I visit enough sites that some I just have to mostly lurk at in order to balance my schedule, but I am always looking for sites that show great potential for open minded exploration of fact based advancement of our craft. The latter is one of the reasons I post less frequently on many sites due to the mind field one can find themselves in for not being P.C. This site shows great promise, especially with Delbert’s guidance, so I lurk here occasionally but only care to give input where I feel it can be helpful, but more importantly welcome. I sincerely hope this post will meet those parameters for all present, if not I will not trouble you more.
I have spent a many years of my career struggling with standardized testing that will yield real reliable numbers, and attempting to identify trouble spots in the interpretation of testing results. An awful lot of misinformation about what happens in knife steel has resulted from innocent misreading of test results, I have often found that our own eyes will indeed lie to us if we don’t very carefully identify what we are actually seeing. I used to focus on single general tests for analysis, but then I found that the more accurate the test, the smaller the part of the overall big picture I could see because reliable, accurate tests must eliminate as many variables as possible outside the specific property measured. So, often what I saw on the Rockwell tester made no sense until that same sample was placed under the metallograph.
If I can offer a few suggestions to help out with your testing, these are some things that I found and overcame in my own testing along these lines:
Hardness testing, when not done
very carefully, can be more misleading than one may suspect. Portable testers are fine for ballparks on the shop floor, but lack the same certifiable results that one could get from a properly calibrated bench top unit. A single test does not provide enough information, and I always do at least 5 readings in order to average. The readings need to be an even sample pattern across the test piece. The test sample must be a flat level surface, bevels can’t give proper readings as one side of the penetrator crater give more than the other, and edges tend to flex and render the reading worthless. There can be no distortion in the test sample that could cause the least bit of movement, and this is just one reason why 1/16” samples would not work, as well as the thin stock taking the penetrator differently than a solid mass. If you would like to test varying thickness in the same piece, grind the sample in steps, I suggest that you start with ¼ stock and remove 1/16” at a time.
The other reason for this stepped sample is that to go with the thinnest possible section is an unforeseen bias in the results. We think the edge is all that counts but any steel with carbon will harden in almost any quench if it is that thin, resulting in no significant differences across your sample data, and making all quenchants appear equal. But you are but testing whether the edge of any given knife will be hard enough, you are testing the quenching ability of various mediums, that will only be reflected in the depth of hardness in increasing thickness. This is how industry measures depth of hardness and the effectiveness of quench.
A real hurdle is the apples and oranges comparisons in the quenchants. Will they all be preheated to their own optimum temp? Or will you try to even the playing field with one standard temp? Will the oil be the exact same temp for the last sample as the first? Will agitation be exactly the same for all? Eliminating agitation will shift the test from heat extraction to vapor formation. These are just a few of the obstacles I have faced in this line of testing.
A serious problem I spent a few years overcoming is precise duplication of soak times and temperatures. If you get this part off your hardness test will indicate miserable performance in the oil, when the oil never had a chance to work its magic on a piece of steel that was not prepared the same as the others. Decarb and oxidation, as well as the process of removing them will skew the results in ways that had nothing to do with the oil, except that the oxides could prevent proper liquid contact.
I think you are doing good to only focus on Rockwell hardness for the first tests, and doing as-quenched is indeed the only way to go. I would suggest distilled water and not brine as your control, water is the standard by which quench speeds have been measured for years with the scale of “H” factor (water being 1, brine =2.0, medium oils = .30).
Your “quench wars” does sound like a lot of fun, and if my technical geek suggestions could interfere with that please stick to the spirit of your exercise. But if the fun factor and the informal nature is a major component, I beg you gentlemen to embrace the results in the same informal spirit. Many people read things on these forums and take them quite seriously and the results you present could greatly influence decisions people make, there may be responsibility with that. Not that I am raining on your parade, I keep many of my own findings for my own consumption simply because I want to be certain they are error free before going public, and I am too neurotic to ever reach that point.
In fact, I have my own predictions for your results based on previous measurements, however, I will leave them from this post for the same reason I would not want them if I were in your position, until I was done. The human factor and biases based on previous information is one of the largest of all obstacles in accurate analysis.
A book I would highly recommend for you endeavors would be “Quenching and Martempering” by ASM, it would be a treasure trove of information on this topic.
I will shut up now, but wish you gentlemen luck and lots of fun.
Sincerely
Kevin R. Cashen