Re: transom and deck

DELFTship forum Hull modeling transom and deck Re: transom and deck

John R. Coil

While it has probably been suggested before: I would suggest defining (in its own layer) a curved surface in either the horizontal or vertical orientation that has the curvature you desire and then rotating and moving it into final position. This greatly simplifies the math involved. For example, to get the values for lateral displacement (Y) for a plane of a given radius formed by rotating around the vertical (Z) you use this formula:

Yd = Rz(sin(acos(Xd/Rz)))

… Yd is transverse displacement, Xd is lengthwise displacement, and Rz is the radius rotatied around the vertical axis.

… which you can use as starting point to plug into the Web 2.0 sci calc here ( ) if you don’t have a scientific calculator of your own. The nice thing about it is that it has a memory so you can pull up the last attempt and input your new lengthwise value and then hit equals.

The last time I used this it was with a radius of 4.8333′ for a forward cabin roof line. I put my control points at 4.7333′, 4.6333′, 4.4333′, 4.2333′, 4.0333′, 3.8333′, 3.3333′, 2.8333′, and 0.8333′. This curve, combined with the point in the middle, defined in a fair surface to work with.

The math for finding the values for control points for curves based on ellipses is somewhat more involved but work more of less the same way and the above calculator is up to it.

Generally, I find that a radius equal to the width of a transom (or even less) is minimal, and can result in dramatic sweep around at the sides — all you really need is enough surface to cover the while thing. One that is half again (or more) than the width of the transom seems to yield more classic looking results. As with all things, this is a matter of personal taste. Ditto for negative camber in a swept forward transom (I think of a classic transom as aftward slope, modern sailboat transoms as forward).

To insure proper placement after defining the basic curved surface I would recommend first rotating said surface to its desired angle (either forward as is common today or aftward as was the norm in days gone by) and then use the values (x,y,z) of 1 control point on the centerline of your surface as the starting point to move the whole layer … matching that up to the corresponding center point on either the sheer line or the keel as desired. If you want to match points both at the sheer line and keel you’ll have to play around with the angle until it’s good enough or you just don’t want to mess with it any more.

Then find the intersection points in your transom and trim the excess. Ditto for your hull.

Naturally, I recommend saving the work in iterations in different files. That way you can back up at any time and not be committed until you want to be.