Retirement Musings

After 36 years at Motorola/Nokia I will be retiring in October. With this major life event, I found myself reminiscing about my 36 years with Motorola/Nokia. We all have stories from our careers, and mine aren’t really that interesting. Nevertheless, I decided to write down some ‘musings’ from those 36 years. Of course, to cover all that time I could write a book. I will just write down some highlights that pop into my head as I write, with no real reason for picking some and leaving out others. This is longer than I would normally like for a post, but I will try to keep it short. I will mention some names and projects along the way, but many, many names and projects are left out in the interest of brevity.

I graduated from Portland State University in Portland Oregon with a degree in Electrical Engineering (EE), and accepted a job in the Phoenix Arizona area with Motorola Government Electronics Division (GED). I remember driving my little green Volkswagen Rabbit from the rain and green in Oregon across the desert from LA to Phoenix, and wondering “what am I doing?”. It seemed so brown and drab to me. I thought I’d work there a year or two and go back to Oregon. But life had a different plan for me. I put down roots, got married, had kids, made friends, got involved in Church, and came to appreciate Arizona, with the Sonoran desert around Phoenix, and the variety around the state. And life at work was also rewarding, which is the topic of this post.

I started in the 6 month Engineering Training Class with 35 other college fresh-outs. Afterwards I was assigned to work on a radar system. I went to our Chief Engineer (Bob Curlee) and requested a different job (pretty bold for a young punk engineer 🙂 ) but was rejected, so off I went to the radar group. I worked on building a hardware radar target simulator for a Ku band (12GHz) radar. We had a rack full of coils of waveguide that were switched in and out with RF (radio frequency) switches that we nicknamed ‘gold bricks’. I remember a technician named Claude Dochow showing me the ropes on how to build and test the test equipment. I worked in R&D (research and development) on future ‘modifications’ (we called them ‘mods’, the last I worked on was Mod9) of the radar. We managed to borrow an HP9000 computer from the test group, and I used HP Basic to simulate flying the radar low over the ocean and simulating the strange antenna beam shapes and the radar return from the ocean and using Doppler to detect the moving target. I remember being in the lab troubleshooting a circuit with an oscilloscope and spectrum analyzer trying to figure out why it was oscillating. Terry Richmond (aka ‘TR’) came out and found some impedance mismatches and leakage across circuit tracks, and fixed it. I told him “I could never have done that in a million years”. He encouraged me saying yes, I could, but I just needed more experience. Somehow those simple encouraging words impacted me, as you can tell since I remember them all these years later (probably because they came from TR, who I really admired). Those radar systems used all analog circuits and we implemented algorithms in analog, including timers (a linear voltage ramp triggering at a threshold), variable trigger points, integrators, combinatorial logic, etc. The late Ken Burgess is another person from that period that I learned a lot from and had a big impact on me. I got my masters degree from ASU (Arizona State University) during that time, and after I graduated I decided to take a different job working on communication systems rather than on radars.

Getting my masters at ASU was a good experience. I think it was more productive for me to get it while working because I learn better when I can apply what I’m learning. Some of my favorite classes were: The Digital Signal Processing series of classes using the Oppenheim and Schaeffer text, taught by the famous Dr. Jim Cadzow. Dr. Cadzow also went to my Church. Detection Theory classes going through the Whalen textbook. The Probability class going through the Papoullis textbook. Transform Theory was taught by my co-worker Dr. Doug Hill.

I interviewed with the Advanced Communication Section (ACS) manager Mark Barota (who later became a Motorola Vice President). He told me his vision was to build a complete communication system from top to bottom. He got his wish years later with the Iridium program. The job was in the Ocotillo facility in Chandler, which at that time was new, and Price Road going to the plant was still a dirt road. I was assigned to lead a half completed R&D project to build an IIR (infinite impulse response) equalizer based on research and a PhD dissertation done by my co-worker Bruce Cochran. It was a rack full of high speed ECL (emitter coupled logic). How times have changed – that rack can now be implemented in a single chip! I spent many hours finally getting the signal constellation points of the modulated waveform to converge. I remember when I told Kent Terrell in the hallway by the cafeteria he let out a loud shout and hugged me. Not exactly engineer-like behavior 🙂 I also worked on a ‘continuous adaptive equalizer’ which was an analog equalizer designed by Dr. Frank Cornet (not sure I remember his name correctly). We couldn’t get it to converge, so I wrote a computer program to simulate it. Sure enough, the program didn’t converge either 🙂 He used a LPF (low pass filter) to estimate an integrator. I replaced it in the simulation with a real integrator, and it worked, and did the same in the hardware (the integrator was a current source with a capacitor, and yes we were still building transistor circuits in those days!), which fixed it. Dr. Cornet was very happy. He later left and went to teach in a university, and asked me if I could work with him in my spare time, but I didn’t have time for that.

I took some post graduate courses during this time. One was Modulation Theory taught by Dr. Carl Ryan who ran the Motorola Communications Research Facility (CRF), or nicknamed the Carl Ryan Facility 🙂 Dr. Ryan is one of the inventors (if I recall correctly) of MSK (minimum shift keying) modulation format, or a derivative of it. I also took Coherent Communications taught by my co-worker Dr. Bruce Cochran. Bruce gave us a file with a sampled waveform. We had to write software to down-convert the signal, and use a PLL (phased lock loop) to decode the message in the file. I spent a lot of time on that class! I still remember one the messages (there were several for different assignments) was “Peace is Joy Dancing” – kind of corny but really cool to see it finally pop out of the program. Bruce was very good at time management. He would rarely work more than 8 hours. We would go to lunch for 1/2 hour in the cafeteria. He was very productive in his 8 hours, and wouldn’t participate in non-work ‘chatter’. Then he accomplished a lot in his non-work time: choir director, teaching classes, getting his PhD. I tried to learn from his example, but I didn’t do very well at it.

I worked with the ACS group for about 14 years. It underwent several name changes and leadership changes. I worked on a lot of different projects. Scott Blanchard became the driving technical force in the group. I think Scott was the best engineer I’ve met in my career, and he had the biggest impact on me as an engineer. He is a brilliant guy, and his talent is amazing to me. We developed a satellite propagator which we used on many projects, and later we used it on Iridium. We also developed a geolocation algorithm that accepted as input many satellite measurements of different types from an emitter on the earth, and used them in an MMSE (minimum mean squared error) algorithm to compute the location and error ellipse of an emitter on the earth’s surface. Multiple FOA (Frequency of Arrival) measurements produced a conical surface where the emitter could be, and TOA (Time of Arrival) measurements produced a spherical surface. The cone intersects the earth in a hyperbola, and the sphere in a circle. The intersection of the circle and hyperbola on the surface of the earth is the location of the emitter. The error ellipse is created by the error in the measurements, and we had to resolve possible ambiguities in the location (the hyperbola and the circle can intersect in two places). I also worked with Dean Vanden Heuvel on these projects. He is an excellent engineer, and he wrote a lot of the software for doing the orbit propagator and geolocation algorithm. Dan Dolan eventually became the manager of the ACS group, and I worked for Dan for the rest of my time at Motorola/Nokia, through various incarnations.

During this time I worked on a very cool classified project, which I wish I could tell you about, but if I did I’d have to shoot you 🙂 I was the technical lead for this project, and worked with Joe Mintzer. Joe was great to work with, and I learned a lot from him. We (along with Greg Twaites and Don Hancock) developed a method for specifying and testing the software that I have used in various forms ever since. We called it Specification Using Responses to Events (SURE). We developed our own notation (SURE Diagrams), and process for writing requirements. We also used an incremental development method (similar in a lot of ways to the latest fad of ‘Agile’ development), which I have pushed for on projects ever since.

For many years I was interested in hardware design and doing analysis (or being an ‘analyst’). Software was a tool to help in those endeavors, but I avoided working on purely software projects. I wrote a lot of software, but it was all simulations and algorithms used as a tool so support my analysis. I never took a software class in college (though I did take Electrical Engineering classes where I learned machine language for the Motorola 6800 processor). Our Circuits Theory professor provided extra-curricular training on Fortran so that we could write programs to compute voltage and current on circuit nodes. (we would submit our program with card decks or paper tape to the main frame computer, and wait to be given the output). Over time as computers became faster and memory cheaper and more dense, software could do more things and became more interesting to me. I learned different software languages by reading textbooks and by the job experience. I started writing software that was deliverable to our customers in approximately the early 90’s. Object Oriented (OO) software was a breakthrough for me. I found it to be much more intuitive and practical, and much better than the SASD (structured analysis structured design) approach, which was the fad at the time. I read the Stroustrup C++ book (and others), and started using C++ for various projects. When I got busy with other responsibilities I always tried to assign myself at least a small software task to keep my hands ‘dirty’.

I worked on the geolocation software for the Iridium program. Iridium is a cellular system, but it is unusual because the cell sites are not at fixed locations on the earth, but instead are flying on a constellation of 66 Low Earth Orbit (LEO) satellites. We leveraged our experience and software in satellite propagation and geolocation algorithms. I was the lead for the Call Processing Location Determination (CPLD) subsystem. When an Iridium Subscriber Unit (ISU, aka the phone) needed to be notified of an incoming call (aka “ring alert”), CPLD was given FOA and TOA measurements from the ISU. It used our MMSE algorithm to compute the ISU location and error ellipse, and based on that computed the LA (Location Area) on the earth where the ISU was located (a geopolitical area). It propagated the satellites to their positions for the correct time instant, and used models of the satellite antenna beams to compute which beams overlapped with the LA, which would be used to send a ‘ring alert’ (RA) message to the ISU. The output of CPLD was a response message indicating which beams to use to ‘ring’ the ISU, and the LA of the ISU. For the LA we invented a way to tessellate the earth with large resolution ‘squares’, and decrease the resolution incrementally where there were geopolitical boundaries. A competitor estimated a solution that used a large hardware system with many CDs full of map data. Our system was fast and efficient and all the data could be stored in RAM. CPLD computed 32 RAs every second. It was designed as a ‘parallel processor’, so you could plug in another card (which was a computer using the VME bus) and increase capacity without any software or configuration changes. I feel fortunate to have worked on this unique project. It was a fun project where we had the opportunity to design cool new algorithms, do lots of fancy math for the algorithms, coordinate transformations for the propagator and ring alert beam calculation, etc, etc. And the best thing was that it worked very well. David Garcia was a young engineer at the time who helped me with this project. And of course I leaned on Scott and Dean a lot, who I had worked with previously on the algorithms. I took some heat for rejecting the ‘ObjectTime’ approach that the Iridium Gateway team was using. I thought it would be a train wreck. Unfortunately, we were proven right in the end (our software was used as is; the software that used the other approach had to be rewritten).

In approximately the year 2000, government contracts were in a down cycle, and the Public Safety Group in Chicago needed a team of System Engineers and Developers. I made the tough choice to go with that group, and change from government to commercial work. Not long after that Motorola sold the government group to General Dynamics. We started a design center in Arizona which was led by Dan Dolan. I was the technical lead for the Trunking Astro Site Controller (TASC) group (I get the credit/blame for that name). We worked on a lot of different projects there. One was the Data Site Controller (DSC). That was really fun project. The best thing, I think, was the process we introduced, and the software mechanisms. I pushed a ‘SURE-like’ process that mandated testable requirements, which was foreign to the Public Safety group in Chicago. It was very successful. We also developed software mechanisms for doing active objects, communication between objects, and state machines, that was very efficient for a real time system. Brad Sanders was instrumental in contributing to this. He could cut through a lot of the clutter, and come up with great ideas and insights. I learned to listen to what Brad suggested 🙂 Another memorable project was Astro LE. We developed a single site Public Safety system from scratch. We worked a lot of long hours, trying to justify our existence as a design center in Arizona. Curt Zenigami was the project leader and I was the technical lead. I remember we had a big problem and Joe Mintzer and I went to Curt and proposed scrapping one subsystem and starting over on it. Curt said to go for it. A small group of us spent every day for a month in a room developing the new solution. Astro LE was a great technical success, and (if I do say so myself) a very good new product for our customers. We achieved ‘virtual zero’ defects for quality. Unfortunately, this didn’t secure our Design Center, because not long after we delivered Astro LE they decided to shut down the design center. So much for all the hours we put in 🙂

We were all preparing resume’s for finding a job when the Cellular Infrastructure Group (CIG) at Motorola decided they could use our team for a CDMA (code division multiple access) cellular system. So we all went to work on the CDMA cellular system. Eventually we were given the AE (Applications Engineering) group. We added many new features, and improved the quality of the existing design.

CDMA is a Third Generation (3G) cellular system. When the standards for the Long Term Evolution (LTE) Fourth Generation (4G) cellular system came out, a small group of engineers in Arizona (along with others from the Chicago area) were assigned to work on a trial LTE system. I worked on the modem Scheduler part. Brian Turner is an outstanding engineer I worked with since the days in the government group, and he was instrumental in working on the Controller part of the trial project. On the Scheduler we determined that the ‘strawman’ software design we were given wasn’t very good (e.g. too many unnecessary software threads), and we started from a blank page. We got a lot of grief for this, but it was the right decision. The real time requirements were stringent, and the design had to be very efficient. We re-used some of the ideas we had developed on the DSC. We again worked long hours with a small group and the trial was successful. We were rewarded by being given responsibility for the LTE eNodeB product. The eNodeB is a critical piece of hardware at the cell site with software responsible for communicating with the phone, managing mobility, scheduling RF resource for it, etc. There was no let up on the long hours because we needed to keep up with the competition, and our staff was very lean. I was the lead software architect for the eNodeB. I think in some ways this project was the culmination of all the others I worked on up till this time. We were allowed to define the process for specifying and developing the eNodeB. We used a SURE-like process that used UML (Unified Modeling Language) instead of SURE diagrams, and incremental development. There is much I could write, but in the end the eNodeB has performed very well with excellent quality. It is out in the field at over 50,000 cell sites and takes part in every cell phone call made at those sites. Here in Arizona we had an outstanding team that can be very proud of their accomplishments. I worked with mostly the same development group through Public Safety, CDMA, and LTE. This includes Curt Zenigami, Brad Sanders, Matt Hunsaker, Tim Curtis, Jenifer Petras, Eric Wang, Tom Swendseid, and Doug McBride. The system engineering group led by Shawn Hogberg is an impressive team that gets credit for much of the success of the eNodeB. In the early days I leaned a lot on Bill Shores from Shawn’s team – he is an excellent engineer who was the ‘answer man’ for LTE standards questions. Others that stand out are Brian Turner, Vinayak Hegde, Jamil Shihab, Lukasz Turon, Tomasz Sztuka, Pawel Mackowiak, Suresh Kalyansundaram. Many others! I would mostly like to thank Dan Dolan for leading our group through the years, and trusting us with this important project, and, from a personal perspective, trusting me with the role of lead software architect. We were successful because of the outstanding team we had. It was a great opportunity.

In 2011 Motorola sold our group to Nokia Siemens Networks (NSN). NSN has an eNodeB, so they are phasing out the eNodeB that we developed. Eventually NSN was purchased by Nokia, so then I worked for Nokia (based in Finland). In addition to helping maintain the ex-Motorola eNodeB, I have worked on projects for the Nokia eNodeB. I made several trips to Ulm Germany and met some great co-workers there.

Nokia has decided to do some consolidation, which resulted in reducing staff and closing our plant in Arizona (the remaining Arizona employees are all teleworkers). When Tom Swendseid left Nokia he related a story I really liked. I won’t repeat it, but the moral was that in the face of these changes we can be bitter and angry, or we can be thankful for what we had. I choose to be thankful to Motorola/Nokia for many years of employment with interesting and challenging projects, and providing a means for me to support my family.

I’ve traveled a little over the years. In the government group I traveled to various sites in Washington DC, a site near Denver, regular trips for a few years to Baltimore, a conference in Monterey, CA, to present a paper titled “LPI Detection Using Spatial Correlation”. In the commercial group there were trips to the Chicago area (Schaumburg for Public Safety and Arlington Heights for Cellular). Tim Curtis and I were in Schaumburg the day the Diamondbacks won the World Series. A number of trips to Krakow Poland. Krakow is a wonderful city and I loved it there. Some highlights are Wawel Castle, the salt mines, Czestochowa (took a train there), Auschwitz. I especially loved all of the very old churches. Maybe Mary Ann and I will make a trip there. I also visited Ulm Germany several times. Ulm is a beautiful city along the Danube river (I have some posts on this website about my trips there).

The ride with Motorola over the years has been interesting. At its peak, Motorola employed 30,000 in Arizona, and 150,000 globally. There was a time at the Ocotillo plant where you couldn’t find a parking spot, or a free table in the cafeteria. Over time the new Board of Directors at Motorola decided to break the company up, either by spin-offs (e.g. Freescale, ON Semiconductor) or sale (e.g. General Dynamics, Google, Emerson, Nokia). I worked in several Motorola facilities in Arizona (e.g. Hayden, Roosevelt, Ocotillo, Diablo), and was involved in some others (e.g. Price and Elliot, ASU research center). Now those buildings have all been sold, and there is no more Motorola employment in Arizona. Global employment for Motorola is now about 15,000.

Well, that’s enough reminiscing! If you made it this far in the post, I’m surprised, and who would want to read more? 🙂

Many thanks to all who have put up with me over all these years. You have worked with me (and some against me), supported me, encouraged me, inspired me, and become friends.