Manufacturing Researchers Report
Old issues and new developments are the subjects of papers presented at this year's conference of SME's North American Manufacturing Research Institute
By Brian J. Hogan
The thirty-second annual North American Manufacturing Research Conference, (NAMRC 32), was held at the University of North Carolina Charlotte June 1 - 4. During his keynote address to attendees, Jimmy Williams, director of development, Alcoa Technical Center (Pittsburgh) noted that direct labor cost in the US is about $30/hr and in China it's $2/hr. But despite this much-publicized difference, during his address on June 2 Williams dismissed those figures. He believes the problem with manufacturing in this country is not worker wages, but the lack of innovation. He believes US manufacturing needs to take advantage of the many manufacturing technologies now under development.
Williams says a four-step progression is followed by most products: ferment (growing interest in an idea), dominant design (several try to develop the idea and one becomes the leading concept), incremental innovation (the leading concept is improved), and maturity (the accepted design begins to lose dominance). He noted a parallel to this idea in the development of the airplane. First there was an early interest in flight, then the Wright brothers had a successful design. This was followed by the airplane's evolution from wood to aluminum, then composites as the key structural material.
Williams believes, "When any given technology is improving at a decreasing rate, it's likely a new technology will emerge to supplant it." Therefore the key to success is to get into the market with a new idea at the time of transition when it's possible to supply something that no one else is offering.
But, he notes, technology is not always the major issue. "Managing the change in customer preference is often the hardest task."
Sometimes a major obstacle to innovation is what Williams calls "the competency trap." It's when a company has a core competency that has been the key to its success in the past. Managers are, therefore, reluctant to take the risk of pushing into new technologies.
A second stumbling block he described is the reluctance to take full advantage of new technology. He cites the current push for HSM where, too often, a company will concentrate on a specific problem and ignore the larger opportunities. For example, HSM not only leads to faster production, but can often mean lighter parts, fewer tools needed, the chance to make modules rather than multipart assemblies, plus the many positive changes in machine tool and cutting tool design.
To take advantage of developing technologies Williams suggests:
- Watch for signals of a changing technology.
- Both component and total system changes offer the opportunity for gain.
- Don't rely on tradition. You can't count on a company's established channels for gaging a market because they were not designed for the new technology.
- Be ready to manage change. Timing is everything.
At the Founders Lecture, also given on June 2, Alan T. Male, professor of mechanical engineering, University of Kentucky (Lexington) reviewed some of NAMRC's history, particularly its academic versus industrial makeup.
He noted that in the organization's early years, the split between the two groups was about 50/50. Today, after 32 years, it's about 15% industry and 85% academics. This shift is due to several factors, the most important being the decline in company research. Many organizations that initially supported large research staffs have closed these facilities or greatly reduced staffing. There was also a migration from industry to teaching after layoff/retirement by some active individuals.
One potentially negative effect of this shift was the emphasis on nonspecific research. "For our own survival, we need to focus on research activities that have industrial relevance as well as academic rigor. If no one can use the results of our work, what good is it?" he asks.
The main focus of NAMRC is always the presentation of research to the manufacturing community. At NAMRC 32, papers were presented on a range of subjects related to manufacturing. The following summaries are representative of the work being done to resolve problems facing manufacturing engineers and managers.
A model for comparing and evaluating toolholder/spindle interfaces of machine tools was presented in the paper A Methodology to Measure Joint Stiffness Parameters for Toolholder/Spindle Interfaces. John S. Agapiou of the Manufacturing Systems Research Laboratory at General Motors R&D Center (Warren, MI), points out in this paper that the tooling structure is the most critical element of a machining system, and is also often found to be the weakest link in the system.
To select the right toolholder interface for a range of applications, manufacturing personnel have typically relied upon data from bench tests, which have always been used in the past to compare interfaces. Agapiou's paper demonstrates that bench tests don't properly represent the interfaces in a machine tool spindle, and offers a proposed methodology using a machine spindle to produce more accurate results.
In his work, Agapiou considered two styles of toolholder/spindle connections--the CAT-40 style and the hollow-shank HSK-A63. In his paper, he describes the experimental test setup, which used simulated tools, a very stiff spindle, a spindle nose, and an instrumented drawbar. The CAT-40 and HSK-A63 test tools were made monolithic. Static and dynamic tests were performed on this experimental setup.
Agapiou compares results from bench tests and experimental results obtained from the machine-tool spindle. When tested in the spindle, tool stiffness is about equal for both interfaces, especially below 600 N radial force at the tool end. The results from bench tests, however, indicate that the HSK interface is much stiffer than the CAT system. Natural frequencies are much lower when the toolholder is in the spindle than in the bench fixture.
| Comparison of static bending stiffness between CAT-40 and HSK-A63 toolholders on a bench fixture and in a machine-tool spindle.
The methodology developed to experimentally determine the toolholder/spindle interface characteristic joint stiffness parameters uses 2-D Finite Element Modeling (FEM). A two-degree-of-freedom (2-DOF) spring with linear and rotational stiffness components is extracted from a 2-D FEM of the bench test fixture. It was evaluated against two other approaches:
- The FEA in ABAQUS finite-element code, and
- A method based upon rigid-body dynamics and frequency response function measurements.
This new method can be refined and adapted universally for generating common data for all available interfaces, according to Agapiou. These data can be used during the design of a spindle interface, for selection of the proper cutting conditions in an application, or to compare toolholder interfaces.
This work shows that the static and dynamic differences seen at the tip of the cutting tool depend upon the stiffness of the tool, spindle geometry and bearings, the housing, and the overall machine structure. The methodology developed by Agapiou provides a means of evaluating several styles of interfaces in a spindle without performing actual tests on the spindle.
White layer, sometimes formed during grinding or hard turning, may be a serious problem depending upon the application. In a paper entitled Microstructural Characterization of White Layers Formed During Hard Turning and Grinding, authors Y.B. Guo (Department of Mechanical Engineering, the University of Alabama, Tuscaloosa, AL), and G.M. Janowski (Department of Materials Science and Engineering, University of Alabama at Birmingham) set out to understand the nature of white layer.
They specifically looked at formation of white layer during hard turning and grinding of AISI 52100 steel. It's known that a white layer usually consists of two layers: a white layer and a dark layer, with the bulk material beneath. The white layer is harder than either the dark layer or the bulk material. A turned white layer is usually less than 12 µm thick, while a ground white layer may be as much as 100 µm thick.
In the studies done by Guo and Janowski, ten turning samples and ten grinding samples were prepared. They found that the ground white layer etches readily, while the turned one does not. In their studies, the thickness ratio of dark layer to white layer is 2.5:1 for the turned surface and 5.3:1 for the ground one--likely a result of differences in strain-hardening and heat generation between turning and grinding.
Data demonstrate that the microstructure of the turned white layer is significantly different than that of the white layer produced by grinding. A turned white layer is much more strained than the ground layer, and looks severely deformed in the cutting direction. Undissolved cementite particles are present in the matrix. The turned white layer is featureless, while a ground layer shows clear microstructures. Turned white and dark layers have much more retained austenite (10 - 12%) than ground ones (0 - 3%). White and dark layers experience about the same tempering in turning, while a ground white layer has much less tempering than a ground dark layer. Also, microstructural evidence indicates that mechanical deformation plays a larger role in white layer formation during turning, whereas thermal processes dominate white layer formation in grinding.
Milling tool stability was the subject of a paper by John C. Ziegert, Charles Stanislaus, and Tony L. Schmitz of the Department of Mechanical and Aerospace Engineering, of the University of Florida (Gainesville, FL), and Robert Sterling of The Stellar Group (Jacksonville, FL). For any milling operation, the limiting depth of cut (DOC) for chatter-free machining is essentially set by the dynamic stiffness of the most flexible mode of vibration of the tool/holder/spindle system. If the dynamic stiffness of this mode can be increased by some multiple, chatter-free DOC at any speed, and the material removal rate (MRR) will also increase.
For many high-speed milling operations, particularly in aerospace work, it's necessary to produce deep pockets with small corner radii, which requires the use of long, slender, and relatively flexible tools. Increasing the typical tool damping (usually 1% or less) will have the same effect on chatter as increasing stiffness.
In the work described in this paper, researchers increased end-mill damping by hollowing the tool body and inserting a multifingered damper into the center opening. Fingers are created on the insert by using EDM to cut axial slits along most of the length of a cylinder whose OD matches the ID of the tool body, thus forming multiple fingers.
When the tool bends, the fingers bend with their neutral surface passing through their own centroids. Thus the axial strain experienced on the outer surface of the fingers is different than the axial strain on the inner surface of the tool body, causing relative sliding between them. When the tool rotates at high speed, centrifugal forces press the fingers against the inner surface of the tool body, creating Coulomb frictional losses.
| Stable cutting depth versus spindle speed using solid, hollow, and damped tools.
Two HSS end mills were made during this study. They are three-flute mills, both 19.05-mm in OD and 125-mm long. One has an internal blind hole 9.5 mm in diam and 105-mm long. An eight-finger damper insert made from a 9.5-mm OD tungsten carbide blank was slit down 76-mm of its length to form fingers. Its solid portion provides a light press-fit into the end mill.
Data obtained during experimental cutting studies indicate that for most cutting speeds, the damped tool outperforms solid and hollow tools. Maximum stable cut depth for the damped tool was 53% higher than for the solid tool (2.3 mm versus 1.5 mm).
The damped tool does not provide deeper cuts at all spindle speeds due to differences in the natural frequencies of the three tools and the shifts in stability lobes that result. But in high-speed milling, the spindle speed is selected to maximize stable metal removal rate. When this choice is made, the researchers report that the damped tool can provide as much as a 53% increase in MRR versus a conventional tool.
Micro parts and micro products call for the use of micro-EDM. In a paper entitled Tool Wear Compensation and Path Generation in Micro and Macro EDM, Jayakumar Narasimhan, Zuyuan Yu, and Kamlakar P. Rajurkar of the Center for Nontraditional Manufacturing Research at the University of Nebraska-Lincoln describe the development of a theoretical model to create the toolpath for generating a desired workpiece surface profile. They illustrated the application of the model to micro and macro-EDM by machining flat slots.
The researchers explain that in die-sinking micro-EDM, tool wear compels the use of multiple tools with increasing dimensions to produce the desired geometry and accuracy. Such tools are expensive and making them is time-consuming. Micro-milling by EDM, which employs a single, simple square or cylindrical tool guided by CNC, is becoming a more economic method for micro-EDM. But tool wear still results in geometric errors, so it's necessary to compensate for wear during machining.
A proposed model and its integration with the Uniform Wear Method (UWM) of compensating for tool wear are discussed in the article. The model development assumes that tool wear factor and discharge gap remain constant during machining. This model was verified by doing micro and macro machining of flat slots. After micro machining, the bottom profile of the machined slot was measured. The bottom surface flatness error was less than 1 µm, far less than the total feed of 10.5 µm for each layer. Macro slot machining was also done to validate the model. Using UWM with linear compensation develops a surface with higher dimensional accuracy than that achieved without linear compensation.
This article was first published in the September 2004 edition of Manufacturing Engineering magazine.