Raptor 3D Project - Edits 01

Get your 3D workspace set up right, before beginning.

The drawings to the right are the ones I've chosen to use at this time, as my primary reference. aided by others, and photos. These drawings match the plain black and white "general arrangement" drawings seen around the net, but have more information, showing some internal details... Nice! I don't know where this image is located, on Lockheed-Martin's site, because they move things around, but you can find it via Google. If you do that search, turn on the "extra large images" choice in Google's image search, to get the biggest one you can.

I sliced these up into four separate images, and checked them against each other. (to be sure, for example, that the top view matches the side view) They do match, so I then removed all the margins, (placing the edges of the actual aircraft right at the edge of the images) and scaled them to the same overall size. (top view same length as side view, for example) That process lets me them use them as texture maps on properly-sized "planes" in my 3D software, as in the image below.

Here are the sliced-up and scaled drawings, used on three different flat planes, in my 3D setup. They're helpful in referencing things like the overall shape, position of the tail and landing gear, etc.. The image above was made after the first sets of edits, which are described below.

You may notice in the image above, that there's a small gap at the bottom front of the side view drawing. That's intentional. The drawing shows the aircraft as it would appear on the ground, with the "squat angle" of about 1/2 degree, due to the slight flattening of tires under load, and the compression of the nose gear strut. To align that drawing while keeping the model (horizontally) level, required tilting it slightly, and that's why you see the gap. This comes from a simple assumption that, when in cruise flight, the aircraft's control surfaces should not be deflected. (That would create drag.) So, drawing a line through the center of the tailerons and wings gives a horizontal datum line that is reasonably trustworthy. A simple detail like this will bite you later, if overlooked, because you will have built all the other angles wrong. So, the most important part of the working setup is to define what "zero" is, in every axis, and align the model, based on that.

With the model at "0,0,0" in my 3D workspace, I can compare the overall exterior shapes in 3 different axis. These "planes" are grouped together and labeled "Isometrics", (including the ISOMETRIC camera) so they can be made visible or invisible with just a mouse click. In the image above, you're looking at what Carrara calls a "conical" camera, meaning that it has variable focal lengths, and shows both convergence and divergence, as in real life. That is NOT the way you want to work with details. For those, I switch to the isometric camera, which gives a view along each axis, directly in line with the drawings, and shows no divergence or convergence. (next image, below)

Opacity helps a great deal...

In the image above, viewed through the overhead isometric camera, I've made the colors of the model's exterior semi-transparent, and tinted them a reddish color, to make it obvious where the model is, relative the the underlying drawing. The model itself has no texture maps, so the paint scheme you see in on the drawing, under the model. This image was created after my first few days of work on the model, and you can see that most of the major parts now match the drawing. Notice in this view, for example, that the shape of the left rear taileron on the model still doesn't exactly match the drawing. (It's forward inboard edge needs to be extended just slightly.) Of course, this whole process, while important, will only verify the exterior outlines, and not the 3-dimensional shapes that will be refined later. I think of this as "setting boundaries" for that future work.

There are four versions of the documentation to compare.

At this point, I had Lockheed-Martin drawings, cross sections from my sliced-up plastic display model, lots of photos, and the existing 3d model, to compare with each other. I've decided to give the Lockheed-Martin drawings the highest "believability factor", and defer to them when there's a discrepancy.

Here's the Lockheed-Martin color cutaway, with the opacity turned down to 50 percent, so you can see that the (Lockheed-Martin) line drawing underneath it does indeed match. (aside from the landing gear "squat angle", mentioned earlier) That's what made me believe that the color cutaways are a good as anything else, as a reference. The red line passing through the image shows the cruise "datum line" that I believe is valid for alignment of the major parts.

How do the plastic model's cross sections compare to the Lockheed-Martin drawings?
And... How do they compare to the existing 3D model?

For the next test, I scaled my cross-section drawings from the sliced-up plastic display model to the same size as the others, put each section on an invisible "plane" in 3D (PNG images with Alpha Mask) and here are the surprisingly-good results.

My original scans, produced from Revell's (discontinued) 1/48th scale F/A-22

The plastic model's cross sections, mirrored in the horizontal axis, and stacked up.
(The small lines sticking out from the canopy simply indicate there the canopy frame is.)

The cross sections overlaid on the Lockheed-Martin drawings.
These match too closely to be a coincidence. They must have used the same drawings.

A 3D test of the same items...

There are 18 cross sections here. So, I placed each of them on a transparent plane in my 3D software, and rendered them alone.

With the existing 3D model visible, the differences between it and the Lockheed-Martin drawings begin to surface.

Another view...

Prepped and ready for surgery!

With all this evidence in hand that I had good, matching documentation, it's time to move on with the modeling. I want to mention again that my rebuilding of this model is in NO way any sort of negative comment on the original author's work. He never intended that his model be the same type as what I want to create, but even so, it's quite excellent work. (!) Without his very nice model to start with, this project would be more tedious. Thank you, SIR!

A "Top-Down" Approach:

Simply put... This project will (mostly) follow a hierarchal approach, where the obvious parts of the general arrangement and shape will be edited first. Next we start to add various "bumps and bulges", interior items, and refine the existing shapes to match the known documentation. It makes sense to wait until late in the process to do all "openings", like the cockpit area, weapons bays, landing gear bays, and so on. You wouldn't like it if you modeled a landing gear door, then changed the fuselage, because that door would then not fit. This model was delivered to me with all these openings already done, so you'll see how I solve that problem, in the article that follows.

These are the issues I need to consider BEFORE moving any polygons around. As always, one thing affects another. Move the canopy, and you then have to move the canopy frame, cockpit interior, and the underlying fuselage, to fit. Do that, and you affect the subtle curvature of the fuselage, requiring some polygon-level corrections. It can snowball quickly. Think ahead, and go slowly, for best results!

(Click the "Edits 02" link below to continue.)