Assuming it's a symmetrical lens, which the notion that it's supposed to be convertible suggests, it would be a 180mm lens. You can verify this by focusing your camera on infinity and measuring the distance between the nodal point of the lens and the film plane. It doesn't have to be super exact, of course. Then the aperture is pretty easy; assuming your 90mm is f/5.6 natively, this means a 16mm physical aperture. Let's say that the 'converted' lens is indeed 180mm, then the max. aperture would be 180/16 = f/11. No surprise there, since it's exactly a one stop difference - but the calculation will also work if you find the focal length is somehow very different. The other aperture values can be as easily derived.
I've used such 'converted' lenses in a pinch - regardless whether they were supposed to be used that way or not. The results can be quite satisfactory.
It's not quite that simple, although what you propose will get within a fraction of an f-stop which is moderately close.
- Half of a symmetrical or near-symmetrical lens is not double the focal length, and the aperture also doesn't scale exactly. For example, for older convertible Schneider Symmars that are fairly symmetrical and marked for full lens and the rear-cell-only, you have eg 150/5.6 converts to 265/12; the 210/5.6 converts to 370/12. The ratio of focal length of cell to total lens for a Symmar is usually about 1.75 and the ratio of f-stops is about 2.1.
- There are two reasons for that: 1. Focal lengths of the pair don't double precisely, any asymmetry and the separation of the principal planes also matter. 2. It's not the physical diameter of the aperture that matters, but the apparent diameter viewed from the front, and the aperture is usually magnified by the front cell by some modest amount. That's why the f-number gets slower-than-expected when you take the front cell of a Symmar off. Both of these factors depend on lens design, so the factors for the Symmar won't tell you exactly how the Angulon behaves.
- The formula for the combined focal length of simple thin lenses is 1/f = 1/f1 + 1/f2 - d/(f1*f2), where d is the separation of the lenses. For complex lenses, it's the same but d is the separation of the principal planes. You can't tell where the principal plane is just by looking at the lens, although you can do experiments to find it (such as tilting the lens and looking for zero image shift). Anyway, nonzero d is one reason the Symmar has that focal length ratio of cell/total = 1.75, not 2. For more on principal planes and nodal points, see
https://en.wikipedia.org/wiki/Cardinal_point_(optics)
- When you take off one cell, the remaining cell of a Plasmat, Angulon, etc is highly asymmetrical. The principal planes of such an asymmetric lens can even be outside the lens body. For a fairly symmetric lens like the full lens, the focal length is usually fairly close to the distance from lens center to film when focused at infinity (which is close to the flange-focal distance). But for the very asymmetric lens, that assumption is less accurate. A way to measure the focal length of the cell more accurately is to focus at infinity, and then measure the extension needed to focus at some reproduction ratio, like 1:1. The extension past infinity to 1:1 is exactly one focal length. You can then use the apparent size of the apertures to figure out a new aperture scale.
- Symmetric lenses have several aberrations always corrected. Technically this is only true at 1:1 and for fully symmetric lenses, but in practice it works pretty well for other focus distances and not-quite-symmetrical lenses. Using just one cell gives that up, which of course may be useful pictorially, but it's one reason that the old manuals suggest using a yellow filter on a convertible lens for B&W, to reduce aberrations.
Now, all of the above might be making 10-30% differences in inferred focal length and f-stop number of the individual cell, which might not be that significant for pictorial work. I think it's useful to understand the optics foundation, though.
