In previous post I created a simplified model of the SBD fuselage that helped me to identify the eventual troubles in the modeling process. In this post I will create the mid-fuselage (more precisely: its upper part).
I always try to think ahead about the mesh topology required for a given shape. In the case of the subdivision surfaces that are used here, this approach is extremely useful. When you place vertices of the initial bulkhead in the proper places, it greatly simplifies further modeling. To mark some “longeron” edges as “sharp” (Crease = 1), I started with a thin mesh “strip” instead of a single contour (Figure 23‑1):
As you can see in the picture above, it is built around the cockpit sides. Looking on the photos you can notice that one pair of the main longerons forms the side edges of the cockpit. It will be the upper edge of my fuselage. (The part below the windscreen seems to be a separate assembly, riveted over the longerons (see Figure 23‑2a). I will create it later:
Figure 23-2 Details of the cockpit edge
I recreated the small fillet around the cockpit edge using two parallel edges placed close to each other (Figure 23‑2b). I could obtain similar effect using a single, partially sharp edge. However, in such a case the contour of this fillet on the fuselage cross section would have a shape that significantly differs from the circular profile in the real aircraft. What’s more, I will split these double edges at the rear edge of the cockpit opening. I expect that in this way they will help me to shape its rear, rounded corner.
I extended the initial mesh strip from the firewall to station 140 (station locations — see previous post). After fitting vertices of this “bulkhead” edgeloop around station 140, I inserted in the middle of this mesh a new edge, just at the end of the skewed station 54. Then I removed the bottom fragment from the rear part of the resulting mesh (Figure 23‑3):
To avoid the curved contours of this mesh in the top view, I directed the lengthwise edges little downward in the second segment of this fuselage (Figure 23‑4a):
To verify if the cockpit sides are really straight in the top view, I placed (on the tools layer – 10) many straight “stringer” probes (Figure 23‑4b). All of them are horizontal, arranged like the real longerons in the airplane (Figure 23‑5a):
For the properly shaped surface, these probe objects should minimally protrude from the fuselage skin (as in Figure 23‑5b). I used them to apply small adjustments to this mesh.
When the cockpit sides are ready, I recreated the upper part of the station 140, and extruded it toward the cockpit. In this way I obtained the initial strip of the tail upper surface (Figure 23‑6a):
It is relatively easy to prepare in this mesh a rectangular opening for the gun doors. The general rule of the subdivision surfaces is that their sharp edges (i.e. edges which Crease = 1) have the same shape as the free edges (for example — opening borders). Thus, if I incorporate into a smooth surface an area encompassed by sharp edges, I can later remove its inner faces without altering the shape of the outer mesh faces.
But how to obtain a smooth surface around a sharp edge? It is simple: place it in the middle of a flat face of the control mesh. I did so. As you can see in Figure 23‑6b), it is enough to make the three vertices on every bulkhead collinear. (In practice, small deviations from the theoretical line still produce acceptable results).
You can learn more about this and other useful properties of the subdivision surfaces in Vol. II of the “Virtual Airplane” guide.
In the next step I cut in this mesh strip the skewed rear edge of the cockpit opening (Figure 23‑7):
After removing the unnecessary vertices, I created a few new faces that finally joined this fragment with the rest of the fuselage mesh. As you can see in Figure 23‑8, I also inserted another edgeloop just after the cockpit rear edge (Figure 23‑8):
This additional edgeloop and the few vertices in the corner control the surface curvature around of the cockpit opening. As you can see in the picture above, one of the resulting mesh faces has five vertices (so-called n-gon). In general, it is possible to decompose it into a triangle and a quad. However, I carefully examined the resulting surface and decided that this additional vertex does not deform in any way its smooth shape. Thus I decided to leave it “as it is”.
Note that there is a single vertex in this mesh that controls the shape of the fuselage skin in the corner of the cockpit opening (Figure 23‑9):
I modeled it to resemble the original shape as I can see it on the photos. However, I will return to this fragment during the detailing phase of the modeling. With the cockpit canopy in place, I will then re-examine my photos and decide about the further details (for example — cutting the smaller openings for the ammunition feeders on the gun doors sides, which were introduced in the SBD-3).
In this source *.blend file you can check all details of the model presented in this post.
In the next post I will continue working on the fuselage.