Sapim CX-Ray spoke question

[daveornee]

On 24 Apr 2007 12:47:48 -0700, Ron Ruff <[email protected]> wrote:

>On Apr 23, 10:28 pm, daveornee <daveornee.2pj...@no-
>mx.forums.cyclingforums.com> wrote:
>> I demonstrated a 5% lateral stiffness difference in changing from 14 g
>> to 14/15 DB spokes in an otherwise identical build. In that particular
>> comparison the lateral deflection difference was .07 mm at a 25 pound
>> lateral deflection load.
>
>I'm interested in more details about this. I've tested a few wheels,
>but not with spoke gauge as the only variable. Also, my rig isn't very
>stiff... but it seems like the spokes should have a greater effect
>than that. In the FE model I made, spoke stiffness mattered more than
>the rim stiffness.

Dear Ron,

Dave may be getting the kind of precision needed to measure 0.07 mm
differences in lateral rim deflection with his 25-lb rig.

But the top of Damon Rinard's table mentions that his values are
"average lateral deflection +/- 0.05mm" for what sounds like the same
~25 lb weight hanging from a rim:

http://www.sheldonbrown.com/rinard/wheel/data.htm

Similarly, Damon Rinard's table for deflection in inches (not mm) for
1/4 turns looser shows deflection changing ~0.004 inches (~0.10 mm)
for the first few 1/4 turns looser:

http://www.sheldonbrown.com/rinard/wheel/index.htm

(Page down to Questions 1. Does stiffness vary with spoke tension?)

That suggests that very small tension changes can affect the very
small deflection.

Weirdly, Rinard's deflection _decreased_ according to the measurements
for the first eight 1/4 turns that the nipples were loosened.

Cheers,

Carl Fogel

Carl,
You discussion reminds me of part of the challenge I/we had in making precise and repeatable measurments: truly smooth, true, and flat rims that stay smooth, true, and flat to within 0.01 mm (or better). From my days in Metrology and with my uncle's guidance he/we decided that to get what I was trying to do .... and seeing such small changes we really needed 0.001 mm smoothness/flatness/trueness. We also needed repeatability in that range. Challenges like this were at first interesting to my uncle, but his interest diminished when I asked him to machine away the "errors" in wheels that I built.
I suppose that I could have a better discussion with Chalo regarding these and other issues revolving around how to do this practically without loosing significant accuracy, but my uncle didn't have the patience and I didn't have the time to persist or make it into a paying proposition.
David Ornee, Western Springs, IL

 

[Kinky Cowboy]

On Mon, 23 Apr 2007 12:33:00 -0700, RS <[email protected]> wrote:

>Will Sapim CX-Ray spokes have more side to side deflection then
>Sapim Race (14/15) spokes? I know the CX-Ray are supposedly
>stronger and lighter (and substantially more expensive) but I wonder if
>there is any disadvantage to them. I'm about 190lb and will have a
>set of wheels built up with DA7700 hubs fr: 28/2x rear-oc: 32/3x w/
>Velocit Aerohead rims.
>

FWIW, I'm 200lb+ and ride Goldtec track hubs laced 2x into 32h Open
Pro rims with CX-Rays at the 142m Calshot velodrome. They're stiff
enough for that. Also bear in mind that your 28h CX-Ray build will
have way more lateral stiffness than any 3/4 spoke wheel (Hed3,
Corima, Xentis etc.), and you don't hear people complaining about
those.

Obviously, the bracing angle on the rear drive side is much lower on a
7700 hub than on a symmetrical track hub, so that's the place for the
extra stiffness a larger spoke cross section brings

Kinky Cowboy*

*Batteries not included
May contain traces of nuts
Your milage may vary

 

[jim beam]

daveornee wrote:
> Ron Ruff Wrote:
>> On Apr 23, 1:33 pm, RS <[email protected]> wrote:
>>> I ask because I notice the American Classic 350 Wheels have a weight
>>> limit of 220lb. with 14/15 spokes but only 180lb with Sapim CX-Ray,
>>> though they use a radial lace front.
>> AC wheels are "flex challenged", because their rear flange spacing is
>> quite narrow. Some large riders have issues with this, and the lighter
>> CX-Rays will flex more. If you use a hub with wider flanges, it is
>> less of an issue.
> Low number of flexible spokes on closely spaced hub flanges is not a
> good combination. Once a spoke goes slack it isn't doing the job
> intended.
> It is interesting to see the approaches of different manufacturers.
> Cane Creek used to make (and maybe still does make the Cane Creek
> Crono) a front wheel with 18 each 2.0 mm diameter spokes on a very
> widely spaced hub. That wheel had significant lateral stiffness,
> decent aero characteristics, reasonable bearing seal drag, and
> relatively low radial stiffness.
> You can see other characteristics of 7 different wheels that Fran�ois
> Grignon tested at:
> http://www.sheldonbrown.com/rinard/wheel/grignon.htm
> Where he shows the 18 spoke wheel having a little more lateral
> stiffness than a 32 spoke wheel (yes, I know it has 2.0/1.8 butted
> spokes), but only 2.2% less lateral stiffness than a 36 spoke wheel
> with 2.0 straight gague spokes.
> David Ornee, Western Springs, IL
>
>
interesting. but i find it more interesting that folk follow the
misconception that stiffness is purely a function of spokes and
disregard the rim...

 

[jim beam]

Kinky Cowboy wrote:
> On Mon, 23 Apr 2007 12:33:00 -0700, RS <[email protected]> wrote:
>
>> Will Sapim CX-Ray spokes have more side to side deflection then
>> Sapim Race (14/15) spokes? I know the CX-Ray are supposedly
>> stronger and lighter (and substantially more expensive) but I wonder if
>> there is any disadvantage to them. I'm about 190lb and will have a
>> set of wheels built up with DA7700 hubs fr: 28/2x rear-oc: 32/3x w/
>> Velocit Aerohead rims.
>>
>
> FWIW, I'm 200lb+ and ride Goldtec track hubs laced 2x into 32h Open
> Pro rims with CX-Rays at the 142m Calshot velodrome. They're stiff
> enough for that.

they will be with that flange spacing.

> Also bear in mind that your 28h CX-Ray build will
> have way more lateral stiffness than any 3/4 spoke wheel (Hed3,
> Corima, Xentis etc.), and you don't hear people complaining about
> those.
>
> Obviously, the bracing angle on the rear drive side is much lower on a
> 7700 hub than on a symmetrical track hub

indeed it is.

>, so that's the place for the
> extra stiffness a larger spoke cross section brings
>
> Kinky Cowboy*
>
> *Batteries not included
> May contain traces of nuts
> Your milage may vary
>

 

[jim beam]

daveornee wrote:
> Ron Ruff Wrote:
>> On Apr 23, 10:28 pm, daveornee <daveornee.2pj...@no-
>> mx.forums.cyclingforums.com> wrote:
>>> I demonstrated a 5% lateral stiffness difference in changing from 14
>> g
>>> to 14/15 DB spokes in an otherwise identical build. In that
>> particular
>>> comparison the lateral deflection difference was .07 mm at a 25
>> pound
>>> lateral deflection load.
>> I'm curious about your test setup. Have you tested many wheels? I've
>> been trying to test a few, but my "rig" is not very stiff. On the
>> other hand I'm getting bigger differences than that. It seems to me
>> from calculations and experiments that most of the lateral stiffness
>> is provided by the spokes, and there is ~23% difference in stiffness
>> between those spokes (2.0 vs 1.8mm). Between 1.8mm and 1.5mm there is
>> ~44% difference.
> When my uncle owned a machine shop I used his milling machine base + a
> couple of fixtures he milled for me to insure that the locknuts would
> not move when the wheel was laterally loaded. The setup was almost
> identical to Damon Rinard's setup when he had access to the machine
> shop at University of California - San Diego.
> My fixtures were made from CrMo steel and I used a vinyl coated 1/8"
> steel cable looped over the rim to hold the 25 pound weight. I
> intended to test a bunch of wheels, but my time and shop time became
> obstacles.
> Yes, the spokes themselves have a contribution to the lateral
> stiffness, but the rim, hub, and how the spokes are laced also
> contribute to how the wheel responds to lateral loading. I also had
> plans to see how various wheel configurations respond in radial
> loading.... and then in combination of radial and lateral loading.
> Again, my time and my uncle's patience/machine shop avaialbility became
> obstacles I could not overcome.
> Check out what Damon says and measures on his site:
> "7. How does spoke gauge affect stiffness?
> Thicker spokes make a wheel stiffer, if all else is equal. A typical
> 32 spoke wheel built with 2.0mm spokes is about 11% stiffer than a
> similar wheel built with 2.0-1.45mm swaged spokes.

butted butted butted. swaging is a particular manufacture method -
sapim and wheelsmith are not swaged, but they're still butted.

>
> Compare the deflection of two wheels: numbers 39 and 47. Wheel 39 is
> built with 2.0-1.45mm swaged spokes, but wheel 47 is built with 2.0mm
> straight gauge spokes. Hub dimensions are effectively identical, spoke
> count is the same and the rims are the same make and model, so the only
> structural difference is the spoke gauge.
> Result? The wheel with thinner spokes deflected 0.051" (1.30mm) in font
> and 0.067"1.70mm) in the rear, but the wheel with thicker spokes
> deflected less: only 0.046" (1.17mm) and 0.055" (1.40mm) for front and
> rear, respectively. That's an 11% increase in stiffness for the thicker
> spoked wheels.
> Interestingly, wheel stiffness depends on more than just spoke
> thickness; the rim and other factors also contribute, so only part of
> the increase in raw spoke stiffness shows up in measured wheel
> stiffness. The thicker spoke by itself is nearly twice as stiff axially
> as the thinner spoke!"
> From:
> http://www.sheldonbrown.com/rinard/wheel/index.htm
> Item #7
> and data table at:
> http://www.sheldonbrown.com/rinard/wheel/data.htm
> I did a few tests by only changing one variable, such as spoke guage,
> to get the details of what goes on there. I varied hubs, such as
> changing from Shimano Ultegra to Phil Wood.... but I also had to use
> slightly shorter spokes with the PW hubs. My tests are old and my
> uncle has since sold his shop and retired to play golf.
> David Ornee, Western Springs, IL
>
>

 

[jim beam]

daveornee wrote:
> [email protected] Wrote:
>> On 24 Apr 2007 12:47:48 -0700, Ron Ruff <[email protected]> wrote:
>>
>>> On Apr 23, 10:28 pm, daveornee <daveornee.2pj...@no-
>>> mx.forums.cyclingforums.com> wrote:
>>>> I demonstrated a 5% lateral stiffness difference in changing from 14
>> g
>>>> to 14/15 DB spokes in an otherwise identical build. In that
>> particular
>>>> comparison the lateral deflection difference was .07 mm at a 25
>> pound
>>>> lateral deflection load.
>>> I'm interested in more details about this. I've tested a few wheels,
>>> but not with spoke gauge as the only variable. Also, my rig isn't
>> very
>>> stiff... but it seems like the spokes should have a greater effect
>>> than that. In the FE model I made, spoke stiffness mattered more than
>>> the rim stiffness.
>> Dear Ron,
>>
>> Dave may be getting the kind of precision needed to measure 0.07 mm
>> differences in lateral rim deflection with his 25-lb rig.
>>
>> But the top of Damon Rinard's table mentions that his values are
>> "average lateral deflection +/- 0.05mm" for what sounds like the same
>> ~25 lb weight hanging from a rim:
>>
>> http://www.sheldonbrown.com/rinard/wheel/data.htm
>>
>> Similarly, Damon Rinard's table for deflection in inches (not mm) for
>> 1/4 turns looser shows deflection changing ~0.004 inches (~0.10 mm)
>> for the first few 1/4 turns looser:
>>
>> http://www.sheldonbrown.com/rinard/wheel/index.htm
>>
>> (Page down to Questions 1. Does stiffness vary with spoke tension?)
>>
>> That suggests that very small tension changes can affect the very
>> small deflection.
>>
>> Weirdly, Rinard's deflection _decreased_ according to the measurements
>> for the first eight 1/4 turns that the nipples were loosened.
>>
>> Cheers,
>>
>> Carl Fogel
> Carl,
> You discussion reminds me of part of the challenge I/we had in making
> precise and repeatable measurments: truly smooth, true, and flat rims
> that stay smooth, true, and flat to within 0.01 mm (or better). From
> my days in Metrology and with my uncle's guidance he/we decided that to
> get what I was trying to do .... and seeing such small changes we really
> needed 0.001 mm smoothness/flatness/trueness.

you'll never get that, but you could make sure you're always measuring
the same set of spots on the rim - should be repeatable then. if
everything is properly bedded in.

> We also needed
> repeatability in that range. Challenges like this were at first
> interesting to my uncle, but his interest diminished when I asked him
> to machine away the "errors" in wheels that I built.
> I suppose that I could have a better discussion with Chalo regarding
> these and other issues revolving around how to do this practically
> without loosing significant accuracy, but my uncle didn't have the
> patience and I didn't have the time to persist or make it into a paying
> proposition.
> David Ornee, Western Springs, IL
>
>

 

[Ron Ruff]

On Apr 24, 2:09 pm, daveornee <daveornee.2pk...@no-
mx.forums.cyclingforums.com> wrote:
> I wonder how you come up with the data on rim stiffness to enter into
> your FE program.

Well, I thought it would be easy. I just calculated Izz and Iyy for
the cross section, adjusting the wall thickness until the mass came
out right. As it turns out though, this apparently over predicted the
stiffness by a lot. The best test I could think of was to remove a
spoke from a built wheel and measure the deflection, and then compare
this to the model's prediction. The real deflection was about 70% more
than predicted, and to match that I would have had to reduce the rim
stiffness to ~1/3 of the calculated value! At that point I was stuck,
because I couldn't figure out why the rim would be so flexible. You
are probably familiar with HP Gavin's paper on wheels. He measured the
stiffness of the rims independently, and also determined that the rim
stiffness was lower than you'd think.

Why would this be? I'm no structures expert, but I can't see how those
little holes could make such a difference.

And your result (and Rinard's too) that spoke gauge makes little
difference is equally puzzling. If the spokes are not resisting the
lateral force, then what is? Even with my rim stiffness set 3 times
too high, the spokes had the greatest effect. And if the spoke
stiffness increases by over 20%, how could the wheel stiffness only go
up by 5%?

 

[[email protected]]

On 25 Apr 2007 21:12:46 -0700, Ron Ruff <[email protected]> wrote:

>On Apr 24, 2:09 pm, daveornee <daveornee.2pk...@no-
>mx.forums.cyclingforums.com> wrote:
>> I wonder how you come up with the data on rim stiffness to enter into
>> your FE program.
>
>Well, I thought it would be easy. I just calculated Izz and Iyy for
>the cross section, adjusting the wall thickness until the mass came
>out right. As it turns out though, this apparently over predicted the
>stiffness by a lot. The best test I could think of was to remove a
>spoke from a built wheel and measure the deflection, and then compare
>this to the model's prediction. The real deflection was about 70% more
>than predicted, and to match that I would have had to reduce the rim
>stiffness to ~1/3 of the calculated value! At that point I was stuck,
>because I couldn't figure out why the rim would be so flexible. You
>are probably familiar with HP Gavin's paper on wheels. He measured the
>stiffness of the rims independently, and also determined that the rim
>stiffness was lower than you'd think.
>
>Why would this be? I'm no structures expert, but I can't see how those
>little holes could make such a difference.
>
>And your result (and Rinard's too) that spoke gauge makes little
>difference is equally puzzling. If the spokes are not resisting the
>lateral force, then what is? Even with my rim stiffness set 3 times
>too high, the spokes had the greatest effect. And if the spoke
>stiffness increases by over 20%, how could the wheel stiffness only go
>up by 5%?

Dear Ron & Dave,

For convenience, here's Professor Gavin's wheel article:

http://www.duke.edu/~hpgavin/papers/HPGavin-Wheel-Paper.pdf

Cheers,

Carl Fogel

 

[daveornee]

On Apr 24, 2:09 pm, daveornee <daveornee.2pk...@no-
mx.forums.cyclingforums.com> wrote:
> I wonder how you come up with the data on rim stiffness to enter into
> your FE program.

Well, I thought it would be easy. I just calculated Izz and Iyy for
the cross section, adjusting the wall thickness until the mass came
out right. As it turns out though, this apparently over predicted the
stiffness by a lot. The best test I could think of was to remove a
spoke from a built wheel and measure the deflection, and then compare
this to the model's prediction. The real deflection was about 70% more
than predicted, and to match that I would have had to reduce the rim
stiffness to ~1/3 of the calculated value! At that point I was stuck,
because I couldn't figure out why the rim would be so flexible. You
are probably familiar with HP Gavin's paper on wheels. He measured the
stiffness of the rims independently, and also determined that the rim
stiffness was lower than you'd think.

Why would this be? I'm no structures expert, but I can't see how those
little holes could make such a difference.

And your result (and Rinard's too) that spoke gauge makes little
difference is equally puzzling. If the spokes are not resisting the
lateral force, then what is? Even with my rim stiffness set 3 times
too high, the spokes had the greatest effect. And if the spoke
stiffness increases by over 20%, how could the wheel stiffness only go
up by 5%?

H. P. Gavin's paper is interesting and as you know heavily loaded with math and figures. Even though I appreciated his efforts, I like Rinard's approach of making actual measurements. I made only one set of measurements that I can correlate with his. He made measurements on 140 wheels, while I made measurements on only 2 variants of the same wheel. It took me 5 days to get things set up and tested for reliable and repeatable measurements. When I found out how exacting I would need to be it pushed me back into further work in making more accurate truing and tension balancing. I learned a lot through the excercise and appreciate how much time can be spent. I just don't have the time to make more measurements and now that my uncle sold his business I don't have the place. If I were 40 years younger and Gavin is still at Duke I would go there and spend some time with him in the lab doing more work on the subject and maybe illustrating his work to the level of understanding of non-engineering types so that some of the complexities could be shown and understood by all (self included).