techs-mechs: A Conversation with Hed's "Repository of Knowledge"

Steve Hed refers to Andy Tetmeyer (http://wheelguru.blogspot.com/ ) (a Paris-Brest-Paris finisher) as Hed's corporate "repository of knowledge." His other titles are "wheelbuilder to the stars" and "lackey." Recently, Tetmeyer and I carried on an e-mail discussion about Hed's design process, the general use of carbon fiber, and some other stuff. While Tony Martin's and Judith Arndt's Worlds Time Trial winning clinchers will make a cameo appearance, the main focus is definitely techs-mechs. Hmm. I'm getting hungry.

(Thanks to Andy Tetmeyer for being super-helpful, and for all of the photos).

RMc: What sort of general concerns drive the design process at Hed?

AT: Well, the simple answer is speed - duh. Obviously that is a very simplistic answer. For instance, a lot of people confuse speed (and maybe quality) with light weight. While the two are not mutally exclusive, weight is about our fourth concern.

First is aerodynamics, and more specifically, aero drag at wind angles from 7 to 20 degrees. When we test a bunch of wheels in the wind tunnel, whether we're are looking different wheel depths, or comparing our wheels to other brands, we see that most wheels don't have much drag difference at zero degrees. All the way up to seven degrees there are not giant differences. After seven the differences between wheels becomes much more apparent. At 15 degrees wind angle there are very big differences between different depths, and between brands. The practical application here is that some wheels are a large percentage better than others in low wind angles, but in practical terms it only adds up to a few grams of drag (or a few seconds in 40K). At higher wind angles, a well designed wheel will crush one that was made to go fast in headwind or tailwind conditions - the difference can be more than 100 grams of drag, and over a minute time savings in 40k. Aerodynamically we're looking for the most bang for the buck.

Our next two concerns are handling and stiffness. Handling doesn't need much explanation except to say that you'll go faster in a straight line, and if you can easily manage a deeper wheel then you are going to be faster. Stiffness needs a little more explanation. We strive for stiff and efficient wheels, but it is just as important to consider the tire as part of the equation. A squirmy tire on a stiff wheel is a poor combination, but a tire that is pumped up to very high pressure is not a fast idea either. Unless you are dead, you have noticed wheels getting wider over the past few years. On a clincher, a wider rim spreads out the tire beads and changes the profile of the tire. The sidewalls straighten out and stiffen up, and the load capacity of the tire increases due to the larger air volume, so less PSI is necessary. The tire is stiffer laterally under hard efforts, corners more predictably, and rolls both faster and more comfortably. With a tubular tire air volume can't be increased by using a wider rim, but there are still benefits. A wider rim supports more of the tire, and stiffens the structure by decreasing the unsupported area of the casing. As with a clincher, cornering is improved as is rolling efficiency - due to the simple fact that a greater percentage of the casing is glued down.
We brought out our first wide tire bed in 2006, though it was not our first wheel with a profile wider than 19mm. Steve's original CX wheel (in 1987) was 24mm wide, though it was made for a 19mm tire, not a 23.

Durability comes next. We have always strived to build wheels that will be suitable for everyday use. Race use is much harder on a wheel than commuting or training (or even touring). If a wheel can't stand up to daily use, it has limited value in a race.

Next, weight. It does not equal speed. If all things are equal, there is no question that a lighter wheel is better. However we think that a more aero, stiffer, better handling wheel that will get you all the way across the finish line will beat an uber light wheel made of the latest cotton candy/graphite/titanium alloy every time.

RMc: In particular, what factors support using carbon fiber instead of aluminum alloy?

AT: One, complicated shapes. Unless you are making a large volume of parts, aluminum is going to be extruded, cast or forged. Castings and forgings can have complex shapes, but not so much with extrusions. In the case of rims, we use an extruded aluminum clincher rim (it is pushed through the extrusion die and the rolled into a hoop) and a molded carbon tubular rim. The tubular stingers have a curved brake track that makes the wheel more aero. Trying to make that shape consistent in a rolled aluminum rim would be very difficult. Molding the stingers into that shape is much easier than it would be to extrude and roll the same profile.

Second, weight.

[Tetmeyer elaborated here in a follow-up e-mail:]

OK, went back and looked at what I wrote originally. Forging, casting, and extruding don't add much to the question. Complicated shapes do: Tubular and clincher are each problematic shapes, but in different ways. A clincher is complex because the hook has to hold the tire bead, the structure has to resist outward forces from tire pressure, and the aero profile has to work with a clincher tire. (in regards to shape... Ideally the tire does not bulge past the rim, and the whole thing has to fit into existing brakes and frames).

-or- Complex brake track curvature won't achieve much on a clincher because the tire can't nest inside the rim. You have match the tire and rim interface closely to maintain smooth airflow. Aluminium is strong and cheap, can be extruded with the necessary tire bead hook, and easily rolled round. This does not imply that carbon can't do the same things, but we don't think that carbon can do equally well at competitive cost and weight.

Tubular rims are very well suited to all carbon construction because the tire shape is independentt of the rim. It nests inside the tire well, which means the brake track can be shaped to enhance the aerodynamics of the wheel. With curvature from nose to wheel edge, it makes a lot more sense to mold the complete rim section, even though labor and material costs are many times higher.

As a side note and blatant plug, I note that both the mens and womens UCI TT champs last fall were won on clinchers - with alloy brake surface. That they were our wheels won't add to your article, but it may dispell the myth (or add pour gas on the fire) that " carbon is better, because it costs more - right?" Just like your article, I would argue that carbon is not automatically better, and can be a poor substitute for other materials.

R Mc: Are there uses where Hed feels that aluminum is a better choice than carbon fiber?

AT: We don't make a carbon clincher rim - not because it is not possible, but because by the time you make a rim with a hook bead that is strong enough to resist the pressure of the tire, and a brake wall that is strong enough to resist both braking forces, and internal air pressure that occurs as a the wheel heats during braking, and resists damage from road hazards. The result is a rim that is a few grams lighter than aluminum, and a few hundred dollars more retail price.

Second, crash damage. In the pro ranks, if you crash hard but your bar is not snapped, you dust yourself off and finish the day. The mechanics will generally replace your bars before the next stage or race. in case there is cracking that is not immediately apparent. Pros crash, and frankly replacing carbon bars after every mishap would be very costly. Quite simply the benefit of a part that weighs only a few grams less is not worth the cost.

RMc: What kinds of investments (work-hours, testing costs, and production processes [??]) does Hed make to bring a new design to market?

AT: We start with a mock-up. If the proposed design is similar to something that we already make (as in the case of the wider C2 Jets in 2007) - we just make a one off and start riding. With the wider rims it was obvious in just one ride (actually within a few blocks) that we would be refining the wheel and bringing it to market. Even in a crude mock-up it was clearly better. Where we don't have the luxury of a rideable mock-up, we make full scale wooden models and start with wind tunnel testing. If it proves out in the wind tunnel then we start nailing down the particulars of manufacture. Prototypes go through destructive testing, then we start riding them. If a wheel makes it a full season and we're satisfied that the product is viable, pro teams might get a it 6-8 months before it is for sale to the general public.

As for "how did we come up with the Stinger shape?"

A lot of it was prior experience. Like in any successful endeavour, when you stick with something for a long time... you tend to get really good at it. Steve has been building aero wheels since the late 80s, and one of his very first models, the CX (circa 1987) was very similar to the 2012 stingers.

At the time, the strangely curved brake tracks and a blunt nose of that wheel were not as widely accepted as they are today. Call it zen magic, or call it the knowledge and skill that comes from 25 years experience. As a company we have lots of experience to draw on and help guide product development. Spread among staff we have: two former continental pros, a national elite TT champion, at least 10 state TT wins, over 20 ironman finishes (including a 3rd place finish in the pro womens field), the MN hour record holder, PBP finishes, and the 2011 Tour De Gruene ITT champ.

We also do considerable computer modeling, but not the same stuff that our competitors do. Instead of spending our time and money trying to model wheel shapes (and Zipp is correct when they tell you that it is extremely complicated) we model courses. Three years ago I started building a program to model race courses. It is quite complicated, but not nearly as theoretical as trying to model a spinning wheel. My (our) program starts with GPS data of a course, broken down into segments a few meters long. ( I started modeling the London olympics TT yesterday, it has 411 segments for 40K). The segments all get a heading so we know which way the rider is pointed, and are assigned the correct up, flat, or downhill grade. We add the wind speed and direction, bike/rider/wheel drag, and wattage algorithm for effort up or down hills. When it is all done, we can predict finishing times within about 2% of real life.

...so, why is this important? Among other things, our model calculates and compiles apparent wind speed and direction. We now have a library of courses models with various rider positions and powers. From that, we know what wind angles we need to zero in on in order to make faster wheels. A wheel that crushes at 10 degrees yaw isn't really that good on a course where 60% of the wind encountered is 15 degrees or more. With accurate wind conditions in hand, we set about making the wheel that we want. We might make 10 different mockups for a single wheel depth and run them all in the tunnel to see which one is best. Since the wind tunnel is only one step removed from the road, and highly repeatable, we spend our resources on machining a building full size models for testing instead of relying on computer modelling.

Again, thanks to Andy Tetmeyer . . . and thanks to you if you've scrolled this far through a fairly dense post.