“What has been is what will be, and what has been done is what will be done, and there is nothing new under the sun.” -Koheleth
Now that the latest, greatest, most expensive, fighter jet is coming online the critics are piling on. It’s too slow! It can’t fire its guns! It can’t dog fight! The software does not work! Unlike most of these critics I am rather hopeful that the F-35 Joint Strike fighter will eventually be alright. It is truly a triumph of engineering in the face of overwhelming bureaucracy. Sadly, the cost of overcoming bureaucracy is far greater than that of overcoming design challenges. The Pentagon’s procurement system is seriously broken. It needs to be fixed now because in the near future the Pentagon may have to respond to rapidly changing global threats. New weapons platforms and technology will have to be rolled out with astounding speed to respond to innovations in war fighting. Far better to fix development processes now than during an emergency.
Despite my twenty years in the world of IT security I still maintain an interest in engineering, product design, and aerospace systems. I learned fight mechanics from an Admiral in the South Vietnam Air Force. I learned rocket design from professors who worked on the Trident I submarine launched ballistic missile (SLBM). But most importantly, I learned how to engineer products with tight constraints for cost, weight, and timing, as an engineer in the automotive world.
As the first degreed engineer working for Hoover Universal, before it became Johnson Control’s Automotive Systems Group, I designed car seats for GM, Saturn, Ford, Chrysler, and Jeep. You can look up my patents on seat frame structures.
There are three rules for effective engineering product development: prototype , prototype, prototype. Build the first one out of paper. Build the next one in the shop out of materials on hand. Make temporary tools for the next prototype. Discover every problem you can, get as complete a representation of the final product as possible, before committing to production.
My job in 1982 involved running between my computer simulations, the drafting board, and various prototype shops in the Metro Detroit Area. I could get anything built in Detroit back then. Practically every machine or tool shop I visited was working on the Space Shuttle or B1-B bomber at the time. I would use the same acid baths to give me multiple prototypes of different gage metal as were used to create optimally milled manifolds for the B1-B engine inlets. One shop formed the injector plate for the Shuttle main engines out of a single forged billet of Inconel 718. It was electro-chemically etched until there were 300 injectors sticking out of a four inch plate of solid nickel-chromium alloy. Each injector plate cost $2 million.
These shops were founded by what we used to call business owners, today we call them entrepreneurs. They would get their start making early prototypes and their business would thrive as space and weapons systems ramped up for production. Sadly they are all gone, along with most of Detroit’s tool and fabrication plants, victims of off-shoring. the end of the Cold War, and the hiatus in manned space flight.
While I witnessed first hand the sprawling supply chain for airframes, fuel pumps, and electronics for government programs, I worked on some of the biggest automotive manufacturing projects, including the launch of a new car company, Saturn. I even participated in the Manhattan Engine Project, a secret effort on the part of Pontiac and Chevrolet to develop the next generation four-cylinder engine.
But the program I contributed the most to was the re-design of the full size Chevrolet sedan, dubbed GM200. You may remember how GM turned the proud square lines of the Caprice Classic into the jelly bean shape their market research predicted would appeal to their buyers.
I was brought on to the project to prepare prototypes, first for crash testing, and then for full ride-and-handling testing. Crash-test prototypes were easy; bring in a fleet of 22 brand new Caprice Classics, carve some holes in the frames to improve crushability and add some beams to the upper fenders to reduce pitch during impact to avoid throwing the occupants up against the roof. Then I would ship them off to Milford, Michigan, where they would be instrumented with accelerometers and high speed film cameras. Then we destroyed them by hurling them into a solid concrete and steel barrier at 30 mph.
The real challenge came when I was tasked with creating full scale prototypes of the new design. I needed stampings for 250 different sheet metal parts. I needed to assemble them into a body, mount it on a modified chassis, and prepare the interior for road testing.
At the far end of the GM200 project’s main design room was a glass wall behind which was a full size mock-up of the new vehicle. Every sheet metal shape was created in clear plastic and the vehicle was assembled with plastic fasteners. Design teams would meet around the mock-up and work out real estate compromises inside the engine compartment. One day I asked someone how those plastic parts were made. They were vacuum formed. In other words, somewhere in the building was a set of wooden forms carefully machined from the CAD models. Warmed sheets of plastic would be draped over each form and the air would be sucked out through tiny holes to form the plastic into an exact shape.
I found those forms and arranged to have them moved to GM’s Kirksite foundry where a small team of artisans would take impressions of each form. Using precise thicknesses of wax to represent the final sheet metal, they would create large blocks of plaster with the shape of the male and female dies required to stamp the needed shape. Using green sand they rammed these models into copes and flasks and poured molten Kirksite in the cavities they left in the sand. The final product would be a matched set of solid Zinc alloy that when put in a press could hammer out up to 2,000 components before needing to be re-worked.
I would have these tools moved next door to the sheet metal shop and request two finished stampings for each of the 250 different parts I needed. When they were done with the tools they recycled the Kirksite.
How did I get hundreds of skilled craftsman at GM’s Tech Center to work for me? I was a lone wolf that hacked the system. GM had a staff of prototype buyers that were in the process of trying to starve their internal resources by giving all the prototype business to outside shops. The story from the union guys was that the outside shops provided fine dinners, and possibly “gifts,” as they vied for the work. The internal workers loved their craft and loved the fact that this crazy 28 yr old engineer in a suit was flooding them with work. I just signed work orders with a charge code that had been assigned to my project. I was done and gone before anyone in accounting ever thought to ask what I was doing.
In a union shop engineers are not allowed to pick up a tool or handle any material. In order to get a finished part from the sheet metal shop to my final assembly buck I was supposed to fill out a “move order.” If I was lucky someone would pick my part up, strapped to a wooden pallet, and take it to the shipping dock the same day I requested the move. At the shipping dock I had to fill out a “shipping order” to have the part loaded on a truck and driven a quarter mile to the next building over. When it arrived I would meet it and write another move order to have it taken to my team for spot welding into the vehicle.
My team, all union guys, gave me advice on how to beat the system. Pick up each part from the sheet metal shop and hand carry it between buildings. When I passed the shop steward’s office walk real slow, pretend I was looking for someone to help me. Don’t rush, don’t appear to be on a mission. Amble. It may have helped that I was limping from a swollen knee.
I carried two sedans and a full size station wagon from one building to another that summer.
After I had developed my prototype technique for the sedans I was asked to produce the station wagon. Some of the designs for the parts had not been completed yet. I took the drawings to an outside shop that had experience making cardboard mockups for architectural models. I had them make models of the dies out of cardboard and took those to the foundry. By this time I had the system down. I assembled a complete station wagon, shipped it to Milford (where test engineers were allowed to use tools) and mounted the body on a complete chassis. I transferred all the interior trim from an existing vehicle to my prototype and handed a complete vehicle off to the test engineers. Total time from start to finish? Four weeks.
During this time I became fascinated with the idea of rapid prototyping and how it could be used to shorten development times. One day I read about the Chevy II, a vehicle that General Motors built in 18 months from concept to production. The project I was working on was supposed to be groundbreaking: resigning a vehicle in 2.5 years. The first year of production of the Chevy II was 1.6 million vehicles in six configurations: coupe, sedan, station wagon, soft and hard top convertible, and differing seating options. The most popular vehicles of today barely exceed one million in production volumes.
While I wandered around the GM Tech Center in Warren, Michigan, I asked every old guy I met if they remembered the Chevy II. One of the team helping me on my builds remembered coming to work on his first day in 1960. He said “I thought I had come to Ford’s by mistake” there were so many Ford Falcons in the hallways and on the platforms. An older engineer said “Oh that? We cheated. We copied the Ford Falcon bolt for bolt.”
When I posed the question: Why is our program taking 2.5 years when the Chevy II only took 1.5? I got the answer “Cars are more complicated today with all these regulations for safety and fuel economy.” I did not buy it. We had CAD/CAM, Finite Element Analysis (my speciality), and simulators. We should be able to beat the old way.
Every new car program I was on started with a tear down of the competitive vehicles in the market. GM had a huge warehouse to hold what they called the Mona-Lisa project. At one side of the cavernous space was a line of complete competitive vehicles from Honda, Toyota, and Volkswagen. Stretching away from each car was 50 yards of folding tables. Every component, tagged with its weight, was laid out for inspection. An engineer tasked with designing a new strut, control arm, or engine mount, would spend a day at Mona Lisa taking notes and armed with calipers and micrometers. (I went to the Henry Ford Museum to go back farther in time so I could see the evolution of each component.)
Every new car, engine, suspension system, seat, or frame, is an evolution from an earlier day. There are no pure clean-sheet-of-paper projects. The reason for this is that no engineer can take all the variables into account. An existing component has passed all the federal mandated tests, survived 100K cycles of fatigue testing, has been through the requisite tweaks for manufacturability, and most importantly, has survived years of real world testing in every environment. Starting with an existing design and changing one or two variables, like material, or gage, while adding the latest feature, is the fastest way to create a new design. Otherwise the product gets shipped and as problems are discovered, very expensive recalls and retooling ensues.
The biggest problem with Pentagon procurement is that every major project is a clean-sheet-of-paper. Designs are not allowed to evolve. Mission profiles and specifications are determined 20 years before the first system is fully operational, and those systems are expected to meet all requirements for the ensuing decades before the whole exercise is started over.
Long design and development time scales lead to costly changes as requirements are modified to meet current demand. Changes further delay production. Invariably a new weapons platform does not meet the current needs of our military.
In Part II: How the DoD has failed in major weapons platform designs. We will look at the Littoral Combat Ship and lessons learned from WWII construction of Liberty ships. Then we will look at what is proving to be the most expensive weapon system in history, the F-35 Joint Strike Fighter, and compare it to some amazing aviation successes from Lockheed Martin’s famous Skunk Works.
