But maybe they did, and either didn’t bring it up (horrible for them), or were ignored by GCN (horrible for them). There is enough to go around to embarrass everyone, but in the end the facility they used could just say ‘We warned them, and they wanted to continue’.
Running a business, I remember when our accountant inserted a disclaimer (CTA) note that the data used in these included documents are derived from information provided by our client. We make no representation to their accuracy’. But, um, they combed through our records every quarter. If there was anything wrong, they would have discovered it I’m sure. So if they thought they could hide behind that statement, it’s not at all likely to stick. Sure, ‘garbage in, garbage out’ but they went through the container and pulled it all out. But anyway…
This has been, on one level, rather hysterical. And it took them how long to retract it?
I use carbon soles because of comfort. Plastic soles give me hotspots both in XC and road setups. For power losses to occur one would have to effectively compress the sole material (I just don’t see this happening) or cause some bending of the sole when pedaling (ok, maybe if one uses a very rearward cleat position, bridging between the ball of the foot and the heel).
I have very low expectations for GCN “tests”, but, like others mentioned, my biggest question mark here is how Silverstone let this one go through. They either didn’t spot the protocol error (awful), or didn’t care at all about what GCN was doing and provided no guidance (bad) or knew about the problem but still let GCN publish their content (bad). I don’t see any scenario in which Silverstone comes out of this in good light.
because some people were mentioning AI here, just for fun I suggested to ChatGPT their test protocol and their objective. The systems first answer to the prompt was positive. With a second prompt I questioned it if the protocol would indeed work since the first power meter was after the sole. ChatGPT proceeded to give me a comprehensive answer on how that wouldn’t work and suggested some different test setups. So AI is not really there yet but very close
I like GCN, and I really miss GCN+, but their notion of “testing” is laughable. It’s a shame, because they do tackle some interesting questions. With only a tiny bit more effort, they could be providing some useful information.
Yeah, it’s bad enough now that if people put them in charge of anything that can hurt people, there will be people getting hurt. The early AI generated images were some level far removed from laughable. What’s worse than horrific?
The problem with AI is that there are people that are so out there, they will ride AI into a wall, turn it around, and keep going, and that’s the problem. Stephen Hawking was right, AI IS a clear and present danger to humanity because most of the people that seem so dead set that it is THE answer are so ‘damaged’, and could be argued aren’t really human, or humane. shrug And like programming in general, the product is only as good as the people that created and ‘taught’ it. If you aren’t at least concerned about AI displacing human beings yet, just wait.
And there already is a company named SkyNet. They aren’t in AI. Yet…
No, Im not saying to build a power meter into the shoe. If you simply measure sole deflection and have an accurate pedal power meter then you can get the hysteresis value of the sole and calculate power loss. You dont need to worry about the material of the sole, youngs modulus, etc. You also get a direct value of sole deflection which is interesting on its own. An analogy would be using a ruler to measure the deflection of a coil spring. In this analogy you put a scale on the output of that spring ((pedal power meter). So even though the force input into the spring is unknown, you can still calculate the power losses in the spring.
Can you provide more details? (I have a background in physics, so no worries to use formulas.)
I think I understand your point in a static situation, then you can compare the deflection to the measured force at the power meter pedals. But I don’t understand how that’d work in a dynamic situation, because the static picture neglects (and can neglect) kinetic energy.
Wouldn’t an energy balance work better and more reliably?
All springs behave according to Hookes Law which is F=kx where F is force, k is your spring constant and x is spring displacement. Normal high school physics right. The sole of the shoe is also a spring and follows Hookes Law. But in a shoe, the spring isnt a coil but just bending like a leaf spring. In a perfect world it wont matter how springy your sole is as all the energy is perfectly returned. But in our case the spring isnt perfect and there are losses. So as you deflect the spring you will measure its bending (x). You will also measure the force coming out of the spring (F) (powermeter pedal). You will plot that force against the bending. In that perfect world this is a straight line but because of the losses in the spring you will get a slightly different line during compression and expansion of that spring. The difference between those compression and expansion lines is your losses.
On your energy balance question, Im not sure how you could measure power from the rider into the shoe with enough accuracy. I may not be following your question tho.
In addition and worth consideration, high end carbon soles also have better heel/foot support as opposed to a flexible heel cup. Described well at 7:40 in the below video…
There are two flaws: this only works if the situation is (on the relevant time scales) static. In this case it is now. Moreover, what you are describing is a way to measure Young’s modulus (= the spring constant), but I don’t see how that would give you the losses. Those come from “friction”.
I guess what I am saying is that I don’t see any easy way to measure the losses.
To get where GCN thought they were going, they have to control way more of the process, and make some concessions. The thing to do seems is remove the whole ‘riding on a bike’ idea, and use a ‘last’ to push down on a static pedal in a fixture that eliminates rotation. The force could be measured on the ‘foot’, and the shoe, and the foot and insole/outsole could then be heavily instrumented to accurately determine any flex in the system. You can’t accurately measure the output of the human foot without hurting the owner, but you can ‘fake’ it and more accurately measure deflection of the shoe.
Such a jig to hold the ‘pedal’ could be used to compare pedal brands too, maybe finally answering the question: Should SPD pedals be used for road riding/racing.
Oh, one recommendation from my in-house engineering staff: use piezo strips attached it to the outsole of the shoe with a last or a human foot. My idea was to substitute strain gauges also attached to the outsole. And another idea is shoes being loaded in a rig that would take images of the outsole as vertical pressure is increased and measure the deformation as it happens. Testing the shoes for flex should be a lot easier than testing the flexion in an active environment with a real human foot. I worked with a company that made industrial fixtures. I’m sure they could come up with something that would answer the questions GCN couldn’t.
