OK, I have regrouped after my dismal failure.
It boils down to not having heated up the spine of the blade enough during the draw-back. I had forgotten the exact procedure taught to me at the ABS Introduction to Bladesmithing class I took five or six years ago. In researching for this testing I found a website that describes the procedure and in the description the bladesmith suggests heating the spine to a blue/white color. I did that and found it did not soften the spine of the blade enough resulting in the broken blade of a few days ago.
I have no doubt that the procedure described by the Mastersmith on his website works well for him. it just didn't work for me. With so many variables it is impossible to exactly re-create what another smith does. We all have to find what works best for ourselves.
By all the testing I am doing my memory was jogged and I remembered that in my ABS class we turned off the lights during the draw-back of the spine and heated the spine to a dull, dull red, then let it cool to room temperature. One smith I talked to suggested doing that three times. The difference in temperature from blue/white with the lights on to a dull, dull red with lights off could only be a couple of hundred degrees but it was enough to do the trick.
I drew back the spines on three knives, the first one I cycled once, then I cycled one twice and finally I cycled one three times. I want to see if multiple thermo-cycles would make much of a difference.
This is a blade I did two draw-back cycles to. It's almost to 90 degrees. The objective is to soften the blade to about spring hardness to distribute the bend along the length of the blade. If softened too much the blade could kink at the vice jaws and cause a crack of more than the permissible 1/3 width of the blade.
Angel said "smile". That's my version of a smile.
I kept going past 90 degrees just to make sure I was making progress.
Here's the spring-back after the bend.
No cracks.
While we're in the subject of testing, here's some shots of the rest of the performance test.
I tried several different angles for my secondary edge, I used bevel blocks on my belt sander to maintain consistent bevels. This one is at 12 1/2 degrees. Added to the 2 1/2 primary bevel it makes for a 15 degree edge bevel or 30 degrees inclusive.
It can be a little risky sharpening a blade with the edge up but, I could find no other way to do it with a bevel fixture. After the bevels were established the hard edge between the primary and the secondary bevels was softened to make it a convex edge.
The thirty degree bevel cut the one inch hemp rope very easily.
It's a lot of fun cutting rope, six inch pieces pile up fast.
Of all the angles I tried, the thirty degree bevel handled the 2 X 4 cuts the very best. It cut the wood multiple times with no damage.
I ran out of arm hair a while ago, and a new crop was too slow to grow in, so I made up these test sticks with bear hair, caribou hair and fox fur stapled to them. I was going to say that my wife has copious amounts of leg hair I could test my edges on but that probably wouldn't go over well, besides, I would never be able to get her to stand still long enough. This is just better all the way around.
The thirty degree edge bevel cut all of my samples really well. This is the caribou hair.
The bear hair probably resembles human hair the closest though it is somewhat courser.
The fox fur, being the softest, and most apt to fold over, was the most challenging to cut. The thirty degree bevel did the job surprisingly well.
I'm pretty happy with this outcome. I'll do the same thing with two more blades to make sure it wasn't a fluke. If the results are the same I will make my test knife with the same heat treat and sharpening procedure.
Stay tuned, more testing and the designs for the presentation knives next time.