Darren:
I did a break with a very deep bend on a new carved back a couple of months ago at full thickness without removing the big channel of wood like the Chandler book instructs or as Arnold mentioned. The results were very pleasing. PM me and I'll send a few images if you'd like.
I tried this in the past with just mediocre results using a simple center support beam. I originally bent the break on this back about 18 months ago in Portland, Oregon in the spring. I had a couple of things going on in life that year and knew that the bass would be in storage in the high desert for about six months and then there was a good chance that it would be moving with me to North Carolina. Because of those things, I chose not to join the back or top to the ribs during that period for fear of them splitting.
A year later, my hunch was correct and the seam from the bend to the button split open and the break bend had relaxed to almost nothing. Time for some majic and to rebend the entire thing. The center seam support is a good idea, but since I realized I only had one good shot to get it perfect on this one, I went a lot further.
I made a complete mockup of the desired finished back curvature out of MDF (medium density fiberboard) glued up in 3/4" layers until I had a giant, 70 lb., 24" x 30" by 8" thick block. I cut out a slightly oversized shape of my ribs (don't forget to add the back button!), and then spend about two hours with a large plane and carved out a mock up of my finished back with a slightly larger bend than I was after, to compensate for some springback.
I screwed the whole beast down to my workbench and starting at the button, I clamped my unbent carved back to the end of the "mold". If I was to clamp the back to the mold near the c-bouts and go the other direction, I would have a lot of tension to deal with; working from the button back, a gentle downward push allowed gravity to do it's thing.
Using the standard wet cloth over the top of the maple, I used a clothes iron on the steam setting and worked the entire piece for about 25 minutes, doing my best to minimize the area over the joint. As the back would gradually bend into position, I clamped the perimeter down to the mold. I let the entire setup cool completely overnight and then repeated the process two more times, but never removed the back from the mold until it was finished.
The results were just what I was hoping for. I started out with the carved part of the beack ending right where the bend begins. The finished effect visually appears as though the two morph together. The bend goes from right about 8" rib thickness at the upper corners down to 4.75" at the button / heel junction. Pretty extreme, but that is what I was after. The bass has a few other fun things going on also. I'll post images of it next month.
After all of that, I'm glad I did it and have sweet bragging rights when I show it off, but I have to agree with Arnold in that I probably won't do another one. Getting that done was a major crux and stressor for me, even after having built 157 instruments over the years (this was the first complete carved bass). All things considered though, it was sooooo much easier than doing a difficult restoration when somone brings their bass in to me in pieces stuffed in a large hockey gymbag....
I don't have any more need for the mold I used, so it is available if anyone is interested.
j.
www.condino.com
Do you think the same amount of concern would apply to a roundback as well, like is being discussed here? 
