Evaluating the "Usefulness" of force calibration equipment.
Usefulness may be the one word we often neglect to consider when purchasing force-measuring equipment, or any test equipment. Let's think about that for a bit and let it sink in. Most people I have run across seem to be concerned with price, a manufacturer's specification sheet, and maybe the physical size of the force-measuring device. However, when we discuss "usefulness," we start to think about what matters most. Is the equipment able to accomplish what my organization or I need it to do?
Figure 1 S-beam or S-type load cell
Out of the three main concerns; price, physical size, and manufacturer's specification, none of these give the full picture of if the device is useful or not. Price is important when budgeting and making any decision. I think many of us want our expectations met when we buy something; however, that does not always happen. What does happen is that lower prices tend to mean mass production, and mass production tends to raise some quality concerns. That is not to say the lower-priced equipment is bad; it's just to say that it is more of a buyer beware depending on the application. Now there are several low-priced solutions that, when properly tested, accomplish what is needed or, they may be useful for the intended application. In my 20 plus years in force measurement, I have rarely encountered a bargain. Morehouse used to sell S-type load cells more frequently. That is until the time it took to test each load cell thoroughly outweighed the $295.00 price tag. On bad lots, we ended up rejecting about 70 % of all the S-type load cells we tested as they failed in rotation. On good lots, the rejection rate was about 25 %. To this day, we still have about 100 some S-Type load cells that are okay for general testing applications where the accuracy of 0.25 % or better would be good enough. However, we find that just good enough still carries a lot of risk, and almost all our customers do not want the risks associated with equipment that may cause rework.
Figure 2 Acceptance limits of a 0.5 % tolerance with a device known to within 0.25 %
Think about it this way, I have a device that is known to within 0.25 % and I want to make a measurement that is within 0.5 % if I take measurement uncertainty into account, which I definitely should do if I'm using a decision rule that requires accounting for measurement uncertainty, my window for passing a device with a tolerance of 0.5 % is low. Figure 2 above shows the acceptance limits of between 9975.679 lbf through 10024.321 when taking the measurement uncertainty into account. It means I may have to adjust what I am testing as my device does not give me enough of a window to "pass" without adjustment. Then the customer has a potential failure that is caused by my equipment that was deemed just good enough. I would not call this equipment very useful if it costs more of my or the companies time to frequently make adjustments. Furthermore, we have simplified the equation as a device that is accurate to 0.25 % at the time of calibration often carries a much higher uncertainty when stability, repeatability, reproducibility, environmental conditions, resolution, and other CMC uncertainty parameters are not considered in these examples.
Figure 3 Acceptance limits of a 0.5 % tolerance with a device known to within 0.1 %
What if we change the reference equipment to a better load cell and meter combination that could achieve a realistic 0.1 % tolerance? Figure 3 shows that acceptance limits will increase if the reference standard uncertainty decreases. In this scenario, the end-user can "pass" more instruments as the acceptance limits with less than 2 % risk increases to 9960.286 to 10,039.714
Figure 4 Morehouse 600K lightweight load cell
The physical size is important as you need to know if the instrument will fit, and if the technicians can carry it or if other devices may be needed to help set the equipment up. If I am purchasing a load cell for my field technicians to go out and calibrate concrete machines, or other machines, is the person going to be able to carry an 80 lb load cell? It is a consideration of what may or may not be useful. If my technician cannot carry the load cell, I doubt it is very useful. Morehouse developed a 600,000 lbf (can go to 3000 kN) compression-only type load cell that weighs less than 30 lbs because of this very issue. Sometimes it's a trade-off as the smaller load cell may not perform as well as needed. The above load cell is usually much better than 0.02 % of full scale. That translates to about 0.2 % at 60,000 compared with a shear web type load cell that can achieve 0.005 % of full scale and 0.05 %; however, the shear web cell will weigh about 300 lbs at this capacity. A weight that only a few people I have watched on a late-night strong man competition could maneuver.
Figure 5 Morehouse Precision Load Cell Specification Sheet
The third and probably the most asked for item is the specification sheet. The specification sheet may tell you how well the cell repeats, how it reacts to certain temperature conditions, how linear the device is, it's side-load sensitivity, which is very useful, however, I have yet to see a specification sheet talk about interactions with various adapters. I have not seen a specification sheet say the results are only achieved if this load cell is loaded through the shoulder in compression with a material hardness of X in a machine that is plumb, level, rigid, square, and has low torsion. I see very few that tell you how reproducible the load cell is. Reproducibility spec is very dependent on the equipment it is being used with. It the equipment has bending, misalignment, and torsion, the load cell is not going to perform as expected. Therefore, almost every specification sheet does not tell you what you need to know. Morehouse puts a note in our literature about what can be achieved, but that makes assumptions that the end-user may be using our force equipment and machines with our recommended adapters. If you have not read my paper on adapters, I encourage you to do so as the wrong adapters can turn a device known to within 0.01 % into a device with 0.3 % error, as we have seen with top blocks of different hardness. That paper that won a best paper award can be found at https://www.mhforce.com/Files/TechnicalPaper/16/TechnicalPaper.pdf.
Figure 6 Morehouse Precision Shear Web Cell