The Stiffness Saga, Pt. 1: Rims

If you've never read it, Edward Tufte's book The Visual Display of Quantitative Information is definitely worth a look.  With all of the info that we're shaking loose from our wheel stiffness testing and measuring, it's definitely a challenge to parse it out into pieces you can actually chew.  This project has actually made me need to go back and digest what it is I've thought I've learned several times.  

Since the wheels we make are really nothing if not assemblies of components, it's worth it to talk about the components themselves.  You have rims, hubs, and spokes.  Each affects the system, but they all work together.  You do a bunch of horse trading back and forth to try and tease out the biggest spike of the features you want to maximize, while paying the lowest cost in negative features like weight and drag.  

Once you have the right fixtures, it's easy to measure rim stiffness.  You support the wheel at 3 o'clock and 9 o'clock, and load it at 6 and 12.  Measure the deflection under load et voila, there you go.  A few pieces of aluminum extrusion, a weight, and a dial gauge were all we needed for this one.  

We measured every type of rim we have on hand, except the Iron Cross rims because some dummy already built them into wheels and can't wait to use them.  Mike channeled his inner Edward Tufte and this came out:


All of these are 700c/29" rims (which is why the Stan's Crest isn't there - some dummy ordered the wrong size of those).  Note that we have plotted them against weight, so the bias line you see is stiffness to weight ratio.  On the line can be said to have average stiffness to weight, above the line can be said to have good stiffness to weight, and below the line can be said to have less good stiffness to weight.  

Rail 34s do quite well measuring this way.  

Now, of course, stiffness to weight isn't the only important thing in a rim.  The Rail 52 is fundamentally a pretty similar structure to the Rail 34, but pays a weight cost for its depth and shape.  That depth and shape come with the benefit of making it among the fastest wheels anywhere near its depth (and any clincher near its weight).  Conversely, the Stan's 340 comes with a very light weight.  Light riders who aren't slapping out watts might not need all that stiffness, and can benefit from lighter weight.  And a Stan's 340 with a good hub and an appropriate spoke count can be a quite stiff wheel - a 28 spoke Stan's 340 was a standout in whole wheel testing.  In the case of the Arch EX, we're building those with 32 spokes, not the 20 spokes that a Pacenti SL23 front might have.  And it needs to durability as well as stiffness - the two are neither mutually in- or exclusive.  The Kinlin XC279 which has been the basis for our FSW23 wheel has a nice mix of all of the attributes.  

We'll talk about hubs and spokes individually before we get into the whole assembly, and note that this is lateral stiffness.  Radial or circular stiffness is another thing that we'll talk about.

All of these measurements take massive mounts of time - building wheels, configuring the test apparatus, etc etc etc, so this will come over some time period.  But we're happy to be out of the land of guessing and into the realm of knowing.  

Back to blog


Excellent analysis and clearly presented. Now I know why I like my Stan 340, really improved my cycling especially on hills, but also why I need Rails for strict racing performance.


Point taken but grams should be self-explanatory and the vertical axis won't mean anything because its our protocol on our fixture. What would thousandths of an inch mean there? Squat. Already people play the dead-end game of comparing one guy's wind tunnel test against another. We're discouraging that behavior.


Interesting stuff! One comment: If you REALLY want to effectively use the techniques of visual display of quantitative information the first thing to include might be AXES! Lab 101.


The absence of axis labels is my fault. Dave was running some stiffness testing of rims and said, "whoa, look at these numbers!" I quickly dropped them into Excel and made the chart you see, just for our own internal edification. We then realized it would make a good blog topic so we just ran with it. Most of our blogs come about the same way – we learn or realize something interesting and then post it on the blog before we do any self editing. I'll put it into a fit-for-public-consumption template and repost it soon.

Mike May

MDBiker – Thank youJack – You sound angry! pyf – We'll get to that, but thanks for bringing it up.


Leave a comment

Please note, comments need to be approved before they are published.