To  Each 

His Own Parameter

Custom surface finish parameters

can create more problems than they solve

--unless you know what to watch out for.

______________________
by Alex Tabenkin

Parameter Clarity: Adapting to New Technology An example of a "justified" series of special parameters is the one used to ascertain the ultra-clean condition of tubing used in microchip manufacturing. These processes require the delivery of neutral gases through ultra-clean tubing and fittings. In order to achieve high yields, it is essential to minimize contamination of the process gases, so the absolute cleanliness of the inside surfaces of the tubing is critical. Because the smallest surface imperfections may harbor contaminants and resist cleaning, component and system manufacturers consider surface roughness a critical quality issue. Ra, the most common surface finish parameter across most industries and applications, is typically measured over five cutoffs of data. Manufacturers of ultra-clean tubes have found this to be insufficient, and want to know surface conditions of the whole surface. They have, therefore, developed a special series of roughness parameters and measurement methods to generate three Ra values.  RaCH (continuously high) continuously shifts the segment under analysis by a single data point and identifies the largest Ra evaluated across the entire part length.   RaCL (continuously low) continuously shifts the segment under analysis by a single data point and identifies the smallest Ra evaluated across the entire part length.  RaCA (continuously average) continuously shifts the segment under analysis by a single data point and takes the average of all the individual Ra measurements. While this methodology is not recognized by any national or international standards body, it has proven so useful that it has been accepted as a de facto standard by this particular industry. In response, at least one gage manufacturer developed special software to perform these measurements, and incorporated it into existing surface finish gages.   Parameter Confusion: Will The Real Rz Please Stand Up? You might think that a simple surface finish parameter like Rz, mean roughness depth, would be reasonably straightforward. It's not. A classic example of parameter confusion, Rz has at least three flavors:   The old international (ISO) version --10-point Rz, which averages the distance between the five highest peaks and the five deepest valleys, based on an unfiltered profile  The old German (DIN) version, which is now the new ISO/ASME version --an average of distances between maximum peaks and valleys in five cutoffs, based on a roughness profile  The recent Japanese (JIS) version --10 point Rz which averages distance between the five highest peaks and five lowest valleys based on a roughness profile This situation presents numerous opportunities for confusion:   The specifying engineer might write a specification based on an old or a new standard.  The machinist may be working from an old part print or a new one.   The print may have originated in an engineering department operating under a different national or international standard.  The measuring instrument might incorporate any or all of the algorithms under the same or different names. It's not clear how many potential errors that adds up to, but it's too many in any case. You can be clear on one thing from this discussion, however: If your print calls out an Rz value with no further explanation, warning bells should go off. A time bomb may be ticking.

Human nature being what it is, engineers take great pride in inventing new surface finish parameters specific to parts that their organizations manufacture. Thus, out of something akin to pride of ownership, new surface finish parameters are born, even though there are many existing parameters that may have done the job satisfactorily.

 There are currently more than 100 parameters related to two-dimensional stylus-type measurements alone, with no end in sight. Those parameters, most of which are included in various national and international standards, describe a wide variety of surface conditions and are capable of satisfying the great majority of metalworking applications, including many of the most complex.

 There is nothing wrong with custom parameters per se; however, when they are substituted for standardized parameters that are already in effect, they may harm the user and the industry in ways both subtle and highly tangible.

Describing unique surfaces

 Sometimes, standardized parameters are insufficient for describing surfaces with very special technical requirements. In this case, development of a new parameter is justified. Such development is also justified when a new parameter is required to improve the quality or performance of an existing application.

 In other cases, especially in new or rapidly changing industries (such as computer chip manufacturing and biotechnology), new applications require surface conditions unlike those that occur in the traditional metalworking fields for which most existing parameters were developed.

 Occasionally, new measurement technologies, or refined techniques, such as improved filtering, lead to the development of new, more informative parameters. In some cases, these "custom" parameters have been accepted as de facto standards by the larger industry, and some have even been adopted officially into national or international standards. Some examples of "justified" special parameters include:

  Crown drop measurement --performed with a skidless stylus surface finish gage, to characterize the radius of curved surfaces on parts such as aircraft bearing rollers and piston rings

  "Inverse slope" parameter (Dx/Dy) --developed to evaluate housings in which bearing surfaces or sealing faces are of different materials, with different hardness values

  Wear control for crankshafts (WeAn) --developed to control wear by comparing new and worn surfaces on the same sample

  Dental wear parameter (WeAr) --used by researchers and product developers to assess the abrasiveness of toothpaste and toothbrushes and to optimize dental filling materials and finishing methods

  Special tubing surface finish parameters (for tubes, valves and connectors) --used in microchip manufacturing

Reinventing the wheel

 While some custom parameters help manufacturers come to grips with unique problems, many others do not. New parameters are often created simply because engineers and quality control personnel aren't familiar with the large base of existing ones. This reinventing of the wheel seems to be particularly prevalent outside the metalworking field, where there is only a passing familiarity with surface finish measurement. Some of these new parameters are merely redundant, while others are somewhat less well-conceived than existing ones.

 In the United States, it is often part designers and manufacturing engineers who initiate the development of new parameters. The responsibility for algorithm development, however, is often not in the hands of a metrologist but those of a software engineer, who may not be familiar with existing standards, appreciate their value, or have the discipline or the specialized education to research and follow them. While the PC has had a tremendously positive impact on manufacturing, it has also made it just a little too easy to create algorithms for new parameters and to incorporate those algorithms into the application software of PC-based measuring instruments.

Unhappy consequences

 What's wrong with an overabundance of surface finish parameters? Parameter proliferation may cause confusion among designers, manufacturing engineers, machine operators and inspection personnel. It can lead to economically significant disagreements between suppliers and customers over what constitutes a proper testing method and what kind of results are acceptable. Here are some examples:

  Mean roughness depth. Rz is a widely used parameter to measure mean roughness depth. Unfortunately, there are three different Rz parameters still in use. When manufacturing engineers, machinists or parts inspectors encounter the specification on a part print, they don't necessarily know which one it refers to. This situation leads to all sorts of confusion, some quite costly.

  Ra and Rq roughness. There is frequent confusion over the Ra and Rq roughness parameters. To start with, Ra used to be called AA or CLA, and Rq used to be called RMS. Rq results happen to be 11 percent higher than Ra results when measuring the test patch with a pure sine-wave profile. However, there is actually no mathematical relationship between the two parameters and, depending on the manufacturing process and the resulting surface profile, the ratio between Rq and Ra can vary from 20 percent to 200 percent (see Table 1).

Table 1: Ratio of Root Mean Square to Arithmetic Average Roughness

 Nonetheless, there are numerous gages still used in the United States that will "convert" actual Ra/AA measurement results to Ra /RMS results by simply flipping a switch. However, the instruments are programmed to apply a multiplication constant of 1.11, so the conversion is only accurate when measuring a sine-wave test patch and inaccurate on all real-world manufactured parts.

 Some astute gage users noticed that the conversion occurred as a constant factor, so they naturally concluded there was a direct relationship between the parameters. Unfortunately, this has become a commonly accepted fallacy. You'll occasionally find machinists or inspectors who, when needing to measure the parameter other than the one the gage measures, will manually apply the same multiplication constant of 1.11 to make a conversion and get incorrect results with great consistency.

 There are numerous other examples of parameter confusion. As the development of new standards has progressed, several basic terms and parameters have changed "names" and symbols. In other instances, the name of the parameter remained the same though the algorithm changed, as is the case with Rz. Policies that require all in-use part prints to be up-to-date with existing standards may help minimize confusion within a company, but any time specifications are exchanged from one company to another, there is a possibility for error.

What to do

 The decision to develop new surface finish parameters should be approached with great caution to avoid the introduction of superfluous, or even erroneous, methods. The great majority of applications can be satisfied with existing, standardized parameters, which should be used whenever possible.

 As mentioned earlier, the decision to create a new surface finish parameter is neither intrinsically wrong nor right. It all depends on whether parameters that already exist are sufficient for the application. Of course, it takes a bit of knowledge and some work to identify the appropriate surface finish parameter. However, the additional effort will eventually be rewarded, most likely with savings in time and money and by avoiding some of the costly consequences of "parameter confusion" outlined above.

 On the other hand, if you have done a responsible search of existing parameters and find them lacking for your purposes, by all means customize one. Such a decision may prove to be far more than self-serving. A technically sound new parameter developed by academic or manufacturing entities to satisfy special requirements may eventually find its way into broader use. This is good because it improves the coherence of quality and engineering efforts worldwide.

 In most cases, it's best to start out assuming that the parameter you need already exists. Your organization, and the industries in which it participates, will be all the better for it.

About the author

  Alex Tabenkin is a surface finish and form metrology consultant for Mahr Federal Inc. of Providence, Rhode Island. E-mail him at atabenkin@qualitydigest.com .

Useful Surface Roughness Tables

B46.1--1995 vs. ISO Parameters	Comparison of Basic Terms Between ISO 4287-1:1984 and ISO 4287:1997
Comparison of Parameter Symbols Between ISO 4287-1:1984 and ISO 4287:1997