My first EE job, the Naval Research Lab

I didn't have a job secured when I left Berkeley in June 1962 with my new MSEE degree. However, I did have three job offers: Hughes Aircraft in Fullerton and General Dynamics and an antenna R&D company in San Diego. They were well paying jobs, well-paying at the time, but after my exposure at Berkeley to Maxwell's Equations, antennas and Rumsey's Reaction Concept (Reaction Concept), I wanted to get into a research position. The Hughes job was a good possibility but I remembered that the U.S. Navy was seriously into electronics and the Naval Research Lab (NRL) in Washington D.C. had sent recruiters to Berkeley in the Spring of 1962. I hadn't met with the NRL recruiters but their name plus their heavy involvement in electronics got me thinking about applying for a job there. So after proudly going through the graduation ceremony, I drove home to Brea with no job to look forward to. At home, I decided to write a letter (snail mail back then) to NRL to see if they would hire me in the Antenna Section of their Radio Communications Division. Luckily, Martin L. Musselman, the Section Leader of the Antenna Section, was on the hiring committee and the Civil Service Board for NRL and he got back to me right away with job opportunities. It was a low-paying offer, but it was NRL and the possibility of a research job, so I immediately dashed off an acceptance letter to Musselman and in a couple of weeks had been hired at NRL. In early August my parents drove me to LAX to catch a plane to Baltimore for my first real EE job.





NRL viewed from the Potomac River.


Arriving in Washington D.C. in August, 1962, I checked in to the Y and notified Musselman that I was ready to go to work. I was a few days early for my starting day, so began searching for a low-cost apartment in the SW area of D.C. It was a hot and muggy time, of course, but I was a young and healthy kid, so I trudged around with newspaper in hand, looking for apartments. It was sweltering hot and then it began to rain. Being from California and not used to the swamp-like weather of D.C., I thought "it's going to cool off!" But no, it just became more miserable. So I promptly decided to find the first air-conditioned apartment that was available and that I could afford, finding a cozy studio apartment for only $130/month at 1001 3rd St., SW, quite close to the Capital building.


So the following Monday, I showed up at NRL fresh-faced and eager to go to work, hoping to do some kind of researchy sort of work in antennas etc. However, I hadn't really thought of working on ships and that didn't really appeal to me. I had been thinking about doing wonderful research in a nice office and was a little taken aback when I walked in the Antenna Section area and immediately was startled to see several large brass ship models, each extremely detailed with tiny antennas arrayed on masts on the ships. They were about 5 or 6 ft long models at about 1/50th scale and quite impressive in their construction and finish, built in the NRL Model Shop for testing of the antennas at the turntable on  the west side of the building. I am embarrassed to say that I foolishly blurted out to Musselman that "I didn't want to work on ships!" Musselman, whom I later knew as Marty, fortunately didn't boot me out of the office but rather just ignored my foolish outburst and ushered me into a large office that I was to share with W. Dale Long (Dale), another antenna and radio communication EE in the Section. I met and was to be supervised by Richard K. (Dick) Royce who had an office next to Marty's office. I learned that Dale and Dick were both highly competent antenna and radio communication EE's. Dale was more heavily involved with High-Frequency (HF) and Ultra-High-Frequency (UHF) antennas and communications and Dick was more into Very-Low-Frequency (VLF), Low-Frequency (LF) and HF antennas and communications. I learned much and enjoyed working with both of them in the following years, working at NavCommStaWashingtonDC (NSS Annapolis) at Annapolis, Maryland, at the FAA Radio Station on west end of Molokai, HI (now defunct), at the FAA Radio Receiving Station at Molokai, HI, with HF antennas on Navy ships and the TRSSCOM (TRSSCOM) Moon Relay systems.


Initial work and Finite Element Analysis with Rumsey's Reaction Concept





The Bendix G-15 Computer.

I didn't yet know what Musselman had in mind for me to do, but Dick Royce had a small electronics project to get me started: to build a very-low-capacitance VLF receiver amplifier. I completed that project, concluding that the capacitance of the amplifier couldn't be reduced sufficiently to be worthwhile and then started working on my own research-ey projects employing Rumsey's Reaction Concept. We had a Bendix G-15 (G-15) vacuum tube and drum memory computer located in a nearby office that was offered for my use on whatever and whenever I wanted. So I started working to apply the Reaction Concept to Finite-Element-Analysis (FEA) of dipole antennas, approximating the current distribution on the dipole and other such as antenna radiation patterns. The G-15 used its own interpreter program, Intercom, and also a compiler called Algo. Intercom was quite useful for simple tasks but the G-15 and Intercom couldn't handle the 50 or 100 simultaneous equations for my finite-element-analyses. Later on I used the G-15 to compute dipole antenna radiation patterns for some Navy brass at their BuShips office, showing the effects of the types of ground (dirt, sand, saltwater, etc) for the dipoles in horizontal and vertical orientations and for various heights above the grounds. These analyses, using the Reaction Concept, were easy to carry out and were quite illuminating as they showed that the antenna beams would lift up from the ground quite surprisingly, at least it was surprising for the Navy brass. Here is a youtube video (Antenna Patterns) of a monopole at various heights above perfect ground, but my analyses were for perfect and imperfect grounds. The Navy had always had problems communicating between ships mainly because of obstacles on the ships, even though the ships were visible. That problem was understood but a vexing problem for the Navy, often causing them to resort to signalling with flags or by flashing lamps. However, the effects of imperfect ground were not known by the Navy brass and they didn't have a clear understanding of the variability of antenna patterns depending on the height of the antennas. I mentioned to the brass that I could compute antenna patterns for dipoles above various kinds of ground such as dirt, sand, water etc and we left, planning on returning in a week for more discussions. Back at NRL I did my antenna radiation pattern calculations for vertical and horizontal dipoles at various heights above several different types of ground. However, I did not try to compute the effects of ocean waves on the antenna patterns. The next week we returned to BuShips to meet again and I showed my antenna pattern graphs. The brass were surprised to see that the antenna beam would lift off the ground and that the effect could be quite dramatic, strongly suggesting that this could be a significant cause of their ship-to-ship communications problems. However, of course, it didn't show them a solution to their problems and a real solution wouldn't arrive until years later with other means of communicating between ships. I made use of J.R. Carson's Reciprocity work (see Reciprocity#1 and Reciprocity#2) and Rumsey's Reaction Concept to compute the antenna patterns. The process is a bit involved to describe but firstly Carson's Reciprocity work shows that a transmitting antenna pattern can be computed as if it is a receiving antenna. Secondly the receiving antenna would "see" a plane wave arriving from a great distance. This means that the electromagnetic near-fields of the transmitting antenna don't have to be considered, resulting in a terrific simplification of the antenna pattern calculations. Then the Reaction Concept shows that the calculations are completed by integrating the product of the plane wave and the assumed current distribution of the antenna as if it is transmitting antenna, the plane wave possibly reflected from a perfect or imperfect ground. As Rumsey pointed out in a note to the IRE Antennas and Propagation, the Reaction Concept is particularly powerful since it is a Variational analysis technique, with the consequence that the current distribution on the dipoles had only be an approximate dipole current in order for the antenna radiation pattern calculations to be accurate. Roger Harrington's text, Time-Harmonic Electromagnetic Fields, provides more information on Reciprocity and the Reaction Concept and probably a more lucid description of the whole process but trust me, the calculation of transmitting antennas is greatly simplified by using Reciprocity and the Reaction Concept.


To get more computational power for FEA analyses, I next started using NELIAC (NELIAC) on NRL's vacuum tube computer. I think the NRL computer was called the NAREC (NAREC),



NRL's NAREC Vacuum Tube Computer System.







shown in the photo, but I don't remember the computer room being so tidy. I recall two or three technicians working in a mess of cables and stuff to replace vacuum tube modules in order to keep the computer running. But it was a fast computer for the times and the NELIAC language was powerful. I used the NAREC for a time and then NRL installed a CDC3600 computer which was quite a powerful and reliable solid-state computer, programmable in Fortran and using punched cards in batch mode, of course. I made a lot of use of the CDC 3600, successfully computing the impedance and the current distribution of a dipole antenna using the Reaction Concept and 100-equation FEA analysis techniques. It was quite exhilarating to be able to complete the analysis and I planned to do more analyses on more complex antenna structures but got busy with other tasks that Musselman came up with for us.




The CDC 3600 Computer System.



Martin L. Musselman
Marty was always digging for projects such as the FAA Radio Station project and the Moon Relay Project that would help the Navy, often projects on a shoestring, using surplus equipment from the Navy and also from the Air Force to complete the projects on bare-bones budgets. They were interesting and worthwhile projects and we loved working for him and with him as he was a hands-on guy, working alongside us EE's and technicians to get the job done. He was on the Civil Service board at NRL and used his influence to get promotions for us EE's, getting us promoted at the earliest possible dates. Marty worked at NRL for 65 years, finally retiring in 2008 at 87 years old, contributing for all those years. I didn't know it when I arrived at NRL in '62, but learned much later that he was involved with the Moon Relay Projects. See reference #9 of Moon Relay. Amazing guy!


Moon Relay and TRSSCOM
I was unaware of NRL's Moon Relay history when I arrived at NRL, that James H. Trexler had conceived of the idea in 1945 and proved it in 1954 with Communication Moon Relay. Then in January 1960 a facsimile of the U.S.S. Hancock was sent from a transmitting station near Honolulu, HI by Moon Relay to a receiving station in Cheltenham, Maryland. I didn't become aware of NRL's involvement until Musselman acquired a surplus Navy analog sonar computer and asked me to see if I could get it to control a shipboard 16 ft. diameter antenna, mounted on a surplus 40 mm anti-aircraft gun-mount (USS Jamestown, TRSSCOM). For the facsimile transmission of the U.S.S. Hancock a frequency of about 400 MHz was used but a much higher frequency of about 2300 MHz for the TRSSCOM system because the shipboard antenna was only 16 ft in diameter, requiring the S-Band frequency. Also, greatly improved receivers, the Parametric Receiver or PARAMP, was available by the mid-1960's. I understood that the TRSSCOM Project was NSA sponsored and that the NSA preferred the Moon Relay project over communication satellites supposedly because they, the NSA, didn't trust the security of data transmitted via communication satellites. The first TRSSCOM ship was the U.S.S. Oxford. I worked on two different TRSSCOM installations, the U.S.S. Georgetown and the U.S.S. Jamestown. A later TRSSCOM ship was the ill-fated and heroic ship and crew of the U.S.S. Liberty, attacked during the Six-Day-War. 


Musselman acquired some 10 kW S-Band transmitters, the PARAMPs, some duplexers so the antenna could be both a transmitter and a receiver, waveguide and rotary waveguide joints. Two different frequencies near 2300 MHz were used, one for transmitting and the other for receiving and the duplexer had to be tuned for the two different frequencies so the signal from the 10 kW transmitter would go out to the antenna and the minuscule receive signal would be passed from the antenna in to the PARAMP. Since there was such a disparity in the power levels, the duplexers had to be very carefully tuned. Dale Long and I were able to learn how to tune them to ensure that the 10 kW transmit power would not leak into the PARAM receiver and destroy it. The duplexers were large S-Band waveguide devices, about 5 ft x 5 ft x 1 ft.  For the antennas and controls, Musselman acquired some 16 ft diameter mesh dish antennas, some 40 mm anti-aircraft gun-mounts (without gun barrels!) and some ancient analog sonar control computers, Korean War era computers. They were the Mk 59 Mod 6 sonar computers comprising several synchro and servos that were wired from one to the next to compute coordinate conversion equations. Ship's heading, pitch and roll were provided by synchro signals from the ship's SSQ-14 gyro system. I wired the SSQ-14 signals into synchro receivers of the Mk 59 computer. The Mk 59 computer's several synchro and servo units all had dials showing the input signals, the intermediate calculated results and the final output signals for controlling the gun-mounts. Three of the synchro and servos were used for the ship's gyro info and we used others for the ship's longitude and latitude, the Moon's position relative to the Earth and the antenna azimuth and elevation positioning. I also cobbled up a sort of stepper motor affair to update the Moon's position at its longitudinal motion rate of 0.55 degrees per hour. The Moon's motion varied somewhat depending on its elevation above the equator but the 0.55 degree per hour value was adequate for the short-term TRSSCOM communications. The gun-mounts came in two flavors, hydraulic and electric drives. The hydraulic drive mounts were fearsome beasts, highly accurate and very powerful and very quick-acting but subject to hydraulic fluid leaks that would spew oil on the deck surrounding the gun-mount. The electric drive mounts were equally accurate, slower but were preferred as not being subject to the hydraulic leaks and not being so quick and powerful as the hydraulic mounts. The sonar control computers utilized plug-in vacuum tube amplifiers that failed often and required that a "bushel basket" of the amplifier modules be kept on hand for the frequent failures. I left NRL in August of 1966 and I understood that later a new EE would develop some solid-state amplifiers for the sonar computers, greatly improving their reliability. Once I understood how they worked, the Mk 59 computers were easy to rewire to do the coordinate conversion equations. Musselman gave me a document presenting the coordinate conversion equations to compute the azimuth and elevation commands for the steering the antennas. 





U.S.S. Georgetown at sea on a shakedown test to Key West, FL. Note antenna on aft end of ship, pointed at the Sun.

We obtained complete documentation for the Mk 59's and I was able to rewire one to command the gun-mount to point the 16 ft dish antenna at the Moon, at the Sun or at Cassiopeia for testing. With the Parametric Amplifier/Receiver, we could pick up static noise and verify both that the



R390A Receivers and Analog Computer on the U.S.S. Georgetown.


antenna was pointed correctly and that the PARAMP was working properly. When we received the surplus Mk 59 computer, we had it moved next to my desk on the second floor after we verified that the floor would be strong enough for the several hundred pound computer. We had one of the anti-aircraft gun-mounts with the 16 ft dish antenna set up outside on the ground, just below my office window so we could crack the window open a little in order to run cables to the gun-mount's servos for me to control from the Mk 59. Once I had the Mk 59 




Analog Computer and Teletype model 28's on the U.S.S. Georgetown.

rewired, I could set the Sun's position into a pair of the computer's servos and control the antenna to point at the sun, verifying the computer's operation. So the computer system development (rewiring the Mk 59 syncros and servos) and testing was a simple operation. I was able to get a trip on the U.S.S. Georgetown from Norfolk VA to Key West, FL for the shake-down cruise of the antenna control system. We didn't attempt to operate the Moon Relay TRSSCOM system, just to evaluate the operation of the Mk 59 computer control and the gun-mounted antenna. During the cruise I frequently set the computer to point at the sun and could verify that the antenna was correctly positioned by looking at the shadow from the antenna's feed-horn onto the surface of the dish antenna.




10kW Transmitter and R390A Receivers etc on the U.S.S. Georgetown.



For the U.S.S. Jamestown installation, we purchased a Computer Control Company DR-20 Digital Resolver that very rapidly computed the coordinate conversion equations using the CORDIC algorithm. Computer Control was purchased by Honeywell and who continued to produce other digital processing systems, the DDP series. The DR-20 was programmed by means of a diode pin-board. The diode pin-board was quite easy to program, not requiring any keyboard or processor to store a program as the position of the diodes plugged into the pin-board established a binary word that could be a command and/or a data word. So it was quite easy to program the DR-20 and I was able to quickly set up the program to do the coordinate conversion equations to steer the 16 ft antenna, much easier than rewiring the Mk 59 analog sonar control computers! The ship's gyro system, the SSQ-14, produced analog synchro signals, so we used some spare




DR-20 CORDIC Digital Computer and R390A's etc on the U.S.S. Jamestown.



 synchro units and some digital rotary converters to provide digital inputs for the DR-20. Also, the anti-aircraft gun-mounts required synchro signals requiring us to purchase two digital-to-synchro converters from GAP Instrument Corporation located in Long Island. As with the development of the Mk 59 analog sonar control computer, we set up the DR-20 next to my desk and I could conveniently program and test the overall system with an antenna mounted gun-mount. Once the 3C DR-20 Digital Resolver was all set up to steer the gun-mounted dish antenna, we shipped it to the Naval Shipyard at Portsmouth VA and installed it on the U.S.S. Jamestown. The DR-20 was very well built but not really a ruggedized system and I don't know how it held up on the shipboard environment.




R390A HF Receiver for the downconverted S-Band signals.



Another task that Musselman challenged me with is to build an encoder for the TRSSCOM Quadrature Frequency Shift Keying (QFSK) system. In the TRSSCOM system, two channels of data were decoded into four separate frequencies by modulating the S-Band transmitter with the four different frequencies, very close together in frequency near 30 MHz. The received S-Band signal was down-converted to the four frequencies and demodulated with four R-390A radio receivers. Consequently there were four outputs from the R-390A receivers that Musselman wanted me to encode back into the two channels of data as shown in this 4 to 2 encoder chart that shows, for example, when the input is from R-390A receiver #3: F3 is 1, CH1 output is 1 and CH2 output is 0.



F4	F3	F2	F1	CH1	CH2
0	0	0	1	0	0
0	0	1	0	0	1
0	1	0	0	1	0
1	0	0	0	1	1


The R-390A receivers were excellent vacuum-tube amplitude modulation receivers that we commonly used for HF signals. I developed a simple decoder using operational amplifiers to amplify the baseband signals output from the R-390A's and then a simple diode encoder circuit. The encoder's output provided on-off data pulses for the two received data channels. So my encoder was able to retrieve the original two channels of data and the data, if unencrypted in cleartext tests, could be




Teletype model 28.

printed on a teletype machine, probably a model 28 such as in the photo. If the data were encrypted, it would be sent on to the cryptologists in an isolated guarded room on the ship.


Besides Musselman, Dale Long and myself there were at least three others working on the TRSSCOM RF systems and the mechanical aspects: L.H. Feher, V.W. Graham, W.E. Leavitt.












The FAA HF Receiving Station at Molokai
Early in 1964 Musselman came up with another project at the FAA Radio Receiving Station at Molokai. The Navy was interested because TRG Incorporated, a subsidiary of Control Data Corporation, had installed a new type antenna for the FAA at Molokai, the HF Luneberg Wire Grid Lens Antenna and the FAA wanted an evaluation of the antenna. TRG was the primary test organization for all the antenna tests and they subcontracted to SRI, then attached to Stanford University, for the antenna pattern tests. SRI personnel flew a B-19 airplane towing a transmitter around the site while we received the transmitted signals and recorded them on a chart recorder. TRG Incorporated, which was a subsidiary of Control Data Corporation, reported the results in the IEEE Antennas and Propagation Society International Symposium, 1966. and in a more complete report to the FAA in 1966. TRG's sketch of the antenna is below.



TRG's sketch of the Wire-Grid Antenna at the FAA Receiving Station, Molokai.

Musselman, Dale Long, Dick Royce, an EE technician who's name I have forgotten, and I traveled to Molokai mainly to test the Wire Grid Antenna but also to check out and test the several Rhombic Antennas on the FAA site. I suggested that a 15 ft tall Monopole Antenna be erected on a hill and be included in the testing to compare with the Rhombics and the Wire Grid Antenna, so with assistance from the U.S. Navy personnel from the Pearl Harbor base, we set up a 15 ft tall Monopole antenna on a hill close to the FAA building. For data testing of the antennas, I had cobbled up a gadget at our NRL office in Washington D.C. to send a simple "RY" Baudot "two-character" code to the FAA transmitter located near Washington DC for character error rate testing of the antennas on the FAA site. The RY code is a simple binary "0101010101" "two-character" Baudot code that was used for communication system testing, the RY code transmitted at 10 characters-per-second for periods of 20 minutes with intervening periods of 10 minutes when the Woodbridge FAA transmitter was turned off. The NRL tests were conducted between 8 MHz and 20 MHz as reported in the TRG report to the FAA. We installed R390A HF receivers in the FAA main building and received the transmitted signals, passing the R390A output baseband signal to a bit test system and error counter to count the bit errors so we could determine the bit error rate for the various antennas. The HF signal transmitted from the FAA transmitter at Woodbridge, VA near Washington DC was carried to the Molokai site by ionospheric propagation that was available only during daylight hours as the ionosphere, gradually fading away at nightfall. One test that we did was to determine how long the RY code could be received in the antennas, thereby providing a comparison of each antenna type. We determined that the bit error test results for the Rhombic antennas were quite a bit better than for the Wire-Grid Lens antenna. TRG reported that the results of the FAA bit error tests for the Wire-Grid Lens antenna was 4.5 x 10^3 while the character error rate for the Rhombics was 3.5 x 10^-3. My recollection is that the Rhombic antennas could communicate the longest time each day and were best by about 3 dB over the Wire Grid Antenna and the Wire Grid Antenna was better than the monopole antenna by about another 3 dB. However, that was after Dick Royce detected some problems in the Rhombic antennas, some defects in the insulators where the antenna wires were attached to poles. The defects or damage to the insulators was probably caused by lightning strikes, resulting in low impedance paths from the wires to the wooden poles. Royce replaced the defective insulators, resulting in significant improvement in the operation of the Rhombic Antennas. As I say, the overall results were Rhombic Antennas best by about 3 dB, the Wire Grid Antenna next best by about 3 dB better than our Monopole antenna. Fortunately, TRG thoroughly reported our tests, the FAA tests, the SRI antenna pattern tests (described below) and their test results. My recollection is that we concluded that the Rhombic Antennas bested the Wire-Grid Antenna by about 3 dB.


Musselman sent me home from Molokai in March 1964 as I was getting very concerned and anxious about my wife's pregnancy and imminent delivery of our bouncing baby boy. When I returned to work at NRL, Musselman phoned me from Molokai and told me he was sending me many paper chart recordings of the antenna patterns from the SRI tests described above. Musselman had also contacted BuShips, asking for several technical people to be sent to NRL to help with reading the values from the charts and transcribing field strength values to yellow-lined pads. So here I was supervising some fairly high-level people that BuShips had sent to me for transcribing the data from the paper chart recordings. They all worked very well as a team and I felt fortunate to have such a good group of people working on the task. Later on a couple of computer scientists from SRI arrived to collect the transcribed data and would produce contour radiation plots for the Rhombic and the Wire-Grid Antenna. TRG noted that the resulting antenna patterns were probably degraded by perturbations in the terrain and also mentioned that the inefficient feed systems on the Wire-Grid antenna likely resulted in less expected performance that would probably be corrected with recently developed feeds. Also TRG noted that the bit error rate of the Wire-Grid antenna was about 20% higher than that of the Rhombics. All in all, our testing at the FAA site was quite successful, determining that the Wire-Grid Antenna had good possibilities but was not as effective as the Rhombics in this installation and with the Wire-Grid Antennas inefficient feed systems. The Monopole antenna fared reasonably well, operating about 80% to 90% as long as the Wire-Grid Antenna during the ionospheric propagation time. Also, another side effect of our work at the FAA site was Dick Royce greatly improving the Rhombic antennas by replacing several defective insulators. Fortunately I arrived home to be with my wife a few weeks before she gave birth to John William Baker.


Other Projects
I can't possibly give adequate credit to my supervisor, Dick Royce, and co-worker Dale Long as I was and am more concerned about my own projects and experiences but I do recall a couple of instances that Dale and I worked on together, besides the Moon Relay and Molokai FAA Radio Receiving projects. One project was Dale's to develop a UHF over-the-horizon Tropospheric Propagation transmitting antenna. Dale's antenna comprised four spiral antennas in an array. I helped him with the testing atop our building at NRL. It was a good design and more than met the Navy's needs. Another project was developing impedance matching circuits for HF receiving antennas on the TRSSCOM ships. I developed a knack for combining various lengths of various impedance coaxial cables using Smith Chart techniques (Smith Charts). We had to match the antennas to a 3:1 Standing Wave Ratio (VSWR) or better, in order for the receivers to adequately receive the HF signals antennas . Then Dale and I built the matching circuits and traveled to the Naval station at Portsmouth VA and installed them on one of the TRSSCOM ships. Dale was the leader on this task and we worked well together. But like I say, I am mostly concerned about the projects that I was directly involved in and so cannot possibly give adequate credit to my coworkers.


Personal Stuff
Except for the swamp-like weather in the summer and the miserable snowy damp winters I enjoyed living in Washington. I met my future wife Lydia Cecelia (Celia) Baker in early 1963. She was a nurse in the Pediatric Department of Children's Hospital and looked terrific in her starched white uniform with the little nurse's cap. We had a lot of fun running around Washington, often going to the National Gallery and then getting ice cream cones at the downtown Hot Shoppe. It was an exciting time in Washington with John F. Kennedy and Jacqueline in the White House and then Caroline and John-John, Camelot and all that. Tragically Camelot came to an end with Kennedy's assassination in November 1963. Celia married me in the summer of '63 and moved to Alexandria to the LandMark Terrace apartments up South Van Dorn. Celia still worked at Children's Hospital for a while after we were married but then stopped working at Children's when she became pregnant with our son. Joyfully, our son John William Baker was born in 1964 soon after I returned from Molokai. I had really enjoyed working for Musselman and with Dale Long and Dick Royce at NRL and was lucky to have such good projects to work on but got sick of the swampy weather in Washington and became homesick for California. So I searched for job opportunities in California at U.S. Navy facilities but settling on a job at the Lawrence Livermore National Laboratory in Livermore, CA. So we packed up and moved to Livermore in August 1966 after telling Celia that there were 3 or 4 trees in Livermore, a church and a gas station. So she was prepared for the worst and was nearly delighted to discover that there were a few churches, not just one, a few gas stations and several trees in Livermore.