So, you haven't answered any of the questions put to you above. Do I need to repeat them?

Just start with why you think it reasonable, if you want to examine the internal losses in a system, that it is perfectly ok to use a model that uses imaginary materials that prohibit internal losses? How does one use a model that doesn't look at something to reach accurate conclusions about that very thing? I am all ears perfessor. Anyone with an answer can chime in here if they want.

---

--------------
Frank,
An original Ironman and the Inventor of PowerCranks

Well, what I would do is quantify the overall rate of energy production (release, really) using indirect calorimetry, then compare that quantity to:

1) the power that is generated by the limbs (determined using inverse dynamics), and

2) the power that makes it to the pedals (using either force pedal data or simply the power setting of the ergometer).

If you call the first quantity above A, the second B, and the third C, what I would predict* is that:

A > B ~ = C

demonstrating that the primary source of inefficiency is "upstream" of the point at which the limbs begin to move, and not "downstream" as you have continually emphasized.

Then again, what do I know? I'm just a "scientist"...


*Of course, this is a safe prediction to make, since it is what has been shown in the research literature.

Because the materials in question don't deviate far from perfection.* Specifically, neither joint friction nor limb bending represent a significant energy "sink" - if they did, models based on assumptions of perfection wouldn't be able to predict actual data with the accuracy that they do.

*BTW, if you want to drive a biomechanist mad, start asking questions about "floppy masses". ;-)

Again, how do you use a system that prohibits limb bending (or soft tissue deformation) to evaluate a system that does. One can eliminate friction, or material deformation, or anything else, in the model if one wants to try to isolate the various areas of loss. But, isolating an effect requires that effect be kept in the model. It seems unreasonable to use a model that eliminate every possibility of loss to try to examine the magnitude of the various losses that the 2nd law dictate must be there. Using such a model dictates you get an answer (zero losses) that violates the second law (even though your starting point may not. No one, except me, seems to have noticed this little "problem."

The purpose of modeling is to help one better understand the real world. Using an unsuitable model, as you and everyone else has here, has simply made everyone's understanding worse because it has come up with an answer that cannot be true (it didn't look at anything "real world") and violates a fundamental law of thermodynamics and yet, "everyone" insists it is right.
Very little is not zero. And, the data of McDonald suggests there are substantial losses, the unloaded losses are almost 300 watts under certain conditions. The smallest losses they encountered was about 70. You, and everyone else, have had to go through all sorts of mental gymnastics to make the data fit your unsuitable model (loading the chain suddenly makes the losses go away). It is bizarre to me that you would still be supporting this model for determining internal losses.
Ugh, which data is that?

---

--------------
Frank,
An original Ironman and the Inventor of PowerCranks

Well, unless your model allows you to specify the degree of "perfection" of your material one cannot know how much effect there will be. You simply think the materials are almost perfect, which in your mind means the losses must be small, so you have decided to use a model that ensures you get an answer that is close to your bias. Some science that is.

For a model to be good it should give results that are close to experimental data. Let's examine the results your model would give for a real world cyclist and compare them to the experimental data. We will use their unloaded data.

condition, MMF losses, experimental losses
low pedal speed, 0, 70 watts
high pedal speed, 0, 290 watts.

Hmmmmmm. Correlation doesn't seem to good here. Maybe the model needs to be "tweaked" a bit.

---

--------------
Frank,
An original Ironman and the Inventor of PowerCranks

The way you use any model (no model is ever a perfect representation of reality, but that doesn't mean that they are useless/provide incorrect answers).


As I said before, you shouldn't be looking at data for the cost of unloaded pedaling in the first place. Be that as it may, an increase in metabolic rate from ~80 W at rest to ~290 W when pedaling at 100 rpm against no load using 190 mm cranks is hardly what I would consider "substantial", at least given the task. (By comparison, my metabolic rate when TTing is 1200-1500 W.)

The predicted metabolic cost of "unloaded" pedaling is not zero.

You mean like this? <g>



(From Neptune et al., 199X)

Note: J1 and J2 are two independent criteria used to optimize the model by altering the timing of muscle activation, whereas "subjects" represents actual experimental data (n=12, IIRC).

That is what your model predicts it would be. A chain attached to a wheel does not require the wheel to be loaded. The KE is available to move back and forth, even when the wheel is unloaded because of the energy contained in the wheel. No different than the condition used in the lab.

A model, if valid, should work in all conditions. Einstein's model of the universe is expected to work under all conditions. So far Einstein has held up pretty well. Much better than the MMF model. One comparison and it doesn't look too hot as a predictor. Tell me again why this model is so good?

---

--------------
Frank,
An original Ironman and the Inventor of PowerCranks

Is that from the MMF model? I suspect not since it is referring to muscle activation and according to the MMF model, not muscle activation is required once up to speed. Can we restrict this discussion to the MMF model?

---

--------------
Frank,
An original Ironman and the Inventor of PowerCranks

Not "my" model, and no, it does not. That is, once again you are confusing physics and physiology.

Sheesh, Frank...haven't people been asking you to do exactly that for about 6 pages now?

You might want to put a PowerCranker on that graph and see how close the model comes. Something tells me there would be a substantial deviation from that predicted line. A good model needs to predict correctly under all conditions. That looks good for the two cases studied. Need a lot more to say it is a great model.

---

--------------
Frank,
An original Ironman and the Inventor of PowerCranks

Just above you said you needed to use that model to analyze pedal losses. I consider it your model. You may not have "invented" it but you are defending it and you are telling me you use it. And, it does predict zero losses for the unloaded condition. If not, how is that the case? Show your work.

---

--------------
Frank,
An original Ironman and the Inventor of PowerCranks

From that predicted line? Of course, because said line assumes that the cranks are coupled.


N=12, actually.


Then I suggest that you search PubMed for articles by Hull, Kautz, Neptune, Martin, Broker, Ingen Schenau, and/or any of the other dozens, if not hundreds, of biomechanists who have applied inverse dynamics to the pedaling motion.

Again, you are confusing physics and physiology: the little stick person pedaling with his/her perfectly rigid limbs and his/her perfectly frictionless joints in a complete vacuum has no muscles, but once set in motion by an outside force will pedal forever. OTOH, the model that Neptune used to generate the data in that figure includes 14 different muscles, each one of which has its own unique activation dynamics, speed of shortening, etc. While AFAIK it has never been used to predict, e.g., the pattern of force application or the energetic cost of "unloaded" pedaling, it will obviously not be zero as you have assumed, if only due to the "floppy" ankle joint.

What on earth are you talking about? Compare the predicted losses of the MMF model to those experimentally measured by McDonald. Defend your use of the MMF model to predict the energy cost of simply moving the pedals. The MMF model predicts zero energy cost. Experimental data measures substantially different values. Defend your continued use (and defense) of that model despite the fact that the data suggests it sucks as a model (at least for this purpose).

---

--------------
Frank,
An original Ironman and the Inventor of PowerCranks

I don't have a problem with inverse dynamics. I have a problem in using inverse dynamics in a model that prohibits internal losses to predict real world internal losses. The fact you can make the math work is not evidence the results have any validity in the real world. Defend the use of the MMF model in predicting internal losses in the real world. You have told us it is the way you and, apparently, everyone else does it. It is the model you used to "debunk" my claim that there are internal losses. Defend the basis of your criticism of my point. There is zero evidence to validate the model for this purpose.

---

--------------
Frank,
An original Ironman and the Inventor of PowerCranks

Inverse dynamics doesn't prohibit internal losses.

Now why would I want to do that, since they are two different things?

No, the MMF model that you are using does. Have you figured out your problem here yet?

---

--------------
Frank,
An original Ironman and the Inventor of PowerCranks

Correct, why would one want to use a model that really doesn't predict anything in the real world to predict real world results? Of course, I guess it could be useful to try to shoot down the opinions of others who are trying to comment on real world results before a group of people ignorant of the deficiencies of the model. Could make one look smart and important until the fact that the emperor is not actually wearing any clothes happens to get mentioned.

---

--------------
Frank,
An original Ironman and the Inventor of PowerCranks

Frank,
The MMF was modeled without losses. But one could model losses into it and obtain a MMF-cum-losses just by introducing a viscous element between two sticks. This is what tigermilk was trying to explain to you when he talked about modeling software such as ADAMS. Engineers (and researchers) do this every day: they keep adding additional features to their models to see how they behave.

---

Giovanni Ciriani
http://www.GlobusSHT.com

Let's try going back to the beginning...

This thread took off in this direction when the question was raised as to why efficiency decreases at some cadence above the optimum. As you have for many years now, you claimed that this was due to an obligatory loss associated with the intra/interlimb exhange of potential and kinetic energy that scales with the square of the cadence. The little stick figure with infinitely rigid limbs and frictionless joints pedaling in a vacuum was then introduced to demonstrate that your claim is based on an incorrect understanding of the physics involved. That is the sole purpose of this "thought experiment", i.e., it is not meant to reflect reality, and is not the model used by, e.g., Neptune.*

Ass u ming that you have now accepted that you are wrong, you should be able to at least appreciate that it is possible to accurately quantify the energy "flow" while pedaling based on fairly simple Newtonian mechanics. What such inverse dynamic calculations reveal is that of the power that is produced/absorbed across the hip, knee, and ankle joints, >90% of it makes its way to the pedals. Since thermodynamic efficiency is <<90%, this demonstrates that the primary energy loss occurs (as I have said several times now) "upstream" of the point at which the limbs are set in motion, rather than "downstream" as you have claimed.

*The data of Neptune that I posted are actually derived using forward dynamics.

A model is used to try to understand things that are otherwise difficult to understand and predict. We have models for the weather. Models for electronics. The only useful purpose for a model is to help one better understand the world around them and the things they work with. Bicycles and riders are made of real world materials. While a model that eliminates all friction and all real world materials might hold some intellectual curiosity (might even be useful as a "character" in a science fiction novel) it it totally useless to predict any behavior in the real world.

If you take the perfect MMF and add friction to the model one might be able to understand how much friction is contributing to pedaling losses. If they are found to be small then we can understand that there may not be upside to spending a lot of time trying to reduce friction losses (probably not much can be done about them anyhow). Or, we can take the perfect MMF and add material losses to see what happens. Then, we would know how much this contributes to the loss. Is there any benefit to making the bike out of a "stiffer" material or one that is less stiff but has more energy return (a better spring, so to speak). Or, if the losses vary with the square of the cadence, how much can be gained from reducing the cadence. Put these two models together and one should be able to predict the McDonald data pretty well if the models are any good. If these data cannot be predicted well then either one or both models need to be improved.

What is simply amazing to me is that all you very smart people seem to not have a clue as to how to model anything around the bicycle and to how worthless the perfect MMF model is to this discussion, which is about real world losses, and how it was being used to attack my point. It is like you are all hell bent on proving whatever I say wrong, regardless of how it is done. (If I am ever allowed to be right about anything it might mean I could be right about PowerCranks, I suppose, is the thinking.) People kept at it even though I pointed out that their solution violated the 2nd law. Unfortunately for you folks, I am not cowed very easily when I am pretty sure of my position. I think I have been proved right here and several of you owe the ST community (I could care less about myself) an apology for forcing this fallacious thinking down their throats for the last several years, whenever this came up.

---

--------------
Frank,
An original Ironman and the Inventor of PowerCranks