You make a two dimensional plot (on graph paper) of the dial reading against the loads applied. Draw a straight line for the group of readings and the slope of the line is your constant. Result: load being applied is the constant (slope of that line) times the dial reading. Of course the dial should be set at zero at the start of the calibration. I used to do the calibration in a compression testing machine that also has been calibrated by a proving ring. In my experience the proving ring calibration is a constant for all loads.
A fairly cheap proving ring can be made from the outer race of a large ball bearing. I picked them up from the scrap bin of a Construction machinery dealer dealing in large bull dozers. They came from bearings that had failed. However, the outer races looked fine and calibrated well. A little bit of attaching device machining needed and a dial ring of good precision is all that you need. Reading to 1/10,000 for the gage.
Probably indicates the maximum load you can apply before the elastic limit is reached. Some rings are described as between a low and a high load, probably the most linear part of the calibration range. Apparently for some rings the initial low readings are not quite linear, in which case use the plot of the calibration results.
since my proving ring has an indication of 3KN... means maximum force it can take is 3KN and my dial guage in proving ring has 100 small division.. so is it right to say that my proving ring constant =(3*1000)N/(100 divisions) = 30N/div.....????
I say again, you need to calibrate the proving ring against a known load measurement device such as another proving ring or a load cell. You can't assume liniear behavior.
You are unlikely to find that any proving ring is of such a size that direct reading of the dial can be called the units of weight, such as a pound or Newton.
In case you are having trouble finding a place to calibrate the proving ring, the least you can do as a start is take it to a bathroom scale and come up with a rough calibration, better than nothing.
That's OK as long s the dial starts out at zero and the reading is multiplied by that constant. What you can't do is just read the load on the dial directly by some even digit multiple, such as 100, or 1,000. If the gage doesn't start at zero for no load, the calibration only is good if you take the gage "throw" range, from no load, times the constant.
