Q: My measurement device can compute dozens of different parameters (Ra, Rq, Sa, Rz, Rpk, etc.), which is the best surface spec to use in my application?

A: The place to start is to look for statistical correlations between any standard parameter and the performance metric that matters in your application. Sometimes this works well, and then you’ve found the best metric. But unfortunately, there are other cases where none of the standard parameters correlate with performance. Here is where it is useful to go beyond the standard parameters.

Q: We have a few different topography measurement techniques available (or we are about to purchase a new one), what technique should I use to measure the surface roughness of my samples?

A: Of course there is no single answer to this question. The best technique will vary depending on (at least) three factors:

(1) What type of topography exists on your sample? Topography is multi-scale, so you will need to choose a technique that is able to detect at the right size scales. Additionally, different techniques vary in their ability to measure at high slopes, measure in deep valleys, preserve fragile surface topography, etc.   

(2) What type of topography controls performance in your application? A car driving on a road is most sensitive to topography at the scale of its tires: centimeters to meters. An implanted medical device might be most sensitive to topography at the scale of cells or proteins: micrometers down to nanometers. So different techniques can measure these more or less effectively.

(3) What are the process needs? Some techniques are very fast, and are compatible with inline measurements in high-volume manufacturing. Other techniques are slower and require offline sampling. Some techniques are non-contact and non-destructive, while other techniques require contact with, or even sectioning of, a sample in a way that it is contaminated or destroyed.

Q: For my process, all I care about is hitting a ceratin Ra (or other surface spec), how important is it to know the “true” topography?

A: For a mature manufacturing process, there may be a certain surface metric that is sufficient for process control. However, sometimes traditional metrics aren’t good enough – for example when an established process exhibits unexplained failures, or in developing a new process. In these cases, it’s useful to know the true topography so that you can discover which aspects of topography matter in a particular application. 

Q: I have an established process. It’s not in my budget to purchase and implement a second measurement technique to use alongside my established technique. How can the combination of multiple measurements help in an established manufacturing process?

A: There are a few ways: (1) A second measurement technique can be used to establish target values and to periodically validate the results from the first measurement method. (2) A third-party measurement lab may be able to assist with periodic measurements, giving you the benefit of multiple measurement methods without incurring the capital expense. (3) Even if measurements are taken using a single device, they can still be taken at various different magnifications in order to compare results to detect artifacts or other anomalies.

Q: My print requires a certain Ra, so that's what I have to measure. How can I implement the best practice to tie my measurements to length scales?

A. At a minimum, report which cutoff wavelengths were used to compute the value of Ra. The reference standards (ISO and ASME surface texture standards) specify certain cutoff wavelengths, and all of the major software packages make this information available if you know where to look. So at a minimum, just report the large-size and small-size cutoff wavelengths that are being used – this ties your measured parameter to a certain length scale, just as we have done in the scale-dependent analysis.

Q: I would like my engineering team to be able to access and share raw data rather than calculated results—I think they could learn a lot about our surfaces if they could explore the data. But our management contend that quality decisions should be made by quality professionals who know how to interpret the data. Engineers and technicians may not know everything that goes into measurements or what they imply, and they're bound to draw conclusions that could lead us to expensive misinterpretations. How can I break down that barrier so that more people can access the raw data? SHOULD I break down that barrier?

A: There’s a difference between sharing data to improve visibility and understanding in production and making quality and process decisions based on that data. Training a wider team of engineers and technicians to interpret data improves capability of the entire manufacturing team, even if the quality team serve as the final decision makers.