I used to visit with Fred when I went to my violin making sessions at Carleen Hutchins' in NJ. In spite of spending a fair amount of time talking about back bracing, I don't really remember exactly what he did. I do remember the objective, though, and worked with a number of different back bracing schemes over the years since to try to realize it.
Fred had done some computer modeling using electrical circuit analogies to get a handle on the low range function of the guitar, in the 'bass reflex' range (roughly up to about the open G string fundamental). He had noticed that guitars that he liked tended to have a monopole resonance ('main back' mode) that was not too much higher in pitch than the top monopole. When that happens you get coupling between the top and back modes that effectively increases the compliance of the box, and lowers the pitch of the 'main air' resonance, which is the other half of the 'bass reflex' action. His objective was to come up with a back bracing scheme that would allow him to 'tune' the pitch of that back mode, and also enlarge the area, which would increase the coupling strength.
You have to keep in mind here that Fred was also working with 'free' plate modes at the time (that's where I caught the bug). Since a normal Torres-braced Clasasical top will produce a very clean 'ring and a half' or 'ring+' mode there was some hope that a similar sort of thing could be done with the back. Again, at that time Carleen was exploring the 'free' plate modes of violins, looking to see if there were relationships between the top and back plates that produced 'better' instruments, so there was some thought maybe guitars would be similar.
IIRC, the first layout that Feed tried was a sort of lattice of four braces; two parallel member that ran along the length of the back, and two that went across. He was able to tune the free back to produce a well formed 'ring' mode, and that gave a strong monopole mode on the assembled guitar. The advantage of this was that you had independent control of the lengthwise and crosswise stiffness of the back, so it was relatively easy to get the correct stiffness ratio to produce the ring. What was harder, apparently, was to get the ratio right and also the overall stiffness, so that the pitch of the assembled back monopole was where he wanted it; about a semitone higher than the top monopole. That lead him into other brace layouts. I don't know if he ever settled on something as a 'standard'. The last time I saw him was, briefly, at the GAL meeting in '92, at which time he told me he'd given up on 'free' plate tuning.
In the years since I've tried out a half dozen or so different back brace schemes, many involving 'X' bracing of one sort or another. One of the tricks with X bracing is to figure out the 'correct' angle for the X for this piece of wood. The long-grain to cross grain ratio of back wood varies a lot: a quick look at my data shows values for Indian rosewood ranging from 3.5:1 to 7:1. Even with that data in hand it's hard to get it just right, and then you still need to figure out how to get the absolute stiffness to weight ratio right so that the pitch is correct.
As it turns out, it seems to be easier to get things to work well using plain old ladder bracing, but not necessarily the way it's often done on Classical guitars. Most makers seem to hold with the 'reflector back' model, which envisions the back as a rigid element that doesn't move appreciably, particularly in the low range. What Fred was after was an 'active back', and that tends to call for a lot more flexibility, particularly in the lower bout, than the usual tall and narrow braces will allow. You can shave them down, and particularly the lower one, to make the back more mobile, but that ends up with some pretty small braces that don't inspire much confidence.
What I found was that using four back braces, instead of the usual three, and making the lower two much lower and wider than usual, worked better. Thus is the back brace scheme that the Martin company has used for a long time, and, although they do now mostly make steel string guitars, the designs before the Dreadnought in the late 'thirties were meant for gut strings.
With this setup it's fairly easy to get an active monopole below the waist. Making the waist brace low helps extent that up into the upper bout some, at the expense of dropping the pitch, perhaps too low. I'll note that I had the same issue with the X-brace back patterns I tried; extending the monopole up into the upper bout to any appreciable degree required making the upper part of the back more flexible than I liked, both acoustically and structurally.
So, I'm sorry I can't tell you more about what Fred did (I miss that guy!). Maybe my own experience helps a bit, though.