This could easily be like 10 topics here, but the base problem is that there is so much marketing hype and crap out there that are confusing everybody. Plus, the great wheelbuilding books of the past 20 years have in many cased reinforced and even hyped the 'black magic' aspects of wheelbuilding. I can honestly say that I have built, tested, and also known others to build and test radial laced rear wheels. The truth is that they work decently, much better than anybody would have thought, but they don't really work as well as tangentially laced rears when tested. There is a testing standard using a bump drum, basically a big ass drum with bumps on it that a load wheel rides against. A standard may be for a wheel to go say 2000km or 3000km with a 100 kilo load and a 1cm bump every 1m or something like that.

All radial wheels will do well on this test as well as the drive forces are pretty constant at the hub, but in the real world drive forces are subject to all sorts of inputs from constant torque input to shock load torque inputs and that is where the problems start. The real world riding of radial rears that I know of has generally produced wheels lasting between 1000 and 2000 miles, not bad considering, but also not good. I would imagine that the higher torque loading of a mtn bike combined with the more severe shock type loads would produce a wheel that didn't last very long, but it wouldn't explode instantly and it certainly wouldn't even die in the first few dozen rides.

In all honesty, the weak point is most likely to be the rear drive hub flange. Even on our flanged hubs, which are specifically designed for radial lacing on the non-drive side, we have drive side flange which is not designed for radial lacing.

To give you an idea why, we use 7XXX a custom heat treat for extreme toughness and crack resistance with yield strength of 80,000psi, our flange is 2.85mm thick and we impact forge the individual spoke holes to locally harden them – and that’s still not enogh to hold the most agressive torque loads in a radial wheel. Now most other hubs use 6061-T6, a material with yield strength of 38,000psi, and since the flange thickness is limited by spoke head geometry, the flange thickness between our hub and the others is the same.

Most hubs just cannot handle radial lacing due to the reduced volume of material over the spoke hole, so most every hub I can think of with very few exceptions will void your warranty for radial lacing, and they will eventually crack. By crossing the spokes you increase the amount of material over the spoke in the direction of load. Furthermore, you decrease shock load because a cross-laced wheel is less stiff radially than a radial laced one, and I think that this is generally the cause of problems with radial rears, the spokes and hub flanges just aren't designed to handle these types of shock loads in the directions which they are experienced.

As far as some of the othe comments made on wheelbuilding and design, you would be shocked to see a modern wheel factory. We lace many wheels using a small machine, but tension and finish by hand as machines have great difficulty building carbon rims. But any and every aluminum rimmed prebuild wheel on the market will be built by machine. These machines are very impressive, they can lace, true, pre-stress spokes, check tension and dish, and even pre-stress the entire wheel through some cycle and then re-true and tension. The machine lines used by Campy or Mavic cost millions of dollars and are very impressive and spit out wheels that are phenomenally round and true at a rate of like 1 per minute...insanely impressive if you ask me.

Much of the reason that modern wheels work as well as they do and much of the black magic of wheelbuilding is gone now is due to our understanding of pre-stressing spokes and eliminating 'creep' from the spoke. In the old days you built and trued the wheel, squeezed the spokes, re-trued it and that was it. This is still how many handbuilts are done, but with time the spokes will creep, or stretch, and lose tension. It is this loss of tension that will lead to breakage, so the wheels need to be re-tensioned and checked. Eventually this creep will stop and the wheel will become stable.

We started using what we call a 'wine press' 6 years ago and it's pretty amazing. It basically uses two 12" discs to press the spokes in the middle until they more or less touch each other. During the build, they tension and true the wheel, then press it 2 times. When they put it back on the truing stand the spokes will be completely de-tensioned as they have stretched so much. They re-tension and true, then press again. In all the press is used 4-6 times or until they press it and the wheel doesn't lose tension. If you remove a spoke from one of our built wheel once it is done you will find that it is 1-1.5mm longer than when you started, but now that wheel will probably go 10,000km without a true or tension.

This process was sort of understood 10 or 15 years ago, but not really, so wheels were much more finicky and needed more maintenance. I also think that some guys were doing this and others weren't which led to some builders gaining reputations for incredibly durable wheels and others reputations weren't so good. Along the way various camps popped up supporting things like spoke washers, relationship of hub flange hole diamtere to spoke diameter, canted flanges and so on, but most of these ideas have been abandoned with modern wheels as it just doesn't seem to matter anymore.

As far as “custom-build” vs. “off the shelf” wheels go, many companies are just putting together pre-existing parts. but if and when a wheel can be designed and built as a system, the result is much improved. We sell lots of rims to people to build on hubs of their choice, but the reality is that we can specifically drill our straight-pull hub holes to a specific angle for a specific wheel, then drill the rim at the same angle and when it is built up, there is absolutely no side load in that spoke. This year we introduced a new hubset which uses 5-axis milling of the spoke holes so that the drive holes can now be located in the hub ear in 5 axes so that there is no side load or bending in the spoke.

The result was an improvement in wheel stiffness of about 2% just from changing the drilling! This is the first use of 5 axis milling technology in a bicycle hub and I suspect this will soon be followed by other US or Euro manufacturers as the results are really impressive. But you just can't design that specifically for parts that are being sold as parts only, so building a wheel from various parts you will always be facing some design compromise somewhere.

Lastly, for my wheel recommendation for your mtn bike, I would build 2x/2x everywhere. This has a nice balance of comfort (radial compliance), torque transfer capability and durability, and be sure to pre-stress the crap out of the spokes. 2x/2x doesn't look super cool, but it works, and at the end of the day, that's what really counts.

The issue her isn’t so much about about whether spokes should be bladed or ovalized, but about the limitations of using aluminum for a spoke material. The reason the Ksyrium and similar wheels have such large side forces in a cross wind is that an aluminum spoke has to be about 2x the diameter of a steel spoke for similar strength, so to negate the very, very negative aero effects of this you really HAVE to blade the thing, round stainless steel spokes will cost you about 10-12 watts PER WHEEL, but round aluminum spokes will cost you nearly double that amount, creating a wheel that is slow beyond belief. Those of you who have ridden the old Spinergy Spox will remember, those wheels actually took nearly 30 watts more than a 32 spoke wheel at 30mph, and the same would be true of a round Ksyrium type spoke. But when the spoke is bladed, you have a lot of material you are moving around so you end up with a lot of side area.

A Ksyrium spoke is roughly 5mm x 1mm, compared to the CX-Ray we are using which is 2.2mm x 0.85mm. A standard round spoke is 2mm diameter, so whereas the aluminum spoke shows an increase in side area (and therefore side force) of roughly 250% the round spoke, the butted-bladed CX-Ray has an increase in side area of only 10%. There is good stuff about this on our website in the note about spoke shape posted earlier: http://www.zipp.com/...%20spoke%20shape.pdf

And actual drag and side force data on the Ksyrium shown in this note on wheel aerodynamics, you should note that the aluminum spoked Ksyrium shows a side force between the 202 and 303, where the 202 has a 3mm deeper rim and 2 additional spokes, and the 303 has a 16mm deeper rim and 2 additional (bladed spokes)...but both wheels have considerably better drag performance:

Rim Shape: (PDF)