Index Home About Blog
From: (Jobst Brandt)
Subject: Re: Rolling resistance....Tires
Date: Sun, 4 Feb 1996 04:12:56 GMT

Jim Papadopoulos writes:

> The simplest, but not the only reason for energy loss, is that the
> work done in climbing a protuberance (which slows the vehicle) is
> not returned when descending the other side, if the wheel had enough
> upwards velocity to 'leap' the crest. Upwards kinetic energy is
> converted to heat and dust when the rigid wheel crashes to earth.

You have created an unreal model that is about as valid as those who
argue that the bicycle wheel hangs from the upper spokes, in spite of
there being no increase in tension when the wheel is normally loaded.
Your model of jumping wheels is equally unreal and assumes a non
randomness in the road, something like a saw toothed road in resonance
with the passing tire.

> In addition, when it encounters the small bump, there is a mismatch
> in velocity at the contact point, which means a collision. These can
> be partially inelastic for the steel wheel, though not for the
> high-pressure tire.

Maybe if you considered why botts dots highway lane markers have such
a jolting impact on the passing tire you would also see other effects.
You may have noticed that these lane markers occasionally break and
are dislodged by passing rubber tires.  There are forces here that may
not be apparent and these arise from the change in rolling radius of a
softer tire that is greater than that of a harder one under the same
load.  This means the wheel must change its rotational speed abruptly
when passing over the bump.  For this reason, these porcelain smooth
dots have black skid-mark rubber on them.  This constitutes a large
loss of energy.  Try it at various speeds and you'll see that there is
a hole in your model even for the circumstances you cite.  Soft tires
at speed do not roll more effortlessly over Botts Dots than hard ones.

Jobst Brandt      <> 

From: (Jobst Brandt)
Newsgroups: ba.bicycles,rec.bicycles.misc,
Subject: Re: Where to find a good selection of 27 inch tires
Date: 26 Oct 1996 23:44:29 GMT

Sheldon Brown writes:

> You are unlikely to find center-ridge road tires in any size...the 
> center-ridge design is a fad that came and went.  Although it does
> provide improved tread life, it degrades handling in corners, so 
> it has been swept into the dustbin of history.

The center ridge tire by Specialized was made under the belief that
rolling resistance was caused mainly by tread scrubbing on the road
and that reducing the contact patch would reduce rolling losses.  In
fact losses are mainly caused by rubber deformation in the tread, tube
and the inter-ply rubber matrix, not by road contact.  The raised
center ridge increased all of these losses besides decreasing road
contact, and therefore, traction when cornering or braking.

The tire wasn't all bad.  It was extremely durable in contrast to
previous clinchers, that burst at the slightest stress or off road
use.  It was the Specialized Touring II that displaced tubulars as the
mainstay of nearly all high performance bicycling and was the advent
of the clincher as we know it today.  It together with Super Champion
who supplied the E-3 rim, known today as the Mavic MA-2 were a
revolution in tires.  I am grateful for them.

It was the Specialized Touring II tires with their raised center ridge
that encouraged Avocet to go into the tire business and offer a smooth
tread tire.  These tires have about half the RR of the old Touring II
for several reasons, one of which is the lack of tread squirm because
they have no voids into which the tread rubber can deform.  Another is
that the casing has finer cords with less rubber between them to
absorb energy by flexing.  In addition to that, these tires introduced
a new rubber compound that doubled the wear life while improving
traction.  Today Specialized offers similar tires and will most likely
never revert to the raised center ridge or the older tread compound.

> No doubt somebody will re-invent the design in the 2010's, and it 
> will have another flurry before people remember why it was discarded.

> Sheldon "There Is No New Bicycle Technology" Brown

That's because there is so little written technical history to go by
and many bicycle "inventors" don't believe in reading, or at least
reading about what has been done.

Jobst Brandt      <> 

From: (Jobst Brandt)
Subject: Re: tubular tires
Date: 19 Mar 1998 18:34:45 GMT

Mike Sullivan writes:

>> I have posted the rolling resistance tests done by IRC

> I'm sorry but I am not familiar with these tests, is it possible to
> repost them?  Where were they performed?  Whose facilities were used?
> What tires were they tested against.  And when were they performed?
> Curiosity.

You can have the data, but I should explain what it shows.  This
information is contained on the data sheet but may not be obvious.
The tires were mounted on the same rim with the tube offered for that
tire by the manufacturer.  Tubulars (2) were glued onto the rim with
standard practice using tubular glue.

All tires were run against a smooth steel drum with the same load
at a range of inflation pressures.  What is shown by fitting a power
curve to the points is that the data is consistent with little scatter,
and that all tires fit into a family of rolling resistance curves that do
not cross and are essentially the same equation.  The exception was the
tubular tires that were offset by a constant from their proper curve
so that they crossed other tires.

This constant offset was causes by rim glue losses that depend on load
alone and are not affected by inflation pressure.  Therefore, tires
with low rolling resistance don't get much better the harder they are
inflated and have a flatter characteristic over the range of inflation
pressures while high rolling resistance tires have a steeper slope,
getting better the less they flex but never as good as a low RR tire
except toward infinite pressure where none would have any flex or RR.
Of course these tires all had smooth or nearly smooth tread patterns.

>> Both tubulars and clinchers are completely round in cross section in
>> the portion free of the rim.  There is no way of causing any other
>> shape with a uniform bias ply casing.

> I'm not sure if I understand this, if a portion of the clincher is
> mounted inside the rim, and a portion of a tubular is mounted on the
> rim, then where is the cross section measured and I assume when you
> say round you mean "literally" round as in an equal distance as
> measured at all points of the inside and outside of the tire when
> inflated and with no weight applied.

The casing is a thin walled fabric that supports the tensile stress of
containing the inflation pressure.  THis casing is circular in cross
section wherever it is not in contact with the rim for both tubulars
and clinchers.  Where the tire is in contact with the rim, it has no
effect on tire compliance of supporting a wheel load.  Therefore, that
open face of the clincher as the underbelly of the tubular have no
effect on tire deflection.

>>  Avocet TT, which is based on the Criterium

> Are these tires made by CST or National?  This is not a knock, just
> couldn't find one to check myself.

Avocet tires are made by IRC.

>> Therefore, tires like the Avocet TT, which is based on the Criterium
>> but with a thinner tread, has the lowest RR and is the most tubular
>> like of these tires.  If it were fitted with a thin walled tube it
>> would not only be better than a glued on tubular that it already is,
>> but it would beat a tubular with hard glue.

Rolling resistance is casued by rubber deformation losses in the
tread, the tube and the casing.  The tube is firmly attached to the
casing by inflation pressure so it is like a tread inside the tire.
The whole tire flexes in three dimensional space, X, Y and Z.  You can
verify this typically by laying a stadard business card between tire
and tube.  It will shred to fine confetti when ridden.  It is this
motion that causes hysteretic loss in the elastomer in the casing, the
tread and the tube.  If the tire has patterned tread, such as a knobby
tire has in the extreme, deformation of the tread into the voids will
cause additional loss.  That is why knobby tires roill so poorly.

Unfortunately, I cannot locate the tire test with that included the
Vittoria CX and Clement Criterium tires.  These must be in the Avocet
archives.  As you see these tests were done 12 years ago and the
following is only one of the series of tests that were performed.
Others tested the effect of various tubes in the same tires.  I only
have this sample now the rest must be on some backup tapes floating
around here.

Jobst Brandt      <>

Tires, AVOCET and SPECIALIZED, 18 Apr 86
Rolling resistance (g) vs. inflation pressure (kg/cm2) @ 50 kg load
         Tire ID        Size   Sample   Nominal wt  Measured wt  Width +-0.2  TPI
         -------        ----   ------   ----------  -----------  -----------  ---
Col  1 - Air Pressure (kg/cm2)
Col  2 - S Turbo/LR     700x25C (A)        205          234          21.4     116
Col  3 - S Turbo/LR     700x25C (B)
Col  4 - S Turbo/LS     700x25C (A)        205          243          21.4     106
Col  5 - S Turbo/LS     700x25C (B)
Col  6 - A Criterium/20 700x25C (A)        225          236          23.1      66
Col  7 - A Criterium/20 700x25C (B)
Col  8 - A Timetrial/20 700x20C (A)        215          214          21.0     116
Col  9 - A Timetrial/20 700x20C (B)
Col 10 - S Turbo/LR     700x28C (A)        225          291          24.9      66
Col 11 - S Turbo/LR     700x28C (B)
Col 12 - S Turbo/LS     700X28C (A)        225          299          24.9      66
Col 13 - S Turbo/LS     700X28C (B)
Col 14 - A Road/20      700x28C (A)        265          272          25.2      66
Col 15 - A Road/20      700x28C (B)
Col 16 - S Turbo/R      700x25C (A)        180          188          21.0     116
Col 17 - S Turbo/R      700x25C (B)
Col 18 - S Turbo/S      700x25C (A)        180          193          21.0     116
Col 19 - S Turbo/S      700x25C (B)
Col 20 - A Criterium/30 700x25C (A)        190          182          22.6      66
Col 21 - A Criterium/30 700x25C (B)
Col 22 - A Timetrial/30 700x20C (A)        165          168          20.2     106
Col 23 - A Timetrial/30 700x20C (B)
Col 24 - S Turbo/R      700x28C (A)        220          248          24.8      66
Col 25 - S Turbo/R      700x28C (B)
Col 26 - S Turbo/S      700x28C (A)        220          253          24.8      66
Col 27 - S Turbo/S      700x28C (B)
Col 28 - A Road/30      700x28C (A)        230          241          25.0      66
Col 29 - A Road/30      700x28C (B)

 c1  c2  c3  c4  c5  c6  c7  c8  c9 c10 c11 c12 c13 c14 c15 c16 c17 c18 c19 c20 c21 c22 c23 c24 c25 c26 c27 c28 c29
3.5 356 365 395 369 367 392 348 347 430 443 423 432 423 406 355 364 382 386 378 387 361 369 419 432 427 442 394 403
4.0 322 329 360 343 337 361 319 316 393 401 391 393 386 376 325 333 352 356 350 357 325 331 390 390 395 401 360 363
4.5 298 302 330 319 316 336 296 290 366 373 362 368 354 347 297 309 326 332 322 329 301 302 364 362 369 375 334 331
5.0 279 282 311 298 294 314 280 271 341 348 335 345 335 320 275 289 311 311 302 307 282 284 340 339 342 351 311 310
5.5 265 268 293 287 279 297 265 255 324 329 319 327 319 305 262 276 295 296 284 295 268 267 320 322 325 333 290 289
6.0 255 253 282 274 267 282 249 242 310 312 306 310 303 293 248 260 283 285 269 281 253 252 305 304 311 318 275 273
6.5 244 242 274 264 254 269 239 230 297 294 292 297 290 283 237 253 275 277 257 270 244 242 292 291 298 309 266 262
7.0 238 232 263 255 247 256 229 221 287 284 282 288 280 273 231 245 267 270 247 261 235 232 281 283 286 299 254 249
7.5 231 226 255 250 238 247 222 213 272 277 272 279 272 264 222 236 260 260 238 249 228 224 272 272 274 289 248 242
8.0 222 219 248 244 233 239 215 205 265 267 266 272 264 258 216 230 253 253 231 241 223 217 263 264 273 281 243 235
8.5 216 212 244 237 226 231 209 201 259 259 259 266 257 252 208 223 245 246 223 233 216 209 256 257  0  274 235 229
9.0 213 211 241 236 223 224 204 195 255 256 259 259 254 245 204 219 245 245 222 231 212 208 252 256  0  273 233 226

From: (Jobst Brandt)
Subject: Re: Rolling resistance data (was tubular tires)
Date: 25 Mar 1998 00:55:57 GMT

Don Winston writes:

>> So how did you like the tubular offset due to glue?

> I plotted the data in Excel, and it makes a strong case for not
> riding tubulars. Jobst, do you have the data for hard and soft glue?
> Or at least the value of the offset?

We had no tires at the time that could be glued with track glue.  For
that you need a "dry" non rubberized base tape.  You probably noticed
that track tires have different base tape than road tires.  I think
you'll find the curves you plotted to be solid proof of the rim glue
effect.  You could even manipulate the curves by subtracting a
suitable constant from the tubular curves until they fit the family of
curves set up by clinchers.  If you do that, the tubulars come in
below the others in RR just as you would expect of a tire that had
thinner tread, casing and tube.

I'm sorry we never tested the glues with the same track tire glued
with track and road glue, but the effort at the time these tests were
made was to show that smooth tread and specifically Avocet tires were
as good or better than other tires offered.  Subsequently the
cornering photos were made showing riders at extreme banking angles on
wet and dry pavement.

Jobst Brandt      <>

From: (Jobst Brandt)
Subject: Re: Rolling Resistance
Date: 28 Apr 2000 20:53:20 GMT

Terry Morse writes:

>> All this data only makes good sense if plotted on a single graph
>> with different line types to identify the tires.  Only then does
>> the significance of testing at different inflation pressures come
>> to light.  This is in my estimation the only credible comparison of
>> RR.

> Well, I took the challenge and plotted the data with different
> colored lines for each tire. Please don't blame me for the color
> choices--they were chosen by Excel:


Thanks, that looks good, although I prefer a smooth curve "power" fit
to take out measurement noise.  This is similar to the graphs shown on
an advertisement at the introduction of smooth tread Avocet tires.  As
you can see, this is a family of curves that more or less do not cross
one another, except where there is another influence, such as in the
case of the two tubulars.  Other anomalies seem to arise from a larger
cross section tire crossing a smaller one, the larger one benefitting
more from higher pressure.

There is more information in these plots that first meets the eye and
careful analysis can logically decipher what it means.

Apparent to me was that the tubulars have a flatter characteristic
than the others.  That is, they are not affected as much by inflation
because they have the thinest and most resilient casings, use latex
tubes and should already have low RR.  Yet they have a higher RR than
heavier clinchers.  This can only be explained by a constant offset,
the rim glue.  Unfortunately a repeat experiment with hard (track)
glue was not possible, but I am satisfied that the tubulars would have
been at the bottom of the chart had they not had they been on hard
glue.  It was this graph that answered my question about "Strada" and
"Pista" tubular glue.  None of the common explanations given, similar
to stories about tied and soldered wheels, had made any sense to me.

The point is that low RR tires benefit less from higher inflation than
their more sluggish counterparts and that good tubulars, as most
people expect, have the lowest RR, except for the glue losses.  The
effect of these losses are apparent to people who have observed how
tubulars wear out from the back side almost as quickly as from tread
wear.  Aluminum rims generally show cloth base tape fretting patterns
while rim glue becomes dark grey from abraded aluminum oxide, both
evidence of tire motion on the rim.

Jobst Brandt      <>

From: (Jobst Brandt)
Subject: Re: Rolling Resistance
Date: 30 Apr 2000 01:13:22 GMT

Terry Morse writes:


>> Thanks, that looks good, although I prefer a smooth curve "power" fit
>> to take out measurement noise.

> I just updated the graph with smooth lines. I also added markers for
> the tubular tires to make them stand out and increased the total size
> for legibility. I also added a source attribution (be sure to hit
> "reload" to see the new chart).

Looks good.  I think that for those who have not followed this, the
source of this data may not be clear.  These data were collected by
IRC for Avocet to assure that the tires were in fact as good or better
than prevailing tires that used a different rubber compound that IRC.
Today most good tires use that compound and these curves are more
interesting for their characteristics than any absolute values.

The tubular anomaly and the crossing lines caused people who saw them
consternation but careful scrutiny reveals reasons for the
differences.  I am at a loss to explain the steel bead to Kevlar bead
differences but then they are not entirely consistent between this and
other tests.

> I was surprised to see how close all the RR's are at higher pressure.

I think that is also an important point to recognize and that this is
a small part of the energy balance of fast bicycle riding.  These
differences are bandied about in commercial promotions as though they
were significant.  They are not.  This is much like the recurring
"rotating mass" justification for the oddest wheel combinations.  On
the one hand aero rims are important (although substantially heavier)
on the other, light tires to the ridiculous for inertial reasons.
Hey, what about that heavy rim that has much greater additional weight
than the tire is lighter?

Don't bother me with facts!  They get in the way of my equipment.

Jobst Brandt      <>

From: (Jobst Brandt)
Subject: Re: Rolling Resistance
Date: 30 Apr 2000 00:57:36 GMT

David Roberts writes:

> I would like to throw in a few things about rolling resistance.

> 1) a wheel & tyre combination has both tread drag and aerodynamic
> drag. As you say, a smaller diameter wheel may have a slightly
> higher tread drag but will have lower aerodynamic drag. There is
> obviously a break point about which where either one will dominate
> and since the top of the wheel and spokes have an air speed of
> double the bike's road speed, I would have thought it is the latter.
> I am therefore very surprised that smaller wheels (say 24/25 inch)
> have not appeared on fast bikes. They would be lighter too.

I think enough of that concept has been blown about and that there is
little doubt that both of these effects are trivial compared to
aerodynamic drag of the bicycle as a whole.  That is the difference
between one wheel and another is something for the world record holder
to consider but not the readers of this newsgroup.  If you think you
are onto something we are not, please present some substantiating
evidence to support your hypotheses.

> 2) as the tyre is loaded and the contact patch develops, the
> material that formerly described a curve is now flattened to a shape
> that if it conformed should have a smaller surface area. Since the
> tread material does not compress, either the sidewall must stretch
> longtitudinally or the contact patch will assume a concave shape -
> just imagine a damaged ping pong ball.

I think your analogy is faulty.  The casing of a bias ply tire can
accommodate these deformations and does so by elastic inter-cord and
inter-ply movement as well as pure bending.  It is the longitudinal
compliance that causes the damaging effect on the "business card"
test that I described.  For bicycle tires, this is a small effect.
So small, in fact, that bicycle tires bulge at the contact area like
radial tires, while bias ply car tires neck in.

> Amongst other things, it is the sidewall's ability to accommodate
> and create a properly flat contact patch that differentiated a
> radial car tyre from the old cross/bias ply tyres. This innovation
> reduced scrubbing at the tyre/road transition to give an improvement
> of at least a factor of two in lifetime and a significant
> improvement in RR. Bike tyres are still basically crossply
> (irrespective of any tread reinforcement).

I think you mean the casing, not the side wall because the sidewall
does not need to change length, it is the part of the casing that
carries the tread that must comply to the geometric changes.  However,
the comparison with automobile tires does not apply because the effect
is primarily one that affects toroids with a small minor to major
diameter difference.  A bicycle tires of interest approximate an
infinitely large radius tube, responding like the flattening of a

Were this not the case, we would long ago have seen a radial ply
bicycle tire for world record attempts.  Michelin would be only too
glad to, once more, show the world the advantage of their radial ply
tire that was invented in conjunction with rubber tired railways and
subsequently showed the same benefits for automobile use.

> 3) should tyre carcasses have a square, circular or egg shaped
> section? I suspect square for RR (like a worn tyre), but since
> motorcycles still favour the circular, presumably that is still
> considered best for cornering

I'm glad you mentioned that this is all presumption at this point.  I
think phrasing this as such at the outset would give your hypotheses
the appropriate relationship to the subject.  By the way, tire
carcasses are the things one finds in worn out tire repositories, tire
casings are the structure of the tire.  Just because Continental
translate this from German as a carcass, doesn't change tire casings
to dead tires.

Jobst Brandt      <>

From: (Jobst Brandt)
Subject: Re: rolling resistance vs load ?
Date: 2 Jun 2000 20:38:08 GMT

Terry Morse writes:

> What's the relationship between tire rolling resistance and load?

It is dependent on how much the tire flexes.  The best estimate would
be to compare tire deflection versus load and versus pressure.  From that
measure you can look up a similar tire in the graph you have at:

and you'll be a lot to the relative effect.  I haven't done this
because the weight on my bicycle is not optional nor is inflation
pressure that is given by performance requirements (no pinch flats and
no mushy tires in curves) at between 100-110psi for the tires I use.

Jobst Brandt      <>

From: (Jobst Brandt)
Subject: Re: Relative rolling resistance of road tires.
Date: 13 Sep 2000 20:44:33 GMT

Scott L? writes:

> Here is the question.  Since I have the option of going fully
> suspended, why shouldn't I find a tire with the least rolling
> resistance and best aero characteristics, even if it means inflating
> them to some rediculous pressure?  Cornering is a minor
> consideration, since 99.9% of RAAM is done riding in a straight
> line.  As far as I recall, the highest rated pressure that I've seen
> on a clincher is Vredstien at 145lb.  I assume that for any given
> tire, the higher the pressure, the lower the rolling resistance, but
> is this true?  Has anyone ever measured RR at say, 145PSI or
> whatever the max rated pressure of the tire happens to be?

I'm not so sure you are doing yourself any favors with those tires
because 145 psi seems to be at the blow-off limit.  The disengaging
force is dependent primarily on the width of the rim, not the cross
section of the tire.  When you get below an actual 25 mm cross
section, I don't think you are gaining anything judging from older RR
data that I have.  Although record attempts on tracks are made with
smaller cross sections, I have not seen any excuse for them other than
smaller is better (which it is not).  I think tubulars of old were
optimized and never got much below 24 mm in cross section.

Jobst Brandt      <>

From: (Jobst Brandt)
Subject: Re: Relative rolling resistance of road tires.
Date: 13 Sep 2000 22:28:52 GMT

Scott L? writes:

> So a 700x25 Avocet 30 was the best in your test?  I like Avocets,
> but they are hard to find.  I've been told that the Ritchey road
> tires are made by the same company(Mitsuboshi?), but they have
> tread, so ????????

The tires are IRC, the last I heard was that Specialized tires are made
by Mitsuboshi.

> I have a perfect three mile downhill grade near my house with no
> crossroads.  From a standing start, my max speed over the three
> miles is about 30mph, average is about 20mph.  I did a test last
> night, which was one of the calmest of the year, and the three runs
> I did were all within 3 seconds of each other on my bike computer,
> which unfortunately has a granularity of one second.

Tire RR is such a small variation among good tires that inflation
differences of 10 psi is about the difference between one and another.
Beyond that, a small gust of wind from a passing vehicle is enough to
change the outcome of a coasting test, quite aside from how you place
you fingers on the bars.  This is an integration of speeds that is not
accurate enough to yield discernable results, This was even apparent
at the Michelin booth at last year's InterBike show where the time to
coast from 30 to 10 km/h was timed.  Slight changes in inflation
pressure made large changes in results.

> I'm going to get a stopwatch and run some tests.  I'll mount my last pair
> of Michelins, and run three runs at 105psi, and three at 120psi, which I
> know is over the recommended limit, but I've run them that high before on
> these wheels with no problems.

Forget it.  You can't control the breeze or position on the bicycle
accurately enough.  Besides, how accurately can you control the speed
at the start of the coasting run.  Starting at zero doesn't work.

> Then if the weather holds, I'll mount up a pair of Avocet 700x25 on the
> same wheels and rerun the tests.

Good luck.

> I'm interested in whether the difference between the tires is measurable,
> and whether the pressure makes a measurable difference.

It is measurable but I doubt that you can get consistent results that

> I realize this won't be laboratory stuff, but it might be
> interesting.  The problem is that it is a rare day when the wind
> conditions remain the same for two or three hours.  And depending on
> the time of day, the traffic will make a difference. Oh well.

That's what I say.

Jobst Brandt      <>

Index Home About Blog