With a Little Help
Autobiography of Rick Wagner

Chapter 6: TRW

For more information on the company that hired me out of engineering school, see the Wikipedia article on TRW.

Starting a Career

"TRW" stands for Thompson, Ramo, and Wooldridge. The Ramo Wooldridge Corporation helped the Air Force get its ICBM program going in the 1950s and then produced spacecraft as a pioneer of spaceflight. The Thompson Products Corporation started out making capscrews in the early 20th century and then moved into engine valves and other products. Their sodium cooled valve for the Merlin engine helped make the air superiority fighter, the P51 Mustang, possible. Thompson Products bought Ramo Woolridge and named the resulting company TRW, Inc. TRW's Space Division was in Redondo Beach, California. Dr. Simon Ramo is the R in TRW and I read his book on tennis (Extraordinary Tennis for Ordinary People) in the late '80s. More recently, I read his 1995 book Meetings, Meetings, Meetings; Getting Things Done When People Are Involved. Both books influenced me somewhat.

I was recruited to TRW from the University of Hawaii by John Otera of spacecraft integration and test (I&T). I filled out an application form and after interviewing me at UH, TRW flew me to the mainland for interviews. So I had a solid job offer before I graduated in December of 1979. I also had job offers from General Dynamics and the Jet Propulsion Laboratory, but TRW offered the highest starting salary, so I accepted their offer. It was four times what I had been making as a Staff Sergeant in the Air Force.

I flew to the mainland in early January 1980, started work on January 7th, and TRW put me up at Barnaby's Pen and Quill hotel on the Pacific Coast Highway (PCH) and Rosecrans Ave. It took me over a month to find an apartment to share in Hermosa Beach, and TRW paid the entire hotel and resteraunt bill. I bought an old Buick station wagon from a TRW employee for two hundred dollars. I rented a room in Hermosa Beach. Andrea came and visited me for a few days and we had a good time together. I volunteered at the South Bay Free Clinic on Manhattan Beach Avenue for a few months in the clinical laboratory. I learned how to do urine tests including microscopic analysis and blood tests including drawing blood. In the springtime Andrea came to live with me and we moved to the lower rooms of a house on Palm Drive, just off the Strand in Hermosa Beach. Robert Wilson, a psychologist, owned the place and lived intermittently upstairs with his wife Lynne, our landlady.


February 1980. Andrea visited me after flying to Dallas for a fashion show. She was working for the Fashion Guild in Hawaii.

I learned how to be a test conductor from Mike Miyamoto. He mentored me in a DSCS II counterweight deployment test. We did cold torque-angle tests and a latch-up test at cold temperature. I was slightly disturbed and disappointed when Mike insisted we throw out some data we didn't understand. To my way of thinking, test data that's not understood should be investigated, not discarded. This was a flight hardware acceptance test, and I had learned in school that it's not good practice to selectively ignore data that is otherwise valid.


DSCSS II, TRW artist rendering.

The DSCSS II counterweight deployment illustrates an interesting bit of satellite technology history. Early communications satellites were of the "spinner" variety. That is, they were spin stabilized, with a de-spun platform for pointing high gain (parablolic reflector) antennas at the earth. If this sounds like a kludge to you, you're right. The solar panels for powering the spaceceaft are on the spinning cylindrical section, so half the solar cells are always in darkness too. Our cross-town spacecraft rival, Hughes Aircraft, continued to build spinners long after the DSCSS II program was discontinued at TRW. The first launch of DSCSS was unsuccessful. The high gain antennas were folded down on the despun plaftorm for launch, and popped up like mouse ears for orbital configuration. That change in configuration caused the spacecraft to wobbble out of control. The fix was to add a counterweight on a boom that offset the center-of-gravity change, another kludge!

So the solution to the spinner problems was to invent the three-axis-stabilized spacecraft, a TRW innovation. The spacecraft was statilized by internal reaction wheels or control moment gyroscopes (CMGs). A hydrazine (monopropellant) reaction control system (RCS) was used to unload the reaction wheels or CMGs from time to time. Stabiizing the spacecraft this way allows solar panels to be pointed continuously toward the sun for the highest power efficiency. The world's first three-axis-stabilized communication satellite was the Fleet Satellite Communications system spacecraft (FLTSATCOM) developed for the United States Navy, which I worked on next. By the way, a FLTSATCOM was the satellite we communicated with when I worked at the Air Force ground station in Wahiawa. We were originally planning to communicate via the Air Force TACSAT, which failed on orbit (it was as spinner and the despun platform bearing locked up, so the whole thing spun), so we borrowed some comm channels from the Navy. FLTSATCOM had been flying for several years when I joined TRW. The ones I worked on were flights 5, 6, 7, and 8.


FLTSATCOM, TRW artist rendering.

FLTSATCOM also had its development problems. It was originally designed for a single high gain antenna for both receive and transmit. Problems with intermodulation distortion led to the addition of a separate receive helical antenna on a separate boom. In testing that boom, I also learned the test conductor game from Max Barham, an old experienced mechanic in space testing who used to be an automobile mechanic at the Cadillac dealer. Max and I did some FLTSATCOM boom proof load acceptance tests at cold temperature. I remember talking to Andrea in Hawaii on the phone while that test was going on. I told her it would be good if she could come over and stay with me.


Andrea came to stay with me on 29th Street in Hermosa Beach before we moved to Palm Drive at Longfellow.

I also spent time in an office with other test engineers planning tests and producing test proposals and procedures. One test I conceptualized and cost estimated was for a docking simulation using a momentum simulator. I designed a simulator using double channels and crane proof weights. It would have been huge, about 60 feet across and suspended from a crane. When the requestor saw the cost they decided to pass up that test. It would have simulated the docking inertia of the space shuttle to the international space station (neither one existing yet). I used hand calculations to simulate the rotational inertia in two axes simultaneously.


Andrea doing some vacuuming at the ground floor flat on Palm Drive in north Hermosa Beach.

I did proof testing on some D-rings for hoisting an Air Force telescope at Mauna Kea. The telescope is used in their space tracking data collection. I had to design a special attachment fixture for pull testing the D-rings in a tensile test machine. It was the first thing I designed at TRW. I talked with the machinist about the drawing and he gave me some tips. I still have that chunk of iron. I got to keep it after the test was over. Otherwise, it would just have been scrapped.


Rick, Robert, friend Peggy, and Andrea at Palm Drive apartment. It was the lower floor of a house across from Bocatto's grocery.

I met Bill Hoffman in the office room I shared with several others. He was running calculations on a programmable calculator to show response and stability of a hydraulic device for a non-electronic active zero gravity simulation machine. It was a somewhat successful IRAD (internal research and development) project that I would take over in its assembly stage. I named it the AZGSS (active zero gravity simulation system). It was an interesting concept but had no real advantages over passive systems, so it was never used in production.


My test crew for the DSP (defense support program, classified at the time) solar paddle development test.
Solar paddles deployed, with test crew, September, 1980. Bill Murray was the measurements technician. Fred Schmidt
was the lead mechanical technician. Jim Dollar and Joe Scardina were mechanical techs. Pete Dejesus was the ordnance
technician, and Kem Rosenberger was my assistant test engineer. We also measured shock propagation from the ordnance
pin puller panel release.

I have a page where you can see some of the unclassified programs I worked on at TRW, now called Northrop Grumman.

Torque-Angle Testing

Bill Hoffman had a sign in his office that read "one test is worth a thousand expert opinions." He was a very sharp engineer. Bill suggested that I use a torque sensor for measuring the torque characteristics of the spring driven hinges for the DSP solar paddle test. Torque sensors were relatively new, and available in the sensitivity ranges we needed. This was an industry first, using a torque sensor for hinge testing. The test was a big success and set the standard for all future torque-angle testing. For a larger hinge, no sensor was available (500 inch pound range), so I researched them and designed my own. It was machined by Mal Stiles and strain gaged by Chuck Wright's strain gage lab. Its response turned out to be within one percent of my prediction.


Andrea took me to visit her Uncle Jack at their beach house at Crystal Cove in Orange County. October, 1980.

Multi-Dimensional Stiffness Testing

I was assigned to do a six-dimensional stiffness test on a prototype stepping motor actuator for moving a payload arm in space. I decided to use micro-LVDTs (linear variable differential transformers) connected to a computerized data acquistion system. This was another first for the industry, leading to fantastic productivity gains and high quality of data interpretation. We acquired eight channels of data and processed them offline to show three dimensions of rotation and three dimensions of translation of the point of interest on the actuator, as functions of applied loads. I programmed the data reduction routine to give graphical output and compute best fit linear functions as characteristics of the actuator. This was the first use of automated data acquisition for stiffness testing.

I was programming the MINC-11 (modular instrumentation numerical computer) with dual 8-inch floppy disks, an RT-11 FB (foreground-background memory swapping) from the Digital Equipment Corporation (DEC), the world's foremost maker of mini-computers. Taylor Nielsen, a physics major working in the measurements laboratory, showed me how to compile FORTRAN programs on the MINC-11. I had learned FORTRAN at UH and had taken a numerical methods course from Professor Kihara, who later became ME department head at UH. Taylor also showed me how to run programs on the DEC VAX 11-780. I found I was able to beat a chess program running on that timeshare system.

Fred Schmidt was a senior technician from whom I learned a lot. He was assigned to support me in several tests, including the stiffness testing. He was also an accomplished racer of model boats.


I had devised a new stiffness test technique using LVDTs and an automated data acquisition system to fill in a
six by six flexibility matrix. This is the brassboard A actuator test for a classified program. Fred Schmidt
suggested I use the two inch thick steel plate as a frame of motion reference. It was a good idea. Joe Scardina was
a machinist who became a test technician. He machined and assembled the aluminum fixturing shown around the actuator.

The MINC-11 had 64k of memory. Not very much. Large FORTRAN programs simply wouldn't compile, and I hit the ceiling fairly rapidly. So I taught myself BASIC and started programming in that interpreted (not compiled) language that had a built-in graphics package.

For the stiffness testing setup, I had invented a cable and pulley windup spring-controlled load system for applying continuously contolled moment or shear loads. That is, by controlling a gearmotor, I could make a test run with load varying continuously from a negative load up through zero to a positive load, and then continuosly down again through zero to a negative load, sampling the eight deflections in micro-inches every second. I could record (and later graph) full hysteresis loops simultaneously in six dimensions, something that had never been done before. Word got out and soon there was a parade of senior engineers and managers through the test lab.

The stiffness matrices I developed with that technique were useful for dynamic models of deployed spacecraft "modes" of vibration. An added bonus of the full hysteresis loops was that the area inside the closed loop represented energy leaving the system as damping. Hence, never seen before models of system modes were possible.


Dinner at Crystal Cove with Andrea, Don, Lenore, and uncle Jack.

Spacecraft Static Load Testing

My security clearance came through less than a year after I began at TRW. I had held a DOD Secret clearance in the Air Force and I had been put in for a special access clearance. If you don't know what that is, I can't tell you. But I was able to go into a building where I could see some of the things I was working on put together into a system. I was assigned to do a static load test on a spacecraft structural model and designed a 40 foot high tower to apply the hydraulic loads. An instrumentation tower was independent from the load tower. Mike Miyamoto was the test conductor and I was his assistant. I designed a screw jack system to put deflections into one end of the test article. The test was the largest of its kind ever conducted and was a very big success.


My friend Dave with my new prototype boogie board. Dave graduated in EE from UH with me and was recruited at the same time
by TRW. He worked on TDRSS (tracking data relay satellite system). Photo taken in the garage at our place on Palm Drive.

Mike had done static testing before, and we had a bunch of Edison unit static load maintainers. He had experience with them, but he seemed unaware that they had a limited rate of load relaxation. We set up something like 40 channels of hydraulic actuator loads (forces), with measured reactions at the screw jacks. We had about 60 channels of deflections, with some ranging 20 inches using string pull LVDTs. During the test, we were simulating a landing condition in the Space Shuttle, so we had to deflect one place with the screw jacks. We had applied the loads and deflections gradually in ten percent increments up to 100 percent. At high loads, something wasn't looking right, so Mike ordered a rapid decrease in load. I was operating the two screw jack motors. I could see that if I reduced deflection rapidly, the loads on several hydraulic actuators shot up rapidly, threatening to break the structure. Therefore I refused to go at the rate Mike ordered. He became angry and still I refused to break the spacecraft. He still believes to this day that I was insubordinate, but I saved his ass because he would have been fired if I had obeyed him. Needless to say, I got bad performance reviews from him, and that set back my career by a couple of years until I got out from under him.

I was interested in validity of test results. For example, the motions of the spacecraft in a static test were measured relative to a tower built separately from the loading structure. But they were connected by the six foot thick concrete "hard floor" (it had embedded steel rails). The assumption was that the floor was completely motionless, so the tower would not sway when loads were applied to the spacecraft and reacted through the floor. So I installed a pair of LVDTs on eadh end of a steel beam attached to the floor only at its center. The result was that we could measure floor motions during the test, but we were able to conclude that they were small enough that measurement tower motion was insignificant compared to the spacecraft motion measurements. That was good to know, and it was included in the classified test report.

A few years later, for a different test for air bearings, we were leveling eight foot long, foot thick, "microflat" slabs of black granite. The slabs were supported by three jacks each for leveling in two directions, and they were supported on the same six foot thick hard floor. I found that, with the help of some good measurements technicians like Tony Demarchi, I was able to measure the tilt of the slab as I walked from one end of it to the other. As Bill Hoffman used to say, "everything deflects." The granite slab weighed about a ton and a half.


Our first Christmas tree in December 1980. I put her engagement ring under the tree and she almost threw it away by mistake.

Boogie Boards

I started building wooden boogie boards with foam cores. Andrea and I spent time at the beach testing them too. My first board was called Panda I, after Andrea's nickname. We used to roller skate on the Strand, a concrete walkway about 12 feet wide running all along the beach cities, right next to the sand.


Beautiful beach bunny Andrea at the beach.

I proposed to Andrea in December, 1980, and we were married June 13, 1981, in a ceremony in her parents' back yard by my step father, Prof. Rev. Mitsuo Aoki.

Vince Vicini had moved into the room I had occupied on Coolidge Street in Moiliili. I asked him to be my best man and he agreed.


Top: Lenore Johnson, Andrea, Rick, Christina Wagner (hidden), Janice Biles, Vince Vicini (best man).
Bottom: Bill Johnson, Dr. Reverend Mits Aoki, Prof. Don Johnson.

Mits, my stepfather, was an ordained minister, doctor of theology, and founder and head of the Religion department at the University of Hawaii. After the vows Mits led the crowd in a three banzai cheer, a Japanese celebratory tradition.


The three banzai cheer led by Mits Aoki.

If you're interested, here's a page with all of our wedding photos.


In front of the Palm Drive house with Andrea, Pollywog I (radio controlled pontoon hydroplane boat), and Donna..

Having enjoyed water skiing when I was younger, I had often speculated that it would be fun to have a boat that one could drive like a car on top of the water. That is, it would have four pontoons that would rotate like skis to adjust their planing angle of attack to the speed of the boat. It would handle just like a sports car and corner flat. So I built one as a scale model for radio control. It worked just as planned, hence the smug look in the photo above. The model fuel outboard motor was fixed in the chassis and steering was affected with a front rudder. Worked like a charm. I chased ducks all over the pond at Alondra Park in Torrance.


Andrea's older brother Bill lived in Palmdale and visited frequently. We lived right off the beach (just across Palm Drive).


Friends Kimo and Dave with Andrea at Palm Drive. Andrea made the curtains.


Having lunch with my brother Walter at an old hotel in Santa Barbara.


Andrea made me a birthday pie for my birthday, August 11, 1981.

We moved to an apartment on Manhattan Beach Blvd. in Manhattan Beach, about four blocks from the beach. We hated to move from Palm Drive but Lynn wanted the downstairs rooms back. It was small and inconvenient, but I still worked on my boogie boards in the bedroom doing the sanding downstairs in the carport. Andrea's brother Bill gave us his green Chevy station wagon because he would be leaving for Saudi Arabia and an engineereing job with Parsons.


On vacation on Kauai in the fall of 1981.


Walter with the boys at Paradise Park way up in Manoa Valley.

House on 10th Street

We moved to a rented house on 10th Street in Manhattan Beach. Every time we moved it was further from the beach!

In 1982 I performed a latch repeatability test on the Main Pivot Hinge. It was a large titanium weldment construction that had a very high preload latch for high stiffness. We wanted to see how much scatter there would be in two pointing angles over a sequence of latchings. For this test I chose to use an autocollimator, a very accurate optical device that looks at a mirror on the test article. This was the first use of an autocollimator for this kind of testing, with very high accuracy.

In troubleshooting some anomalous results in the data with the autocollimator, we discovered something useful. It was the practice then to sometimes use transfer adhesive (like a double sided sticky tape) for fastening first surface mirrors to spacecraft and components for alignment measurements with theodolites. The autocollimator was a specialized optical device with much more accuracy, so we noticed the mirror moving relative to the pivot hinge. It turned out that using that kind of adhesive wasn't a good idea. The motions went away when we bonded the mirrors with epoxy.

Also in 1982 I took the math intensive Dynamical Systems and their Applications, a TRW, Inc. after hours course. The instructor was a PhD physicist working on TRW laser systems. See my Curriculum Vitae for a complete list of after hours courses and certificates.


At my woodshop garage with the boys at the rented house on 10th Street in Manhattan Beach.
The house was just a few blocks from work at TRW and one block from public tennis courts.

In the spring of 1983 I quit smoking cold turkey. I got up one morning, got dressed for work, picked up my packet of cigarettes to put in my shirt pocket, looked at them again, threw them in the waste basket, and never smoked again. The next month Andrea became pregnant with Malia.


Robert's birthday party preparations at the house on 10th Street. Party guest, Tim, Robert, and Andrea. September 1982.

Tim and Robert came to live with us for a while. They shared a room in the two bedroom house on 10th Street. They got into arguments about getting into each other's things, so I put a piece of masking tape down the middle of the room with an access way from the door to Tim's side. Each boy had to stay on his side of the tape. It cut down on the fighting quite a bit.


Robert's birthday at our rented house on 10th Avenue in Manhattan Beach, September 1982.

In 1983 I was asked to do a first of its kind cold (-250 F) deployment test for a solar array deployment mechanism. There's no way to build a cold box high enough for zero-g cables, and it was felt that long slots in the box for them would be problematic. I chose to develop cold gas bearings for support from below inside the cold box. In developing the bearings, I used some two dimensional calculus in a polar coordinate system. I also built several prototypes and measured performance on granite tooling slabs. I designed the solar array test equipment and the test was a success. Test conducting was taken over by another engineer when I got pulled into a classified area for flight test conducting.

Our mechanical test department manager, Jim Rozelle, had invented torque-angle testing years before after a flight failure. An arm failed to deploy because of insufficient deployment torque througout the range of motion of the deployment hinge. So Jim figured out a way to measure and graph the torque as a function of deployment angle. If the deployment torque always exceeds the resistance torque, the hinge will fully deploy. Resistance can come from wire harnesses, latch friction, etc.

Torque-angle testing had become ubiquitous and routine at TRW, and was one of the reasons for the good performance record of our spacecraft. The amount of stored energy available to deploy the hinge is proportional to the area under the torque-angle curve. This energy can become important in dynamic stored energy deployments because kinetic energy in the deploying boom becomes the impact energy when the hinge latches up, possibly breaking the hinge. Using automated data acquisition techiques and post test data reduction, I was the first to measure this torque-angle energy in 1984.

House on 180th Street

In 1983 we bought a small house on 180th Street in North Torrance, just off Hawthorne Blvd., using a GI loan. Andrea and I looked at several houses in our price range. We liked the look and location of the one in Torrance, so in the evening we drove over there and sat in the car across the street and just watched and listened. All was quiet, so the next day we bought it for just over a hundred thousand dollars. It was good to finally own our own home.

In 1984 I was promoted to Staff Engineer (from Member of the Techical Staff) and was Manager of Test Operations, Mechanical Systems, project 5081, with responsibility for all the day-to-day spacecraft integration and test activities, including daily schedules and manpower assignments.

History of Mechanical Integration and Test at TRW

In 1984 I was asked to research and write a history of my department (Mechanical Integration and Test, or MIT), so I interviewed a number of old hands. Bill Rittenhouse had actually kept all the organization charts from the beginning. Later that year I was asked to write an assimilation manual to help new hires and their department sponsors (mentors) adjust to the work environment. The department history was included in the Assimilation Manual.

In 1984 I re-read George Orwell's 1984. Sometimes I would slip into sympathizing with the state, trying so hard to save Winston from himself. Andrea's brother Bill had married Lisa and moved to a house in Silver Lake, a suburb to the west of downtown LA but to the east of Hollywood. Andrea and I would visit them sometimes, and sometimes they would come to our house. Bill told me about a science fiction book he had read by William Gibson called Neuromancer. It was Gibson's first and it made him an instant star. I was intrigued by this new "punk" genre of science fiction with its dystopian future. I read all of his books. Gibson is still writing and don't miss one as they are published.


Taking Malia for a walk on the sidewalk in front of our house, 1984.

There was a pause in the classified program I had worked on in building M4. I was in charge of the mechanical integration and test procedures, so I made sure that they were backed up on the VAX computer system there. We knew it would be a couple of years until we were ready to begin work on the next flight so I verified specifically that the dozens of procedures were backed up on tape. They said not to worry, there was a tape backup. But I did worry, so while I still had access via a PC terminal, I copied the text files to five and a half inch floppy disks. I had a cardboard box about two feet long full of disks which I kept in the bottom drawer of my safe in my office. So the procedures existed in three places: on the VAX system hard drive, on the VAX tape backup, and on my floppy disks. I didn't think I would ever need them, but I was feeling more secure by having personal possession of them. I then worked for a while on other programs. When I came back to the classified program and started getting ready to begin integration and test, I was told that the VAX hard drive had crashed, but that they were loading the backkup tape. Then I was told that the "tape broke" and that my procedures were not recoverable. So I loaded my floppy disk files on to my PC hard drive and they were all intact. By keeping that floppy disk backup, I saved the company millions of dollars.

FLTSATCOM F8 Static Load Test

In 1985 I was assigned to perform the world's first static load test on flight hardware. Previously, a special structural model was used for static testing, and then flight hardware would be built after the design verification by test. In working toward higher efficiency, it was decided to test flight hardware to slightly lower loads so the test article could be flown afterward. I used all the experience from the classified program to do the unclassified test on FLTSATCOM flight 8 structure, with a new EHF hexagonal bay attached below the regular FLTSATCOM structure. In the original test, which I studied in preparation for this one, a hole had to be cut in a flight panel to put a shear load at the apogee kick motor (AKM) interface ring deep inside the structure. I figured out a way to apply a pure sheer load by reaching down into the structure with two actuators in a V configuration, another industry first.

The structural engineer on the F8 FLTSATCOM static test was Gary Furuta, who writes:

That was my first static test at TRW and the project was one of the most challenging because I did not understand the history until later. My Aerospace structural engineer counter part (Carl ...) did not trust me for two years until we developed a mutual trust that became a friendship that extended for many years. Apparently, many other companies that he interfaced with were not completely truthful when dealing with Aerospace and he initially thought TRW was the same. That is why I was directed to keep in close contact with Carl on a weekly basis regarding our situation, good or bad. Eventually, we began to work as a team on solving problems in a collaborative manner. It was the best collaborative working relationship I ever had with mutual respect on both sides.

An additional thing that stands out was the need to develop a technique for ensuring that the installation of the separation band was tensioned evenly. The mechanic, Jim Scarborough, used a technique to with a "transfer rod" to strike the band with a hammer and bar along the circumference of the band during tensioning.

I think the test is the first time I met Cris Corrlett as a "young" mechanic, who became a test engineer himself, on some of my other tests.

Because the F8 structure was flight hardware, there was considerable concern about damaging or contaminating it during setup and test. I devised precautions to prevent spilling hydraulic fluid on the hardware, including running return lines on the unpressured side of hydraylic cylinders, in case of a cylinder seal failure. I also used some wiffle tree arrays for tension loads to minimize the number of cylinders used.

I designed the main wiffle tree spreader to have a welded flange on the compression side for stiffening against out-of-plane buckling under load, and we proof loaded all elements to three times working load. The department manager at the time criticized my design of that spreader, saying the stiffener was unnecessary and therefore wasteful. After the test was over, I had a technician cut V-notches in both sides of the stiffener (to defeat it) and re-proof loaded it. It failed in buckling, just as I had predicted. I refrained from saying "told you so."

Hawaii Vacation

In December of that year Andrea and I took Malia, who had just turned two, to Hawaii to visit Adrea's parents, Don and Lenore. Malia was restless on the airplane and spent most of the flight walking up and down the aisle. We also visited my Mom and stepfather. I had shipped two of my custom boogie boards with the walnut and laminated teak and ash skegs to Hawaii. One I later gave to my brother John in 1990 who was living in Hawaii at the time, and I still have the other one. Andrea and I took Malia and drove to the north shore, and stopped at the beach at Waimea Bay. We put down some towels on the sand, and Malia was playing, when Andrea noticed that waves were starting to break on the reef. It doesn't break at Waimea unless the surf is twenty feet. I grabbed my boogie board and fins and headed out. I tried and caught two waves and then came in, not wanting to push my luck. They were the biggest waves I had ever ridden. It was quite exhillerating. Unfortunately, Malia had headed for the water, and in chasing her and keeping her in control, Andrea was unable to get any pictures! I was the only one in the surf.


John Wagner, Lisa, Johnny, and Pat, November 1985. Andrea and I would visit whenever we were in Salinas, and John and Pat
often graciously allowed us to spend time with them in their house.

Bonsai

In the late fall of 1985, I signed up for a bonsai instruction class at Hamilton Adult School in north Torrance. I had been dabbling in bonsai for a few years and now and while in Hawaii for the holidays I told Mom I had signed up for the bonsai course and I would be "professionally trained." Her husband, and my stepfather, Mits Aoki, loved gardening work and had lots of potted plants, but he didn't do bonsai per se.

I began the bonsai course, one night a week, in January of 1986. Jim Tatsukawa was the instructor, and Penny Jensen was his volunteer assistant. Both were accomplished bonsai artists and teachers. Jim sold bonsai material (plants) and pots in the classroom at cost as a service to students. First time students were instructed as a group, and a smaller group of continuing students were taught as a more advanced group. I learned so much and enjoyed the class enough that I continued as a student in 1987, 1988, and 1989. I told Jim I wouldn't continue in 1990 because I was starting a Master's Program at USC. In 1986, my first year, Jim and Leila Kusumi had just founded the new bonsai club, Dai Ichi Bonsai Kai (number one bonsai club) in Gardena, and I started attending meetings on Friday nights the next year, in 1987.


We bought our first new car when we lived on 10th Street in Manhattan Beach, a white 1982 Pontiac Sunbird.
Here it's parked in front of our small two bedroom house on 180th Street in 1985. We were able to buy the house
in 1983 with a VA loan for just over a hundred thousand dollars. I built the picket fence.

I started subscribing to Byte Magazine and Dr. Dobbs Journal, two magazines for serious computer users and programmers. Pete Bringmann, then on TDRSS project, was promoting desktop publishing for procedures and I started expoloring the idea too. I started using the new graphical user interfaces (GUIs) for personal computers, GEM Desktop from Digital Research in Monterey, and Microsoft Windows 1.0, both on floppy disks. Our PCs at the time had no hard drives.

Later, when we got some 10 megabyte hard drives, I was able to set up Ventura Publisher (for GEM Desktop) and Pagemaker (for Windows) and do some desktop putlishing demonstrations. I wrote the Desktop Publisher User's Guide for test engineers.

In 1986, using the test document library I had set up in our test department, I compiled torque angle data from wire harness tests to create parametric curves of wire harnesses to use to help aid designers of hinges and harness routings. It took the form of a three dimensional graph with axes consisting of number of wires in the bundle, harness length, and temperature. The number of spot ties or lacing was also found to be a big influence. The more tightly bound a harness is, the more it resists twisting and bending. It's a major torque driver and needs to be considered and specified in design.

Also in 1986 I began supporting the Gamma Ray Observatory (GRO) project in planning for a static load test. That task was later taken over by another department engineer, Chris Corlett whom I continued to advise as it was his first static load test. GRO was the heaviest payload (20 tons fueled) ever launched by anyone, including the Russians. I followed the GRO progress for years. A scientific payload, GRO was a typical NASA project, constantly being slipped with funding cut by Congress. It ended up taking twice as long and costing twice as much as originally planned. Congress should stop trying to save money by slipping programs and just fund them fully to get them done on time. It would be a lot cheaper for everyone. When a project is deliberately delayed and slipped, the engineers working on it have to go away to something else. Getting them back again is difficult, and if they're replaced by people without project experience, the inefficiency of re-learning project lessons increases cost.


The Gamma Ray Observatory being deployed by the Space Shuttle.

I was promoted in 1986 to Section Head, Mechanical Test Engineering, DPD I&T Department, responsible for all spacecraft mechanical test procedures, test engineering floor support, and personnel hiring and assignments. It was my first taste of what is called "functional" management. Like many aerospace companies, TRW had a "matrix" management structure. Everyone had two bosses, a project boss and a functional management boss. The functional management provided resources, facilities, equipment, and personnel, to the projects. The functional manager determined the performance evaluationa and compensation for an employee, and was dependent on the project manager for employee performance input.


Andrea in a white blouse in 1986.

In 1987 I did a feasibility study for using helium balloons (I called them "levitizers" to avoid the carnival or circus connotations) for zero-g simulation for a five panel solar array going through a multi-axis depolyment (3D). While the results showed feasibility, it was decided to do the deployment test by parts using an air bearing system that I had also conceptually designed.

Verification "by parts" is often a compromise that is resorted to in ground testing to keep costs down. For example, suppose a deployable system has two motions, one motion constrained to one plane and another motion constrained to another plane. If one had a perfect zero g simulation system, the deployment could be done in one continuous sequence, first one plane, then the other, without stopping. However with a compromise lower cost simulation, one could do one motion, stop the test, reconfigure for the other motion, then continue the deployment. The risk is that a failure mode will be masked by the interrupted test. So a test engineer is constantly trading cost for risk. Quantifying risk (in dollars) is in the realm of the discipline of risk management. People are always balancing cost and risk intuitively. Engineers do it formally.


Robert with one of many wooden-railed boogie boards I made, 1986.

Neural Networks

Around this time I started reading books about neural networks and cellular automata. One PC Magazine columnist, John Dvorak, was convinced that neural networks were going to revolutionize investment and stock picking. I obtained some neural net software for DOS and played with it, and found an interesting application: shimming a booster adapter.

A program I worked on had a 24-strut booster adapter (BA), a truss made of graphite (reinforced plastic) tubes. It had eight vertical struts and 16 diagonal crossing struts. The BA connected eight interface points on an upper stage to eight points on the spacecraft. It was important to shim the BA properly because any mismatch would cause high preloads in the struts when the interfaces were bolted together, and those preloads would add to the launch loads, potentially breaking the BA causing catastrophic mission failure. This was regarded, naturally, as a rather important problem.

I worked with a structural engineer, (name TBS), who had some calculations to predict loads in struts due to shimming configurations. Theoretically, there was some lower (booster end) interface shape (shim set) that would result in minimal strut preloads. The "minimal" case would have the lowest maximal preload. Because the truss loading was statically indeterminate, it was usually the case that some struts would go into tension when the interface was bolted. Intuitively, it can be seen that a well-shimmed BA should share the one-g gravity load fairly equally, with all struts contributing to support the weight of the spacecraft, so that none of the struts would go into tension. However, we learned that due to inaccuracies in manufacturing the radial and tangential positions of the booster, BA, and spacecraft, as the shear pins at the upper and lower interfaces were engaged, some struts would go into tension. The BA was highly sensitive to these inaccuracies because the graphite struts were quite stiff.

We measured strut tension or compression with strain gages on each one. The strain gage data were fed into an instrumentation computer and displayed in realtime, as well as recorded in computer media. We spent days in trial and error, with (name TBS)'s calculated predictions usually being completely wrong. Fortunately, this was all done as a series of fit checks well in advance of an upcoming vibration test where a failure of the BA would be highly unfortunate.

I had obtained a neural network simulator software program for Windows, so I took the shimming experiment data and created a neural network to "learn" it. The network had an eight node input layer, one for each of the shim values in milli-inches, two intermediate layers of nodes, and 24 output nodes, one for each of the strut loads in pounds. I had about 30 trials of data for training the neural network. When training was complete, the neural network was able to take a shim configuration and predict very closely the strut loads from a couple of datasets withheld from training.

In 1987 I was promoted to Subproject manager (SPM), Mechanical Integration Engineering, project 8482, with multi-million dollar budget responsibility. I had the planning and engineering responsibility for all mechanical integration and test activities, including system MGSE requirements, facility modifications, and overall spacecraft schedule.


Rick and Tim on the couch, 1986. I built that small cabinet with glass doors out of koa wood in the garage on 10th Street.

Microsoft Windows (Windows 3.0) was becoming the domiminant operating system for the Intel 386 personal computer. I had been using Windows on a 286 machine since 1985, and when Windows 3.0 came out for the 386 machine, it was quite apparent that with its memory management and multitasking, it was the way to go for desktop publishing of procedures and other applications.

Pete Bringman was the restricted program (for some reason which I never knew, program security personnel didn't like us saying "classified program" in unclassified media) assistant project manager (APM) for I&T and he was a big advocate of computer productivity improvements and had a local area network (LA) installed to facilitate collaboration among disciplines. I began using Visual Basic (VB) at work for custom applications. I created a neural network simulation program to implement a shimming algorithm for a graphite strut booster adapter. It worked beautifully, and was able to specify exact shims based on first fit strut loads measured by strain gages. It was much better than the old trial and error method.

I also started writing Visual Basic programs at home for fun. I wrote a network email program and put it on the computers on the LAN at work. Everyone used it for sending messages and files to others for collaborative productivity increases.

I uploaded some of my VB programs for Windows to Compuserve, a dial-up (telephone modem) bulletin board service. This was before the internet was well known and available. My Network Email program was sold as shareware and was our best seller. My software company was called Ivory Tower Software, and our products also included a relational database for keeping track of software licenses, a freeware 3D crossward puzzle, a file encryption program, a lawyer joke display program, and a freeware cellular automata workshop.

I started studying C programming because that's what a lot of professional programmers used. It would come in handy later when I went to USC to study computer science.


Halloween costumes, Malia as a witch and Becky as a fairy godmother, October 31, 1988.

Toward the end of Ronald Reagan's second term as President, in 1987-1988, the administration started to worry about national budget deficits, so President Reagan appointed a new Secretary of Defense, Frank Carlucci, AKA "Carlucci the Axe," because he was put in specifically as a budget cutter. They completely eliminated the big restricted program that many TRW people worked on. Some last minute heroic lobbying by Dan Goldin (later to become head of NASA) saved two of the program's future spacecraft out of a planned four. But times were getting harder and my career seemed to be stagnating. I started looking into the possibility of going back to school for a Master's degree.


Andrea, Tommy, Becky, Malia, and Lenore, December 1988.

In 1989 I was promoted to Engineering Staff, I&T Operations, project 8482, responsible for creating and implementing automated assembly and test procedures. I worked with the United States Air Force and the Department of Defense to produce the Militarily Critical Technologies List (MCTL, published in October 1992).


Andrea feeding Tommy at Crystal Cove in 1989.

My son Tim had married his high school girlfriend Seba Kennedy and they had a boy named Stephen. Robert was living with his mother in San Jose.


Tim, Becky, Tommy, and Tim's son Stephen, July 1990.

There was a science fiction movie made in the early 1980s called Blade Runner, an addaptation of Philip K. Dick's novel Do Androids Dream of Electric Sheep?. I saw this book listed as recommended in a column by science fiction writer Jerry Pournelle, who wrote a computer hardware and software evaluation column for Byte magazine. I found the book and read it. I was amazed. I had never read P. K. Dick before. I began to read everything I could find by him. In a book of short stories I recognized one about a passenger on a starship on a hundred year voyage. All the passengers were in hibernation and the ship was controlled by a computer. But one passenger woke up due to a malfunction in his hibernation unit. The computer tried to be helpful and guide him in dreams so he wouldn't go crazy during the years of the voyage. It happened to be one of the stories I was reading in the Playboy magazine at the beginning of my first psychedelic experience when I was 18.


Arriving at HNL in December of 1990 for a holiday in Hawaii with Andrea's parents, Don and Lenore Johnson.

I applied for a TRW Fellowship for a Master's Degree at USC in Computer Science in 1991. The Fellowship program wasn't funding mechanical engineering Fellowships at the time, and I had been doing a lot of computer programming in my work, so I thought I would go back to school and get a Master's Degree in comuter science. I was accepted into both the TRW Fellowship program and the Computer Science Master's program at USC.



January 7, 2012, McKinley High School, Honolulu, Oahu, Hawaii.

Note to other Rick Wagners about the use of our name: I am Richard J. Wagner, and I know there are lots of us Richard Wagners who go by "Rick." I use our name, "Rick Wagner," in the title of this autobiography because that's the name I go by. I established my personal home page under that name back in 1995, long before many people even knew the Internet (much less the Web) existed. So I consider I have a kind of Web presence precedence, and therefore am entitled to use my nickname in the title of my work. I know and/or hope you will understand.


Email: Richard dot J dot Wagner at gmail dot com
chapter6.html
Hand crafted HTML code copyright 2002-2017, by Rick Wagner, all rights reserved.

This page created April 9, 2002.
Updated December 17, 2008.
Updated August 5, 2009.
Updated September 2, 2010.
Updated September 16, 2010.
Updated March 17, 2011.
Updated April 18, 2016.
Updated November 24, 2016.
Updated December 31, 2016.
Updated February 19, 2017. Last updated April 30, 2017