A product first engineered to perform a function is never the end product produced for customer consumption. This division between "functiona1" engineering and "production” engineering is a basic requirement since it is not possible in evolving the original design to inject into that design all of the hindsight that develops in processing the product through the various stages of its manufacture.
In every other mass production industry, prototypes must be developed and proven and all experimental drawings carefully redrawn or "product engineered” on the sole basis of production requirements. The penalty of this violation of fundamental design practice in the aircraft industry is expressed in terms of thousands of drawing changes after production starts and the loss of thousands of dollars in reworked and scrapped tooling. Part of this is also due to the fact that many designers work purely from a functional standpoint with a total disregard of how the parts and assemblies may be most economically fabricated. Design of experimental and production type airplanes should be separated since design considerations governing the former, except for first order breakdown, rarely carry over into production, on military contracts at least.
Production design involves adhering strictly to established and proven principles of manufacture. In this respect it should be noted that the functional designer is prone to seize on nebulous or partially developed processes and incorporate them into a new design without sufficient facts to justify their use. Thus, a burden is placed on manufacturing personnel not only to develop a new product, but at the same time to develop new processes without a delay in production. Development and research are absolutely essential in a progressive company but such research must be proven before incorporation in design.
Functional design by groups, each intent on its particular portion of the product, also results in reduced efficiency as regards the application of standards for parts. tools, tolerances, and dimensions. Each designer feels that his way of designing the part is best and since there is no effort by him to re-engineer the part for production after it leaves his hands, the value of standards is totally lost since standards naturally evolve out of production design.
If a functional designer is responsible for the function of the part he designs, the structures engineer for its strength, the weight engineer for its weight, so should the production engineer be responsible for its practicability for manufacture. There are definite limits of accuracy obtainable using various manufacturing processes and these should no more be deviated from than should the structures engineer deviate from required bolt sizes in a structural joint.
Owing to the present day wartime schedules, it is realized that it is not possible to take the time to engineer in this manner and a definite plan of operation is required which will allow for short cuts and a minimum amount of re-design because of error or oversight.
First, let us analyze the factors a designer must consider when he is given a preliminary design and a set of customer specifications. These factors include aerodynamics, structure, weight, interchangeability, function, tooling, manufacturing, schedules, standards, limitations of equipment, materials, service and cost. Those factors relating to purely engineering considerations can be reconciled within that branch. The others require coordination with other branches and departments and, unless their requirements are considered at the earliest possible moment, the result will be confusion compounded.
All this leads to a subject which has been the cause for a considerable amount of controversy, namely, the most efficient breakdown of airplanes for production. The subject has not only been the topic of much general discussion but considerable differences of opinion may be found among tooling and manufacturing personnel who are working on one particular design. It should be obvious that first the breakdown will vary with the size of the craft and second with the given set of manufacturing conditions under which it will be built. Before this, however, it should be pointed out that there are several basic factors which affect any manufacturing consideration desired for the sole purpose of facilitating production and which must be thoroughly understood.
Customer requirements will dictate the type, range, speed and altitude of the plane to be built. Customer specifications will, to some extent, dictate the type of materials and processing to be used and the customers’ wishes as regards spares and interchangeability must be given consideration before manufacturing and tooling requirements can be studied.
After the basic design is determined, the establishment of the first order breakdown is then made on the basis of floor space, line flow, and handling and shipping requirements. Here, the first clash of philosophies occurs. Obviously, for structural and weight considerations, the Engineering Branch must be extremely reluctant to allow the type of joints which will simplify the mastering, mating, and locating problems of the tooling departments.
Speaking structurally, it is a fact that the simplest way to transmit a load is through one solid member of a given size to another of the same size with a minimum of joints and with these joints in the same plane as the load carrying numbers. If there are a multitude of joints, then proper load distribution presents difficult design problems. To stagger these joints will reduce weight and complexity but is impractical from a production standpoint. If a joint must be made, then one loaded in pure shear will be simplest.
For structural reasons, many small airplane fuselages split longitudinally since shear and torsion are the only main factors and the sections are satisfactory to spares and production. When a fuselage becomes too large for such breaks due to material and equipment limitations, it is necessary to break the fuselage across the barrel introducing discontinuity along lines of major stress and all the factors of bending, shear, torsion, and maneuvering conditions are introduced.
The conditions surrounding a highly loaded wing attachment joint are obviously more severe than any other since here a load must be fed to a point or relatively few points and then transfer the load again to a larger area due to contour, taper, and space limitations.
In a particular airplane undergoing major modifications, it was agreed to provide two transverse breaks at nose and aft fuselage to facilitate production. The weight increase for the nose joint was 4½ lb and for the aft fuselage joint 17 lb, a total of 21½ lb.
In another design, a much larger plane, it was agreed to provide a maximum of tension joints in the fuselage to facilitate manufacture instead of the conventional riveted stringer, bulkhead, and skin joints. This resulted in a total weight increase of 126 lb. Since only about 5% of the weight of an airplane is doing the work, the rest, as it has been put, only going along for the ride, such an increase in weight can well become critical as every added pound will make performance guarantees that much more difficult of attainment.
To take members such as main spar caps, webs, etc., and cut them into sections for ease of production is prohibitive from both a structures and weight standpoint. In addition, the splices, doublers, fittings and close fit holes necessary for such breaks also usually present enough assembly and coordination problems to make them undesirable from a fabrication viewpoint. Here the limitations should be one of material size, capacity of shipping facilities and clearances along major railroads.
In yet another medium bomber design of 32,000 lb weight empty, it was calculated that approximately 9000 lb of this weight was in the aluminum alloys used to fabricate the airframe; 7500 pounds or about 22% of the total weight went into the basic structure required by the stress analyses; 15,000 pounds or about 47% went into engines, instruments, accessories, landing gear and other items over which no weight control could be exercised, these latter two totaling to 69% of the entire weight. On this basis, it is estimated that weight control could be had on about 11% of the total, other factors controlling the balance. It may be seen that Production Design control can also only be exercised over a like percentage of the total weight and therefore the margin to work with is comparatively slim.
It may be well to explain what is meant by first, second, and third order breakdowns. [sic] one of size for handling, access and type of construction. First, assemblies requiring dissimilar construction methods should be kept separate so as not to confuse the structures engineers and to permit assembly of details within departments; second, breaking primary members must be avoided such as main beam caps, webs, etc, to avoid doublers, clips, close reamed holes and drive fits and, third, items such as structural panels with bolts in shear required to be frequently removed in service should also be avoided since the holes in such members will gradually loosen thus weakening the surrounding structure.
The design should provide that panels and sections can have a maximum of plumbing, wiring, soundproofing and miscellaneous parts and items of equipment installed on them prior to the assembly into a major component. Probably the greatest man-hour cost is that incurred in assembly of such items in shells or between shells where employees must climb over one another and work in cramped, awkward positions.
For the past year or two, Tool and Production Engineers have been mainly concerned with the production of aircraft designed at a time when production schedules were of little importance and designs were based on functional requirements alone. That there has been a definite increase in the amount of consideration given production on new design, speaks well of the cooperative attitudes in effect between these men and Engineering Branches. This can best be brought out by illustrations. Fig 1 illustrates a design group breakdown which provides such attention to the detail design that the Engineering drawings reflect to the fullest, the actual assembly techniques of the Manufacturing Branch. This, it is believed, was the first time an Engineering Branch had agreed to "Product Engineer" an airplane and, while imminently successful as far as fulfilling its intent, has had to be dropped on subsequent projects due to the critical shortage of Engineering time; that is, beyond establishing the first and second order breakdowns.
Originally (as shown in Fig 2), the design groups on the projects were divided functionally as shown with little or no attention paid to the need for production assemblies incorporating manufacturing wishes as regards second and third order breakdown,
Fig 3 illustrates the type of assembly methods made necessary due to a disregard of efficient production methods. Little planning to most efficiently use labor and production processes can be used and assembly costs are high. This is graphically illustrated by Fig 4 which shows the huge load placed on the final assembly departments. An attempt at a production breakdown is illustrated in Fig 5 yet the mating difficulties due to design are obvious.
Fig 6 illustrates a refinement of the original plane brought about through the design of tension connections at nose and tail cone at a saving of 15 man-hr of manufacturing time. A further refinement was made whereby tension connections at nose, aft fuselage and wing were incorporated along with other changes for production, and Fig 7 shows how assembly sequences may be planned when design is broken down with production in mind. Fig 8 shows an original tail surface design in which the stabilizer is built in three parts and mounted to contoured formers by a total of 139 shear bolts. Protruding stabilizer tips, while not shown, are also part of this design. The re-design produced a straight leading edge on stabilizer, eliminated stabilizer tips and the attachment of stabilizer to aft fuselage was made through four bathtub fittings and tension bolts in comparatively loose holes. The redesign resulted in a saving of approximately 25 man-hr of assembly time.
Fig. 9 should give some impression of the cost entailed in weight by breaking of primary structural members such as main beam components. The cost of tooling for full coordination and the assembly problems of close reamed holes and drive fit bolts brought about by such breakdowns must. be weighed carefully. If shipping and factory space permits, the full center section of Fig 10 considerably reduces the aforementioned costs and is also satisfactory to the customer. The weight saving on a given airplane of a solid 30-ft center section over one broken at the fuselage sides is 162 lb, all in complicated fittings, doublers, and splices. The number of nacelles in the section only enters the picture on the basis of allotted floor space per man and time in the jig and, if designed for a minimum of attaching points, should provide for a maximum of work prior to installation on center section.
As stated before, the Engineering Branch cannot be content with merely providing first order breakdowns of structure, such as shown in Fig 11, nor are such assemblies the most economical from a labor hour viewpoint (Fig 12). Whenever possible, a second and third order breakdown must follow to permit best arrangement of production flow and best utilization of working area. Skin sizes and stringer, bulkhead arrangements, etc, should permit the panel assembly sequence (shown in Fig 4) to be utilized to its fullest extent followed by such arrangement of equipment, wiring and plumbing as will permit fullest installation of such items prior to joining of the major assemblies.
One type of breakdown is that which routes the control system, plumbing and wiring around the structure and in and out of small holes in that structure so that any saving in sub-assembly and fabrication of detail parts is lost in final assembly and schedules upset. If the structure can be tooled in sections and panels even with closely coordinated lap and shear joints made necessary by more direct routing of these systems, this should be done.
Last year, Lockheed Tool Engineers suggested that more consideration be given to the type of design depicted in Fig 13. The structural problems of discontinuity, torsion, and local transfer are somewhat difficult. However, considerable testing has been going forward on this type of structure since the manufacturing advantages can be so readily seen.
The progress that has been made in this direction can be noted in Figs 14 and 15. The first depicts a tail cone assembly designed in the conventional manner with use made of the stamped inner frame technique on the non-structural doors only. A redesign produced the bulkhead and side panel design shown in Fig 15 and saved man-hr in both fabrication and assembly departments.
None of the foregoing simplifies the design of major jigs. Locations and tolerances are just as critical and require as accurate design. The number of jig designs and the coordination problem are less involved but other problems will take their place.
Basic dimensions are needed to definitely establish the points of major interchangeability, such as between wing and fuselage and stabilizer and rudder. These same points are the ones to be used as locating points in the various jigs and fixtures in which those components are built.
It is at this time that the designer first runs into the bugaboo of tolerances; tolerances to meet the functional needs of the design, tolerances to allow for production and tolerances to permit interchangeability. The designer’s ingenuity is taxed to arrive at a compromise which will satisfy all three conditions. He must understand that the successful tying together of the structure depends on the first dimensions he lays down.
In building jigs for the first order breakdown, Tool Engineering works from close tolerance machined parts for basic location and dimensions from zero butt, water and station lines. Engineering must establish basic dimensions from the same points which are liberal enough to provide ease of mating and arranged so that the inherent warpage in comparatively flimsy sheet metal sections will not cause trimming, filing, or reaming on assembly. While it is sometimes necessary to hold attaching fittings within tolerances of a few thousandths, overall tolerances on first order breakdown should not be less than 1/8", second order 1/16" and third order 1/ 32", and if proper mating is to be obtained these tolerances must be made plus on one part and minus on the other to provide proper clearances. Designers must dimension from fitting and hinge center lines and, in case of multiple joint attachments, the number of tight fits must be held to a minimum while the balance receives sufficient tolerance to permit mating without too many final assembly operations.
In cases where contours change, such as in a fuselage section or where a wing sweeps out, station, butt and water lines must be accurately located and dimensioned at those points for tooling purposes.
When hinge point center-lines on elevators and ailerons are held close, sufficient clearance must be allowed between the adjoining sheet metal surfaces to compensate for the larger tolerance with which this type of part is manufactured.
Lap joints must be such as will permit of firm location of attach members, offsets and breaks of the joints due to incorporation of equipment or doors across the joint avoided. It must be remembered too that full size rivet holes are only practicable in small detail assemblies and that where single planes are available, that coordinated holes are possible and reduce time in assembly jigs.
Serrated attachments and adjustable connections will, in certain cases, eliminate some tolerance and assembly problems. Where groups of holes in mating members are concerned, tolerances on individual bolts and holes are not as critical as on the assembly as a whole.
Aerodynamic considerations are important today on high performance aircraft but designers are apt to mistake extremely close tolerance for the uniformity and smoothness of contour and surface which this department requires.
If the main fuselage section of a bomber were to be built as one unit with all stringers, bulkheads, and other details put into place in one jig, it is obvious, to secure some resemblance of uniformity, and to facilitate the later typing in of all the internal parts, such as equipment, controls, plumbing, etc, that a great multitude of locating points are required. To fix these numerous points in space by transit level and square is a job running into thousands of hours; and when complete, the flexibility of the jig under load at these points makes the job of holding tolerances almost an impossibility. Thus, a master tool is born, a complicated structure whose locating points can be more readily established since it is a skeleton of the assembly itself built to toolmakers’ tolerances. This is set up and the jig built around it, and the jig may be duplicated as many times as necessary from the master.
If a breakdown requires that a multitude of parts be joined in one jig, the time in that jig will usually be greater than that for other components permitting of greater breakdown. To avoid a bottleneck, duplicate jigs are then considered necessary, all of which perform the same operations and which must be made from the same master to assure mating and interchangeability. The tooling cost is materially increased and accessibility is still at a premium. A breakdown permitting full hole coordination, pilot drilling of mating parts, joining in single planes or the use of mechanical fastening means on panels and sections spreads the operations over a number of smaller fixtures and with the master simplified to locate only a minimum number of joints. Conversely, where the structure does not level itself to this breakdown, then the minimum points should still be held in the first jig and these attaching points moved to pickup jigs, and complicated masters are again avoided.
Designers must realize that the human element in assembly is unpredictable as to output unless the job performed is one of operation, the process being automatic. In other words, a fully automatic machine will produce so many parts an hour, a semi-automatic machine will produce so many parts an hour if the operator throws the switch, pushes the button and trips the treadle the maximum number of times the machine can repeat in an hour while the hand-operated tool will only work as the operator.
Since assembly operations are all fastening operations, then the breakdown must provide for mechanical fastening to be efficient. Erco or squeeze riveting or spotwelding in place of gun riveting or a multitude of screws; flashwelding or furnace brazing in lieu of arc welding; rolling or die operations instead of braking. The counter to determine machine and operator efficiency has yet to make its appearance in the aircraft industry.
Through the use of cost estimates, it can be readily determined whether a forging, casting or pressing shall be used and if assembly by flashwelding, projection welding, furnace brazing, or other production processes can be justified. Cost analysis of alternate designs is becoming more and more important and where such alternates present themselves, the designer has no excuse for selecting the most expensive method. Cost control will also serve to prove whether or not proposed redesigns after production starts will justify reworking.
It is not to be expected that the original release for pilot line production, regardless of the care used by the designer, will be 100% perfect; also, due to time, the first release is often affected by delivery dates of component parts. It is therefore necessary that during this release and while tooling is planning for production, that a production design study be made to firmly establish the changes desirable for production and provision must be made in the Engineering schedule for this redesign.
This article was originally published in the March, 1945, issue of Industrial Aviation magazine, vol 2, no 3, pp 57-58, 60, 62-63, 100-101.
The original article includes a thumbnail portrait of the author and 7 photos, and 8 detail drawings and diagrams; all are presented as Figures.
Photos are not credited, but are certainly from Lockheed.
A PDF of this article is available.