What’s beyond average roughness?

In surface texture, Average Roughness (Ra) is still the most widely used parameter, even though it has serious limitations. But what other tools are there that can tell us more about our surfaces? We’ll be exploring that topic in a series of posts, so watch this spot for updates every week!

Looking for more info? Check out our original “What’s Really Beyond Ra” post!

Let’s start with our favorite was to move “beyond average roughness:”

1. SEE the data.

For most of us, music notes on a page aren’t really a song until we hear them played. It’s the same with surface texture: parameters are just numbers until we can see what they’re trying to tell us.

A 2D profile shows a lot more about a surface than just a few numbers. A 3D image takes it to still another level, showing trends in all directions. And these days you can even 3D print data so you can hold it in your hand.

But, of course, our eyes can play tricks on us, too. Are those actual spikes or maybe noise in the measurement? Is the surface really tilted, or is it skewed by a few bad pixels? Seeing a surface is a powerful aid, but there are many more tools in the toolkit as well…

beyond average roughness, beyond ra, see data, visualize data


2. Waviness

Average roughness (Ra) describes the finer features in surface texture. The larger shapes that we call “waviness” can be just as impactful. Waviness may be the source of noise, vibration, and harshness in gears, as well as the cause of leaky gaskets, premature wear, stress concentration, and many other issues.

The trouble is, roughness parameters like Ra don’t tell us anything about the waviness. It’s like describing a roller coaster by talking about the grain of the wood: you won’t get any sense of the humps and dips that make the roller coaster fun! Check out our Notepad Series videos,  Roughness and Waviness and Specifying Waviness to learn more.

Average roughness, surface texture consists of surface roughness and waviness


3. Wear Analysis

To gauge how much a part is wearing, it’s tempting to measure the average roughness (Ra) of the surface before and after use, then treat the difference as “how much the part has worn.”


That’s a lot like digging a hole and trying to judge its depth by measuring the am0unt 0f roughness at the bottom!

surface texture, measure wear, measuring wear, shovel in a hole

To measure wear properly, we need to have a reference from the original surface. In the case of a hole, we can use the “ground” around the hole as the reference and measure the difference between the “before” and “after” depth.

In the case of “micro-wear”— the wear within the original texture—we need to match up the unworn valleys. We can do that by shifting the worn surface up and down relative to the unworn surface. 

Got a wear issue? Talk to us about how to approach it! Want more details about analyzing “macro” and “micro” wear? Read more in this wear analysis blog post.


4. Ra by Wavelength Bands

Average roughness doesn’t always help to describe how a surface looks. Appearance may be affected by features of one particular scale (i.e., wavelength) more than others. In a paint finish, for example, excessive roughness at very short wavelengths may result in haze, or dullness.

Even though the total Ra may not be helpful, comparing the average roughness (Ra or Sa) over narrow bands of wavelengths can help differentiate “good” from “bad” finishes. Monitoring Sa over several wavelength bands can help manufacturers dial in an exact finish. Digital Metrology’s Bandify and Bandify3D software are made for this kind of multi-band analysis.


5. Slopes

Our eyes can detect subtle differences between surface finishes. But height parameters like average roughness may not “see” what  we see, because the difference between textures aren’t always difference in heights.

The slopes in the texture can impact how light scatters on the surface, which will change how the surface appears. The Sdq parameter is a measure of the slopes in surface texture, and it can often differentiate surfaces that Ra/Sa cannot. Sdq is a standard parameter in our OmniSurf3D software.

The dull and shiny sides of aluminum foil, for example, have very similar average roughness values, but very different Sdq values.

average roughness may be the same for shiny aluminum foil side and dull aluminum foil side, sdq parameter, slopes may differentiate two surfaces with similar average roughness values


6. Rk Parameters

Can one parameter like average roughness predict how a surface will support a load, retain lubrication, and wear? The short answer is: no. But the Rk family of parameters can help.

The Rk parameters are derived from the material ratio curve, which represents how much material we have at varying depths as we slice through a surface. Three of the primary parameters in the Rk family are:

  • Rk, the peak-to-valley height of the core material. This is the running surface, which will likely support a load.
  • Rpk, the effective height of the peak material. Controlling the peaks may be important for applications involving sliding surfaces. It is also the portion of the surface that may be most affected by break-in.
  • Rvk, the effective depth of the valley regions below the running surface. The valleys may be important for retaining lubricant and/or moving debris out of an interface.

Digital Metrology OmniSurf3D Surface Texture Analysis Software - Rk ParametersThe Rk parameters are well-suited to describing plateau honed surfaces that have distinct plateaus, core, and valleys.


How the material is distributed between peaks, valleys, and core affects many surface functions. In a combustion engine, for example, controlling the Rk parameters for the surface of a cylinder bore is essential for efficiency, gas consumption, emissions, and wear.

There’s a lot to learn about the Rk parameters! The Rk Parameters Notepad Series video is a great place to start.

See how the parameters are applied to produce efficient engines in this Plateau Honed Surfaces tutorial


7. Rz

Our eyes are good at figuring out how “generally” rough or smooth a surface is. You can look at a lawn, say, and gauge how tall the grass is, despite one or two tall weeds or deep divots.

The Rz parameter, or 10-Point Average Roughness,  tends to “see” more like our eyes do. To get Rz, we divide a profile measurement into equal segments (the segment length is often equal to the roughness cutoff). We find the difference between the highest and lowest point in each segment, then average them together.

Rz Surface Roughness Measurement Parameter not equal to Sz Surface Texture Parameter - Digital Metrology

Rz is easy to measure, so it’s no surprise that, after average roughness (Ra), it’s the second most commonly specified parameter. Rz can give a quick, general evaluation of surface roughness. In a production environment it can be useful for spotting changes in a process over time.

Unfortunately, like Ra, Rz is susceptible to outlying data points such as dirt or deep pits, which can skew the data and make the parameter unstable. As with Ra, very different surfaces may have the same Rz value, so it cannot always discern the differences that affect how surfaces will function.

surfaces with identical Ra, Rz, and Wt values but very different surface textures

Very different surfaces with similar Ra and Rz values.


One last note on Rz: most standards, including the ISO and ASME standards, agree on the 10-point definition described above. An older definition, from Japan’s JIS standard, averages five peaks and valleys per segment, as opposed to five peak-to-valley differences total. While this “JIS” definition is outdated, you may still encounter it. so make sure that you are calculating it correctly.  

Want to know more? Check out our Notepad Series video showing how Rz is derived.

You can also learn about the Rz parameter and its uses in the Surface Texture Answer Book.


Watch for more topics in the coming weeks!