Good point!

And looking at the curves (seemingly confirmed by my little test) the smoother the pavement the wider the band where changing pressure has VERY little effect.. so you pay little penalty for going for comfort, grip etc.. within limits of course.. also given this, it’s probably best to pick your pressure based on the WORST roads you will encounter... interesting 🤔

Different ways of looking at the same thing.

When a bump in the road pushes against the bike with a backwards-and-upwards force, that "backwards and upwards force" can be described as slowing the bike due to its backwards component. But visualize it in terms of where the energy is going: it can also be described as converting the bike's forward motion into vertical motion.

I like the latter description better when trying to describe the phenomenon, because most folks don't really know how to visualize force diagrams, whereas a squishier tire producing less vertical deflection in the bike+rider when it rolls over bumpiness can be fairly intuitive without much physics knowledge.

But the vertical force vector has zero effect on the forward motion/energy.. only the horizontal portion of the vector does... force needs to be in same plane to add or minus the forces.. if you rolled over smooth bumps (like a rollercoaster) you would not lose that energy.. would just go between kinetic and potential then returned as kinetic.. like bouncing on a spring.. no loss in energy though.. when you impact a road imperfection.. kinetic energy from you and your bike is lost to the environment though... heat, sound from force of each little impact in the opposite direction of motion...

I think there is one aspect that maybe we both did not have 100% correct but is noted in the Silca article on impedance. CRR is made of 2 components, one is casing losses (hysteresis) and the other impedance. The first part of the CRR curve is mostly hysteresis casing losses and declines as you add more air pressure, then as a certain point the impedance effect overwhelms the effect of hysteresis in the opposite direction and becomes the dominant portion of the CRR. So how I understand this is as the casing becomes more rigid there comes a point where the tire does not deform adequately over an imperfection and suddenly due to this the force vector points backwards (probably an over statement but you get my idea) up until that time the force vector was mostly upwards because the tire could deform adequately to overcome the imperfection. This, in my way of seeing things, is why a supple tire have a lower CRR and vice versa for a stiff tire casing. It also shows that the impedance effect gets rapidly worse because of the stiffening of the tire due to tire pressure.

So below the break point CRR is mostly caused by tire deformation losses (what I believe is called hysteresis). Above the breakpoint pressure the dominant effect is impedance. The impedance effect is much stronger for changes in pressure above the break point as shown by the slope of that portion of the curve. That is why the vibration concept that Flo is pursuing makes sense. Up to the break point the vibration is absorbed by the casing of the tire, but once the tire reaches the break point the tire gets more rigid and the vibration is not as well absorbed but is transmitted through the system. They have been able to show this already with the data they have presented. It will be interesting to see what they learn using this approach.

I think this is early days in this study and there will be new items added but it is interesting that Silca and FFT do not add a tire quality aspect to their pressure calculator. Based upon this it seems that tire volume is more important that casing quality?

Yeah... I can see the vibration thing working.. depends on how implemented but, I can see it..

Can see cases where

1. the tire is soft enough to totally absorb all the force from impact, deforms and then returns energy as expands with some small losses from hysteresis (essentially heat) = no vibration
2. the tire is completely rigid and system absorbs all the force = STRONG vibration
3. all ratios in between.. where amount some amount is absorbed and returns with hysteresis losses.. but some force of impact make it through to system = magnitude of vibration dependant on how much tire absorbs based mostly on the pressure

But I think the above system could tell you how well the tire is doing at absorbing the bumps... but wouldn't let you know how well it was doing with hysteresis.. so this would isolate for impedance, but not account for rolling resistance.. so how much energy is lost in heat from bending the tire carcass..


FFT actually does let you enter the Crr for the tire which you could get at e.g. https://www.bicyclerollingresistance.com

The circumstances in which the horizontal component exists are relevant. If you're riding steady and straight along a flat road, and the force from the road onto the wheel is only ever pointing perfectly straight up, it means that you're on a perfectly-smooth road surface and you won't ever be bouncing at all.

When you hit a bump, the thing that causes the rearward-oriented force component is the same thing that causes the vertical component of the force to exceed gravity and cause the bike and rider to accelerate upwards. You can isolate them in a force diagram if you want to, but they're not really separate physical phenomena. Again: visualize the situation in terms of the energy. If you don't think that the energy that bounces the bicycle and rider with overly-stiff tires on rough roads is coming out of the forward motion, where is it coming from?


The rollercoaster example is actually a really good way to visualize what I'm talking about.

If you're rolling up a steep slope, the force from the rails onto the car has a strong horizontal component, and forward motion is being converted to vertical motion very rapidly.
If you're rolling up a shallow slope, the force from the rails onto the car is only weakly horizontal, and forward motion is being converted to vertical motion only very slowly.

The reason that you regain the forward motion lost to vertical motion in the rollercoaster case, but not for the bump in the road, is that the rollercoaster rolls back down a second smooth slope later.
What if the rollercoaster doesn't do this? Imagine a rollercoaster than rolls up a 1-foot rise, and then simply flies off the rails onto a set of unsloped rails that's one foot lower down. When it lands, obviously there will be a jarring impact, and anyone in the cart will suffer a painful experience. But you won't get any speed back: the energy that was used to go uphill has now been wasted, dissipated through mechanical friction by rattling the cart and the riders' bodies, and in the form of a loud sound as the cart hits the lower rail.


Some friction obviously happens all throughout the process. But friction requires relative movement in order to happen, and a lot of the energy that gets "lost to heat" is initially converted to vertical deflection of the bicycle and rider. How do you dissipate energy as heat in the rider's body? Shaking the rider up and down is one way to do it! From a raw performance standpoint, the relevant issue is that the energy doesn't have a good way to get converted back into forward motion.

Update... response from Silca

Seems FFT calculator does not do an real impedance modelling... except in a very rudimentary, implicit, approximate kinda way.. assume that a % drop... mm drop in tire from load with given weight has x amount of absorption.. also not incorporating real world data to feed the modelling...


SILCA velo "We are the only calculator that I am aware of that uses calculated impedance values.. so what you have here is the result of thousands of actual field test calculations to find peak minimal rolling resistance, and we then fit an algorithm to that to help interpolate the empty spaces.. so for example, we have more than 120 optimizations across 5 teams, 30 athletes and 4 tire sizes all done in the Carrefour de l'arbre.. so we then curve fit for 26mm tires, riders 165-210lbs system weight.. then we fit for 28mm tire data, then for 30mm tire data and so on. Each of these tests originally having the result of being the actual pressure used by that athlete on that day. Our data comprise more than 20 pro triathletes across 4 years ad 5 tire widths on the Queen K at Ironman and so on.. from that, we've built the algorithm to predict the likely peak minimum rolling resistance likely for your weight, tire size and surface. All other calculators that I know of use either % drop of tire on a flat surface for a given load OR height of drop on a flat surface for a given load. Both of these methods are good starting points to begin the optimization process, but are only indicative of static tire stiffness on a flat surface and have nothing to do with impedance or any of these other factors that prove out to be critical in the real world. Also important to note that most calculators by manufacturers are also constraining their data by the max pressure of the rim or tire being sold.. so many of them will take %drop and then just shift the results or limit the results based on these stated pressure limits.
I love your idea on % of surface per ride, we use this in our pro team optimizations and do plan to include it at some point in a future version of this calculator.
As for tire type, this has proven not to affect pressure in real world testing. Since the pressure is controlling the spring rate and the tire construction is controlling the hysteresis this makes sense, think of tire pressure as being the spring in your suspension and the tire construction as being the damper.. across small changes in damping, the optimal spring rate remains unchanged."


SILCA velo "This is because our minds tend to approach all problems as being linear.. so a 25% change in system mass should result in ~25% change in pressure. However, the tire stiffness and spring rate behavior is non-linear to tire size and is also non-linear to bump input performance...and also remember that the calculator is trying to find the optimal middle ground between impedance which is non linear AND casing losses which are also non-linear. This is definitely an area where the mathematical model has not caught up to the field testing work that has been done, so we aren't yet in the best spot to have tested a hypothesis detailing the specifics, but I can tell you that nothing about any of it is linear. This could also be a place where having the entirety of the data set based on pro athletes could skew the data.. they are all much lower body fat and higher muscle % than the rest of us which theoretically lowers impedance which would also theoretically increase the impedance pressure break point.. but again, this is the data set that we have."

SILCA velo "So this FastFitnessTips calculator is combining % drop and mm of drop data and does not have any functionality to predict impedance. This is the type of calculator that we used years ago to find a starting point to begin optimization and our calculator is based on the results of all those optimizations."

what you are not consider is that.. if you use energy to raise something up against gravity.. when it goes back down, gravity restores all of that energy you put into it.. there are no losses from vertical up and down...

If exert energy to ride up a hill against gravity... when I roll down the hill, I get all that energy back when I roll down the hill.

I addressed this in my post.

Moving downward through a gravitational field returns gravitational potential energy to kinetic energy. But it only usefully gets converted back to forward movement if you do something (such as smoothly roll down a slope) that redirects it to forward movement.
That's why I gave the example of a roller coaster cart that simply flies off of a drop and then lands on a flat section of rail below. The cart will accelerate as it falls, but it's only accelerating downward, not forward. And when it lands on the lower rail, its downward motion gets dissipated in an impact rather than becoming forward movement.
The same happens to a bicycle that pings off of a small surface bump. When the bike+rider get sent upwards, there's no mechanism for that motion to get converted back into forward motion later: it ultimately gets dissipated as sound or converted to heat from frictions within the rider's body and the bicycle.

It's actually not so much what you imply above...it's more about whether or not the energy used to compress the leading half of the contact patch is returned in the trailing half, or not. If that energy gets "through" the tire and is dissipated as vibrational losses in the rider, then the "impedance" losses increase.

I wrote this article back in 2007 about the source of rolling resistance: https://www.slowtwitch.com/...ling_events_226.html

This article was written back in 2009 about "impedance" effects with pressure: https://www.slowtwitch.com/...in_a_tube__1034.html

Both would be good to review for this discussion :-)

---

http://bikeblather.blogspot.com/

HTupolev.. you are correct that the vector is diagonal and some force directed up and some back... this why to understand which forces normally one would split that diagonal vector into one facing straight up.. that you don't need to consider because that portion of the force just going to be returned as the bike goes back down and as the tire returns from being deformed... AND one facing backwards...


I drew diagram below a while ago to try to understand how/why a smaller wheel would be more affected by a given obstacle of a given size.. tend to roll slower as Silica velo mentioned in their video..


First you draw vectors from example a pebble through the centre of the wheel.. the length of the vector is measure of magnitude, so they are all the same length (same weight rider and bike riding at same speed)..


you'll see that the magnitude of the backward facing vector is larger and larger the smaller the wheel compared to the obstacle as observed in real life... ever hit a pebble with a shopping cart? LOL... can imagine if the obstacle was the same size as the wheel (brick wall) ALL of the force would be directed straight back..
[inline Wheelsize&ForceVectors.png]

Nice.. predates Silca blog entry by decade 😎

I do like the example of.. put a broom stick down.. roll up to it up to it very slowly.. it stops you moving forward.. why? because it exerts a force on the bike in the opposite direction of your travel pushing you backwards! That is impedance.

Immediately understandable by just about anyone.

Right, but if you just roll over it, the tire "pushes back" on the other side of the stick by nearly the same amount, pushing the bike forward...and the road roughnesses we're talking about are orders of magnitude smaller than a broom stick diameter. Rolling up to a broom stick is only doing ~half of the action of rolling over a "protuberance".

If you read the Silca blog carefully, you'll see that Josh mentions me ;-)

---

http://bikeblather.blogspot.com/

When we're talking about an over-pumped tire going over road irregularities, it isn't well-returned, because the contact patch isn't doing a good job of rolling down the far side of each irregularity. If it did, the backwards-pointing component of the force wouldn't matter either, because it would have a reciprocal forwards-pointing force on the far side of the bump and you'd just gain all of your speed back on the far side of the bump.

Back to your original point: "I get 25-30psi higher I'm 59kg 68kg with bike... seems EVEN WORSE for lighter riders... if I added 44lbs (20kg) to my weight the difference was 3psi 🙄 this cannot be correct!"

As noted at a very early stage of this discussion Silca has input more data into their calculation than the other calculators and why I tend to believe it to be quite accurate. I also assumed that tire casing did not have a big impact or they would have added this to the calculation, but the quote from Silica Velo is a good way to understand it. So over all they are trying to optimize tire pressure for performance and the changes are not huge for changes in system weight. That seems to mean that back to your original question why if you increase the system weight by 40 kg does the tire pressure only go up a few psi? it would seem that the increase in spring rate is non linear. So yes they got it right and FFT is not correct for the reasons pointed out. Yeah Silca!!!

The only things I would like changes are the type of road and proportion of road and more the actual speed ranges instead of a vague word description such as "moderate group ride) is that 27-30 kph or 30-35 kph????? where does the speed change? even then the changes are small ie recreational to moderate group ride may be 1 psi or less depending on proportion front to rear weight distribution.

In the end I know they put a lot of time, data from real word trials, and calculations into the calculator and I have no reason to doubt it. Josh has done this stuff for a lot of real world situations where it counts to be right. Based upon the fact he is still in high demand and I do believe he is a very smart guy, I am leaning towards believing what he says is likely fairly reasonably correct. Now I need to read Tom's (thanks for the links TA) articles and educate myself on impedance and hysteresis. This thread has been a good place to learn a bit more. Thanks to the expert inputs.

This made me doubt the numbers.. but not really my concern... I really just, selfishly want to know what my numbers should be 😀 I am coming over to the dark side.. a little bit.. and having less doubts 😀

Most significant thing I think I've learned - I though @realbdeal had a really good point, when he pointed out that... if you look at the graphs (it's there but it didn't really register for me).. and the little field test I did showed it too.. that the smoother the road the flatter the effective Crr to rec. pressure curve before the breakpoint.. where there is a band of pressures that almost give the same Crr.. so from a real world practical perspective it's really if you're in this ballpark, within this band, you're good and there is really no performance penalty for a little above or below a precise optimal breakpoint pressure.. this is why I think in some way could show this that would be awesome... if you can drop the pressure 15psi (as shown in my field test) and be WAY more comfortable and that results in a 0.5W difference or no difference at all.. that would be good to know.. I would go for the 0.5W increase for the comfort gained by a 15psi drop in pressure... so if they can spit out the optimal then allow you to drift away from that and see what that means in terms of W or speed decrease that would be awesome from a practical, use experience point of view... of if they could just show the curve based on the things you've entered, then you could hover over the points of curve either side of breakpoint that would work for this use case...

[edit...I did also suggest a multi surface feature.. so could enter 50% smooth, 25% chip and seal, 25% goat track.. lol.. said they do that for clients and were thinking of adding to calculators...]

welcome to the Dark Side... LOL Josh actually has a Darth Vader sticker on his computer on the video about the calculator...I agree it would be wonderful if they did an interactive graph, that would be awesome. I agree with you knowing where you are on the graph would help make a rational decision, like you said, less pressure, slight wattage loss, and increased comfort, as well they could indicate the danger zone for wheel damage/ pinch flats, this would be ideal.

I did notice the mention when I looked back yesterday...

Hmmm.. but I would think that any bumps of the size that a tire can completely deform around are taken care of, really by rolling resistance and the losses are from hysteresis loses.. just ones that are not accounted for in smooth drum because the tire carcass will do more bending around all the cracks, smaller bumps and texture etc than accounted for on a smooth drum... but for everything that is only partially taken care of by tire deformation, there will ALSO be a collision that takes place and that collision will push that bike back... if the tire was completely deforming around the obstacle would you feel the collision? So some of the force is making past the tire and into the bike.. so we know that, that portion is not going to be returned when the tire recovers from deformation.. correct?

Yep...any energy making it past the tire is going to be mostly absorbed and not returned.

Here's the thing about roller testing...it's a GREAT way of determining the hysteresis losses from deformation and ranking tires. If you add in additional "roughness" to the test setup that increases the deformations, it's just adding additional hysteresis losses in a proportional manner, and the rankings don't change. This is why I'm not a fan of adding "roughness" to a roller test. It doesn't add any additional information.

---

http://bikeblather.blogspot.com/

Yup.. the rankings don't change.. but as you say the magnitude of the wattage changes.. you obviously have more experience with this than me, so I assume this to be generally true, is strictly true in your experience..?

Interestingly, I noticed that generally tire rankings don't change with pressure... BUT, some actually do.. some tires change their ranking based on tire pressure...

https://www.bicyclerollingresistance.com/...reviews?orderby=rr60

so wonder, could there be tire characteristics that allow it to isolate the bike better or worse from road.. and that don't vary proportionally with the its hysteresis characteristics...?

Wondering if you could create a rig where you quantify the hysteresis losses on a textured drum... but also quantify the isolation characteristics and therefore the non-hysteresis losses of the tire with an accelerometer and/or strain gage attached to the wheel itself.. then have maybe 4 standard textured drums that you could use to test.. then get a better TOTAL picture of how a tire is actually going to roll.. what pressure a tire rolls best at on given terrain etc..?

Interestingly it seems even the shape and size of the obstacle seems to impact the forces in non-tribal ways.. maybe even using a drum, through practical, might not realistically mimic real world closely enough for this type of testing.. maybe the on bike, black box type testing where you just look at input and output is best..? [..too many moving pieces..?]

https://blog.silca.cc/...r-is-stiffer/harsher

But, there has gotta be a better way of testing tires... what is done now is really inadequate and we know only captures about half of the story.. as some people were talking about for a decade already 😉

I'd be careful about drawing any conclusions based on values from over such a wide pressure range (especially considering the magnitude of differences shown, i.e. tenths of watts). Although that wide range might be interesting...I find it useful to compare the tires in a pressure range they'd most likely be used for the majority of riders.



That "characteristic" you're talking about is the air spring "stiffness", AKA "pressure"...so, if you need better isolation, choose a wider tire so you can lower the pressure.



I think you might be over-thinking it ;-)

The way tires are tested now IS pretty good...it's just that too many people don't have the insight to understand how it relates to "real world" conditions...and they keep trying to add in surface roughness, etc. to the testing.

It's actually pretty simple. Smooth roller testing give you an idea of the hysteresis performance on hard surfaces (smooth or rough). With field testing, one can then also find out where the "breakpoint pressure" of the system is...and then just make sure you stay far enough away from that for the given speed, surface, and load.

Now then, if you want to talk about how tires perform on dirt...that's a whole 'nuther ballgame...but, not impossible :-)

---

http://bikeblather.blogspot.com/

Weird... never heard that before 😉

So I've tried my hand at a Chung analysis... was fairly painless and I was already a user of Golden Cheetah for my performance management software and it's just included in there as one of the tools.

What would you suggest as a user friendly method for finding breakpoint?

The way I found it originally...VE ("Chung") analysis at different pressures.

But, THE most user friendly method would be to just use the Silca pressure calculator ;-)

---

http://bikeblather.blogspot.com/

Makes sense.. thanks so much for your input, it's been quite educative,

...and thanks for "overthinking" this 😉 we wouldn't have our current understanding without people like yourself tinkering around with this stuff and the out of the box thinking 🍻

Just came across this thing, figured I'd try it out and see what I was doing wrong.

Interestingly, it spit out almost exactly the same pressure that I'm using now. I've been using 100psi on nominal 23mm, 24mm actual width. calculator says 100.5 psi front, 103 psi rear.

Also, I wanted to see what it came up with for my race wheels, if we ever get back to racing. 21mm width (nominal and actual). If I pick Cat 1/2/3 racing - I'm up to a whopping 125/128 at my current weight (triggered the pinch flat warning, but not an issue for tubulars).

I really should get wider tubulars, but these are NOS, never even been glued yet. given that they are tubulars, I could run a bit lower pressure than that, probably 110-115???

---

Swimming Workout of the Day:

Favourite Swim Sets:

2020 National Masters Champion - M50-54 - 50m Butterfly