Process optimization is an essential exercise for today’s evolving and adapting machine shop. Increasing international competition, short part runs and need-it-yesterday delivery requirements demand it.
Unfortunately, optimizing a camshaft lobe grinding process has never been cut and dry.
Its degree of success has largely depended upon operator experience and gut instincts. Computer programs, which take into account known machine dynamic constraints and lobe profile to suggest an “in-the-ballpark” work speed, do exist. Still, many test grinding iterations, paired with opining skilled operators, have been required to dial-in the process. And in cases where lobe burning occurred, some manufacturers chose to decrease the wheel feed increment, while others slowed the work speed. These sort of “seat-of-the-pants” changes usually eliminated grinder burn, but didn’t necessarily yield an optimized process.
Digital modeling is an optimization tool that product designers have relied on for years to refine new designs and assemblies. The technique is increasingly finding favor among manufacturers hoping to fine-tune their metalworking processes. Such a predictive computer software tool gives them the opportunity to immediately see the results after adjusting process variables (playing out “what if” scenarios) while removing some process guesswork and trial and error before any material is removed from a part.
A digital modeling tool for camshaft lobe grinding is now available. New thermal modeling software takes existing computer work speed generation programs for camshaft grinding a step further, actually predicting heat amount, location on lobe perimeter and depth reached under the lobe surface during grinding. Likely problem areas can immediately be identified from simple colored graphs, and process variables can be massaged to determine the fastest possible work speed that won’t thermally damage the lobe. Landis Grinding Systems (Waynesboro, Pennsylvania) has added such a thermal modeling module to its Tetra4000 camshaft grinding analysis program. The program is available for use with the company’s 3L CNC camshaft lobe grinders that have linear motor wheel feed drive.
This timely modeling tool comes at a critical juncture, as increasing numbers of skilled workers are reaching retirement age. Not only does thermal modeling optimize the camshaft grinding process, but it also serves as an educational tool for, and bridge between, camshaft designers and manufacturers.
Camshaft lobe grinding presents processing difficulties not found in concentric grinding operations. The contact area between the wheel and lobe (known as arc of contact) is continuously changing as the wheel passes around the perimeter of the lobe. Contact area is greatest in the relatively flat flank area, versus a camshaft’s more rounded base circle and nose. For that reason, it is in the flank areas that burning is most likely to occur, and where many manufacturers slow the work speed to prevent it from happening. However, educated guesswork has typically dictated how much to reduce the work speed.
Lobe geometries are also becoming more complex. Many of today’s roller camshafts have a re-entrant (concave) profile in the flank areas. This feature, also referred to as a negative radius of curvature (NROC), is designed to optimize valve opening and closing for greater engine power and reduced emissions. However, it introduces additional changing contact areas, making an already difficult grinding process even more hairy. In addition, new roller camshafts experience higher contact stresses than previous designs, which means thermal damage must be closely watched.
Lobe grinding is typically divided into roughing and finishing stages, even though they occur during one processing cycle. The purpose of roughing is to remove as much material as possible. Here, the concern about thermal damage isn’t as great, because successive roughing passes are taken at a wheel infeed depth that is sufficiently deep enough to remove any previously damaged material layers. However, in the final roughing passes, thermal damage must not be so deep that finishing passes at a smaller wheel infeed can’t remove it.
On the surface, the exercise of modeling a manufacturing process as intricate as camshaft grinding might not seem to be intuitive. However, it is a relatively simple process of entering known and published values for machine, wheel and coolant.
The thermal modeling module piggybacks onto an existing work speed optimization and acceleration smoothing program, which considers traditional machine performance variables, in addition to material removal rate and lobe lift profile. Camshaft designers provide lift profiles in terms of the amount of lift per degree around the perimeter of the lobe. Grinder manufacturers provide data about machine dynamic limits such as wheel feed acceleration and jerk, and headstock velocity and jerk.
The three primary parameters required to generate the thermal model module are:
• Cr value—This experimentally obtained value represents the material removal capability of the grinding wheel, and it depends on camshaft material and the type of wheel material and bond (cubic boron nitride, or CBN, with vitrified bond currently is the most popular wheel type for camshaft grinding). These Cr values are derived from machine force and power measurements. The “C” component represents grit density (ratio of abrasive pieces to bonding agent), and “r” represents wheel surface topology (scratch width versus scratch depth).
• Thermal partition constant—This is a percentage of the amount of heat energy that will go into the part versus heat removed by the coolant. The thermal partition constant differs for oil- and water-based coolants.
• Feed increment—This is the depth that the wheel plunges into the lobe for each new pass (usually at the nose where wheel/lobe contact area is smallest), and it will differ for roughing and finishing passes. The wheel does not gradually feed inward in a spiral fashion down to that increment depth, but rather it immediately feeds in that increment amount and will maintain that depth around the entire lobe. This feed increment typically is larger for roughing cycles than finish cycles, and it is directly related to the material removal rate.
There are two types of color plots that the thermal model generates. One depicts the actual lobe shape, with colored temperature bands that show the thermal distribution under the surface of the lobe during one camshaft revolution (shown in Figure 1 on page 66). Temperature spikes toward the center of a lobe show where problematic areas are located.
Another plot illustrates temperature distribution in terms of a graph of camshaft rotation in degrees versus surface depth. This graph
shows to what depth various temperatures reach into the lobe at a particular point around the perimeter of the lobe (Figure 2 on page 68 is an example of this graph for one camshaft revolution).
Typically the entire grinding cycle is plotted for analysis, including roughing and finishing passes (the software is capable of modeling up to 20 camshaft revolutions). Benchmark data dictates at what temperature levels and penetration depths the process should be held, depending on camshaft material. After evaluating the heat values and penetration depth, parameters can be adjusted and the grinding finish cycle can be determined to ensure that any previously thermally damaged layers are removed.
For camshafts that have re-entrant profiles, two grinding cycles would be modeled. One model would be generated for roughing the overall lobe shape with a large-diameter wheel, and another for a small-diameter wheel (typically 80 percent of the re-entrant diameter) for finishing and grinding the re-entrant profile. In order to complete both roughing and finishing on one grinder requires a machine with a subspindle for the small wheel.
Bridging Engineering Disciplines
Thermal modeling for camshaft grinding helps join design and manufacturing engineers to deliver optimal camshaft design and manufacturing processes. Manufacturing engineers’ prime concerns revolve around throughput, productivity and quality—essentially how to make a good part as quickly as possible. The camshaft design engineer must make decisions about material type and lobe profile based on camshaft loading conditions. Metallurgists may also enter into the picture, having concerns about residual stresses and the amount of heat that occurs during grinding. Thermal modeling allows the manufacturing engineer to model a new camshaft design and report back to the designer and metallurgist what the model predicts will happen to the camshaft during grinding, and whether this is acceptable based on material and design.
Thermal modeling can also serve as an educational tool for the camshaft grinding process. For example, prior to the development of this modeling software, the consensus was that work speed had the biggest influence on lobe burning and cracking, which is why it often was the first variable to be changed. It was determined that work speed is actually one of the most insensitive parameters, compared to wheel characteristics and wheel feed increment depth.
This thermal modeling technique has also been used to help identify a problem in which no process variables had been changed, but grinding burn or cracking started to occur. One example is a manufacturer that was unknowingly receiving batches of camshafts with a hardness rating that was higher than anticipated. After the problems started to occur, seemingly for no reason, a thermal model was performed on camshaft lobes purported to be hardened to 60 Rc. The thermal model suggested that no processing problems should be occurring, so material hardness tests were then performed. These tests revealed that the new batches of camshafts were actually rated at 65 Rc. Once this was determined, a second thermal model was generated, and a new work speed for the batch of harder camshafts was found.
Crankshaft crankpin grinding is another process in which thermal modeling is being applied. Thermal modeling is also being investigated for centerless grinding of concentric diameters. Like camshaft grinding, this modeling capability won’t completely eliminate test grinding routines, but it does offer the chance to drastically reduce the number of test grinds and develop an optimized lobe grinding process in a more scientific way.