Okay, so there's been a few discussions about half-radial rear wheels (where the non-drive-side spokes are laced radially, whereas the drive-side spokes are the normal cross pattern). I don't mean to repeat these discussions, but rather want to raise a specific question.
First, the background: Sheldon Brown makes a plug for half-radial rear wheel near the bottom of his
wheelbuilding page. Here's what he says:
Originally Posted by Sheldon Brown
More and more rear wheels now are built "half-radial" with semi-tangent spoking on the right side and radial spoking on the left. Radial front wheels offer mainly esthetic benefits, but half radial rear wheels can be substantially more durable than conventional ones, in cases where the wheel is highly dished. The high amount of dishing called for to make room for more and more sprockets has caused an increase in spoke breakage on the left side of rear wheels. This is caused by metal fatigue.
A spoked wheel relies on having all of the spokes in constant tension. A highly dished rear wheel starts with very light tension on the left side spokes. The torque of hard pedaling combined with cyclical weight loading can cause the left side "leading" spokes to occasionally go completely slack momentarily.
Repeated cycles of tension and slackness cause these spokes to fatigue at the bends, and ultimately break.
With half-radial spoking, the amount of dish is very slightly less to begin with if you run the radial spokes up along the inside of its flange ("heads out.") In addition, since there are no "leading" spokes, no amount of torque on the hub can reduce the tension on any of the spokes. In fact, if you have an old wheel that has been breaking left side spokes, "half rebuilding" the wheel into a half radial will solve the problem once and for all.
I used to think that this was exotic, cutting edge technology, until I happened to look at a couple of Model A Fords in a local parade. Their wheels were highly dished inward, and were laced in the same half-radial pattern, for the same reason.
One of the major problems with radial spoking is that the spokes pull directly away from the spoke flange, and can more easily tear out of the flange than with tangential spoking, because with tangential spoking there's a lot more aluminum resisting the pull of the spoke on the flange hole. For this reason, radial spoking voids the warranty of a lot of Shimano hubs (although Shimano makes some hubs that are warranted for radial spoking, with somewhat altered design).
Radial spoking the non-drive-side of a rear wheel would
technically void the warranty of most Shimano hubs, but should not be a practical danger because non-drive-side spokes of dished rear wheels are significantly lower tension than drive-side spokes.
Sheldon notes elsewhere that even non-warranted Shimano front hubs can typically be built up radially, at lower tension, and with no danger, and I assume that this would translate to the rear non-drive-side spokes. But I'm not sure -
is there good reason to think that radially-laced non-drive-side spokes pose any danger to the flange of a forged hub shell (in my case, used Shimano 105)?
Sheldon also specifically recommends (in the text that I quoted) half-rebuilding a rear wheel with radial non-drive-side spoking, so I don't think the worry of stress risers from spoke-indentations from previous tangential lacing on used hubs is much of a worry.