In the previous post I introduced a new method of using photos. I fit the projection of my 3D model into the contours of the same airplane depicted on a high-resolution photograph. I can use such an arrangement as a precise reference. It is a good idea to verify the basic body of the fuselage in this way, when there are no additional details. All the differences that I will find now will save me a lot of troubles in the future. For example — what if I would find that the base of the cockpit canopy in my model should be somewhat wider, when this canopy was ready? I would have to fix both shapes: the canopy and the fuselage. And what if I would already recreate the inner fuselage structure — the longerons and bulkheads — before such a finding? I would also have to fix them all. This is a general rule: the later modifications require much more work than the earlier ones! Thus I have to check everything when the model is relatively simple. You can compare the differences I will find in this post with the plans I published earlier: they contain various minor errors! Just as every drawing.
Last week (see Figures 29-5, 28-7, 29-8) I discovered that the bottom contour of the tail was somewhat lower than in my model (Figure 30‑1):
The main problem with mapping the tail shape was that its bottom part behind the wing was wide and completely flat. On every photo that I have the airplane is more or less deflected toward the camera, so the precise bottom contour of the SBD tail in the side view is an average of multiple estimations. That’s why it can be wrong on my scale plans! I also found a minor difference in the forward part of the fuselage below the wing. However, its forward part on the photo above is obscured by the truck. Nevertheless, I guess that the forward part of this cover it had a straight side contour, located minimally below than this contour in my model. To check this I mapped another photo of the firewall (Figure 30‑2):
This photo confirms my observations from the side view picture. Although the bottom of the fuselage here is lacking the bottom covers, the corner of their mounting flange “touches” the bottom contour of my model. It means that the real contour was somewhat lower, more or less along the yellow line that I sketched on this picture. However, you can see here another difference: the upper part of the firewall is little wider than the elliptic contour that I assumed (It seems that the shape of the firewall was not an ellipse, as I assumed in one of the previous posts).
To make sure that this is not a mistake in the matching the model and the photo (or the effect of a barrel distortion), I also used another picture, from other restoration (Figure 30‑3):
The photo above confirms that the fuselage was little bit wider at the cockpit edges than it is in my model. The trace of the bolt seams on the wing reveals another difference: the wing root fairing was also wider (at least its forward part).
To make sure that this difference is true, I have to find it on every photo that I map onto my model. Thus I mapped two other photos. They come from Pacific Air Museum. I can use them to verify the width of the mid-fuselage and the span of the rear part of the wing root fairing (Figure 30‑4):
The good news is that the maximum width of my fuselage perfectly matches the photo (a good luck!). I found that the width difference at the cockpit edges found at the firewall is (approximately) constant along the whole length of the cockpit (Figure 30‑4b). It disappears behind the cockpit (i.e. in the front of station 140). The wing root fairing was somewhat wider at the trailing edge (Figure 30‑4a).
As usual, I used another photo to confirm these findings (Figure 30‑5):
Because this photo depicts the whole fuselage, I had to check these details using higher zoom factors. This photo confirmed what I have found in the previous one. In addition, it seems that the width of the fuselage in my model matches the real contour of the tail up to station 271.
Once I confirmed all these differences, I had to fix my model.
The wider wing root fairing behind the trailing edge can create impression of lower tail contour on the photos taken from the side (like this from Figure 30‑1). This is because none of these photos is an ideal side view shot. In each of them the camera was located above or below the fuselage centerline. That’s why I decided to begin by fixing these differences in the fuselage width (Figure 30‑6):
Once they were corrected, I could fit the side contour, matching it to the horizontal photo of the tail (Figure 30‑7):
As you can see in this picture, I also minimally modified the upper contour of the fuselage. (Because the upper arc of its cross sections was looking like a part of a flat ellipse, while it should be a regular arc).
Figure 30‑8 shows the ultimate differences between the reference drawings I created several month ago and the contour obtaining from matching the 3D model to the photo:
As you can see, nobody is perfect, so I also did some mistakes. However, I was aware that the bottom contour of the fuselage was a guess: I did not have any photo where it was directly visible. All the pictures were taken from below or above, leaving some space for various assumptions (which often results in some errors).
Finally I fit the covers on the bottom fuselage (below the wing) to their contours in the photo (Figure 30‑9):
I moved slightly downward the forward part of these covers. As you can see in Figure 30‑9a) it was a relatively small difference. Initially I assumed that the cross sections of this bottom fuselage were elliptic arcs. However, in such a case, for the given width and height (from the side and bottom views), the edges of the wheel bays would appear a little bit lower than those visible in the photos. Thus I think that the contour of the middle cover in the front view had a slightly different shape (as depicted in Figure 30‑9b).
In this source *.blend file you can evaluate yourself the model from this post.
While matching the model to the photo from Figure 30‑1 (PAM-3) I noticed slight differences in the wing rib shape: it seems to be a little bit thicker than on the picture. I will analyze this and other differences of the wing in the next post.