Last issue, we discussed the 10-to-1 rule as it applies to gage performance. It says the gage should perform to a level better than 10 percent of the tolerance. While this rule has mostly been replaced by the more scientific and standardized gage repeatability and reproducibility (GR&R) performance test, other rules of thumb are still in play.
For example, we gage-builders take this 10-to-1 rule a little further. In order to build a gage that performs better than 10 percent of the tolerance in use, we actually need to design a gage that performs to 4 percent of the tolerance with the things we can control: namely the gage components and the master. That’s pretty demanding, but it has a foundation in logic.
That logic comes from our SWIPE principle, in which the measurement process has five parts: the standard (master), the workpiece, the instrument (gage), the people and the environment. We have a chance of controlling two out of the five parts of the process: the master and the instrument. If we can control those two items to achieve 4 percent part tolerance performance, we will achieve the 10 percent measuring process performance. This is where we start when trying to solve a customer’s application.
Part staging establishes the basic relationship between the measuring instrument (typically a dial or digital indicator) and the workpiece. Any error in the fixture inevitably shows up in the measurements. Many fixtures are designed around a C-frame shape and, as such, have a substantial cantilever that is subject to deflection. This problem is greatly reduced if the fixture is a solid one-piece unit.
Most fixtures also consist of a minimum of three pieces—a base, a post and an arm—that must be fastened together with absolutely no play among them, as any movement will be magnified at least tenfold at the workpiece. Play of only a few millionths in a couple of joints can easily accumulate so that measurements to ten-thousandths become unreliable, regardless of the level of discrimination of the instrument. Thus, in designing gage structure, the most simple is often the best.
Surface Parameter Conversion
In today’s global economy, machined parts are being made and shipped around the world. As a result, manufacturing and quality control engineers are often forced to decide whether or not to accept a part when the print requirements are not consistent with measurement results of the surface gages at the local facility. Some quality control engineers simply assume that if a part is checked and passed using the local parameter available, the part will also pass other checks.This assumes that a constant correlation, or ratio, exists between different parameters.
This assumption is true, to a point: there are rules of thumb that can be used to convert Ra to Rz or Rz to Ra. Using a ratio range for Rz to Ra between 4-to-1 and 7-to-1 is a safe conversion. However, if the manufacturer uses Ra but the customer uses Rz, then the conversion ratio would be much higher, possibly as high as 20-to-1. Also, the actual shape of the part’s profile will have a significant impact on these ratios.
Communication at the outset of the project can avoid most surprises. Approximate and sometimes questionable comparisons can be avoided with an understanding of exactly what a parameter on a print means and how the various parties involved in the production plan to check surface texture.
Air Versus Contact Gaging
The response of air to surface finish, however, is more complicated. Think of a jet of air. The measurement “point” is really the average area of the surface the jet is covering. Now consider the finish, or roughness, of that surface. The measurement point of the air jet is actually the average of the peaks and valleys the jet is exposed to (see illustration). This is not the same measurement you would have if a contact-type probe was used. This difference is a source of real gaging error, one that is most often apparent when two different inspection processes are used.
For example, let’s say we have a surface finish of 100 microinches on a part, and we’re measuring with an air gage comparator and two-jet air plug that has a range typically used to measure a 0.003-inch tolerance. The typical gaging rule of thumb says you should have no sources of error greater than 10 percent of the tolerance. In this example, that’s 0.0003 inch. If we used this plug on the 100-microinches surface, the average measuring line would actually be 50 microinches below the peak line. Double this error for the two jets and you get 0.0001 inch, or 30 percent of the allowable error. That’s pretty significant, and air would probably not be a good choice for this part. As a general rule, the limit for surface finish with an air gage is about 60 microinches, but it really depends on the part tolerance.