The Ship Model Forum

Thanks guys, a lot to ponder, much appreciated.

I have just upgraded from 3DS Max 7 to Autodesk 3DS 2011 and once I have run through the "what's new" tutorials I will rework the hull. I particularly like the inclusion of layers on this new version which I am hoping will ease handling the complex superstructure. be back later.

Rob makes a very good point about software differences. I use DesignCAD, a fairly simple "traditional" CAD program. It is not a "parametric" program and does not allow simple trimming of surface grids. You have to create the grid by "stretching" it over a set of template lines. If it isn't correct you delete the grid, rework the templates, and create a new grid. This goes pretty fast when you get used to it. This is why my technique differes from Rob's. It is purely a way of working with differences in the programs.

He makes another good point. Stitching together two grids is a chore. You have to plan them carefully if you want the same number of "facets" at the common edge (makes joining them much simpler).

Another trick I use is to throw together a quick first pass at template lines and then generate a surface grid. I can snap to every intersection of lines in the grid, and I use this first-pass surface as a pattern for creating another more accurate set of template lines. This is a good way to fill in gaps between adjacent wide spaced template lines where you need to have greater control over surface curvature. For example, the Table of Offsets gives points on imaginary "station" lines along the length of the hull. Stations are much more widely spaced than the frames in the hull (15 feet on Clevelands, 16.5 feet on Baltimores). After creating a hull surface on these station lines I then generate frame lines at 4 foot spacing, using the original grid as the pattern for the frame lines that are used as template lines for the final hull.

I should note that DesignCAD does allow editing surface grids, but I find it awkward. It has a "hammer" tool that you can use to reshape the grid just as a metalsmith hammers sheet metal to reshape it. This is really useful for creating irregular folds and wrinkles in cloth syrfaces, like the bloomers around guns on a turret or canvas attached to railings.

Phil describes the approximate method I use as well. I only have a few things to add. First of all, extend your surfaces well past where you need them and trim them all down after the major surfaces are in place. We differ in one aspect. I try to do my major surfaces as a single surface. This is a software thing, however. SolidWorks deals better with single surfaces when doing later operations. On Scharnhorst I had a split in the hull surfaces near the outboard prop shaft fairings and the gusset between the fairing and the hull failed because it had to cross two surfaces. So now I have to spend the extra time to make sure the hull form is a single surface as you can see from Z-39 here:
[img]http://www.tankwerkz.com/images/1.jpg[/img]

Like I said, here I have made the surface extend well past the normal borders, it make sit much easier to trim with other surfaces rather than trying to deal with knitting a seam.
[img]http://www.tankwerkz.com/images/2.jpg[/img]

Phil talked about the lines of the hull. When you are using the hull lines from a drawing, those contour lines that run vertically top to bottom are called water planes. Here you can see Z-39's water planes.
[img]http://www.tankwerkz.com/images/3.jpg[/img]

The lines that run horizonatally front to back are waterlines. These control the smoothness of the hull and define the loft between waterplanes. Here are Z-39's water lines.
[img]http://www.tankwerkz.com/images/4.jpg[/img]

If you have ones that run vertically front to back, they are called sheer lines or buttock lines.

Earlier Phil was talking about smoothing the surface. It is pretty complicated to do but usually involves making the waterlines relax. They are defined by the waterplanes and relaxing them allows them to smooth out. Graphical programs make this easy, but in CAD it is a lot more difficult. After you relax the water lines you need to re-attach the waterplanes. I generally have to do this process 2-3 times in order to get a perfectly smooth surface. Like Phil said, if you can get your hands on the table of offsets, it is a lot easier. The only playing around you have to do is with the waterlines to get them to make the program interpret the waterplanes properly. I used the TOO on the last project I did and got the smoothest hull I ever worked with just from using them. Like I said, the only playing around I had to do was with the waterlines because the stern was several feet off the baseline and I wanted to do a single surface.

Notice how I have more waterlines lower down. More waterlines equals tighter control on complex areas. I notice on your upper hull there is a transition area that is creasing. This would be a good spot to add more waterlines to define it better for the computer.

Standing back you can see the imperfections in the hull, it looks okay until rendered then the creases start to show.

The Junk Bashers in place

I am using 3DS Max, the hull just evades me, there is some knack I am sure.

As to the level, I am simplifying but I have gone down to Upper Decks Comms unit so far

I second Rob's thoughts. From the start, working from hull section lines, to a finished hull surface takes me about a week. The first surface is done in about two days, but close inspection reveals many defects. Several days of fiddling with the lines and redrawing surfaces are needed before I am satisfied.

I have not been very happy with the results when I tried tracing scanned hull line drawings. Even large scale (1:96, etc.) drawings have significant errors in the lines. Since the CAD model is 1:1 scale these defects can be very large.

I prefer to work from the Table of Offsets. I enter these numbers into a spreadsheet and convert them to decimal inch values. Then I export an XYZ coordinate file and import it into the CAD program. This gives the most accurate (but still not perfect) station lines for forming the hull surface. Of course, this takes another day or two to type and recheck the table values.

I generate many waterlines (intersections of horizontal planes with the hull surface, also called contour lines) on the hull surface (on a separate layer) and then look at them end on (down the long axis of the hull) to see defects in the hull shape. Even the tiniest ripple in the surface stands out with this method. After revising the station lines I regenerate the surface and create new waterlines. I repeat this until all waterlines come out as nice smooth curves. You can also do this with butt lines (intersections of longitudinal vertical planes with the hull surface).

I do not try to model the entire hull surface with a single grid. Of course, I build one side and mirror it. But I also create many separate grids. That way I can tailor each to meet the requirements of complex surfaces like a bulbous bow, or the transition at the end of the flat keel, etc.

How much detail do you intend to put into your CAD model? What is the smallest object that you intend to model? The amount of time and effort goes up exponentially as the detail size decreases. I am building a CAD model of the USS Oklahoma City CLG-5, in the 1971 configuration. I am working from original blueprints and photos, and the smallest detail is 1/4 inch (1:1 scale). There are many, many thousands of screw heads, bolts, nuts washers and rivets at this level of detail. My hull was built up with more than 500 individual plates like the real ship - that alone took six months to finish! Right now I have just finished the exterior wiring on a missile tracking radar. It took several days for this detail. All of the external lights on the ship have wiring, switch boxes, etc. Radars have wave guides and cables running through the towers. Lifelines and life rails are another very time consuming bit. I have been working steadily for five years, and I think I am about half done. At least half of that time was spent finding drawings and photos to work from. Note: I am NOT modeling individual wires in cables, the twists in ropes or the threads on screws. I may be crazy, but I am not stupid!

So, what is your goal and how much effort do you want to put into it? If you leave off the nuts and bolts and skip the wiring details you can do a pretty decent job in a year or two. If you want to generate images that look like photos taken on the decks of the real ship you will need to add all the details that a camera would have captured. These details will add a few more years to the job.

I am not trying to discourage you, far from it! It is best that you really understand the magnitude of the project from the beginning. That way you won't be disappointed part way into the project when you realize it will take much longer to finish.

And please, post more images of your work. I love to look at other people's CAD models. I learn something from every one.

Phil
http://www.okieboat.com/CAD%20model.html

The hull is always the hardest part. It usually takes me a solid week of work just to get the hull form done, that isn't counting details like scuttles or props. What CAD software are you using? I might be able to give you some pointers.

Looks like a good start!

After a long absence I picked up mouse again and resumed my dormant project to build a virtual Leander Frigate, the hull continues to drive me mental but I decided to try some smaller details and was surprised how easy it all became. This is a screen shot of a Mk7 40mm Bofors.