We were taught all the basics of perspective at college. Horizon lines, vanishing points, grids and all that. We'd spent weeks drawing all sorts of combinations until it stuck in your head. I think it was when we started drawing actual products that the sphere came in. As Tony says, you can't really draw an accurate cube without it. It just sort of happened. I was shown how to do it, understood it and didn't have to be told again, and I've used it ever since. You can work out the ellipse angles and dimensions from it. Very clever.
I was lucky to work with quite a few Technical Illustrators and would pick up ideas and techniques from what they did. You never stop learning and I wish this thread had been around 30 years ago
This might help!
I learnt perspective drawing at college, the basic methods were well taught, though measuring point perspective was of no use whatever.
When I left college I went to work and your methods evolved from other more experienced illustrators.
The basic methods we used on Flight International are probably not unique, but they are born out of experience and needs of this particular job. Perspective grids are of little use, as the viewpoints are too restricted and vanishing points are impractical (never get them on the board), neither are they necessary.
The key to most things is the ellipse, accurately drawn and related ellipses can produce accurate perspective bases for drawings, although one still needs a good "eye" and some experience to make the most of it.
A much more precise method is illustrated in these diagrams. To start, use a general arrangement drawing, take for example this side elevation of the BF-109 (fig1), using a suitable dimension, in this case the total depth of the fuselage to produce a square, the square is then projected to cover the total fuselage length. Draw the ellipses (fig 2) as described earlier to suit the viewpoint, inscribe a square around the ellipses following the fuselage axis, this then becomes a square of the correct proportions for the viewpoint (fig 3).
Extend the top and bottom lines into perspective vanishing lines A and B (fig 4). This is where experience counts, as the amount they converge must be judged according to the size of the subject and the view taken. Next draw a centreline C between them and then project a diagonal D. This will give the position of the vertical that defines the next square; continue for the desire number of squares.
This is the grid from the "squared-up" side view, but in perspective and with the correctly related vanishing and foreshortening. Upon this grid it is easy to plot the side elevation of the BF-109 in perspective (fig 5), add to this a number of cross-sections (the width can again be determined by the ellipse on the lateral axis). A sufficient number of cross-sections will give the true outline of the fuselage (fig 6), the wing can be plotted simply by projecting from the centre line, the span measured from the appropriate ellipse.
This may seem a bit complex, but in fact is all quite simple and not very time consuming. The grid must be drawn accurately as inaccuracies can be compounded. Perspective based on ellipses and not on vanishing points can be very accurate.
Having produced a basic drawing the next stage is a fairly demanding one, using the frame and rib stations drawings, all of the ribs, frames, spars, control surfaces, engine positions etc have to be plotted in. It can take time to get for example, 80 or so unequally spaced frames to work out! At the same time frames and ribs are drawn in the correct section but just as single lines.
During this procedure the accuracy of the original drawing is proved, things just don't fit if it’s not right. It is probable that small corrections will have to be carried out to accommodate subtleties in shape.
There are various ways of plotting, checking and cross-checking detail accuracy, the most useful tool is a pair of proportional dividers, using the "goes into" method you can proportion one part of the drawing to another.
Next all the details are pencilled in using engineering drawings for accurate placing, shapes etc, and photographs for character, texture and small details. All this depends on the availability of the various kinds of information. No two jobs are the same.
The appearance of a finished drawing in no way reflects the effort that has gone into it. An apparently simple drawing, because of the circumstances may have required much more skill and effort to produce than many a more impressive one. In every case though, when the drawing is finished and cleaned up and the pencil rubbed off, one wonders what took so long!