[Andrew Garrett]

Thanks for the input Samurai.

Yes, I agree. Testing on the hollow..., not optimal. But..., I'm at a loss for options. I guess I could go back to the old file tests.

In any case, not satisfied with where this thread was going (making it look as if I'm a heat treating idiot), I took my broken blade back out to the shop when I got home from work and shed my uniform. I did about 15-20 more tests at different spots along the hardened edge. Some, where the hollow allowed space between the anvil and the blade, failed terribly--those readings were in the 30s or 40s. Some however, closer to the edge and those tested on a quickly ground and modified spare pipe anvil with a smaller diameter, measured as high as 64. I got that number 4 times.

Also at issue is my finish. As you see in the photo, it is a 60grit finish. This would soften the numbers too.

Additionally, I calibrated on my testing block several times and it too was inconsistent. These Rams Rockford testers use a spring rather than weights and temperature, humidity, etc. might alter results. I'm not sure, but I've never been totally happy with the device. When the numbers were consistent tonight, they were routinely about 1.5 points under test value (55.8) The standard variance is plus/minus 1.

Then, I went to the tattoo studio and retrieved my 12x loupe to closely examine the break and grain size. I was given a cheap coffee cup recently and decided to sacrifice it in the name of science. I broke of the handle and compared the textures--very similar. If baby powder is #1 on a scale from 1-10, and playground sand is #10, my grain looks to be about a 3 or 4. More scientifically, I estimated the break surface to be similar to a 320g grinding belt (not the finish they leave--the belt itself).

In conclusion:
I feel that my testing methods are flawed as my equipment is not adequate for testing edge-quenched hollow-ground blades, and my finish is not fine enough for reliable testing as evidenced by inconsistent results.
I feel that my HT is solid, but my testing is not.