Frame flex and climbing

I LOVE this quote!! I've seen bicyclist do all sorts of crazy things and spend all sorts of crazy money to save a few ounces here and there. I always thought if I lost 5 pounds it would be much better, cheaper and healthier for me! When I ride, I kinda do the opposite. I take EVERYTHING. I have an aero bag behind my seat (Bacchetta Strada) and it is filled with a bicycle pump (versus CO2 cartridges), full repair kit, 2 spare tubes, latex gloves talcum powder, power bars, lots of water rain jacket and a camera. My bag probably weights 10 pounds but I am happy to tote it along because my goal is to enjoy the ride not necessarily go fast like a scalded dog. And as long as I am on level ground, I really don't notice the difference. (thought I am sure there is some power difference.)

 

Bold emphasis is mine. This is very True. My gravel bike (Burley) weighs a ton & I don't care.

In a BROL thread, sometime in the past 2-3 years, someone mentioned why weight matters to them and after many years of having people show me the charts and explain things in terms a 3 year old could understand, it clicked. The science says that for every X pounds I'll be Y seconds slower on a specific climb (same power output). In those explanations, the time mentioned always seemed trivial. And it is, as long as I'm riding alone or casually with friends. Where it isn't trivial is when I'm trying to push above my pay grade on a group ride of vipers who love to watch folks get dropped. On those rides, the 'trivial' seconds matter - I need every benefit I can possibly find or I'm doing the remainder of the ride solo. When the other rider mentioned this in that thread, it was a major light bulb moment.

 

I probably would stop short of carrying the talc.

It's hilly around Knoxville, isn't it?

Congratulations for reading all of it, or at least as far as that quote.

 

Well, as I think I have said before, I am all for taking weight off the bike so long as nothing more important is compromised significantly.

 

I must have accidentally scared TymberMan away.

 

No, I've not been scared off (yet), Steamer. Your last post gave me a lot to think about. During my century ride yesterday, as I ground endlessly up the, some might say imperceptible, hills I pondered the issue. First, let me say I appreciate the time you spent in your very clear answer, and the thought you put into the mechanics and physiology involved in frame flex and climbing. If you are not an educator, you should be. You have obviously read and thought alot about the problem, and what you wrote makes complete sense to me, at my level of understanding. Though I'm not a mechanical engineer, along the way to becoming a dentist I've had to learn a bit about physics, metallurgy, yield strength, beam deflection with force and such, as well as a fair amount of physiology. The nub of the issue, is, as you said:

This would be ideal, but unless we get Elon Musk to fund it, it's probably not going to happen. Though who knows. Maybe if we get him to try a recumbent he'll get into it. In short order, SpaceX will be producing the recumbent of our dreams. I digress. Given that the lab tests you describe aren't in our foreseeable future, can we still manage to move the knowledge needle a little forward? Sure, we can ride a bunch of bikes and see how fast we ride them as you suggest, and I would love to see people post those numbers. But I think we can get do better than that. Specifically, anyone interested can do A/B trials that attempt to factor out many of the confounding variables you listed. Here's what I plan to do.

There are places in Florida that actually have hills. I did a century in Clermont Fl that had 4,771 ft of climbing. Unfortunately that's about an hour and a half from here, so not great for my plan. The best I can do is the highest local bridge that has about 70' of vertical at a 6-7 percent grade. I can load two bikes in my van and do a few test runs. I will start by comparing the Carbon Aero with the Mystique. Both have the same tires, and importantly, the same seat and cushion. They currently have the same seat angle, but for this test I'll need to equalize the hip angle. The gearing, is the same I think, but i'll have to check. While this won't isolate frame flex as the only variable, we can at least say something about what each bike as a whole does with net power. Now what about gross power? Well, I'm going to ride both bikes, so the engine will be the same. I will keep each bike in the lowest gear, and try and keep my cadence constant, so speed should be constant-ish. Torque and power should vary between bikes if speed is constant. Other trials can be done with higher power values or even max sprint, but this is a place to start. So if speed is the same, and slow to minimize wind, how can we compare the gross power produced to achieve this? One way is perceived exertion, which I can easily record (although those levels have always seemed a bit fuzzy and subjective), another is heart rate. I'll need to read up more on how reliable HR is as a measure of energy utilization. But I'm not looking to calculate a precise calorie consumption number, I just want to see how comparable an effort it takes from bike A to bike B. And maybe heart rate is not the best metric. Maybe I need to use impedance plethysmography. I need to do more reading and I'm up for suggestions.

If I do multiple runs, say doing A-B-B-A one day then B-A-A-B the next then A-B-A-B etc, some of the variables related to fatigue, energy source utilization, etc. should be pretty well controlled, as will the tendency of heart rate to rise with exercise duration. I should also standardize my breakfast. Time of day standardization is easy. In Florida, you go early before it gets hot.

One of the biggest problems with my proposed test is that the climb is so short. The test method may be pretty sound, but it may still not pick up differences between bikes even by summing repeated runs. In medical terms the test may have precision, but lack sensitivity. But who knows? Until we see some numbers we won't know. Another test I'd like to run is comparing my Bacchetta Bella seats. I have a RANS Sling Mesh seat, which is very comfortable for long tours, but is about as squishy a seat as you can find. I've replaced it with a Thor carbon seat. Those would be fun to compare. These are projects that will take me a long time to complete.

Now would I, after all this, reach "universal truths"? Probably not. We could say nothing about HCR (hill climb rating) as it relates to hair color, ethnic origin, basal metabolic rate or a million other things. As you say, the results would be individualistic - I can maybe make conclusions about old, fat Italian men. But we would learn something. Compare this with what the NHTS did with crash test ratings. If you break down the physics and mechanics of surviving a crash, it is impossibly complex. Every crash is different, and an analysis of one crash has only limited applicability to another scenario. Still, they started with something: a frontal impact crash. They learned things, made recommendations and cars improved. They then added side impact, rollover, partial frontal, on and on. Now cars are so safe, we have all ceased worrying about crashing and can just text on our cell phones instead of paying attention to driving. If we are successful in creating HCR's, we will end up improving the climbing ability of recumbents so much that any of us can attack an Hors Categorie climb without even working up a sweat.

It would be great if there were a way for folks to compare their findings of similar comparisons in one location, so we can benefit from the sum of our efforts. I'm thinking of how Weight Weenies (which seems not to be active now) allowed people to list component weights. A whole lot of information was quickly assembled. I could make a shared Google Sheets spreadsheet to list our results, but there may be a better preexisting solution. A lot of information would need to be included about each trial to be informative. Not to mention that there is probably no interest in going through this nonsense, except for nutcases like myself.

 

The protocol you described is pretty much exactly what I had in mind. Although I wouldn't try to control cadence - that should probably be freely selected. Frames of different stiffness may induce different natural cadences. Lots of studies show that experienced riders naturally select the best cadence, or at least close to it, all on their own.

And conducting short, maximal efforts in lieu of trying to control by RPE or heart rate should help eliminate unintended variability due to pacing error. All out is all out. The higher power output should also help reveal differences in frame flex that affect performance. Find the steepest hill you can. It would best if the chainring and crank length used on each bike were the same size.

But you already see a problem with comparing different bikes in a generally applicable way - between the CA and the Mystique, there is no way to have both the same body angle and seat angle. That sort of variable will have to be accepted. I suggest holding seat angle constant rather than body angle.

There are probably some good statistical methods that could be brought to bear that would help as well. A guy like Robert Chung would be the ideal person to ask.

The trouble with crowd-sourcing this type of data lies in some people's stupidity. I would ignore results from anyone I didn't know the capabilities of in terms of being conscientious to perform the tests in consistent manner. In this kind of test there are a lot of ways for people to screw stuff up. Then you have the fan boys who want to see a certain result who can't be trusted. I think no data is better than bad data.

Anyone can use a digital scale to weigh bike parts. And there aren't as many egos involved in such a simple task. That's why it works for WW.

By the way , a crank/pedal power meter would be useful if one does maximal tests. Look at the average power produced, and not just the times to the top of the hill. This is especially important if the hill is not particularly steep and aero effects will creep in.

My guess is that the tests you are proposing between the CA and the Mystique are not likely to be super-revealing. Both bikes are likely quite stiff. They might not be much different from each other. I have test ridden a CA but not a Mystique, so this could be wrong. If a quick and dirty flex test like I talked about earlier produces similar results, then don't expect lots of interesting and conclusive data.

I suggest you just work on answering for yourself the questions you want answered and be satisfied with that. That is exactly where I ended up when I, very similarly, was trying to get people to do Virtual Elevation tests. That was about 10 years ago.

Expect a lot of people to push back and say your test method here is flawed or even that it's un-scientific, and I guarantee every one of those will say something about using a power meter. Some people think all questions can be answered with one. A lot can, but not all. This type of approach we are talking about, for it's flaws, is still a nice way to capture the effects of all performance factors of a given bike / platform (both demand and supply side of the power-speed equation). And it's something that anyone can get their head around. It's also very accessible. Just about anyone has a bike computer, and it seems that a lot of people just aren't interested in doing much more than simply going for a ride with a Garmin bike computer turned on. Things like VE tests, for all of their great value, only look at the external demand side. And power meters don't usually directly tell you anything about how the bike affects the athlete (outside of maximal tests). It's as if people think a given athlete can produce the same power on literally any bike they are plopped on top of. Now, both crank/pedal and a hub meter running at the same time, if accurate and zero'ed beforehand, can tell you about elastic hysteresis and other internal demands in the bike, but again, I suspect that's pretty small. The little bit of (seemingly flawed) data I have seen from that sort of test suggests that its small, and the result makes engineering sense too. Part of the problem is that the magnitude of the losses isn't that much bigger than the sum of the error of the two devices used to generate the result. But that's still a good type of evaluation to do (why not?) - it's just not even close to being the whole story, I suspect.

 

I realize I wasn't totally clear in what I was recommending. Instead of holding power or speed constant at a relatively low level, and trying to measure or rate gross power by RPE (perceived exertion) or heart rate, I was suggesting that you rather do the tests at a maximal effort (in a sense, holding RPE constant, but at a level that's easier to repeat - all out effort) and measure speed and power. If the hill was steep, speed/time would be as good as power, but since it's 6 or 7 percent, you might be moving a bit too fast to use speed (aero differences would mess with the results), so net power is a better measure. Speed/time would still make for a good comparative metric, but you'd have to recognize that it was a combination of different performance characteristics producing the speed / time result.

The idea here is that maximal / near-maximal efforts will improve the resolution in the data - amplifying the size of the relatively subtle difference you are trying to measure, and not have other variables pollute. (signal to noise ratio type of concept)

Heart rate is slow to respond to a change in effort, and that may make things tricky. It would be hard to consistently "reset" your pulse to a baseline value before the start of each run. Noting the values can't hurt, but I tend to think its a bad metric. Heart rate is also influenced temperature. When body temperature rises, pulse rises relative to power or RPE. I guess that's the body trying to move blood at at faster rate to extremities to liberate heat.

If you do maximal tests, It seems you'd you need to spend a good 10 minutes or more between runs riding very easily.

ABBA and BAAB are good. I am not sure I would throw ABAB and BABA in the rotation or not. Seems like more work switching bikes that way - switching 3 times instead of 2.

 

All right. I like that approach. I'll try maximal effort, gather power, speed and time data and see what it looks like. And see what it feels like. I do some interval training, and it sure isn't fun. It's hard to really do "maximal effort", at least for me, if I don't have a pit bull chasing me. But this will be a good excuse for some training I guess. It seems like a good part of what your maximal effort is is your ability to tolerate pain.

Now that you point out the mechanics of ABAB, where I have to not just switch bikes, but switch and re calibrate the pedals, I think I might try it this way: Go up the bridge, stop the ride at the top on the bike computer: A. Coast down the other side, rest, then restart the ride back up for A or maybe A'. Depending on how tired that makes me, maybe do it again. If it is really hard, rest and switch bikes. So A,A', A,A', B,B', B,B' or A,A', B,B'. That way I only have to switch once. Next time I'll start with B. I'll post this when I do it. I haven't ridden this bridge, so the bike lane might be crap, or there is some other reason that this won't work at all. But it's a start.

I see your points about using heart rate. Still, it will get recorded so it's worth looking at.

My first power meter was a G3 PowerTap rear wheel. I was just getting into going faster, and got a beautiful, almost unused Gold Rush, expecting it to be a big step up from my Bella. I got the G3 hub and enjoyed using it to train and push myself. As I got other bikes, I tried to use the G3 wheel on them, only to find, to my surprise, that there are multiple dropout width standards. They Gold Rush has, I think, 170mm, which made it different from all the bikes I then had, and subsequently bought. I got the Assioma pedals, converted them to SPD cleats, and they have been great, since it is pretty easy to switch them. At least the crank arm/pedal thread pitch and direction standards have not changed over the years. One of the first things I did was to compare the Assioma to the G3. I was excited to see if the G3 would show around 3% less power than the pedals due to transmission losses. I set up an account with DCRainmaker. He has a program that lets you compare simultaneously recorded data. I set up a Garmin to record the G3 and a Wahoo to record the Assioma and plugged the files into his program. I don't have the printout in front of me, but I was surprised to see that the G3 showed around 6% HIGHER power readings for most of the range than the Assioma. That jumped to 11% peak power difference. That just reminded me that I have to disregard all my old ride data, or at least couldn't compare it to the new. It also meant that my meager power output just got meager-er.

I'll see if I can attach the file,

 

YES!!! But I mostly ride by myself and I have a granny gear on the front chain ring to help me out.

PS I read your WHOLE post though I won't say I understand it all. Above my pay grade.

 

Did you manually zero the PT hub before the ride? http://thebikegeek.blogspot.com/2015/03/how-do-i-calibrate-my-powertap-g3-with.html

I think there may be a method for actually calibrating the hub based on hanging a known weight off the crank with it positioned dead level, but I am not sure if that's actually possible, or how to do it.

 

When I switch the Assioma pedals between bikes, I calibrate them using the Assioma iPhone app, which makes it very easy. I just sold the hub, so I don't have the PowerTap instructions here any longer, but it seemed that PowerTap recommended spinning the wheel backward, or coasting and the hub would auto-zero, so that is what I would do. Now that you mention it, (and now that I've looked those links up), I do remember reading that there were other tests to truly calibrate it. Given that those methods are complex and obscure, I got the impression that auto-zeroing was all the manufacturer was suggesting you do. If I was going to rely on both those power meters, I might have put more effort into calibrating them. As it was, I just wanted to be able to compare my previous rides with the PT on the Gold Rush, with the Assiomas on other bikes. My initial rides with the pedals seemed like I was generating less power on the other bikes than on the Gold Rush. My max power was down, as was average power. I initially chalked it up to less power in more laid back positions. It took me a while to find DCRainmakers program which allowed me to truly compare the two, and even then I couldn't compare them until I bought a new Wahoo to replace my wonky old Garmin. But once I had two working head units, I paid the few bucks to run his software and it confirmed that either the PT is optimistic, or the Assiomas are pessimistic about my power.

 

Zeroing is good practice, and is normally good enough when you are talking about a single given power meter, and you are only worried about relative results, as we are on the proposed climbing tests. In other words, precision matters, but accuracy doesn't.

As you can see if you wanted to use a crank/pedal meter and a hub meter in tandem to try to determine internal demands (losses), both would also need to be accurate, not just precise. That's where a true calibration by imposing a known torque would be needed, should one not have faith in the meter(s). In your case, as you essentially noted, you had an impossible situation where the hub registered higher power than the pedal meter. Clearly, at least one of the meters was off.

 

If I ever get strong enough to flex my frame it will probably be from losing 10 pounds and getting those elusive "bent legs". When I get there I'll be one happy camper.

Cheers, MAC

 

Tom wrote:

By the way, here is one theory about why a flexy frame messes up your ability to put power into the pedals during high torque efforts.... The pause between muscle contractions is critical for blood to nourish the muscles. Blood flow is impeded by muscle contractions, and fuel is delivered and wastes removed during the pause. When you have a stiff bike, you can use a strong, concentrated contraction to add power, and you don't need to maintain pressure on the pedals as long to make sure the wind up in the bike gets redirected into the chain once the pedal force is reduced, later in the power stroke. This results in a longer pause that allows the muscles to be better nourished, and get ready for the next contraction. The opposite happens with a flexier frame.

I have been thinking about this theory booth as I ride current bikes and recall bikes that are not ridden as often. The perceived effort on stiffer bikes is much less than the more flexy models. I ask whether one would include seat and all related attachment hardware when discussing frame flex.

Thanks for a superb write-up.