Aero wheel tests in Germany
Rick Denney
This article was posted on rec.bikes.tech several months ago. I lost
track of it, but it was posted again today. Originally, it was
partially translated from German, and rather difficult to get through.
I've attempted to take the translation the rest of the way.
Fortunately, numbers defy translation errors, but any other errors are
probably mine.
The table will need a fixed-spacing font in case your reader isn't set
up that way.
I like the approach taken in this test. Unlike what we would see in
U.S. magazine, these guys spent some effort devising a bias-free test
that is repeatable and verifiable. Too bas they couldn't get hold of
Specialized Tri-Spokes.
Some of the conclusions in the text are a little flaky, in my opinion,
and I've tightened them up under the assumption that it was a
translation error. So, call this a paraphrase of a translation.
I also note that my favorite Spinergys peform within one or two
percent of the best non-disk wheels tested. Because this is a power
test, and not an aerodynamic drag test, the weight of the wheel should
be considered. Of course, the test was performed at constant speed, so
the weight of the wheel has no effect. But the test closely
approximates flat time trials in the real world.
All you technogeeks, enjoy.
[Taken from the German bike magazine Tour, who tested 10 Aero-Wheels
in their September '96 issue.]
The Test
The wheels were tested on a Faggin frame, which was equipped with the
SRM Power Meter. The crank used is the well known SRM crank which
employs 24 strain gauges between the crank and chainwheels to
precisely measure pedaling forces and the cyclist's power output. The
computer at the handlebar records the pedaling power of the cyclist
and the speed.
The test cyclist (Lars Teutenberg, a German professional) had to ride
several times around a track at exactly 45 km/hour. After every test
the data of the SRM Power Meter were transferred to a computer. The
power required to maintain 45 km/hr with each of the tested wheels was
measured in this way.
To ensure equal conditions, all wheels used the same tires with the
same pressure. A special fitting on the saddle forced the Mr.
Tuetenberg to maintain just one consistent position for all the tests.
The tester was alone on the course to avoid air turbulence resulting
from other riders. The bike was fitted with a 53-tooth chain-wheel and
a 15-tooth cog, providing a cadence of about 102.5 revolutions per
minute. Temperature was constant at 22°C.
Test Results
Disks
Disks performed the best of all the test wheels, requiring 12.7 % less
power (at 45 km/hr) than conventional wheels with 36 round spokes (2
mm straight-gauge) and conventional box-section rims.
Conventional wheels required 403 Watts of power output from the
cyclist, and disks only needed 352 Watts.
Deep-Section Aero Rims
Not quite as good as the disks were the Campagnolo 12-spoke Shamals.
The high-profile rim with fewer spokes saves 9.7% power compared to
conventional wheels during the test. Following very closely behind the
Shamals were Mavic Cosmic, Citec and Spinergy wheels. The similarities
between the Cosmic and Shamal wheels explain their similar
performance. The rims are equally deep, but Mavic uses 16 instead of
12 spokes. On the same level are the 12-spoke Citec wheels which use a
much lower profile rim. Therefore, the height of the rim is apparently
not the only requirement for good aerodynamics. Spinergy wheels (with
a deep section and eight radial carbon spokes) and HED wheels (very
deep rims with 28 carbon encased spokes laced two-cross) tested not
quite as well, with the Spinergys about in the middle between the
Shamals and the HEDs.
Shallow-Section Rims with Bladed Spokes
The Rigida DP-18 rims (shallow, but still aero profile) with 18
(front) and 24 (rear) bladed spokes performed surprisingly well in the
test. These wheels performed as well as deeper, more aero rims with
round spokes. Wheels with bladed spokes are more expensive, because of
the difficulties in fitting and tensioning the spokes, but still
cheaper than some of the carbon constructions.
Until more precise tests can be done it seems that 18/24-spoked wheels
will be sufficient for cyclists up to 165 pounds with conventional
rims. More spokes on the back wheel are needed for heavier riders or
rough riding surfaces. The depth of the section for most aero rims
used on18/24-spoked wheels will improve the strength of the wheel in
addition to the aerodynamic efficiency.
Not many examples of monocoque tri-spoke and four-spoke wheels were
included in the test. Delivery problems (Specialized, Habo),
production problems (Spin, Mavic, Zipp) and other impediments left
only three wheels to be tested. With 6.7 % less power requirement than
a conventional wheel, the Spengle wheels performed most poorly among
the complete sets of wheels. The single Corima front wheel, tested
with a conventional back wheel, performed not quite as well as a
combination of the Shamal front wheel with a conventional back wheel.
One or Two Wheels?
So, how does an aero front wheel and a conventional rear wheel compare
to a complete set of aero wheels? The test results show that the back
wheel is very important for the aerodynamics of the combination
bike/cyclist. The actual test shows that the complete Shamal set saves
39 Watts, the front wheel alone with a conventional back wheel only 18
Watts. These results might indicate that the back wheel is even more
important than the front wheel, but the test results are too close to
support that notion fully. The test results do indicate, however, that
the back wheel is just as important as the front wheel.
The measurements were made at a constant speed of 45 km/hr. For the
other speeds the values were calculated, assuming a constant combined
rolling resistance of 4.5 Newtons and a chain efficiency of 98%.
Power (W)
Speed (km/hr) 30 35 40 45 50 55
Disks 126 183 258 352 468 609
Campy Shamal
HPW-12 129 189 266 364 484 630
Mavic Cosmic 130 190 268 366 487 634
Citec 12-spoke 130 190 268 367 488 635
Spinergy 131 191 270 369 491 639
HED Jet 132 194 274 375 499 650
Rigida DP-18 132 194 274 375 499 650
Spengle Tri-Spoke 133 195 275 376 501 652
Campy Shamal front/
Conventional rear 135 199 281 385 513 668
Corima VR front/
Conventional rear 136 200 283 388 517 674
Conventional 36-spoke
with box rim 141 207 294 403 538 701
Rick Denney
Take what you want and leave the rest.
Treemoss2
Thanks for that post. It was very interesting. I ended up getting HED Jets last
year. Now I'm reading this and thinking back to my original thought -
Spinergys.
Oh well.
Ken Lotze
Rick Denney
On Tue, 14 Oct 1997 14:38:13 -0600, "Andrew R. Coggan"
<aco...@utmb.edu> wrote:
>Seriously, what was missing from the article that I'd really like to
>have seen was information on the reproducibility of this approach. If,
>for example, there were a +/- 10 W error in the measurements, this would
>reverse some of the orderings in the ranking.
I agree. If I read German, then I could possibly have reviewed the
original article. The "translation" I worked from is really a
transliteration, and I may have screwed up the grammar. Also, the
translator may have left out parts of the article. But it's still a
better, more honest and informative test than I've ever seen in a U.S.
publication. It doesn't stand up to the standards of a peer-reviewed
journal, but it tells me that German consumers expect a better product
in their magazines. (My favorite recent example is the test conducted
by Bicyclist of carbon forks. They talked about how they measure
longitudinal and lateral stiffness using Kestrel's machine, but then
they never reported those numbers. Typical).
And I also agree completely about the repeatability. If the SRM is
accurate indeed to the nearest Watt, then that's one thing. My
question is how precisely the cyclist maintained his body position.
Any variation in body position would have changed his power
requirements a little. That's where the larger sample size would be
useful. Again, they may have done it, and it just wasn't reported in
the article. I don't think maintaining speed would have been much of a
problem. A pro working at his normal time-trial pace would probably be
able to maintain a speed very precisely. My impression is that they
only too samples when his speed was correct, but I may be reading that
into the article.
One thing that was in the article was greater precision than could be
justified. For the test condition of 45 kph, they reported power to
the nearest Watt, which is the precision and stated accuracy of the
SRM unit. For the calculated speeds, however, they reported power to
the nearest 0.01 Watt. Ooops. I rounded the values to the nearest
Watt.
>
>We are currently experimenting with a similar approach, except that
>rather than using a velodrome and assuming that wind, hills, etc., are
>non-existent, we are trying to measure all of the relevant variables out
>on the open road use an appropriately instrumented bicycle. Only time
>will tell whether this approach is sensitive enough to be really useful.
Oooooh. I'm glad I don't have to do your statistics. My concern is
that it would be too sensitive, and I'd spend the rest of my natural
life trying to collect enough data and analyze it sufficiently, with
the appropriate sensitivity analysis, to cover all the typical
conditions. I get more than four or five variables, especially those
nasty nonlinear kind, at one time, and I'm in panic-land.
>
>> I also note that my favorite Spinergys peform within one or two
>> percent of the best non-disk wheels tested.
>
>Which isn't all that surprising - under headwind conditions (i.e.
>cycling around a velodrome in still air) almost any aero wheel on the
>market is markedly better than a traditional wheel. Where the Spinergy
>wheels have been reported to perform poorly compared to some other aero
>wheels is under crosswind conditions, something the Le Tour magazine
>tests didn't address.
Then why did the Spengle tri-spokes perform so poorly? They present
only the width of the tire and the hub to the wind. Even in a pure
headwind, it seems like the eddy current following the leading portion
of the rim would cause drag to the spoke. The fewer the spokes, the
better, but also, the thinner the spokes, the better. The Spinergy
spokes are as thin as steel spokes in a pure headwind, and no thicker
than monocoque spokes in a crosswind. But we'll know the answer to
this when you've complete the Mother of All Regressions (see above).
>
>> Because this is a power
>> test, and not an aerodynamic drag test, the weight of the wheel should
>> be considered. Of course, the test was performed at constant speed, so
>> the weight of the wheel has no effect.
>
>Well, a very slight effect, due to the influence of overall bike+rider
>weight on rolling resistance. But I'm being very nit-picky here. A
>bigger issue is just how well the "constant speed" criteria was achieved
>- small variations in starting and ending velocities (representing
>energy added to/substracted from the system) could confound the
>comparisons. Unfortunately, they didn't discuss how they handled this.
Agreed. But I'd still rather see this information than not see it.
>
>Andrew ("Hoping for an SRM of my own from Santa") Coggan
Keep hoping.
Andrew R. Coggan
Rick Denney wrote:
> I like the approach taken in this test. Unlike what we would see in
> U.S. magazine, these guys spent some effort devising a bias-free test
> that is repeatable and verifiable.
Bias-free, I will believe. Repeatable and verifiable? Well, I suppose
so, if you have access to a velodrome, an SRM, and a rider of putting
out 400 W! ;-)
Seriously, what was missing from the article that I'd really like to
have seen was information on the reproducibility of this approach. If,
for example, there were a +/- 10 W error in the measurements, this would
reverse some of the orderings in the ranking.
We are currently experimenting with a similar approach, except that
rather than using a velodrome and assuming that wind, hills, etc., are
non-existent, we are trying to measure all of the relevant variables out
on the open road use an appropriately instrumented bicycle. Only time
will tell whether this approach is sensitive enough to be really useful.
> I also note that my favorite Spinergys peform within one or two
> percent of the best non-disk wheels tested.
Which isn't all that surprising - under headwind conditions (i.e.
cycling around a velodrome in still air) almost any aero wheel on the
market is markedly better than a traditional wheel. Where the Spinergy
wheels have been reported to perform poorly compared to some other aero
wheels is under crosswind conditions, something the Le Tour magazine
tests didn't address.
> Because this is a power
> test, and not an aerodynamic drag test, the weight of the wheel should
> be considered. Of course, the test was performed at constant speed, so
> the weight of the wheel has no effect.
Well, a very slight effect, due to the influence of overall bike+rider
weight on rolling resistance. But I'm being very nit-picky here. A
bigger issue is just how well the "constant speed" criteria was achieved
- small variations in starting and ending velocities (representing
energy added to/substracted from the system) could confound the
comparisons. Unfortunately, they didn't discuss how they handled this.
> The Test
>
> The wheels were tested on a Faggin frame, which was equipped with the
> SRM Power Meter. The crank used is the well known SRM crank which
> employs 24 strain gauges between the crank and chainwheels to
> precisely measure pedaling forces and the cyclist's power output. The
> computer at the handlebar records the pedaling power of the cyclist
> and the speed.
FWIW, the manufacturer claims that the SRM is accurate to +/- 1 W (for
the Pro model with 24 strain guages - only +/- 3 W for the Amateur model
w/ only 4 strain guages). When the SRM has been compared against a
Monark ergometer (by Peter Keen and Jim Martin, and undoubtly others),
or against a Schwinn Velodyne (by me), the differences have been more
like 5 W or so. This difference, though, could easily be due to
drivetrain friction (in which case the SRM data are actually a better
measure of the rider's true power output).
All in all, the SRM system seems to be quite robust...and well it should
be, I suppose, at over $4000...
Dan Empfield
In article <19971014222...@ladder01.news.aol.com>,
tree...@aol.com (Treemoss2) wrote:
you are too easily swayed. there has been a plethora of tests proving the
efficacy of the hed wheel, while also favoring deep sections in general.
you chose the right wheel. the only better wheel you could have chosen,
and only by a very marginal amount, would have been the hed cx or deep.
but then you'd have paid a lot more.
QRman
Bill Mason
> Thanks for that post. It was very interesting. I ended up getting HED
Jets last
> year. Now I'm reading this and thinking back to my original thought -
> Spinergys.
>
> Oh well.
>
> Ken Lotze
How can you feel bad?
The difference between Spinergy and Hed Jet at 40km/h is only 2.1% !!!
(And the report posted didn't give many details about the HED Jets which
can be built up with a number of spoke, spoke count, and hub variations).
Also, as Andrew ("Hoping for an SRM of my own from Santa") Coggan points out:
>> I also note that my favorite Spinergys peform within one or two
>> percent of the best non-disk wheels tested.
>
>Which isn't all that surprising - under headwind conditions (i.e.
>cycling around a velodrome in still air) almost any aero wheel on the
>market is markedly better than a traditional wheel. Where the Spinergy
>wheels have been reported to perform poorly compared to some other aero
>wheels is under crosswind conditions, something the Le Tour magazine
>tests didn't address.
Other tests have shown sidewinds to cause a difference between wheels of
more than the 2.1% referred to above, and have shown that wheels which
perform better in a headwind can perform worse in a sidewind (both
relative to other wheels, and in absolute terms).
Also, the 2.1% difference can easily fall within the experimental error of
the tests, a factor which was not discussed.
--
Bill Mason, aka BilMas; bma...@autoiii.com
__o o
_ \<_ <\
__/\o_ (_)/(_) />
"Train hard, Race often"
Lidocaine
Thanks for that post. It was great, and it happily confirmed my buying
decisions from last season. I bought a Hed Disc and a Hed Jet for racing. Fared
pretty well according to the charts.
Super!
Andrew R. Coggan
Rick Denney wrote:
> If I read German, then I could possibly have reviewed the
> original article.
Fortunately we have an Austrian post-doc, so she was able to give me the
gist of the article.
> But it's still a
> better, more honest and informative test than I've ever seen in a U.S.
> publication. It doesn't stand up to the standards of a peer-reviewed
> journal, but it tells me that German consumers expect a better product
> in their magazines. (My favorite recent example is the test conducted
> by Bicyclist of carbon forks. They talked about how they measure
> longitudinal and lateral stiffness using Kestrel's machine, but then
> they never reported those numbers. Typical).
Looking at it from that perspective, I'd agree with you - Tour should be
commended. You'd certainly never see something this detailed in
Bicycling, for example. However, there have been a number of wind tunnel
tests of aero wheels reported on in Cycling Science, and I'm not sure
the Tour data quite match up to those - the article gives the impression
that the data were collected with care, but I wonder whether such
field-type tests are as good as a wind tunnel for specifically studying
wheels.
> My
> question is how precisely the cyclist maintained his body position.
> Any variation in body position would have changed his power
> requirements a little.
An excellent point. Merely tipping your head or dropping your chin can
have a slight but significant effect on drag. We're doing our best to
standardize body position in our experiments, but it is difficult.
> I don't think maintaining speed would have been much of a
> problem. A pro working at his normal time-trial pace would probably be
> able to maintain a speed very precisely.
In our first study (currently in review), we timed people (epxerienced
riders all) as the rode a ~500 m stretch of airport runway to get the
average velocity. There were, however, always small differences in
initial and final velocities, which we had to correct for. These
corrections weren't much, but they could obscure differences between
wheels.
> Oooooh. I'm glad I don't have to do your statistics. My concern is
> that it would be too sensitive, and I'd spend the rest of my natural
> life trying to collect enough data and analyze it sufficiently, with
> the appropriate sensitivity analysis, to cover all the typical
> conditions. I get more than four or five variables, especially those
> nasty nonlinear kind, at one time, and I'm in panic-land.
Technically, I should have said that we may not have adequate
statistical power using our method to detect e.g., differences between
different wheels. As statisticians always emphasize, though, the alpha
and beta levels should be adjusted with the overall consequences of the
outcome. That is, to convince others that one position or one wheel is
better than another or for publication purposes I may need a P value of
less than 0.05. However, for fine-tuning my own position and equipment
(which is my primary reason for pursuing this research), the standard
doesn't necessarily have to be as high. After all, if I pick what
appears to be the faster setup and there really is no difference, I
haven't gone any slower for it.
> Then why did the Spengle tri-spokes perform so poorly? They present
> only the width of the tire and the hub to the wind. Even in a pure
> headwind, it seems like the eddy current following the leading portion
> of the rim would cause drag to the spoke. The fewer the spokes, the
> better, but also, the thinner the spokes, the better. The Spinergy
> spokes are as thin as steel spokes in a pure headwind, and no thicker
> than monocoque spokes in a crosswind. But we'll know the answer to
> this when you've complete the Mother of All Regressions (see above).
I'm not familiar with the Spengle wheel, so I won't try to speculate
there. For the Spinergy wheels, though, my understanding is that the
limitation is thought to be in the *shape* of the spokes. They may be
thin, but they are completely flat, rather than having a nice aero
profile. Both shape and frontal area are important - for example, the
early Zipp 3-spoke composite wheel and the Trispoke brand wheel really
aren't much better in the wind tunnel than a traditional wheel, even
though they resemble the Specialized wheel. The difference is thought to
be in the shape (profile) of the three spokes - the Specialized have a
nice aero profile, while the other two are quite blunt. The importance
of the shape can also illustrated by comparing the original Hed Cx
(glossy black finish) to the '95 and later model - by fine-tuning the
shape a bit, Hed was able to slightly but significantly reduce the drag.
Treemoss2
>you are too easily swayed. there has been a plethora of tests proving the
>efficacy of the hed wheel, while also favoring deep sections in general.
>you chose the right wheel. the only better wheel you could have chosen,
>and only by a very marginal amount, would have been the hed cx or deep.
>but then you'd have paid a lot more.
>
>
It is always a difficult choice when buying bikes and their other expensive
parts. You do the best you can with what you know and can afford.
I have no regrets on buying HED Jets. I did buy the deep rear one, so maybe I
did good there. It sure is strong. I flatted at Dinosaur tri and rode the last
five miles in without any damage to either the wheel or the tire.
Ken
Ken
Skippy
Check out the australian triathlete magazine for a test.
www.oztri.com.au/docs/articles/bike_BM.htm
andrew
university wisconsin triathlon team
www.sit.wisc.edu/~uwtri
andrew
Somebody at work played a trick on me
here is the correct address
andrew