Last week, I found myself sharing complaints with another epee fencer concerning the sort of grip we both use on our weapons, the Negrini “Visconti” model. Visconti grips have been available for years and years; they are massive, in terms of the thickness of the metal composing the various elements, by comparison with other pistol-type grips. They are principally characterized by exaggerated ridges for the lower three fingers and a rough diamond-pattern ground on the top barrel and the projecting side member. As far as I can tell, they are the result of straight-forward sand casting, with some post-casting grinding to clean the cast seams and produce the diamond-pattern grooves.

I like my Visconti more than the other grips I have tried, but something still isn’t quite right. Although I have a large-size grip, somehow it only seems to absolutely fit my hand when I am not wearing a fencing glove (which would never be when I fence). This seems to be a common problem:  Visconti must often be filed and otherwise reshaped by their owners; some fencers buy them apparently because the exaggerated bulk of the stock piece lends itself to customization of that sort.

The fencer with whom I was speaking had a less amendable problem: he believes that the hook-like shape that projects up from the grip (which fits between the index and the thumb when the grip is in use) is in a slightly less-than-optimal place, making it difficult to perform certain fencing moves he is practicing.

Since I actually have some understanding of metalwork and casting, it occurred to me that I should look into the possibility of designing and manufacturing some alternative to the Visconti. Of course, the first task is to essentially reverse-engineer the grip so that I can study it virtually…in other words, I have to create a computer model of it. This turned out to be an extremely difficult task, even with a NURBS modeling program such as McNeel Rhino. The compound curvature of the metal in the grip is more irregular than it looks, and it is difficult to simulate with the typical smooth and precise surfaces generated by a modeler.

It occurred to me that this would be an interesting test for the my hacked Kinect…let’s see how good (or bad) a three-dimensional mesh can be made from a Kinect scan, using the Brekel Kinect program, of something as small as this Visconti pistol grip.

So I mounted a Visonti on a piece of wood attached to the rotating head of a camera tripod and stuck a piece of my daughter’s Lego right below it (to use as constant for scaling the resulting scans).

This first attempt was not terribly rigorous…I pushed the the grip so that it was as close to the Kinect as it could be and still be visible as a 3D object in the Brekel Kinect application window, and then I made four successive scans, rotating the grip approximately ninety degrees each time. (I tried to align “by eye” the appropriate face of the Lego to the glass face of the laser-emitter on the Kinect.)

To describe the resulting meshes as vague approximations of the original grip is something of an understatement. They were also much too small…the “grip” portion of each scan measured approximately 3/16″ long, as opposed to the 6″ of the original. Still, it was possible to find the outline of the Lego block target in each capture.

By applying a universal scale to the four captured meshes so that the representation of the block measured the same as the original block, I forced the entire set of captured meshes to scale to the correct size. The quads forming the correctly-scaled meshes have a diagonal measurement of .0016 to .0018 inches, which implies that no detail smaller than that can be captured at the Kinect’s minimum range. In McNeel Rhino, I rotated the four captured meshed and trimmed away the stem and block target to create a rough model of the original grip.

…a very rough model indeed. Ow! (The remainder of the epee was modeled directly in McNeel Rhino, from measurements made with a caliper of one of my weapons.)

Still, that’s enough of a general three-dimensional model of a Visconti for me to begin work on an alternative.

May 28th, 2011
Categories: Design, Featured, Technical, Works-In-Progress
Tags: , , ,