PickaPara

IMPCO.MACHINE.TOOLS

Pick a Parameter...But Not Just Any Parameter | Manufacturing Engineering, October, 2002

Norm Judge, Director of R&D, IMPCO Machine Tools, Lansing, Michigan

Little things, like the microscopic features of a machined surface, sometimes have an enormous impact.

This is especially true in the case of precision rotating, load-bearing components such as automotive crankshafts and camshafts. That's why the ability to evaluate a bearing surface and the process used to generate it is vital to assuring the bearing's ability to perform its intended function. But it isn't as simple as just checking average roughness (Ra)

Over the years our company has refined microfinishing - the finest possible metal removal process-to the point of controlled removal of just microns of material, typically steel or iron, from camshafts, crankshafts, and other precision load-bearing journals. Surface texture evaluation is critical.

When you examine a precision-ground part under a microscope, you will see fragmented material. This amorphous, unstable surface material layer is part of every turned or ground part. On non-microfinished bearing journals during normal operation, this material loosens, erodes the mating component, and breaks up the supporting film of lubricant, causing poor performance and even bearing failure.

Most people accept that the ideal bearing journal features a finish that is as smooth as possible. Some wear surfaces, such as cylinder bores require a small amount of texture to retain lubrication, since a journal-to-bearing interface requires a cushion of oil under pressure to keep the two mating surfaces separated. The smoother the surface, the thinner the lubricant cushion can be before rough spots on mating surfaces come into contact, and the higher the load capacity of the bearing. A logical proposition as far as it goes, but the measurement and documentation of journal surface and load-bearing ability is key.

And, not all surface finish measurements are helpful-some are actually misleading or do not accurately characterize the surface.

Because precision part performance requirements are high, manufacturers must understand and use the surface measurement parameters and evaluation techniques that provide the best information for any application.

Parameters are numerical values assigned to a surface, the most common being Ra, which

averages the peak and valley displacement from a mean line but provides no information about the height of the peaks and valleys or the ability of the material to bear a load (bearing ratio), a vital characteristic for crankshafts. Two different surfaces may have similar Ra, but could function differently. A bearing surface has quite different requirements than a surface to be painted. That's why it is essential to think about the purpose of the surface when trying to decide which surface finish parameter to use in measurement.

Surface finish parameters can be classified in three categories:

Amplitude parameters (peak or valley heights), spacing parameters (spacing of the irregularities on the surface), and hybrid parameters (determined by a combination of amplitude and spacing). Amplitude parameters reveal variations in the profile height of a surface. Spacing parameters are sensitive to variations in the wavelength of the measured profile. Hybrid parameters measure both.

Choosing the right surface finish parameter will help you control part performance, manufacturing process, and machine, particularly during production of bearing journals. The parameters of interest in this context are amplitude parameters.

One parameter that is more sensitive to occasional high and low peaks than Ra is Rq. It is the roughness root mean square (RMS) of a profile calculated over a single sampling length, or the mean result of a measurement over five consecutive sampling lengths. It gives more weight to higher peaks.

Peak-to-valley height can be shown in several different parameters. For example, Rt, the vertical height between the highest and lowest points within as assessment length, can be used to characterize a finish that may be subject to subsequent metal removal operations, such as honing or lapping. Rt shows how much material can be removed before the part size reaches a limit. It can also be used to determine when processing a part would be ineffective.

Rmax, the largest of five roughness depths with a sampling length, can be affected by surface dirt or imperfection, so it's more likely to use an average peak to valley height of five consecutive sampling lengths.

Profile height is characterized by more than one parameter. The maximum profile height from the mean line within the sampling length is Rp, and its mean value determined over five consecutive sampling lengths is Rpm, Rtm is the mean value of Rmax in five consecutive sampling lengths. Rv is the depth of the lowest point below the centerline within the sampling length.

Skewness, Rsk, is a measure of the symmetry of the profile about the mean line. It can be used to distinguish between profiles having the same Ra or Rq values. it shows whether a cast or porous surface will provide a useful Ra value and can be useful in judging bearing surfaces; a good bearing surface would have a negative skew.

Ten-point height, Rz, is the average height difference between the five highest peaks and the five lowest valleys measured over one sampling length from a line parallel to the mean line but not crossing the profile. Some manufacturers use the ten-point count with Ra.

Bearing area ratio (Tp) is the length of the bearing surface expressed as a percentage of the assessment length L at a depth, or "slice level," below the highest peak. It is useful because it simulates wear at various cutting depths of the surface. it can be thought of as a lapping plate honing off the peaks, leaving a flat bearing area. Thus Tp is useful whenever bearing surfaces must be analyzed for wear or lubrication properties. Bearing area ratio depth is critical for machining processes that have characteristic peaks, such as grinding, hard turning, or milling.

Tp height uses the mean line as a reference, making it possible to ignore an occasional peak or valley and evaluate the bearing ratio compared to an average of the entire surface. Htp measures the distance from the upper bearing ratio to the lower bearing ratio. This is valuable because it will tell you what the oil film thickness must be at a practical bearing load, potentially an important consideration in the design of a critical part.

This parameter can be expressed as the Abbott-Firestone bearing curve. The bearing ratio curve is generated by plotting a curve at various depths. The slope of this curve can be useful in determining how fast a surface will wear and the dimensional size change likely after wear-in. For higher bearing loads, a flat curve is desirable. This type of curve represents a surface that requires very little depth to obtain a supporting surface.

Rk, Rp, and Rpk are surface finish measurement parameters best used to evaluate the bearing area curve Tp% (Abbott-Firestone or material ratio curve). The height of the curve is termed the roughness total, and the width is the percent of material at different profile amplitudes. Rpk, reduced peak height, measures the portion of the surface that will wear off during initial loading, while Rk, core roughness depth, is used to describe the portion of the surface that will support most of the load - the long-term running surface influencing the life of the bearing component. Rvk, reduced through-depth, is the oil-retaining capability of the toughs machined into or inherent in the material.

Some surfaces and materials present a special challenge. Creating a satisfactory bearing surface is trickier with nodular iron, which is used for many automotive crankshafts and engine balance shafts. Measuring the surface is tough, too. Microscopic ferrite and graphite inclusions up to 0.002" (50 um) in size are left in the iron, and some protrude above the ground surface.

While grinding may improve the Ra of the cast iron surface, it exposes ferrite caps or pulls them out. leaving tiny peaks or craters in the bearing surface. Caps also may be smeared over during grinding, effectively disguising them in the surface. Subsequent operations such as microfinishing would then pull them out, thus appearing to make the surface finish worse if only Ra is considered. For bearing surfaces, it is essential that Rpk be reduced as much as practical to eliminate undesirable wear and potential bearing failure in service.

Case in Point. Occasionally customers send parts back, suggesting that there is something wrong with the finish on the parts. These are typically nodular iron components with bearing journals, In one case, we examined a microfinished bearing journal where the Ra readings were sometimes the same as those for the incoming ground surface - ostensible no improvement. Microfinishing is supposed to vastly improve ground surfaces. Other journals on the crankshaft were within specification.

For comparison, we then took a new as-ground shaft, microfinished the five main journals in an increasingly fine three-level microfinishing process, and then carefully examined the No. 3 main and the No. 5.

We found that with nodular iron parts, better finished can actually result in worse Ra values. This is due to graphite nodules that remain in the surface or microscopic pits in the part surface. The Ra readings on the crankshaft looked bad because the measuring instrument needle dropped into holes during its trace. When it came out it skipped, suggesting a peak on the chart. The holes could be 50 um across and as deep - plenty of room for a skid to drop into.

We found that the sizes of the graphite nodules were two to three times as large in the No. 3 journals as in the others, providing a relatively large, deep hole into which a measuring instrument stylus would drop. We checked this idea by applying two levels of microfinishing to the part, increasingly fine, and then a third level of extremely fine 9-um grit to the bearing - small enough to get into the craters and remove the graphite. This step made it very obvious in the photomicrographs that the stylus would drop into the cavities, producing a poor Ra reading.

Ra on microfinishing level 2 went from 0.06 to 0.13 um, out of specification where 0.10 um or less is acceptable. After level 3 microfinishing, Ra actually increased again, to 0.28 um. Ra of the journals as they come from the grinder is typically 0.21 - 0.26 um. After level 2 microfinishing, one bearing measured 0.28 um Ra, where the Ra after grinding was 0.26 um. Ra measurements would suggest that the part was not even touched by the microfinisher. So we have to look someplace else for a useful measurement parameter.

We started checking Rk and Rpk, because this leak-bearing journal will be supporting the bearing shell in the engine and subject to high leads. As ground, the journal (#1 main) had Ra of 0.27 um and Rk of 0.75 um. After finishing the surface with 40-um microfinishing film, 15-um film, and then 9-um film, Ra was improved to 0.15 um from its ground state. More important, Rk was improved from 0.75 um as-ground to 0.15 um, creating a surface that will reliably support a bearing load - one with no peaks and many valleys, which provide the important added benefit of retaining lubrication.

On many parts, peaks are unwanted because they break through a lubrication film, sticking straight up into the bearing shell. If left on the part, these nodules would wear away after the first few revolutions, but they would be left inside the engine, possibly damaging other components.

For microfinished bearing journals, we are usually asked to hold 0.2 um or better Ra statistically; Rk, in conjunction with Ra, is to be preferred because it indicates the strength of the supporting material available for the journal. Ra alone is not adequate for characterizing nodular iron surfaces because of the ferrite cavities. If Ra must be used, a different parameter should be used to verify the surface, such as Rk, Htp, and Tp also are universally helpful in characterizing a surface that must support a load. Relying only on Ra can be misleading, suggesting a machining process in incapable when it is not.

Return to What's New page

IMPCO Machine Tools

3417 West St. Joseph Street, Lansing, Michigan USA 48917

Phone: (517)484-9411 Fax: (517)484-0502

Email: Sales @IMPCO.com