The old tubular vs. clincher debate - some evidence?

[David L. Johnson]

On Wed, 05 Oct 2005 22:17:30 +0000, Wayne Pein wrote:

> Jens wrote:
>
>> I forgot to mention.....I wanted to leave it up to you lot to guess
>> which
>> trials are for which tire. ;-)
>>
>
> Since the wheels are also different, this experiment is not able to make
> any inferences on tire performance.
>
Not really. As far as this measurement goes, differences in wheels will
be an order of magnitude below differences in tires/attachment.

--

David L. Johnson

__o | And what if you track down these men and kill them, what if you
_`\(,_ | killed all of us? From every corner of Europe, hundreds,
(_)/ (_) | thousands would rise up to take our places. Even Nazis can't
kill that fast. -- Paul Henreid (Casablanca).

 

[David L. Johnson]

On Wed, 05 Oct 2005 23:51:07 +0100, Kinky Cowboy wrote:

> That
> extra 8mm of spoke, at the fast moving end, could account for more
> than the tyre difference.

How do you figure that? 8mm out of 280?

--

David L. Johnson

__o | "What am I on? I'm on my bike, six hours a day, busting my ass.
_`\(,_ | What are you on?" --Lance Armstrong
(_)/ (_) |

 

[Jens]

David L. Johnson wrote:
> On Wed, 05 Oct 2005 14:57:17 -0700, Jens wrote:
>
> >
> > I posted this over on biketechreview.com. But I know you
> > lot love this topic too. ;-
> >
> > I was testing out some rollers today. I couldn't
> > resist a little comparison. I tried a 404 with a
> > Conti-glued veloflex carbon and a Hed Alps clincher
> > with a Vittoria Ks, both inflated to 120 psi.
> > Here are the trials, in order:
> >
> > 1) 30.6 mph, 128 watts
> > 2) 30.8 mph, 125 watts
> > 3) 30.7 mph, 136 watts
> > 4) 30.6 mph, 138 watts
> > 5) 30.8 mph, 125 watts
>
> which was with which wheel? You have two wheels, and 5 tests. I can't
> put them in any order.
>
>
> --
>
> David L. Johnson
>
> __o | Some people used to claim that, if enough monkeys sat in front
> _`\(,_ | of enough typewriters and typed long enough, eventually one of
> (_)/ (_) | them would reproduce the collected works of Shakespeare. The
> internet has proven this not to be the case.

 

[Jens]

David L. Johnson wrote:
> On Wed, 05 Oct 2005 14:57:17 -0700, Jens wrote:
>
> >
> > I posted this over on biketechreview.com. But I know you
> > lot love this topic too. ;-
> >
> > I was testing out some rollers today. I couldn't
> > resist a little comparison. I tried a 404 with a
> > Conti-glued veloflex carbon and a Hed Alps clincher
> > with a Vittoria Ks, both inflated to 120 psi.
> > Here are the trials, in order:
> >
> > 1) 30.6 mph, 128 watts
> > 2) 30.8 mph, 125 watts
> > 3) 30.7 mph, 136 watts
> > 4) 30.6 mph, 138 watts
> > 5) 30.8 mph, 125 watts
>
> which was with which wheel? You have two wheels, and 5 tests. I can't
> put them in any order.
>
>

#3 and #4 were the 404 tubular.

--jens

 

[Ron Ruff]

Jens wrote:
> OK. Here you go....
> The columns are PSI, Power in Watts, Speed in MPH.
>
>
> Vittoria Evo Ks (Clincher) on Alps
> 120 119.8 30.75
> 100 123.9 30.71
> 90 126.5 30.75
> 80 130.3 30.74
> 70 135.0 30.75
> 60 138.5 30.57
>
> Veloflex tubular on Zipp 404
> 120 133.3 30.67
> 100 136.9 30.67
> 90 135.8 30.69
> 80 137.3 30.66
> 70 141.8 30.61
> 60 146.9 30.71
>
> Vittoria Evo Ks (Clincher) on Weyless Korsa
> 120 115.3 30.91
>
Thank you... looks like it could be very interesting data to me! Please
run more tests if you can... the more points the better... even repeat
the same ones to get a feel for repeatability errors. Also try higher
and/or lower speeds, other tires, etc.

BTW... what is the roller diameter?

Don't have time now, but hopefully I can look at this in more detail
soon...

 

[Qui si parla Campagnolo]

Jens wrote:
> OK. Here you go....
> The columns are PSI, Power in Watts, Speed in MPH.
>
>
> Vittoria Evo Ks (Clincher) on Alps
> 120 119.8 30.75
> 100 123.9 30.71
> 90 126.5 30.75
> 80 130.3 30.74
> 70 135.0 30.75
> 60 138.5 30.57
>
> Veloflex tubular on Zipp 404
> 120 133.3 30.67
> 100 136.9 30.67
> 90 135.8 30.69
> 80 137.3 30.66
> 70 141.8 30.61
> 60 146.9 30.71
>
> Vittoria Evo Ks (Clincher) on Weyless Korsa
> 120 115.3 30.91
>
> The last test was for those who think the wheel makes
> a big difference. I used the same model tire and the
> same tube on a $60 "Weyless" wheel with 20 bladed
> spokes. The tire was newer than the one on the Alps,
> which might account for some of the difference.
>
>
> -- Jens

I think the next 'test' ought to be researching the areas where tubies
'may' have an advantage, why they are used. Like cornering speeds,
safety when flat and weight...

 

[Ron Ruff]

The shape of the data looks about right, but it would be nice to relate
this to actual performance on the road. To do that we need to account
for transmission losses (they won't be small at these low chain
tensions, and they'll have to be estimated) and roller belt losses (?),
and factor out the aero drag from the tire rolling resistance. The
later can be done fairly well by varying the speed... aero drag power
varies by V^3 while RR power is proportional to V. Then we need to
determine what the RR on the road would be based on the RR on the
rollers. If we know the roller diameter and bike+rider weight, this can
be done as well.

So anyway... I'm hopeful that a meaningful conclusion can be had using
this method.

 

[Jens]

Ron Ruff wrote:
> The shape of the data looks about right, but it would be nice to relate
> this to actual performance on the road. To do that we need to account
> for transmission losses (they won't be small at these low chain
> tensions, and they'll have to be estimated) and roller belt losses (?),
> and factor out the aero drag from the tire rolling resistance. The
> later can be done fairly well by varying the speed... aero drag power
> varies by V^3 while RR power is proportional to V. Then we need to
> determine what the RR on the road would be based on the RR on the
> rollers. If we know the roller diameter and bike+rider weight, this can
> be done as well.
>
> So anyway... I'm hopeful that a meaningful conclusion can be had using
> this method.


That all sounds well and good if your goal is to actually
calculate crr. For my part, I don't really care about that.
I just want to know which tire is fastest and which tires to
avoid.


--jens

 

[[email protected]]

Ron Ruff writes:

> The shape of the data looks about right, but it would be nice to
> relate this to actual performance on the road. To do that we need to
> account for transmission losses (they won't be small at these low
> chain tensions, and they'll have to be estimated) and roller belt
> losses (?), and factor out the aero drag from the tire rolling
> resistance. The later can be done fairly well by varying the
> speed... aero drag power varies by V^3 while RR power is
> proportional to V. Then we need to determine what the RR on the
> road would be based on the RR on the rollers. If we know the roller
> diameter and bike+rider weight, this can be done as well.

The concept of testing on "real" pavement comes up from time to time
because these tests are done on large diameter smooth steel rollers.
By this method the tire that is tested, rather than some arbitrary
road texture that may be hard to define for repeatable tests. Rolling
resistance is made of hysteretic losses (rebounding with less force
than compression force) in elastomers (mainly rubber), of tread,
casing and inner tube. Tread patterns other than smooth also show up
as bulging losses as rubber deforms into tread voids such as the
conspicuous ones of knobby tires.

A smooth test surface gives the internal losses of the tire, the
principal cause of RR, to give a comparison among tires being tested.
A rough test surface would add tread deformation that would occur at
asperities on its surface but would not reveal any comparative tire
losses that weren't revealed on a smooth surface. Tires would have a
proportionally greater RR on a rough surface but would not switch
their relative positions among other tires. I think this should be
evident when considering the causes rolling tire losses.

Scrubbing between tire and road was believed to be a significant part
of RR but that is not the case. If it were, tires would wear much
faster than they do. That they don't scrub much was made apparent by
the Specialized Touring II tire that had a 1/4" wide raised center
ridge in an attempt to reduce road contact and scrubbing, and
therefore, rolling resistance. In reality its greater casing
deflections gave greater RR than similar tires with smooth tread.

Jobst Brandt

 

[dvt]

Qui si parla Campagnolo wrote:
> I think the next 'test' ought to be researching the areas where tubies
> 'may' have an advantage, why they are used. Like cornering speeds,
> safety when flat and weight...

Huh? Peter is talking about light weight being an advantage on a bike?
Someone better go over and check his temperature.

--
Dave
dvt at psu dot edu

 

[Ron Ruff]

Jens wrote:
>
> That all sounds well and good if your goal is to actually
> calculate crr. For my part, I don't really care about that.
> I just want to know which tire is fastest and which tires to
> avoid.
>
Precisely... we want to know which is are fastest and which to avoid.
To do that we need to quantify the difference... I mean, if the
difference only amounts to .01 mph then it is not significant. If it is
..1+ mph then maybe it matters for racing... certainly if you are a pro.


We also need to know what the errors are in the testing, and if the
results are repeatable over a wide range and lots of samples.

Testing for power levels on narrow rollers should emphasize the effect
of rolling resistance, ie the differences that you are seeing should be
much greater than what would actually occur on the road. This is good
for taking the measurements because you are seeing differences that
appear to be much larger than the error "noise"... but the raw values
that you are getting don't mean anything. Say you see a 10 watt
difference on your rollers at 31mph; does that mean you will see the
same difference on the road? Not at all... the conditions are very
different. That is why the Crr needs to be quantified as well as
possible and related to conditions that we care about.

 

[Michael Press]

In article <[email protected]>,
[email protected] wrote:


[...]

> Scrubbing between tire and road was believed to be a significant part
> of RR but that is not the case. If it were, tires would wear much
> faster than they do. That they don't scrub much was made apparent by
> the Specialized Touring II tire that had a 1/4" wide raised center
> ridge in an attempt to reduce road contact and scrubbing, and
> therefore, rolling resistance. In reality its greater casing
> deflections gave greater RR than similar tires with smooth tread.

That scrubbing is not a factor became evident to me many
years ago riding a new front tire that came out of the
mold with a thin membrane of rubber on the center of the
tread. It takes hundreds of kilometers to wear off the
membrane. Today the Avocet Fasgrip has this membrane.

--
Michael Press

 

[Qui si parla Campagnolo]

dvt wrote:
> Qui si parla Campagnolo wrote:
> > I think the next 'test' ought to be researching the areas where tubies
> > 'may' have an advantage, why they are used. Like cornering speeds,
> > safety when flat and weight...
>
> Huh? Peter is talking about light weight being an advantage on a bike?
> Someone better go over and check his temperature.
>
> --
> Dave
> dvt at psu dot edu

tee, hee, not 'light weight' but lighter weight.....cornering and
safety are much bigger advantages to me...and the 'curmudgeaon' factor,
of course.

 

[Jasper Janssen]

On Sat, 08 Oct 2005 00:37:44 GMT, [email protected] wrote:

>Scrubbing between tire and road was believed to be a significant part
>of RR but that is not the case. If it were, tires would wear much
>faster than they do. That they don't scrub much was made apparent by

If you ride wide-ish tyres underinflated, they sure do scrub on corners,
though.

Jasper

 

[[email protected]]

Jasper Janssen writes:

>> Scrubbing between tire and road was believed to be a significant
>> part of RR but that is not the case. If it were, tires would wear
>> much faster than they do. That they don't scrub much was made
>> apparent by...

> If you ride wide-ish tyres underinflated, they sure do scrub on
> corners, though.

That isn't scrubbing but rather walking from tire side deflection,
similar to pavement cornering on knobby tires, where the "fingers"
walk. The ultimate case of such walking is attempting to ride a flat
tire with slime in it. This will walk you right off the road as the
tread contacts the road more or less centered and then slips to one
side on lubrication.

Jobst Brandt

 

[[email protected]]

[email protected] wrote:
> Ron Ruff writes:
>
> > The shape of the data looks about right, but it would be nice to
> > relate this to actual performance on the road. To do that we need to
> > account for transmission losses (they won't be small at these low
> > chain tensions, and they'll have to be estimated) and roller belt
> > losses (?), and factor out the aero drag from the tire rolling
> > resistance. The later can be done fairly well by varying the
> > speed... aero drag power varies by V^3 while RR power is
> > proportional to V. Then we need to determine what the RR on the
> > road would be based on the RR on the rollers. If we know the roller
> > diameter and bike+rider weight, this can be done as well.
>
> The concept of testing on "real" pavement comes up from time to time
> because these tests are done on large diameter smooth steel rollers.
> By this method the tire that is tested, rather than some arbitrary
> road texture that may be hard to define for repeatable tests.

However, there does seem to be some evidence that on imperfect surfaces
lower tire pressures are preferable to higher, that rr relative to
pressure decreases as pressure increases, but then at some point rr
begins to increase again. Probably this effect could not be tested at
all on rollers, unless they were altered to provide a less perfect
surface, something that would not be impossible to do if one were
interested in quantifying the effect on a repeatable test.

The effect might be important in comparing the real world utility of
tubulars if you accept the premise that tubulars can be ridden safely
at somewhat lower pressure than clinchers, but OTOH, the effect, if it
exists, might be outside the range of difference between lower safe
tire pressures for clinchers and tubulars.

 

[Jens]

[email protected] wrote:
>
> However, there does seem to be some evidence that on imperfect surfaces
> lower tire pressures are preferable to higher, that rr relative to
> pressure decreases as pressure increases, but then at some point rr
> begins to increase again. Probably this effect could not be tested at
> all on rollers, unless they were altered to provide a less perfect
> surface, something that would not be impossible to do if one were
> interested in quantifying the effect on a repeatable test.
>

FWIW, I did a road test of Tufo S3 lites on a fairly rough chipseal
road at 125 and 145psi a while back. At 145 psi, it took 2 watts more
to go 4.13 m/s (slow). This was the rear wheel only.


--jens

 

[Ron Ruff]

Jens wrote:
> [email protected] wrote:
> >
> > However, there does seem to be some evidence that on imperfect surfaces
> > lower tire pressures are preferable to higher, that rr relative to
> > pressure decreases as pressure increases, but then at some point rr
> > begins to increase again. Probably this effect could not be tested at
> > all on rollers, unless they were altered to provide a less perfect
> > surface, something that would not be impossible to do if one were
> > interested in quantifying the effect on a repeatable test.
> >
>
> FWIW, I did a road test of Tufo S3 lites on a fairly rough chipseal
> road at 125 and 145psi a while back. At 145 psi, it took 2 watts more
> to go 4.13 m/s (slow). This was the rear wheel only.
>
I can certainly believe that... rough surfaces seem to matter a lot...
and there should be an advantage to lower pressures. In fact I would
suspect that on many real road surfaces a pressure of less than 100 psi
would be optimal. Road roughness that is not absorbed by the tire must
be absorbed by the rest of the bike/rider, where it is "lost". I don't
know how people determine Crr on rough surfaces... is this taken into
account? If not, then there seems to be an important term missing from
everyones "equations of motion" for a bicycle.

 

[Mark Hickey]

"Jens" <[email protected]> wrote:

>FWIW, I did a road test of Tufo S3 lites on a fairly rough chipseal
>road at 125 and 145psi a while back. At 145 psi, it took 2 watts more
>to go 4.13 m/s (slow). This was the rear wheel only.

Cool... I extrapolated your data, and figure if I let ALL the air out
of my tires, I'll go 2mph faster! ;-)

Seriously, thanks for posting a very interesting data point (one
that's at odds with some "conventional wisdom").

Mark Hickey
Habanero Cycles
http://www.habcycles.com
Home of the $795 ti frame

 

[Andrew F Martin]

I want to see what sort of extra wattage you see by using Mr. Tuffy
tire liners compared to regular clinchers. It's especially noticeable
at lower pressures, but the tire just doesn't deform the same way with
Tuffys. I have had real good luck with them on my rain bike, but the
rest of the guys on my team roll without them. All winter I feel like
I'm killing myself to keep up and I'm wondering if it's my bulletproof
(read pig-heavy) 36o wheels or my Tuffy's that are causing me this much
pain.