Making a case for Rz
Surface finish has been discussed quite often in our industry so this isn’t going to be an article covering what you already know. Instead, we will be looking at machine tools you use to bore, hone, mill and grind with and surfaces they generate. Each tool or machine process produces a unique surface signature complete with unique defects. One thing is certain; every machining process either mechanical or chemical creates surface defects.The purpose of this article is to explore these surface defects and offer solutions to reduce or eliminate them. Additionally, we will cover single surface parameter measurement values as they apply to quality determination. We will not get into the more complex suite of Rk parameters. Rk and other 3D imaging are best left for another article treating those technologies separately.
The average engine machine shop has choices when sizing cylinder diameters. If stock removal is large, boring is indicated. If stock removal is small (usually less than .010”), honing may be selected. Each leaves behind a different surface structure and depending on final desired outcome, may need additional surface treatment. Let’s look at boring first.
Most shops use boring equipment that ranges from a single spindle, deck mounted boring bar to a fully automated machining center. The deck mounted boring bar runs at relatively low spindle rpm’s (around 800 rpm) and equally slow feed while a machining center can go as high as 32,000 rpm with blindingly fast feed rates. This spindle speed/feed disparity limits the kind of cutting tools to be used and correspondingly, surface finishes created. Low speed spindles were originally designed to work with common cutting bits made from high speed steel and carbide while high-speed spindles are optimized for silicon nitride, CBN (Cubic Boron Nitride) or some other high tech cutting material or coating. Regardless the cutting tool material or insert edge profile, each of these tools leaves behind an identifiably different surface structure. Why are these differences important to understand? Rarely is a bored-only surface satisfactory for engine components. Some post boring machining process is generally required to improve component compatibility under load. If we treat all bored surfaces the same, we risk making the type of surface improvement we desire worse or at best, marginally better. So what does an average bored surface look like under magnification?
As can be seen in Figure 1, quite a few pits are evident in the metal and account for some of the deeper valleys exhibited below mean line in Figure 2 surface tracing. Some of these pits are normal voids or tears in parent material but some are graphite inclusion pull outs. Most low speed boring using high speed steel or carbide inserts are responsible for these tears. Left alone, this surface is too rough and contains too little graphite (used for lubrication) to be considered suitable as a finished cylinder wall.
Using much higher spindle speeds, fewer tears and graphite inclusion pullouts are evident. If a profileometer tracing were conducted on this surface, much lower valleys would be seen indicating a smoother surface. However, even though this image depicts a smoother surface, boring marks and potential for sub-surface material fracture caused by high point contact cutting insert pressure,
renders this surface unsuitable as a final cylinder wall finish. A post boring process is required.
As with any machining process, there must be a first step and honing is no different. Did you ever think it possible to hand off too smooth a bored surface for the hone to work? It is possible and if a bored surface isn’t rough enough for honing stones to stay open and active, glazing may occur. Hone stones and especially metal bonded superabrasive stones need a rougher cylinder wall to wear metal bond down and expose abrasive crystals. If metal bond material is not worn, abrasive crystals will not protrude past the bond surface and no cutting will take place. Once a glaze is created, all attempts to correctly finish the honing process are futile. You know glazing takes place when honing pressure increases dramatically while stock removal stalls. The cylinder wall will also take on a dark grey to black look. If glazing does occur, you can etch the stalled finishing stones by scrubbing stone cutting faces with a course abrasive. This scrubbing action erodes the metal bond sufficiently to expose new abrasive crystals. Often, this is all that is required to restore free cutting but if not, replace abrasives in the hone head with roughing stones and remove a few tenths of stock. Repeat with finishing stones and you should be fine. Note: In extreme cases, you may have to bead blast stalled abrasives but this can be tricky and abrasive damage can occur. Do so only when everything else has failed to restore free cutting metal bonded stones.
As mentioned earlier, there is no need to cover basics of honing. You already know this but if not, contact the manufacturer who sold you honing equipment and they should be able to teach you. Generally not discussed are the types of surfaces we create by honing and reasons we measure surface finish the way we do. We take it for granted that a specified surface value is correct without really knowing what it describes. Images of bored surfaces covered earlier used photomicrographs and profileometer tracings to represent what a surface looks like. Many surface parameter values exist to define different parts of a surface structure but for our purposes, boiling this down to a single measurement parameter value is essential. Simply put, an average shop is far too busy to go through extensive surface evaluations for every honing job. What is within the grasp of an average shop is to qualify the honing process and make it a shop standard. To that end, we need a single surface value that best describes the level of quality we want.
Many shops rely on Ra to meet surface specification requirements. Not surprisingly, Ra is still the number one surface parameter in use today. But Ra has limitations and it is important you know what they are.
For decades, Ra (arithmetic average of all measured peaks and valleys) has served to demonstrate a level of quality commensurate with a manufacturer’s specification. Where Ra fails as an indicator of quality is to isolate critical parts of the surface that are good from those that are not. It is possible to have the same Ra value for an extremely smooth surface compared to an extremely peaky one (see Figure 4).
Many of you have seen the Figure 4 surface comparison before in other publications and know that the three depicted were generated by different machining processes (grinding, boring and honing). In fact, each yielded exactly the kind of surface normally associated with that process (see photomicrographs in Figure 4). In these terms, each process was performing correctly yet, surface topography for all three was decidedly different. How then can we rely on Ra alone to describe surface attributes for all machining processes we employ in an engine shop if the surface structures sharing the same Ra are so different? Additionally, how can we be sure to create a correct Ra repeatedly for the same machining process job after job if it is possible for Ra to be so different? The simple answer is that we cannot. Some other parameter must be used as a better surface descriptor.
Michigan Metrology supplied images for Figure 5 and 6 that go further to demonstrate surfaces derived from the same honing process may possess quite different surface topography even though they share identical Ra values. From these honing examples, it is possible to see surface disparities if honing processes themselves are not qualified and controlled. Loss of process control can come from something as innocent as changing coolant, abrasive manufacturers, operating pressures, operators or machines/tools. Any time a change in process is implemented, that change should be qualified to meet your requirement.
Figure 7 further illustrates the need for a surface measurement parameter beyond Ra. This high magnification photomicrograph taken in a scanning electron microscope is of a cast iron cylinder liner honed with metal bonded diamond abrasive at high tool pressure. Resulting sheet cover formation from the honing process has completely covered all visible cut tracks from diamond crystals. If Ra were used to measure this surface, it would average all surface disparities together and may indicate a relatively smooth surface. Rz measuring the same surface may display a lack of normal valleys associated with free cutting abrasives and cause the machinist to look further. It should be noted that this degree of sheet cover formation generally comes from a stalled metal bond abrasive that is not honing (cutting) properly. In this case, metal bonds applied under high pressure smear cast iron over cut tracks and must be resolved.
Figure 8 is a lower magnification of the Figure 7 surface. Note that very few cut tracks are visible and most of the surface remains in sheet cover formation. Some secondary honing operation is indicated or at the least, a freeing up of the metal bond abrasives to generate more free and open cutting. Once again, Ra may measure this surface as smooth and acceptable when in fact, it is not.
Original Equipment Manufacturers (OEM) have wrestled with this same problem for years. Since OEM’s manufacture millions of parts each year, multiple surface parameter evaluations are conducted randomly. The balance of time, Rz is often used as a single parameter measurement to gage adherence to a qualified process. But honing of cylinder walls by an OEM may include many different surface specifications for various families of engines. In this case, Rz alone may not be sufficient.
An example of analytical tools used by an OEM to qualify honed cylinder wall surfaces can be seen in Figure 9. These 3D images were provided by Michigan Metrology and VEECO Instruments, Inc. and generated through non-contact optical profiling technology. The corresponding color scale below the 3D surface image identifies height and depth of all surface features. 3D imaging offers real world depiction of surface in a way profileometers cannot. All surface features and disparities are clearly identified. This example is measured in meters as are most OEM surface specifications for cylinder honing.
Aftermarket engine shops are in a different position and if the honing process is qualified for that shop, Rz may prove highly beneficial over Ra. Let’s take a look at how Rz may help.
Rz is defined as the average maximum profile height derived from the difference between five or ten highest peaks and five or ten lowest valleys found along a sampling length. Surface displaying peak and valley extremes are captured and isolated in Rz and have much greater influence on the final value.
Figure 10 is a chart exhibiting five Rz standards used around the world. All iterations of Rz share much of the same genetic makeup but how peaks and valleys are treated differentiates one standard from another. For our purpose as engine machinists and builders, Rz found in profileometers sold in North America serve us well. Most common Rz standard in North America is ISO 468, JIS B0601. Figure 11 is a larger representation of this standard and should help you understand what the parameter is measuring. It is advisable however, to determine which standard is used in your profileometer and if it is suitable to meet print specification.
By studying groupings of either five or ten highest peaks and similarly for valleys, it is possible to determine if a machining operation created a surface populated mostly in peak or valley. If Ra says you have met surface specification but Rz indicates an overly peaky surface above the mean line (X), this should raise a red flag as to quality of that surface. Conversely, in some applications, more or higher peaks may be exactly what you are looking for and Rz could be capable of telling you how much adjustment is needed in your machining process to make that happen. Ra alone cannot accomplish this. One additional note regarding Rz; although some have tried using a conversion equation to make Ra into Rz, this really isn’t advised. Ra does not account for outlying spikes found in a surface and simply averages them in. As mentioned earlier, missing these outlying or prominent peaks (or valleys) could prove detrimental to component
Finally, I would like to underscore one of the many important messages David Metchkoff makes in his article on honing (found on page 28); make the cylinder round and straight before worrying much about the optimal surface finish. Without a round, straight cylinder, it doesn’t much matter what finish you put on the bore. Rings do not seal well in an out of round and tapered bore.
Similar to boring, a number of milling machines and cutting tools exist ranging from very slow speeds and feeds to the extraordinarily fast. As you might expect, this wide range of milling speeds will also yield a wide range of surface finishes. Couple in all the choices a machinist has in cutting inserts and the problem of surface generation becomes even more complex. Figure 12 exhibits four common cutting inserts that produce four completely different milled surfaces. Insert shape, coating and material are chosen to optimize cutting efficiency and surface finish.
Additionally, many machinists today want to mill dry so one more variable now enters the surface finish equation. When selecting the correct R parameter to define quality of a milled surface, waviness is of equal importance.
Waviness may be induced through the machining process for a variety of reasons but is impossible to catch with the naked eye. Machinists routinely use a straight edge and feeler gage to quickly determine deck warp but for milled surfaces, finding areas of waviness with a straight edge is problematic. Only a profileometer or very sophisticated 3D imaging is capable of capturing this surface disparity.
Figure 13 displays waviness specifics found in a single surface tracing. Any profileometer incorporating the W parameter (it is suggested that this parameter be included if you are contemplating a profileometer purchase) will exhibit a similar tracing for you to evaluate. Many OEM specifications for block decks and cylinder head surfaces specify a maximum waviness dimension.
Correcting for waviness may mean checking several areas of the machining process. Generally, most form errors occur through positional errors (part fixture problems). Clamping of components introduces abnormal force into the part that when released, may cause that component to relax back to normal form. This release of energy is sometimes measured as dimensional change to the part. Positional errors can also come from too little clamping force allowing the part to move during machining. Small oscillations generated by the cutting tool while machining can be seen as a sinusoidal pattern (waviness) on a deck surface as represented in Figure 13. Other reasons for waviness can be back cutting of the inserts or some other machine wear issue that threatens process control (see
Figure 14, page 16).
Most grinding operations we do as engine machinists is for crankshafts but many wet surfacing machines use grinding stones of various types. Four major classifications of grinding abrasive tools can be defined as vitrified bonded (the common round stone wheel), electro plated (Figure 16), metal bonded (Figure 17) or resin bonded (Figure 18). Vitrified and resin bond can be made with aluminum oxide, silicon carbide or one of many superabrasives. Superabrasives are collective families of diamond or CBN. Diamond is either natural industrial grade or manmade while CBN is exclusively manmade. Crystals of both superabrasive types can be found in all bonds but generally are the domain of plated, resin and metal bonds. It is important to understand the distinction between vitrified abrasives and superabrasives because each produces vastly different surface topography. We witnessed in Figures 7 and 8 the degree of sheet cover formation that may occur when honing with metal bond abrasives. A similar formation called “white layer” is generated when grinding and hard turning. An example of a machining center grinding cylinder head decks with an electroplated superabrasive wheel can be seen in Figure 15 along with a 3D image of this surface. Wheels of this type are routinely used (most with metal bonded superabrasives but some are electroplated) in our machine shops for flywheel resurfacing. Some have even found their way into block and head wet surfacing machines.
Open bond structure of electroplating or vitrified nearly eliminates white layer due to the aggressive cutting action this open structure possesses. An example of this is sandpaper. Because abrasive crystals are so proud of the bond, sandpaper is very aggressive at low pressure. Taking this example further, crankshaft polishers using coated abrasive belts must often be dulled before use or uncontrolled, unwanted stock removal will result. Same goes for ball hones when deglazing a cylinder. Balls of loosely bonded vitrified abrasives possess an extremely open bond structure and will cut aggressively at very low pressures (just like sandpaper). Because of this, balls wear quickly the way they were designed to. Care should be taken not to use these open structure abrasive tools for too long a time period or undesired dimensional changes to the part will occur.
It should be further noted that all grinding currently done in a typical engine machine shop is done wet. This is true for all commonly used electroplated, resin and metal bonded superabrasive grinding tools.
Through this short introduction to tools and processes for boring, honing, milling and grinding, some understanding of fundamentals was gained. We as engine machinists come in contact with a variety of materials and engine component configurations that sometimes require special treatment. We are fortunate to have such a variety of tools to choose from because one tool does not fit all applications. It is also hoped that the reader considers how measurements are taken with regard to surface. Although Rz is far from perfect, it is arguably better than Ra.
If this article generates questions, please feel free to visit our forum at www.aera.org (click on the “Forum” tab) and ask. If you are an AERA member, give us a call and ask as many questions as you like.n
Acknowledgements go out to the following references for supplying images and technical information to this article:
1. Makino USA – Mason, OH; 2. Michigan Metrology, LLC, 17199 N. Laurel Park Dr. #51, Livonia, MI 48152; 3. Quality Magazine – The ABC’s of Rz – Alex Tabenkin, October 29, 2007; 4. Surface Finish from Ra to Rz – Ryan Bourget, April 2003. A final manufacturing assignment; 5. Veeco Instruments Inc., Terminal Drive, Plainview, NY 11803, Tel 516-677-0200 x-1222, Fax 516-714-1231; 6. Widder, Ed – Federal Mogul.
A special thank you to Mitutoyo, Taylor-Hobson, Federal Gauge and others for their contributions.
For a PDF of this article (complete with photos), go to: