From this website
http://steel.keytometals.com/Articles/Art140.htm
Rockwell Hardness Test
The most widely used hardness test is the Rockwell hardness test. Its general acceptance is due to its speed, freedom from personal error, ability to distinguish small hardness differences in hardened steel, and the small size of the indentation, so that finished heat-treated parts can be tested without damage.
This test utilizes the depth of indentation, under constant load, as a measure of hardness. A minor load of 10 kg is first applied to seat the specimen. This minimizes the amount of surface preparation needed and reduces the tendency for ridging or sinking in by the indenter. The major load is then applied, and the depth of indentation is automatically recorded on a dial gage in terms of arbitrary hardness numbers.
The dial contains 100 divisions, each division representing a penetration of 0.00008 in (0.002 mm). The dial is reversed so that a high hardness, which corresponds to a small penetration, results in a high hardness number. This is in agreement with the other hardness numbers described previously, but unlike the Brinell and Vickers hardness designations, which have units of MPa, the Rockwell hardness numbers are purely arbitrary.
Major loads of 60, 100, and 150 kg are used. Since the Rockwell hardness is dependent on the load and indenter, it is necessary to specify the combination which is used. This is done by prefixing the hardness, number with a letter indicating the particular combination of load and indenter for the hardness scale employed. A Rockwell hardness number without the letter prefix is meaningless.
Hardened steel is tested on the C scale with the diamond indenter and a 150-kg major load. The useful range for this scale is from about RC 20 to RC 70. Softer materials are usually tested on the B scale with a 1/16-in-diameter steel ball and a 100-kg major load. The range of this scale is from RB 0 to RB 100. The A scale (diamond penetrator, 60-kg major load) provides the most extended Rockwell hardness scale, which is usable for materials from annealed brass to cemented carbides. Many other scales are available for special purposes.
The Rockwell hardness test is a very useful and reproducible one provided that a number of simple precautions are observed. Most of the points filled below apply equally well to the other hardness tests:
The indenter and anvil should be clean and well seated.
The surface to be tested should be clean and dry, smooth, and free from oxide. A rough-ground surface is usually adequate for the Rockwell test.
The surface should be flat and perpendicular to the indenter.
Tests on cylindrical surfaces will give low readings, the error depending on the curvature, load, indenter, and hardness of the material. Theoretical and empirical corrections for this effect have been published.
The thickness of the specimen should be such that a mark or bulge is not produced on the reverse side of the piece. It is recommended that the thickness be at least 10 times the depth of the indentation. The spacing between indentations should be three to five times the diameter of the indentation.
The speed of application of the load should be standardized. This is done by adjusting the dashpot on the Rockwell tester. Variations in hardness can be appreciable in very soft materials unless the rate of load application is carefully controlled.