Plateau Honing: Which Parameters Should I use?

All plateau honed surfaces are not created equal. The image below is the heart of the story and the focus of this article. In this image you see three honed cylinder bore surfaces with very different characteristics. As it turns out, each is best controlled using different sets of surface texture parameters.

Three types of honed cylinder surfaces. We will see that each requires different parameters to describe them.

 

Conventional height parameters are sufficient for single honed surfaces

The top profile in the image shows a conventional surface made with a single honing operation. The texture is made up of peaks and valleys of all depths. The valleys might be slightly deeper than the peak heights–but nothing too significant. Over the initial run-in period the action of the engine wears away the higher peak material, leaving a surface with smooth plateaus and valleys in between to retain lubricant.

Traditional roughness parameters such as Ra (Average Roughness), Rz (10-point Roughness Height) and Rmr (Material Ratio) are often sufficient to control the honing operations for a conventional, single honed surface.

Conventional parameters (Ra, Rz, Rmr) are often sufficient to describe single-honed surfaces.

 

Plateau honing leads to the K family parameters

While straightforward to manufacture, single honed surfaces do have drawbacks. The wear that occurs during break-in inevitably creates unwanted gaps/leakage as well as debris. Both factors can reduce engine performance.

As emissions and performance standards grew tighter, manufacturers began creating “plateaued” surfaces in order to reduce break-in and improve sealing. Plateau honing uses rough honing to make the valleys, followed by finer honing to create the plateaus. A “plateau honed” surface is shown in the second profile.

But with these improved surfaces came a new challenge: the Ra, Rz and Rmr could not robustly distinguish a plateau honed surface from a single honed surface. Very different surfaces gave similar values for the parameters when dealing with this class of surface.

In order to better describe and control these new surfaces, researchers developed and applied the Rk family of parameters. The Rk family is based on the material ratio curve, shown below on the right. The individual parameters (Rk, Rpk, Rmr1, Rvk, Rmr2) quantify the peaks, valleys and kernel regimes. They allow for better control of plateau honing as they are more targeted toward the individual geometries within the surface.

The Rk parameters are well-suited to describing plateau honed surfaces that have distinct plateaus, kernels and valleys.

 

High performance surfaces need better parameters: Q parameters

High-performance applications (for example, diesel and racing engines) require surfaces that are more extremely plateaued, as in the third of our profiles.

For these applications, manufacturers create very smooth plateaus with discrete valleys, to optimize friction, clearance and engine temperature, with virtually no run-in period. The primary characteristic of these surface is a strong distinction between the peaks and the valleys. There is a clear visual indication of where peaks meet valleys.

Unfortunately, again, the existing texture parameters proved insufficient to describe these surfaces. When a surface is extremely plateaued, the Rk parameter model does not fit the material ratio curve as well. This becomes apparent in the material ratio curve shown below. Looking at the green area we see that there is no exact position on the curve at which the plateaus give way to valleys (as indicated by the left corner of the large, green triangle.) Thus, the Rmr2 and, subsequently, the Rvk parameters become unreliable.

In a high-performance engine the distinction between plateaus and valleys, is almost non-existent, which makes the Rk parameters unreliable.

 

For extremely plateaued surface we turn to the Q parameters family. Rather than using the material ratio curve, the Q parameters are based on a “material probability” curve, shown below on the right. This curve is a representation of the material ratio curve with percentages mapped to standard deviations. When doing so, we see the two distinct distributions (plateaus and valleys) as two clearly linear regions. Furthermore, these linear regions produce a sharp knee on the curve. The Q parameters (Rpq, Rvq, Rmq) derived from this curve can robustly distinguish the surface regimes for extremely plateaued surfaces.

For extremely plateau’d surfaces the Q parameters, based on the material probability curve, are most robust.

 

When do I use K or Q Parameters?

Today both the K and Q parameters are specified in the ISO 13565 standard as well as other upcoming 2D (profile) and 3D (areal) standards. Digital Metrology’s OmniSurf and OmniSurf3D include both sets of parameters as well. Unfortunately, the standards leave out a “user’s guide” to tell us when to use one set of parameters or the other.

What we find is that each set of parameters has its use. The Q parameters work best on two-process surfaces with discrete plateau and valley regimes. However, the Q parameters, are less reliable when the distinction between plateaus and valleys is fuzzier. In some cases where the surface isn’t distinctly plateaued, the standardized mathematics for the Q parameters may provide no results at all.

The K parameters work well for many surfaces for which extreme plateaus are not required. This includes not only cylinder bores, but gears, bearings, transmission components and many other surfaces required a degree of sealing and/or friction control.

K parameters serve well for surfaces like the top profile, while extremely plateau’d surfaces are best described by the Q parameters.

 

And, let’s not forget Ra and Rz! These basic parameters still provide useful feedback for surfaces with a wide distribution of peaks, valleys and intermediate heights.

In any case, the best way to analyze your plateau honed surfaces is to explore and interact with your data. The numbers alone may not tell the whole story. Digital Metrology’s OmniSurf and OmniSurf3D software packages provide all three sets of parameters along with powerful visual tools to help you see what works best for your surfaces.

We have an excellent video here showing the Plateau Honing analysis in OmniSurf3D software. If you are honing, you should take a few minutes to check out how quickly and easily you can explore your surface data.

View our video here on OmniSurf3D’s Plateau Honing analysis.

 

There’s a lot more to talk about on this topic. In fact, Digital Metrology even offers consulting and seminars on surface texture analysis and plateau honing. If you have questions or would like to know more, please give us a call or email.

Using OmniSurf3D’s Pit/Porosity Tool

One of the underlying concepts of OmniSurf3D software is to make everything visual, so that anything you do with the data to produce a number has some kind of image to support it. That’s the way our brains and eyes were designed to interact with surfaces. Our analysis should have a visual basis as well, to make it the most informative.

A great example of this is the Pit and Porosity analysis tool. It’s highly interactive—you can slice through a surface to look at the voids/pores at every level and see the results both as a plane through the data and as results in a table.

We have a great video available here that shows how the OmniSurf3D Pit and Porosity Analysis looks and functions.

The Pit and Porosity analysis is shown below. As you’ll see in the video, as you move the slice level up and down you’ll see the cutting plane moving through the left pane, the corresponding level on the material ratio curve in the middle, and on the right, the features that are being sliced at that particular level.

 

A little more about that plot on the right. The features shown in red are considered “closed pores,” meaning we can see their entire perimeters. Blue features are “open pores,” meaning they open onto the edge of the data set. As you move the slice level upward some of the closed voids become open voids.

If we slide the slice level up even further we reach a point where the open pores actually connect. These “edge-connected” pores are shown in green, and they are potential leak paths. The visual here is quite helpful to highlight the features that link up to form leak paths.

While the visuals can be quite helpful, we also need numbers to support what we are seeing. At any level we can get a count of total pores, and of open and closed pores. Or, we can calculate the “pit density,” the number of pits per square centimeter. Pit density can give a better indication of the surface in general, irrespective of the size of the measurement area.

The analysis also gives us tools that let us compare surfaces at some set slice level. We can do this by either manually setting the Height Above Meanline, in the lower left. In the image below, we set it to 50 microns below the meanline.

 

In some cases we may want to use a standard material ratio reference instead. We can enter, say, a 5% reference level (the Smr0 value that we talk about in the ISO standards). In the image below we set the 5% reference level to 0 and then dropped down to about a 150 micron depth.

 

OmniSurf3D makes things very visual, so we can see in the analysis what we are trying to describe with the numbers. I encourage you to check out the video, and if you have more questions, shoot us an email.

Can I measure waviness on a CMM?

Many Coordinate Measuring Machine (CMM) providers are beginning to incorporate surface roughness measurement into their CMMs. Digital Metrology has had the privilege of collaborating on many of these projects.  However, one question often arises:

Can we also measure waviness on a CMM?

Before we go too deep, we need to first be clear on the type of measuring system that is being used on the CMM: skidded or skidless. 

A skidded system has a large radius in contact with the surface which provides a reference and keeps the diamond stylus in range during measurement.  The skid slides along the surface during the measurement. The advantage is that a skidded system can be moved via the CMM over any measuring length.  The disadvantage is that skidded probes can be used to measure roughness, but not waviness. 

Photo courtesy of Renishaw.

Skidless system, on the other hand, can measure both roughness and waviness.  However, in order to overcome CMM inaccuracies, noise and vibration these CMM-based skidless systems typically incorporate an independent traverse unit that is mounted on the CMM head.  This independent unit accurately moves the stylus tip while it is held in place by the CMM.  The disadvantage is that the traverse lengths are limited by the relatively small roughness traverse unit.

Photo courtesy of Zeiss.

But what about “waviness”?

Often a waviness analysis requires a longer evaluation length.  Thus, it would be ideal to use the CMM itself for a waviness analysis – much like a typical straightness measurement.  Can it be done?

That depends on two things:  1. Is my CMM accurate enough, and 2.  Can a CMM stylus (ball) actually detect waviness?

The CMM accuracy must first be addressed.  For example, if the CMM’s straightness errors for the desired measurement are a major percentage of your waviness (Wt) tolerance, then the CMM is probably not the best choice.  The CMM’s ability to measure waviness can be determined by measuring the straightness (Wt) on an optical flat at the desired position over the desired length.

Can a CMM stylus ball actually measure waviness?

ISO 12780-2:201 indicates that, for a typical 0.8 mm waviness cutoff, a 0.5 mm tip radius (1.0 mm diameter) is acceptable.  However, many users are skeptical about this and some companies even have internal standards saying that waviness must be measured with a roughness (diamond stylus) system. 

Much of this confusion or misinterpretation of waviness comes from the aspect ratio of the profile graphs.  Typical profile graphs are scaled with micrometers vs millimeters.  This gives the impression that profile features are very narrow and steep… which is not actually the case.

Let’s use Digital Metrology’s OmniSurf software to take a look at a typical surface that would be a candidate for CMM-based waviness.  This is a surface from a milled engine block head deck – a component much like the one used in both pictures above.

The above trace came from a traditional, skidless, diamond stylus-based surface texture measurement.  We can see all of the roughness details as well as the underlying waviness.  The waviness can be shown in OmniSurf simultaneously with the primary profile as follows.  (Note: a 0.8 mm cutoff is shown.)  This would be considered a “traditional” waviness profile.

In order to see how a CMM ball follows this surface we will use OmniSurf’s morphological closing filter.  The closing filter acts like a virtual gasket by rolling a mathematical ball over the waviness profile.

OmniSurf activates the closing filter when any of the virtual gasket parameters are selected (Wvoid, Wvdd).  This is a useful analysis for predicting leakage under a gasket.  For example, a 2000 mm radius can be used on the above data set to show the potential leak areas.

We can also use the closing filter to simulate the penetration of the CMM ball.  To do so, we must keep in mind that OmniSurf applies the morphological filter to the waviness profile.  Thus, we will set the roughness cutoff to a very small value (0.0025) in order to create a waviness profile that simulates the actual surface being measured.

We can now apply a closing filter based on the stylus tip radius.  Let’s start with a 10 mm diameter (5 mm radius) ball:

The black curve above shows that the 5 mm diameter follows the general trends of the surface and actually sees a great deal of fine detail.  So now let’s go to the 0.5 mm radius (1.0 mm diameter) and we see the following black profile as the stylus path:

The 0.5 mm tip radius reproduces a significant amount of the milled surface’s details!  We will now export this (black) closing profile and bring it back into OmniSurf as a new profile.

We can now filter this profile with the 0.8 mm waviness cutoff:

A side-by-side comparison shows amazingly similar waviness profiles:

As this OmniSurf-based analysis shows, the resulting waviness profiles are very similar – indicating that it is indeed possible to measure waviness on a CMM.

It should be noted that this analysis is dependent on the surface having relatively uniform roughness.  In cases where there are sporadic peaks, the CMM ball may be more sensitive to these high points and potentially give higher waviness values.  The waviness generated by the diamond stylus will generally remain in the center of the profile in cases of sporadic peaks – especially when using robust filtering.

For more information on this topic, OmniSurf or to further explore measurement options for your surfaces, contact Digital Metrology today.