The PowerTap wheel build…How do you measure your training?

So I have this idea– get a PowerTap, build a wheel, revise the training routine and schedule, and then enjoy the fruits of my labor. So far so good. In this post I’ll discuss my choice of components, the build process (briefly), and the measured results of the wheel build.

Components

The hub drove the project specifications. In this case, a PowerTap SL+ hub with the accessory electronics pack, and a 24H drilling was my choice. A 32H or 36H would obviously be sturdier but I just didn’t want (or need) that many spokes. PowerTap hubs are also available in 20, 28, and 32 hole drilling. Of course, if I was purchasing a power meter, a wireless system was the way to go…no more wires on my frame. Additionally, should I choose to buy a Garmin 500 head unit in the future the resulting integration would be seamless.

PowerTap SL+ Power Meter

The rim was my next thought. I wanted a rim that was just as light as (or a bit lighter than) my current wheel, a Bontrager RaceLite. Second, the rim couldn’t be a box-section as I wanted comparable aero efficiency too. Third, the braking surface must be CNC machined and the rim joint welded instead of sleeved or pinned. (I’d prefer that the two extruded ends not shift-around under any circumstance.) Last, double-wall construction is preferred. The Bontrager was my baseline rim for the project. It weighed-in at 481g for the 700c size.

I found four candidate rims with the following characteristics:

Brand	Model	Weight	ERD	Color	Joint	Brake Surface	Appearance
DT Swiss	RR585	585g	582	Black	Sleeved & Welded	CNC
HED	Belgium C2	475g	592	Black	Welded	CNC
Mavic	CXP 22	510g	597	Black	Welded	Machined
Alex	R380	400g	591	Black	Welded	CNC

After wading through the reviews, availability, and costs I decided on the HED Belgium C2. What made this rim stand-out to me was the 23mm rim width. HED engineers designed this rim width coupled with a 23mm-wide tire to produce a rim and tire with straight sidewalls. This CyclingNews.com article spells out the advantages. I also read (somewhere) that HED sent the C2 rim to Continental Tires for testing. Results indicated a 18% to 20% decrease in aerodynamic drag and/or a reduction in rolling resistance. (I’ll ferret-out the source article to this as soon as I find it again.)

Edit 1/2/2010: Here’s the reference concerning the testing at Continental Tires. Note the portion of text that discusses the “C2” advantage.

Spokes connect the hub to the rim. Strength, durability, and a reduced aero signature were characteristics that I wanted. My existing RaceLite wheel has 20 spokes which requires slotted holes in the hub flanges. The new SL+ hub wouldn’t have any slots. The selection of candidate spokes would be limited to a diameter/width less than 2.5mm–the diameter of the flange holes.

I could choose from straight gauge, butted, or bladed designs. I removed straight gauge immediately because of the needless weight. Butted spokes would have a lower cost but would have a higher aero drag coefficient than bladed. For me, it is easier to detect spoke twist on bladed spokes when wheelbuilding. I decided fairly quickly that I would use bladed spokes.

Quick research showed two spokes in contention. DT Swiss’ Aerolite and Sapim’s CX-Ray. Both were reputable companies with their product widely used in the competitive arena. This fact addressed my initial requirements for the spoke choice. Cost would be the next filter. Basically, the Aerolite spoke was $3 and change per spoke with the CX-Ray close behind. Additionally, I found the CX-Ray to have the same finish as my Bontrager wheel, e.g. stainless steel versus the black finish of the Aerolite. I found the SS easier to clean

Sapim’s Polyax spoke nipple design also impressed me. The rounded shoulder design allows additional alignment towards the spoke line between the rim and the hub. Other designs don’t allow this adjustment and so some bending of the spoke shaft can occur. I see this at the bike shop all the time. I figure since I’m building from scratch I might as well avoid this issue.

I believe the CX-Rays to be one of the strongest and most durable spoke in the competition arena today. Sapim’s Belgium cycling background and testing process leads me to think that they’re serious about their product and how it performs in competition. My spoke choice here was easy.

Spoke length was an interesting evolution. You must know the dimensions of your hub before using some of these. I examined the results of three different online length calculators:

• DT Swiss’ calculator
• Sapim’s calculator
• Roger Musson’s “Wheelpro” spoke calculator

After making the appropriate entries, DT Swiss’ results were a 279mm non-drive side (NDS), and a 278mm drive-side (DS). Additionally, these lengths included a 14mm nipple allowance. Note that I could not choose Sapim spokes but the next closest size. Their interface looks like this:

Sapim’s calculator is a bit more involved. You must enter certain dimensions such as out-of-hub-nut to center-line of flange for the NDS and DS. Note that there’s no allowance for nipple length. DS was 285mm and NDS was 287mm. (I incorrectly reversed K_a and K_b values upon entry.) Their interface:

And last, Roger Musson’s Wheelpro calculations, note also that no nipple allowance is present, but that lengths for different crossing patterns is present. NDS length was 281mm and DS length was 280mm:

I shouldn’t have been surprised that the results were different. But which length to use? I called the experts at Wheelbuilder.com for advice. Vikki said that their technicians use 280mm on the DS and 282 on the NDS. She also recommended that their desired DS tension was 115kg and not to exceed 120kg. I found their customer service to be courteous, surprisingly fast, and refreshingly efficient. I definitely would use their services again and would not hesitate to recommend them to anyone. 24 CX-Ray spokes with Polyax 14mm nipples were soon on their way to my address.

Edit 3/9/2013: after wearing-out the braking surfaces of the C2 rims, I examined the excess spoke thread beyond the nipple head, especially on the DS. In hindsight, I would reduce the length to 277 DS and 280 NDS.

The Build

With all the parts on-board, I was ready to build. Barb and I reached the lake house without incident the following weekend and proceeded to hide-away for our retreat. Only a single jet ski ventured into our cove that weekend. Total quiet, no phone calls, no traffic noise. Oh man, what bliss that was.

I borrowed the following tools from friends of mine, Tomas and Jayce:

Park Tool TS-2 Truing Stand

Park Tool WAG-3 Dishing Tool

Park Tool TM-1 Tension Meter

And commenced the rather pleasing process of building my wheel. Tomas has a nice set-up with his truing stand as he has a location-adjustable run-out gauge to assist in measuring what the naked eye cannot discern—a definite plus. I considered the run-out gauge and the tension meter as part of my quality control effort. Ensure that you use good tools because if you don’t the quality of your product may not meet your expectations. This way, you can build the strongest possible wheel with your components.

The fun part

An excellent and much-read reference to wheel-building was written by Jobst Brandt.Herein you’ll find the same step-by-step assembly instructions that I used. There’s all the information you might want from history to theory. This book is a must-have in my opinion and you might want to check-out Roger Musson’s book (click his calculator link above) for additional knowledge. I laced my spokes in the recommended pattern from the CycleOps PowerTap User’s manual—2x minimum on both sides. My trailing spokes are on the inboard side of the hub flanges, and the leading spokes on the outboard side. I did not interlace my spokes as the reasons I found while researching the subject were not compelling enough to use for my build. Additionally, my Bontrager RaceLite carries the same pattern and I haven’t experienced any issues to date.

The Wheel

Let’s look at the measured results for the drive-side spokes:

Spreadsheet graphic for spoke tension and measured deflection

The preceding MS Excel spreadsheet illustrates what might be five or six successive  iterations of spoke adjustment followed by measurement as I slowly brought the drive-side spokes up to tension while addressing lateral alignment, radial alignment, and spoke twist. The kg-f column automatically interpolates the amount of force based on the index input. The instruction manual for the TM-1 states:

As a rule of thumb, it is best to set tension as high as the weakest link in the system will allow, which for a bicycle wheel is usually the rim. Therefore, to obtain a spoke tension recommendation for a specific wheel, it is best to contact the rim manufacturer.

Conversation with Tim at HED concerning the design of the Belgium C2 rim and the resulting spoke tension revealed that their target index range was 19 to 22. They did some in-house testing with the CX-Ray spokes to arrive at their target range. However, this was off the scale on the supplied Tension Meter table. I needed a kg force number to work towards. I observed that the majority of my final index readings were less than 19. The available index for a bladed 2.3mm x 0.9mm spoke stops at 18 on the TM-1 tension table. So I needed the last figure corresponding to an index of 19 in order to estimate all the decimal index points between. I used an MS Excel graphic as an easy/quick method to arrive at this number, rather than whip-out some quadratic regression calculation:

So, looking at the 5.1% drive-side tension variance and the very small amount of displacement, I think I have a fairly strong, true, and durable wheel, which should serve me well for quite a long time. Once I installed the new wheel into the frame, I only had to adjust my rear brake for the increased rim width. After riding the wheel for about four days, I made a very small adjustment to two spokes where the blades were slightly sideways to wheel rotation. Other than that, I haven’t had to do a thing except enjoy it.

One thing does concern me. Sapim lists the CX-Ray’s middle section strength at 1600 N/mm², which roughly equals 163 kg-f/mm². My estimated average tension is 172 kg. I hope that I haven’t overshot the allowable stress margin. On the other hand, HED’s target index range ran up to 22, which is far above my results. I suppose I could take one of my extra spokes down to an engineering test shop and test the spoke to failure, then I’d know.

When discussing these types of things with external companies, I’m aware of the perspectives of liability. This is my project and I accept what happens to my wheel as my own responsibility.

The Ride

I can feel the same Continental Gatorskin training tire tracking more securely through familiar sets of turns than when the tire sidewall flexed on the skinnier Bontrager rim. Or possibly the better reason is that this wheel is stiffer than the Bontrager RaceLite. I won’t ultimately know unless I fix the wheels sideways in a jig, apply a constant force and measure the resultant deflection. But this wheel subjectively feels and rides better. I’ve already thought about building-up a matching front wheel. Additionally, I can hardly wait to try out the Hutchison Atom Comp, which received nice reviews. This tire and wheel should make a nice set.

The Training

In short, the SL+ is going to fully revise my training and schedule. Having the ability to record and collect data from outdoor rides as well as indoor training sessions will markedly impact my minor and major goals and the methodology to get there. Additionally, wattage sourced from bona fide strain-gauge data has more credibility than data from fit-curve estimates. I’ve about six months to work to my advantage before the next racing season.

Thinking about the possibilities motivates me. How are you going to get stronger for next season?

12/31/2011 edit: After I built the matching front wheel, I checked the lateral true of this rear wheel. The displacement was less than 0.01 inch…I was pretty pleased with that. The wheel build continues to perform as I had intended it to.