SardonicMeow

Yeah, on some of the build logs for Master Korabel kits I noticed that they use a big lateral piece to hold the other pieces, including some angled pieces. It seemed like a great idea to steal. But on this ship, it doesn't seem to gain me much. It would be more appropriate for a ship where the timberheads form stanchions for tall bulwarks with gun ports. On the Sultana, the bulwarks are really short. So I think in the end I'll be going with conventional bulkheads all perpendicular to the false keel.

I've been trying different ideas for how to deal with the bow. The first option is fairly straightforward. There are holes in the bulkheads for the bowsprit. Blue lines are fairing lines that will be scribed on the parts. Option two is a bit more radical. There are pieces at an angle somewhat analogous to cant frames. Horizontal pieces matching the curve of the bow will hold them in place. It looks cool, but I don't know if there is any advantage to it.

Working on the transom this time. My goals include creating a strong surface that the end transom piece can be attached to. Also, I wanted to ensure that there is some open space behind the transom windows. Here is in innermost piece. And the next two pieces. And some pictures of how the transom pieces fit together with the aft bulkhead pieces. Blue lines are the fairing lines that will be scribed on the pieces. Next, as I have been gently reminded, I have to figure out how to deal with the bow of the ship.

Ha. No, let me see if I can make something usable first. At the end of all of this, if there's interest, maybe I'll post the DXF files for others to use.

You're correct, Lou. I think you overlooked the part near the end of my last update where I mentioned that bulkheads near the bow still need to be done. My last picture is just showing the current state of the design, not the final one.

Several bulkheads were added at the center of the ship. Because little to no fairing will be required, no fairing lines were identified for these bulkheads. And here's another bulkhead in the aft part of the ship. But maybe there is an opportunity here. If the bulkhead is positioned appropriately, then I can do this... And that takes care of making the swivel gun stocks. The same is done for the other locations with swivel gun stocks. And just to get an idea of what the final product will look like, I thickened the bulkhead templates to the appropriate thickness. This is what it looks like so far. I still need to work on the transom and bulkheads near the bow.

Before I add any more bulkheads, I think it's a good time to cut out the notches for the masts and bowsprit. The kit-supplied dowels are 6.4mm in diameter for the masts and 5.15mm in diameter for the bowsprit. Tapering will be required above the deck, but I'm happy to leave them at full diameter within the hull. I edit the sketch for the false keel. Rectangles are added with a width equal to the diameter of the dowel, then rotated and moved in position to line up with the masts on the inboard profile image. A similar procedure is followed for the foremast and bowsprit. Finally, the Sketch -> Trim tool is used to get rid of the lines that aren't needed. I hope that little pointy bit at the front won't break.

Yes, that's something I hope to do. Probably not. The bulwarks are very short -- just a few inches above the deck level in some places. I think I'll do things like in the practicum: end the hull planking just below the level of the deck, then fit a strip for the bulwarks above it, with another strip on the inside.

Finally I will start to find the shapes for some bulkheads. I know that I will want bulkheads at the three points where the deck level changes, so those seem like good ones to start with. I start by making the inboard profile image visible, since it shows the internal structure. Via the construction menu, I create an offset plane at the point where the forecastle deck drops down to the main deck. It's 52.7 mm back from the forward perpendicular. Next I make the hull and deck surfaces visible. I select the offset plane and go to Sketch -> Project/Include -> Intersect. As I move my mouse pointer over the surfaces, anywhere the offset plane intersects with a surface, the intersection curve displays as a red line. I'll click to select the first intersection curve. And the one across the deck. And finally the one on the other side of the hull. When I click Ok the three curves are saved as a new sketch. There is some overlap where the deck and hull lines meet. To edit the sketch any further, I have to break the link between the intersection curves and the surfaces. I select a curve, then right click and select Break Link. This is repeated for each curve. The curves change from purple to blue. Now I go to Sketch -> Trim. When trim is active, segments of lines in a sketch that can be trimmed will turn red when the mouse pointer is over them. I trim off all the overhangs. And I add a notch 25mm above the baseline. The first bulkhead is complete. I can right click on the sketch, choose Export to DXF, and import the shape into laser cutter software. Of course, I need to add a matching notch to the false keel. I would also like to add fairing lines to each piece. The procedure is the same, but I don't need to include the intersection with the deck surface, and there is no notch. For the other bulkheads at places where the deck level changes, I follow the same procedure. For the second bulkhead, the hull lines don't quite reach the deck curve. In this case, I first use Sketch -> Extend to extend the lines, then use Trim to trim any overhang. I edit the false keel sketch and add notches at the appropriate locations for the three bulkheads I've created so far. And here's a view of all the sketches defined in this session. All of these will be exported for use on the laser cutter.

My goal is to produce templates for laser cutting. Hopefully it will be more clear in my next update.

The next thing I need to figure out is the exact placement of the decks and the deck camber. I bring up the inboard profile image. Starting a sketch in the centerline plane, I trace the lines along the underside of the planking on each deck. The lines for the two innermost decks (main deck? and quarter deck?) are very slightly curved. The other two are just straight lines. The sketch continues alone the rabbet and back to the start. I'll be using this sketch for the false keel. Notches for bulkheads will be added to it later. The deck camber isn't obvious from the Chapelle plans. On the Model Shipways plans, there is a diagram showing an arc over a span of 72mm with the highest point of the arc at 2.5mm. I assume this is the camber at the widest part of the deck, but can be applied overall. Applying those measurements, I create arcs on the 3D model passing through the start and end points of each deck. Surfaces are lofted across the arcs. For the two larger deck sections, the line of the deck along the centerline is used as a rail, because it's curved. That gives those deck sections a very slight saddle shape. (This picture makes it look like I'm creating a solar-powered ship. 🙂) Here are all the surfaces to this point.

In order to establish a good ending shape for the hull, I added one more curve at the stern. This is a little guesswork since it isn't a line on the plans, but I think the shape is pretty close. Now the lofting can begin. I click on "Create Form" to enter the sculpt environment. Then Create -> Loft. When I select the start and end curves for the first loft, it looks like this. From above, it's clear that the shape isn't right when there aren't any rails. Once the rails are selected, the shape conforms to them. The surface is constructed in sections by lofting between station lines. Uh oh. The loft doesn't work back here. My solution is to loft this area in two parts. First, I break the longer station line into two parts. The top part stops at the same height as where the smaller station line ends. The remaining half of the station line matches a line along the sternpost. Then the two pieces are lofted separately. There is a little gap between the pieces. The Modify -> Merge Edge operation is used to seal the gap. I finish the lofting and create a mirror of the surface. It doesn't look too bad. The reflection in the aft area has a little wobble in it, showing the isn't quite smooth there, but I can deal with that later. It's neat, but so what? How can I use this? For my next trick, I need a volunteer (or several) from the audience to pick a number between 1 and 200.

The sketching of lines continues with finding intersection points and then sketching fit point splines through those points. Here are the intersection points for the two waterlines, with the waterline planes visible. And then the splines. Drawing the lines is easy -- just a matter of connecting the dots with a little adjustment of the curves at the end. The same procedure is followed for the load waterline. But wait... The intersection points don't seem to match up to the line on the plans. Is this an error in the plans? No. This is because of how sketches work in Fusion 360. When you start a sketch, you choose the plane on which it will be drawn. Then Fusion 360 brings you to a view that is tangent to the drawing plane. Because the load waterline is tilted a little off horizontal, the view isn't directly down onto the plans. Instead, it's skewed a little, so the view doesn't match up. The load waterline drawn on the plans is a projection of the hull line down onto a horizontal surface, so to see it correctly, the view needs to be changed so that the viewport is looking directly down. See the view from directly overhead below. Those of you with sharp eyes may have noticed that aft-most portions of the load waterline curve don't quite match the line on the plans. That is an issue with the plans, as the intersection point is correct in the sheer view. I think it's close enough not to matter much. Do diagonals work like the load waterline? Are they projections down onto a horizontal surface? No, they really do get drawn directly on the tilted planes. Here are all the lines at this point. Now it's finally possible to create the hull surfaces.

Thanks, Patrick. I've been closely following your Sultana build log (and the one by moreplovac too).

Thanks for following along. I used Gimp to convert the plans.

Now I'll create sketches of the various hull curves. (In the 3D/CAD forum, there is a thread where I worked through this method. You can check it out if you're interested in more details about the procedure.) For the topmost line, I will create a curve following one of the bulwark lines on the plan. To ensure correct placement, I create a construction point at the tip of the bow, which will be used as a starting point for the curve. A fit point spline is created in a horizontal plane following the curve as viewed from above. Then the points on the spline are adjusted up or down to fit the curve as viewed from the side and front. In the picture below, the curve is viewed from the front and back showing how it follows the plan. Next is a sketch for the rabbet line. It is composed of several lines and curves. The sketch is drawn in the false keel thickness offset plane. The same starting point at the bow is used to start the bow curve, meaning that the bulwarks curve and bow curve intersect at that point. Here are the two sketches so far. Now I select the construction plane for station 1, then go to Sketch -> Project/Include -> Intersect and click on the existing lines to find the points where they intersect with the plane. This is repeated for each station plane. The intersection points are displayed as purple circles. Using the intersection points as start and end points, a fit point spline is sketched for each station following the curves on the body plan. Why all the intersection points? Later on, I'll be using the loft and rails feature in the Fusion 360 sculpt environment. Loft creates a surface between two or more curves (profiles). Rails can be used to adjust the flow of the loft. However, all rails used in a loft must intersect all profiles. To make sure all the curves intersect as needed, I find the intersection points first, and draw fit point splines using those intersection points. Next time, sketching continues.

Step one is to bring the plans into Fusion 360 as canvas objects. The plans are checked to make sure that all lines are perfectly horizontal / vertical, and that lines in one view match up to the corresponding lines in a different view. The images have been scaled in the 3D environment to match the scale of the model (1:64 or 1 foot = 3/16"). The 3D origin is at the intersection of the baseline, the centerline, and the forward perpendicular. A number of construction planes are defined. Here are the station planes. And the two water lines plus the load water line. Planes for the three diagonals are also defined. And canvas objects for the diagonals are added to the environment. A final, and very important, construction plane is one which defines the thickness of the false keel. I plan to use 1/8" thick wood for the false keel and bulkheads. Converting to metric, that's 3.2mm. But because I am only building a half hull now, which will be mirrored later, I will set the false keel offset plane to be 1.6mm from the centerline. With the reference images imported and the construction planes defined, the environment is ready for the hull modeling to start.

My plan, which I hope is not too ambitious, is to make a plank on bulkhead hull. I have access to a local makerspace with a laser cutter, so I intend to laser cut the pieces. I'd like to develop a workflow of going from plans to parts that I can apply to future projects. Let's look at the plans of the Sultana. Follow this link to see the original plans on the National Maritime Museum website. I suspect most visitors to MSW are familiar with the circumstances of the Royal Navy acquiring and recording the lines of the Sultana. Who exactly did the work of recording the plans, at Mr Randall's dock in 1768? Was it this guy whose name is in the corner? Adam Hayes? You probably know how the plans were cleaned up and recorded by Howard Chapelle in 1935. The Chapelle plans are reproduced on the end papers of the Schooner Sultana book. In white on orange... Thanks guys. Scan, invert colors, levels, convert to grayscale, more levels, crop into separate files. That's better. What stands out about the plans is how few station lines there are. Just nine. Was Mr. Hayes in a rush to meet his mates at the pub that day? Often, the plans of a larger ship will have 20 or more stations, and a modeler can just choose all or some, make bulkheads directly from those lines, and be done with it. But nine isn't enough. Based on my one and only plank on bulkhead experience, I want more bulkheads. I want lots of bulkheads. It's time to find my bulkhead lines. It's time to go to the drawing board start up the CAD software.

Welcome to my Sultana build log. I plan for my Sultana to deviate from the kit in two significant ways. First, I will be modeling the 2001 Sultana replica, rather than the 1767 original. In addition to pictures taken by myself and others, I'll be using the book Schooner Sultana: Building a Chesapeake Legacy by Drew McMullen with photos by Lucian Niemeyer. The book covers the construction of the 2001 ship with lots of descriptions and photographs. Best of all, there is an appendix with measurements of many of the main components. The second significant deviation is this.

Thanks for sharing, Jim. It's good to see someone else doing similar work. I haven't gotten familiar enough with the T-Spline editing tools, but it looks like it's worth the effort.

I have been thinking about how the hull could be lofted in pieces rather than all at once. Each section between stations would be lofted, using whatever rails are common to the two. In the end, there would be an option to combine the pieces together. Here are the first several sections. And so on down the line. You'll notice that I skipped one section, because it's problematic. What to do here? If I loft between the two stations, the back end of the keel isn't covered. I realized I need to break this into two parts. One loft at the top from the smaller station (#12) to the upper part of the larger station (#11). Then a separate loft from the back end of the keel to the bottom part of station 11. To do this, I needed to break the larger station into two parts, while also retaining an intact copy for lofts forward. First, I highlighted the curve of station 11 and hit Control-C. Then I started a new sketch in the same plane and hit Control-V to paste. That was repeated to get three copies of the station line. The original will remain untouched. Copy 1 will be cut leaving only the upper part and copy 2 will be cut at the same spot, leaving only the lower part. Let's work on copy 1 (upper) first. I edit the sketch, and keep station 12 visible. Next I sketch a line from the bottom point of station 12 running horizontally to cross over the curve of station 11. Now I go to Sketch -> Trim. When I hover over the bottom portion of the curve below the line, it's highlighted in red. I click to remove the red portion of the curve. Here is the final result. Working on the second copy, the same procedure is repeated, but the top portion of the curve is trimmed. Only the blue portion remains. Now I can complete the lofting. First, I loft from station 12 to the copy of station 11 containing the top of the curve. And then the bottom part is lofted from the back end of the keel / sternpost line to the bottom of the station 11 curve. The final result is a set of surfaces that is an improvement over my earlier attempts at sculpting the hull as a single surface. You will notice there are some thin gaps between the surfaces in the aft area of the hull. I think if I loft using a larger number of faces, the surfaces will fit better to the curves and reduce those gaps. I did try using the edge merge tool in the Sculpt environment, but I ended up with distortions in the smoothness of the surface at the merge areas. It's not as good as the results I got from importing a mesh built in Blender (see my Blender thread for details) but it's the best result so far.

Two weekends ago I assembled a 3D printer kit. I added a keel to the Ernestina in Blender, then 3D printed the model. I probably should have added the bulkheads to make it look a bit better.

As I mentioned in my Fusion 360 thread, my goal in all of this is to create a 3D surface that can be sliced at arbitrary locations to generate intersection curves, then use those curves for a plank-on-bulkhead model. While Blender has some intersection tools, I found the ones in Fusion 360 to be much easier to use. Then the question is: how to bring the mesh from Blender into Fusion 360? The first step is to export the mesh. In Blender, go to File -> Export. There are several file formats available, but the only one that will work out is Wavefront (.obj). Switch over to Fusion 360 and go to Insert -> Insert Mesh. There are a number of import options, including units, and the location and orientation of the mesh. Fortunately, because I was consistent about the scale and location of the origin in both programs, I don't need to make any changes. Here is the mesh imported from Blender on the left and the one modeled in Fusion 360 on the right. I want to get the curves that result from the intersection with a plane. First, I select a plane. Then Sketch -> Project / Include -> Intersect. There is a problem. I can select the T-Spline body that I created in Fusion 360 but I can't select the imported mesh. Apparently the intersect operation isn't allowed on a mesh body. What to do? I'll have to convert the mesh into a T-Spline body first. I click on Create -> Create form to enter the Sculpt environment. I click on the mesh object to select it. Then Utilities -> Convert and Ok. I want to preserve the sharp edge around the transom, so I'll select the edges and click Modify -> Crease. Finally, Finish Form. The new body is composed of a patchwork of T-spline forms. I tried to create a hull in Fusion 360 made of only one piece. Clearly that was the wrong approach. It might be worthwhile to go back to my Fusion 360 attempt and try to build the hull form in pieces. Anyway, with the mesh converted to a T-Spline, it's possible to take intersection curves. By default I have to select piece-by-piece (as in the picture above) to get the full intersection curve, but I get it all at once if I change the selection filter from "Specified entities" to "Bodies". If anyone knows how to do the same intersection operation in Blender, I'd love to see how.

I have gone back to an earlier save. The mesh is still just a flat plane here. This time, I'm going to apply a Simple (not Catmull-Clark) Subdivision Surface right at the start, which gives me four times as many faces. As before, there are vertical lines at each station line and horizontal lines at each water line, but also new intermediate lines. Now I'll work as before, selecting vertical columns of vertices at a station line, hiding all others, and then pulling vertices out from the centerline to lie along the station line. When all the station lines are complete, I have created the USS Accordian. How do I model the curves of the intermediate lines? Well, I can at least find the points where they intersect with the water lines. I select a water line and view it and the halfbreadth view from above. Points already on the water line are from the station line work just done. The points from the intermediate lines can be dragged out to the water line. This is repeated for the remaining water lines. Now it looks like this. Enough points on the intermediate lines are in the correct place to get a good idea of how the line should flow. I go one by one and adjust points on each intermediate line to create clean curves. There is some guesswork and eyeballing here, because there are no plan lines to follow. When the line is done, it's useful to switch from wireframe to solid view and see if the new lines look uniform with the rest of the body. Here it is a little over halfway done. Note that I also adjust the bottom edge of the object to the rabbet line as I go. Two views of the final result. Note in the aft view that this time I carefully constructed the flat transom area. Is this good enough? I'll apply some smoothing, then check the lines in Fusion 360 again. It's much better now, not 100% perfect, but I think I'm satisfied. At the worst spots, it's a deviation of 2 or 3 inches on a ship that's over 100 feet long. Here's a comparison of attempt #1 vs attempt #2. Next time, I'll bring the mesh into Fusion 360 and compare it to my earlier work.

Here are a few views of the state of the mesh. The general shape is there, but it needs more faces in order to smooth things out. (The little info panel at the middle top says there are only 91 faces.) Fortunately, Blender has a great tool for this: the Subdivision Surface modifier. I select the Modifiers icon on the right, which looks like a wrench, then click "Add Modifier" and choose "Subdivision Surface". Modifiers don't take effect until they are applied, and there can be a stack of multiple modifiers which take effect in order. There are two modes: Simple and Catmull-Clark. Simple mode just divides up the faces without changing anything about the shape of the object. Catmull-Clark is where the magic happens. The Catmull-Clark algorithm subdivides the surface, adjusting the new faces to create a smoother surface. The View setting under Subdivisions determines how many subdivisions are actually performed. The image below shows before and after a single level of subdivision. It's looking better. But look carefully, and there is a problem. The tip of the bow is rounded off, as is the aft end of the keel. To fix this, we need to tell Blender not to perform any smoothing at those areas. I go back into edit mode, and at the bottom, switch from vertex select mode to edge select mode. I select the edges that I want to remain sharp, and on the properties panel where it says "Mean Crease" I change the value from 0 to 1. The affected edges are highlighted in a pink-purple color. Now the Subdivision Surface modifier smooths out the hull, but leaves the sharp areas sharp. Just for fun, here are images of one, two, and three levels of Subdivision Surface. Amazing. Incredible. It can't possibly be this easy, right? How close does this mesh match the lines on the plans? I'll import the mesh into Fusion 360 (procedure to be explained later) and plot intersection lines with the station and waterlines. The results are a little disappointing. The curve of the mesh is close at the bow but deviates from the lines amidships. The problem is that Subdivision Surface flattens as it smooths. Here are zero, one, and two levels of Subdivision Surface viewed from the front on the ship. Black lines are the wireframe of the original mesh. From this view, it's clear how the smoothing operation is also shrinking the surface. So much for the miracle of Subdivision Surface. But what if I manually create a mesh with a higher level of detail, matched to the plan lines? Then it won't suffer as much from smoothing. Next time.