All,
I've written some of this information elsewhere but decided to reproduce it here so that the thread will be complete, please forgive the repetition.
Having recently acquired a 36 inch heat treating oven and a Chinese Rockwell tester at a good price I have decided to do some testing on saw plate materials.
Why? Because it interests me and I am curious by nature. I like learning new things, I like building saws (among other things) and I want to build the best saws possible. This thread is in the same spirit as my threads on vacuum infusion of handles with resin and the use of bronze bushings/bearings for strengthening saw screw holes in handles. Why? Why not.
I have read elsewhere on the web that Disston hardened its saw plates to Rc 52.
As IsaacS and I have described in another thread on this site, the use of the C scale of Rockwell hardness, denoted "Rc", testing with the 120 degree Brale diamond indenter is limited by a combination of material hardness and thickness. Authorities differ somewhat on guidance but they state that the thickness of the material being tested should be at least three times the diameter of the indentation made by the indenter, other authorities call for 10X thickness. Thus, for a saw plate with a hardness of Rc 52 the minimum recommended thickness is 0.032". Rc testing of thinner or softer materials is not recommended as the scale becomes non-linear due to indenter induced stretching of the test article around the test point. However, I maintain that for materials of equal thickness that are harder than Rc 20 or so that the Rc test probably gives a good reading of relative hardness despite the invalidity of the Rc scale. I have decided to denote this scale Rc* indicating the application of Rc testing to materials that are regarded too thin to test by this method.
Thus, after setting up my Rockwell tester and calibrating it, delta = 0.5 U over the range 27.2 to 62.2, dead on at 46.1, I set about measuring the hardness of my personal stock of saws.
My first test was of an old (1917-1928 era I think) Disston D-8 I acquired on Ebay. The plate thickness measured at exactly 0.032" at the tooth line. Rc testing, at five different points, yielded five measurements of Rc 52.
My second set of tests was performed on the five commercially available dovetail saws in my collection. These are from two manufacturers. I have one from manufacturer 1 and four from manufacturer 2. I also have a strip of 0.015" 1095 spring steel that I use for my straight handle gents saws. I used this material as a standard. Like many 1095 strips, this one is reported to have a nominal Rc hardness between 48 and 52. I meausured Rc* 42-43 with my tester. The saw from manufacturer 1 gave back readings identical to those taken on the standard 1095 strip. I did multi-point measurements of manufacturer #2 saws and found that they ranged from Rc* 35-36+. All of the plates at the points tested were 0.015" to the best of my ability to read the micrometer. I conclude therefore that manufacturer 1 used 1095 straight from the roll or something of the same hardness even if the chemistry was not the same. Manufacturer #2, given the consistently lower Rc* readings, used steel that was significantly softer.
I then moved on to another saw, this on from manufacturer #3, that had a 0.030" thick plate at the test points. This plate gave the astonishingly low reading of Rc* 11. I have not been too pleased with the performance of this saw and visual inspection of the indentations produced by testing are consistent with the Rc* readings - it is soft.
Not wanting to spend too much time at this point examining the work of others I then began heat treating experiments. I haven't taken pictures because the process is visually boring so I will just report the results.
I have found that 0.015" 1095 steel can be hardened to Rc* 59+ using a range of soak times at temperatures from 749 to 807 Celsius. Soak times range from 45 minutes to 3 minutes for this material over this range. Given that I could produce adequate hardness I investigated annealing.
Googled values for 1095 and a quick Excel calculation (R squared of the fit line = 0.9957) revealed that 1095 could be annealed to Rc 52 with soaking at a temperature of 342 Celsius for two hours - longer than necessary but I am using residual heat for economy. This worked out according to the book, no problems at all. The hardening and annealing produced nice uniform hardness throughout the material and I was easily able to hit my mark of around Rc* 45-46.
Next comes the quenching. I started out using room temperature water. It was easy to get hard material but the temperature gradients that arose when the metal was transiting the surface of the quench bath caused severe distortion. I then proceeded to try hot water, cool brine, hot brine cool oil and hot oil. The best results so far have come with hot oil but I am still not satisfied.
Unfortunately I am stopped as I am waiting on the components needed to make a molten salt bath. I'll post more results when I have them.
Cheers,
Rob