Delta Sail in A "Wind Tunnel"
Experiences from B.J. Slotboom
His explanation is based entirely on the analogy to a deltawing as used on supersonic aircraft like the Concorde. Such a wing can generate a lot of lift at high incedence angles by means of the vortices that develop along its long sweptback side edges. Marchaj calls this vortex lift. He claims that a sail of this shape can do without camber as there is no middle part with a straight leading edge to benefit from it (supersonic aircraft need thin flat wings, but I am sure some camber along the symmetry axis won't hurt at all at low speeds). I think Marchaj went wrong by focusing his attention entirely on the analogy to a deltawing and it is quite easy to show why.
First of all the vortex lift. As we know already vortices are always there at the upper and lower edges of a sail. They are an unavoidable byproduct of lift and are responsible for the lift induced drag. We have also seen that any really low aspect ratio sail generates high lift at high incedence angles because of these vortices. A deltashaped sail is not radically different in this respect. It may be an extreme case of a low aspect ratio sail in the sense that by letting the upper and lower edge meet in a point the middle part with a straight luff is eliminated, but the lift generating process is not fundamentally different.
Back to the crab claw sail. Marchaj admits that a true delta performs best with its axis of symmetry parallel to the airflow, so in the case of a sail horizontal. However his windtunnel tests show that it performs better in the nose down position it usually has. He attributes this to a favourable hull-rig interaction and the windgradient. His tests also show that pointing ability is improved by tilting the sail to a higher position along the mast (higher yard angle).
The forward tack of a crab claw sail is usually attached to the bow and the boom is free to move outwards and upwards. At low angles of the apparent wind this makes no difference but as the angle of the apparent wind increases to 90 degrees the wind hits the yard at a perpendicular angle (the same as if the yard were in a vertical position albeit that the force vector now not only points forward and sideways but also upwards). So in order to keep the sail in the 'low' position at high angles the forward tack should be able to move in the direction of the apparent wind.
Now I don't know the effects of hull-rig interaction and windgradient as I test without a hull and there is no windgradient in my windtunnel but I think the performance of this sail can be explained by the same factors (camber, aspect ratio and shape) as that of any other sail and in this respect it is in no way radically different from other sails. Of course it is a very ingenious sail. No doubt the most efficient rig that can be made out of a mast, two poles and a piece of flat cloth and it is the only rig that uses variable geometry to get the best performance on all courses.
I was fooled by Marchaj's interpretation into thinking that the performance of a flat sail could be improved by giving it a delta luff instead of camber. Tests showed that maximum lift was improved (due to the vortices) but lift/drag ratio went down (also due to the vortices) so the idea was quite useless. You can still find a lift-drag graph in the 'windtunnel' folder in the files section.