I recently helped a client reverse engineer some carburized machine parts that had an excessive lead time from the OEM. It occurred to me that it would be a good idea to run a test piece that could be destructively tested after heat treatment, but we all know how it goes when we’re in a rush.

I did not think of this useful idea until the ideal extra piece that could have been re-purposed from the part itself had been turned into chips. Thus, we had to get a piece of the nominally same alloy. This was a rather large chunk, intended to mimic the thermal mass of the parts themselves. The heat treater had a smaller test piece of the same grade, yet another heat lot. We ran the parts along with both test pieces, and the test pieces, both large and small, came in with much lower effective case depths than the desired range. After a lot of back and forth, the heat treater told me he thought I wanted total case, not effective.

Then, I asked, why did we sit and pore over hardness profile data so I could show you what I was trying to achieve? Did I show a single microstructure image with an etched case? NO!

I told him that, in my opinion, the heat-treating industry should give up using total case depth for case-hardened steels. It has been proven by multiple round-robin programs to be the least repeatable method of measuring case depth. More importantly, it doesn’t tell you anything about the strength of the part.

Well, once again, we learn from the school of experience. The case on the test pieces was lighter than I had wanted. Finally, I remembered that I wanted to actually look at the certs for the actual part and the two test pieces. The part ended up having over 1.5 times the levels of nickel and molybdenum, the two most potent hardenability elements in 8620. So, the parts most likely have a significantly greater effective case depth than the test pieces, but I am once again reminded of the difficulties inherent in even what I thought was going to be a rather straightforward specification.