottle the other. If there's a third person putting the front wheel on and switching bottles then the rider and driver wouldn't know which is which.

My challenge is that I want to have a significant weight difference at the rim, like 400-500g minimum. I don't want to have a 50g difference, it won't have a great effect on the numbers (because, as you pointed out, I won't have a lot of trials). The lead tape I bought is very light, like 100g worth. I'd either have to get another $80 worth of tape or find an alternate way of weighting the wheel.

I was thinking of a spring weight setup I saw somewhere online. It's meant for giving the rider good inertial weight at speed but when you slow down the weights retract toward the hub. it's illegal for racing use but for the experiment it would have potential. Basically there are weights that slide up and down the spokes. If I had such a setup on a front radial laced wheel, no springs, and two sets of "weights" (one set would be light, one set very heavy), then the front wheel would look the same each sprint - one weight on a spoke near the hub, the other near the rim (one weight would be hollow). The third party person could swap all the weights. This keeps the weights on the wheel (truly measuring rotational inertia, since all the weight stays on the wheel), it avoids switching wheels, no bottle changes, can fit any size bike, etc. This would also allow much more weight, depending on the hub flange size and the diameter of the weight "bullets" that would fit on each spoke.