Click for next page ( 28


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 27
Appendix B Automated Assembly Fixture Drilling The Automated Assembly Fixture Drilling System was conceived as a way to automate drilling of such large contoured structures as wings that would otherwise have to be drilled manually. Like other types of automated assembly tasks it became feasible only when mini-computers became available that could operate in a shop floor environment. The system's task is to scan the piece to be drilled, storing the information in its computer memory, locate and check the bole coordinates, and then drill and countersink holes in a wing skin and understructure, or other large component, mounted In a fixture. The system consists of a CNC drill unit mounted on a vertical gantry capable of f ive-axis movement, a scanning camera to guide and check the work, and a modif fed f ixture to accommodate the automated drilling. Certain conditions related to the technical and market environment for this system need to be understood as background to the case of technology transfer. A discussion of these conditions follows. Technical Conditions 1. Grumman holds a basic patent in Automated Assembly Fixture Or illing because the only developmental work on the device was funded by internal R&D money. 2. Much of the drilling in aircraft manufacture (250, 000-400, 000 holes per average fighter, three times as many per average bomber) could be automated by other means, but existing methods were not adequate for drilling large contoured places that had to be mounted in fixtures. 3. The distinctive concept of the Grumman system was its use of a scanning technique to locate and correct the holes prior to drilling. The embodiment introduced other equipment features as issues--for example, the ruggedness and cost of the system as it was configured by GO unman . 4. The drilling task varies along several dimensions. The needed accuracy and measurement capability of the equipment depends on such f actors as wing conf figuration , structure, and whether holes are tb`rough sk in only or into substructure . The ruggedness of the equipment needed in terms of force delivered, durability, and reliability depends on the volume of shipsets assembled and the type of material (e.g. , aluminum or titanium versus composites) to be drilled. 27

OCR for page 27
5. The main drivers for automating the drilling task have beer increasingly tighter tolerances, the expense of templates, and the time and cost involved in manual wor k when pilot drilling and back drilling are required. Another driver is the need for consistency to reduce the danger of major scrappage. A human driller becomes progressively less accurate during an eight-hour shift. 6. Other automated alternatives to Grumman's device bave been explored in the industry. One has been the location of holes using a laser beam, locater, or some other form of sensor that can look through the wing skin at its substructure. Another alternative is a robotic driller (see below in General Dynamics). Market Conditions 1. Aircraft manufacturers beve a variety of cooperative manufacturing arrangements. Contractors act as prime contractors for some programs and subcontractors for others. Major structural parts are often subcontracted to other firms. In this case, Grumman was to have the subcontract, under Rockwell ' s pr ime contract, for the B-1 Bomber horizontal stabili zer . 2. A number of traditional divisions have existed in the industry that tend to affect relationships among contractors. One is the traditional identif ication with a particular branch of military service. General Dynamics has historically worked primarily for the Air Force whereas Grumman teas historically worked more for the Navy. These historical relationships do not prevent companies from designing planes for either service, but the differences in the ways the two services have dealt with their contractors and the somewhat different design traditions have some effect on company development and manufacturing philosophies. 3. Shortages of skilled labor and large f fluctuations in company workforces between major programs have been driving factors towards automation of airframe assembly, particularly during the 1970s. General Dynamics' Fort Worth division, for instance, has fluctuated between 35,000 workers at the height of the F-lll program and 6,000 workers before the F-16 program began to build up production. 4. Because of the nature of Air Force contracting procedures, major new capital equipment is rarely purchased by an airframe manufacturer outside the time when the company is tooling up for a major new program. Most companies monitor developments in tooling routinely, but the cost of shifting equipment in the middle of a program is generally prohibitive. In 1969 and 1970, the Air Force Materials Laboratory began to call for proposals for new ways to automate assembly. The Sagamore and French Lick Conferences both emphasized the Air Force's interest in 28

OCR for page 27
acquisition cost reduction, a shif t from its traditional emphasis on manufacturing for Estate of the art. performance. Programs that were seen as appropriate for Air Force funding would, according to regulations, have to be technically feasible as demonstrated in the company laboratory, generic (applicable to other Air Force programs with a clear indication of payback I, and beyond the normal risk of industry. Programs should promise cost reduction, materials conservation , or shorter lead times. An assumption behind the funding that was not necessarily reflected in specific contracts was that a technology would be reported in such a way that it could transfer readily. Anyone skilled in the art should be able to practice. Theoretically, only lead-time and experience would separate the originator from the adopters. Transfer was desirable to provide a second source for all forms of manufacture. Although spreading the new cost reduction concepts was recognized as valuable, the one sure evidence of successful payback for Air Force funding would be the physical replication of a system in another company's manufacturing facility. In most cases, then, the AFML considers a transfer successful if the embodiment as well as the concept is transferred. Both Grumman and General Dynamics took up the AFML challenge to apply automation to drilling in the early 1970s. General Dynamics began to pursue the concept in conjunction with its F-16 program, for which the first prototype was produced in 1974. Grumman was not in the early stages of a major new program; its F-14 was already too advanced in 1974. Nevertheless it pursued the concept in anticipation of later programs. Both companies submitted contract proposals for their systems as part of the Air Force Manufacturing Methods Program in 1975. The AFML Manufacturing Technology Group selected Grumman's approach to fund because it was judged technically superior. Assessors at the AFML did not believe that General Dynamics' tripod locating approach was technically feasible. Further, Grumman's proposal to include a scanning device as part of the system was attractive. Since Grumman did not itself have a major airplane program coming up to which the new system could be applied, the ANAL suggested that Grumman should cooperate with General Dynamics and demonstrate its system on an F-16 part . Grumman thus became the originator of this automated wing drilling system and General Dynamics became, in the AFML's view, the designated adopter. Grumman Grumman Aerospace had roughly $1. 2 billion in sales in the mid-1970s, of which all but 3100 million was aerospace business. The company was attempting to diversify to lessen its dependence on the volatile defense industries by producing mass transit vehicles and by subcontracting for commercial aviation houses. Nevertheless it still 29

OCR for page 27
relied on military business, mostly from the Navy. Because it had no prime contracts for major new weapons systems, Grumman was seeking major subcontracting business. Grumman was non-unionized, and it tried to stabilize its work levels as much as possible to avoid laying off skilled workers and engineers. Any new process technologies that it might develop were regarded as potential sources of income and possible opportunities to gain significant subcontracts on a fluid Pro Duo basis. Upper management made known its expectation that any other company that adopted Grumman's technology should be prepared to provide Grumman with a reasonable return on its investment in one form or another. Such a return should not only cover Grumman's development cost but also offset the potential cost and risk to Grumman in transferring its development. The additional costs of transfer included extraordinary amounts of documentation and potential legal liability. Grumman's Advanced Development Group was located in a separate facility, Plant Twelve on Bethpage, Long Island. Its staff consisted of a core of 35 permanent employees and a number of others borrowed from the different divisions that the group served. Advanced Development had a standing mandate to f ind opportunities in high cost areas of production, to anticipate production processes needed for major new programs, and to formulate responses to critical material shortages if they arise. The automated wing drilling device was pursued not only because it was expected to offer savings in direct labor cost, throughput time, and f ixture fabr ~cation, but also because i t would enable Grumman to drill wings with improved consistency. Grumman takes pr ide in its reputation for quality and consistency in i ts production processes. As a result the f irm sought equipment designed for a high degree of accuracy. Grumman's stated objectives for the automated assembly wing drilling system were first to reduce production labor with a minimum of capital investment and, second, to improve hole quality . Grumman demonstrated an early prototype version of its device to representatives from General Dynamics and Fairchild in 1975 before its contract with the ANAL began in May 1975. The Grumman representatives indicated their intention to bold the capital equipment cost to $100,000. They also said that the company planned to charge a royalty as a 1 icensing fee . A f igure of f ive cents per hole was suggested . Shortly after the demonstration, Grumman requested drawings of the F-16 wing from General Dynamics. It received a few documents in response, but the flow of information soon ceased. When it was clear to Grumman that General Dynamics did not intend to cooperate further, the Advanced Development Group shifted to demonstrating the wing drilling system on Grumman's own A-6 program. The results of their evaluation on the A-6 part ~ rated at about one shipset per months showed savings of 40 percent with potential further improvement through the learning curve effect if a larger volume of components were drilled. 30

OCR for page 27
General Dynamics General Dynamics was one of the largest defense contractors in tbe country in the mid-1970s. It" Fort Worth Division bad produced bombers for the Air Force for a long time. The F-lll had been its first f ighter program. The F-lll program had encountered such serious cost overruns that the Fort Worth division had mounted a thorough cost reduction effort in order to sell its F-16 programe The F-16 was designed to be smaller, lighter, and simpler than earlier f ighter designs. Its low cost was partly responsible for its successful sales. said to be the largest single buy in history. At peak rate, production would reach 20 shipsets per month, an unusually high volume that required General Dynamics to rethink its approach to manufacturing in many areas. One of the main ways General Dynamics chose to reduce cost was by reducing direct labor, which had the added advantage of moderating the workforce fluctuation that the Fort Worth Division had typically experienced from one airframe program to the next. As a result the General Dynamics Manufactur ing Technology group was seeking to automate such labor-intensive operations as wing drilling. General Dynamics had strained relations with Grumman in the early 1970s because of a bad experience with the F-lll program. In the early stages of the F-lll the two companies had been partners, with General Dynamics taking the design lead for the Air Force version, and Grumman the design lead for the Navy version. Then the Navy had pulled out of the F-lll program in favor of Grumman ' s F-14 which began in 1968-69, and relations between the two companies were damaged. When Grumman showed interest in subcontracting in the composite production area for the F-16, General Dynamics refused to do business. General Dynamics' loss of the automated wing drilling contract to Grumman did nothing to improve its predisposition to cooperate on its further development, especially when Grumman indicated its intent to charge a royalty for use of its system in what General Dynamics interpreted as a violation of the spirit of Air Force sponsorship. Nevertheless, when Grumman demonstrated its system for the industry in 1976, General Dynamics evaluated the system for use on the F-16. General Dynamics had already abandoned its own former approach to automated wing drilling. Grumman continued to request a royalty for i ts propr ietary interest in the system. The Air Force contract had funded the f ixture and the software development, but it had not compensated the company for its prior investment. The AEML left negotiation of licensing arrangements strictly up to the parties involved unless and until an impasse was reached. General Dynamics manufacturing technology personnel who went to i nvestigate the Gru~Tunan system reported that the equipment had now been designed for the A-6 and would consequently require considerable adaptation for use on the F-16. The following problems were cited: 31

OCR for page 27
1) General Dynamics questioned the ruggedness (structural rigidity) of the Grumman prototype for use on the much higher volume F-16. The 60 pounds of force that the Grumman drill delivered was also lower than General Dynamics needed for its wing application. 2) Grumman had tried unsuccessfully to interest Cincinnati Milacron in building the machine, and General Dynamics was uncomfortable purchasing for use in production what amounted to nothing more than a set of drawings, because the system would still be only a prototype when they had replicated it in Fort Worth. Poor communications with Grumman only increased the anticipated difficulty of getting all the necessary information. 3) Because the F-16 had been designed to be easily manufactured, the Grumman system was designed to be more accurate than was necessary for the General Dynamics application. Taking into consideration the original cost of the equipment and the royalty (;ru~Tenan was ask ing, the General Dynamics evaluation showed that the cost of adopting the Grumman system would be roughly comparable to the cost of designing and developing a wing drilling system using a robot. The General Dynamics organization had been look ing for suitable robotics applications in which to gain experience, and the wing was such an application. A robot would be less accurate than the Grumman system but would have the advantages of lower capital cost and flexibility for use on other tasks. Taking all these factors into account, General Dynamics rejected the Grumman system in favor of developing its own robot driller. Because the economic evaluation was indecisive, the risk caused by the poor relations between the two companies, as well as the existence of attractive alternatives, added up to non-adoption of the Grumman automated assembly fixture drilling system. Fa irchild (not interviewed ~ When General Dynamics rejected the Grumman system in 1976, Fairchild, which was also in the early production stages of a new program, the A-10, agreed to cooperate in evaluating the Grumman system. Fairchild signed an agreement to lease the system and indicated its intention to adopt if the economics proved attractive. Tbe AFRO supported this further demonstration of the system with a new contract. Tooling had already been completed for the A-10 program, but it seemed possible that the savings from the Grumman system would be sufficient to justify the unusual step of changing in mid-program. Grumman was no longer demanding a royalty for use of its concepts; the AFML had involved itself in the discussions at the beginning of the cooperative demonstration program, and licensing terms agreeable to both parties had been stipulated in the demonstration contract. In early 1979 Fairchild rejected the Grumman system, saying that the savings that could be expected two years into the program were not sufficient to warrant changeover. Timing was clearly the decisive 32

OCR for page 27
factor in the non-adoption since the AEML had funded the reduction of uncertainty . McDonnell McDonnell was approaching the S2 billion mark in revenues from military aircraft, mostly for Navy use, in the mid-1970s. It was known as a design house, dominated by engineering and highly conservative in manufacturing matters. No separate manufacturing technology group existed at McDonnell to anticipate future production processes. New equipment was adopted when it constituted a low-risk investment that promised to pay of f in the short term. McDonnell paid attention to the reports issued by the AFML concerning new processes. It rarely competed for development contracts, however, in part because it lacked a separate manufacturing technology organization to focus on such matters. The 1976 Grumman demonstration spurred McDonnell's manufacturing process engineers to pursue the automated wing drilling concept. Until the F-18 program, manual drilling had seemed the most economical approach for wings. But the F-18, McDonnell's first significant composite airplane, required more tooling than previous planes. Ordinary numerically controlled equipment would not be adequate for F-18 wings because the presence of seal grooves on their edges made the boles harder to locate through simple edge distance measurement. Grumman's scanning approach seemed to provide the solution to this unusual measurement problem. After careful evaluation of the Grumman system for their F-18 application, the McDonnell process engineers chose to develop their own system. The Grumman equipment fell short of their needs in several ways. The F-18 wing was a heavier machining task than the A-10, requiring 80-250 pounds of force delivered instead of Grumman's 60. Moreover the complexity of the drilling task on the F-18's graphite and titanium skin was much greater than the aluminum drilling task on the A-6, requiring many more tool changes. Changing the Grumman tool took 10 to 15 minutes, which posed a serious obstacle to adoption. McDonnell's device was to be heavier duty, rated for 12 shipsets per month. As of 1980, the McDonnell device has yet to be adopted in production, in part because McDonnell teas yet to f ind a machine tool builder to build it. This is unusual because McDonnell, operator of the largest numerically controlled machine shop In the Free World, r arely has trouble ga in ing the attention of mach ine tool compan ies when it wants something. Nor throp Northrop Aviation was one of the smaller military aircraft producers in the late 1970s. It was traditionally a Navy contractor, 33

OCR for page 27
and in the late 1970s it was planning for its F-18 subcontract under McDonnell's prime contract. Northrop had a longstanding reputation as a low-cost producer characterized by innovative manufacturing. Lacking the organizational resources to do much researob, it followed a policy of seeking innovative process equipment from a variety of sources and adapting it for use in airframe production. Upper management encouraged this receptivity by not insisting on 'strictly abort-term payback s . Engineers in the manufacturing process organization at Northrop became aware of the Grumman device by reading the interim reports of the A-10 contract. The Grumman scanning approach was attractive to them for use on the vertical stabilizer. In 1978 they began a feasibility study on the Grumman system applied to the F-18 component, assuming a capital cost of S250, 000 for the automated fixture drilling system. Since they were proposing to do 10 shipsets per month there were some doubts about the system's structural rigidity, but they estimated that it would not be too complicated to adapt to their purposes. They calculated expected savings of about 28 percent, yielding a modest but acceptable payback period of nearly four years. Towards the end of the feasibility study, however, Grumman raised the capital figure to $1.2 million. At that point Northrop rejected the Grumman system on financial grounds and turned to less expensive alternatives, such as a laser locater device with a calculated cost of S15,000. No alternative has yet been adopted because the F-18 stabilizer has undergone some design changes. 34

OCR for page 27
Automated Fixture Drilling Chronology 1969-73 Series of Air Force-sponsored acquisition cost reduction conferences. 1974 1975 1976 1977 Automated fixture developed at Grumman and first production panels drilled a year later. Grumman demonstrates Automated Assembly Fixture Drilling to General Dynamics and Fairchild. Grumman performs the Air Force contract evaluating 5-axis operation on A-6 parts, using the f ixture on the A-6 assembly line. Fairchild signs a lease agreement for the Grumman system; General Dynamics rejects the system and opts for its own robotics approach . AEON contract supports application of the Automated Fixture Drilling System to Fairchild's A-10 stabilizer. 1978-79 Discussions between Northrop and Grumman result in rejection by Northrop because of increased capital cost. 1979 Fairchild re jects the system and cites timing as sole reason for re j ect ion . 35