[sdiy] Analog division

[David G. Dixon]

> Straightforward, random, and probably wrong idea :D

Not random at all, actually, and no idea can be totally wrong when it's
based on solid math(s).

I like it!  I'll simulate it and let you know how it all turns out.  I've
got a linearized 2164 submodule already built in Multisim, which I drop into
various circuits as required, so it won't be much of a task.

In the meantime, it shouldn't be too hard to determine whether it will be 4Q
or only 2Q.  I'll give that a go as well.