“High school students could do a better job!” a NASA executive said to us after the launch of the Solar Maximum Mission satellite (SMM). He was angry because the satellite could not tell us its attitude, which is satellite-lingo for how was it is oriented within its orbit. It was February 14, 1980, and we were at the SMM Mission Control Center in Greenbelt, Maryland. All this would have been interesting as many problems are to scientists, except that he was yelling at me and my team — the SMM Attitude Support group. We were far from high school students since all of us had PhDs, principally in physics and astronomy. SMM had a big attitude problem and we needed to do something about it, or else SMM would be a very expensive piece of space junk!
This is a story about SMM from both a personal and technical perspective. It is a story of what we did that day to bring the mission back on track. Ironically, high school students can frequently have attitude problems, but how can satellites? This story is an illustration from both a satellite and human perspective of the fundamental truth that the map (or model) is not the territory. Satellites use internal maps within their on-board computers of what they expect to see to perform its mission, and people use mental maps within their minds to manage their world.
Solving SMM’s problem was critical to me since this was my first position as a physicist in the commercial world after ten years in the academic world. It was also critical to me since I was primarily responsible for the mathematical models at the heart of SMM’s problem. How did I find myself, under pressure, in that mission control center on that day? About three years earlier, I was a newly minted PhD in mathematical physics. I had been in college since 1968, except for a brief two year interlude when I was drafted from graduate school into the US Army and spent a year as a medic, medical lab technician and preventive medicine specialist (in that order) in Vietnam (someone mistook “physicist” for a “physician”).
When I graduated in 1977 with a PhD in Physics, I dreaded spending any more time in an academic environment so I searched for a commercial job. If it hadn’t been for my stint in Vietnam, I might have spent my life as an academic. My experience in the Army, especially when I was in Vietnam, made me feel that there was more to life to experience in the world than from the campus of a university. I wanted to find something new and more exciting (which did not include a war zone). I guess I eventually did, although not as expected. Being put on the spot with the rest of my team to solve SMM’s attitude problem was definitely exciting, maybe more than I bargained for.
What if you were like SMM?
To better understand SMM’s problem let’s first step into a similar experience. Imagine you are launched into space wearing just a spacesuit (well maybe you also have some clothes on underneath). You are in a stable orbit, meaning you are not going to fall to the ground anytime soon. However, all is not well. You are tumbling through the orbit, feet, arms and head pointing in all sorts of directions. You have built-in sensors in your face plate. The sensors feed information into a computer in your backpack to help you orient yourself and navigate. The sensors are supposed to use the earth magnetic field and the sun’s position to help you get your bearings. When they are working properly, you can adjust your thrusters to glide (and not tumble) through your orbit, much like superman or supergirl with hands out-stretched looking down at the earth.
You have a sun sensor which gives you a rough position of the sun, and another sensor which roughly senses the Earth’s magnetic field. This information is not precise enough or complete enough to determine your orientation. If you can’t determine it precisely enough you can’t control it. You do have very precise star sensors on either side of your waist. But, something is wrong and the star pattern they sense does not match the pattern stored in your computer. The world is distorted and you cannot get your bearing enough to stop tumbling. You wonder if you went through a worm hole into another galaxy, where the stars would look vastly different.
The SMM satellite, like you in your spacesuit, was supposed to get a fix on known star patterns to help it refine its attitude. Unfortunately for SMM, the star patterns it was seeing were not the ones it expected to see. That was the problem that angered and frustrated the NASA executive.
To See What No Satellite has Seen Before
SMM’s mission was to study sunspots and solar flares. Sunspots are dark spots on the sun caused by fluctuations in the sun’s magnetic field. Solar flares are energetic explosions on the sun can last from minutes to hours and come from the release of magnetic energy associate with sunspots. Solar flares can send energetic particles toward the earth. These particles can cause the colorful northern lights (aurora borealis). These particles can also harm astronauts, airline pilots, sensitive instruments on satellites, and cause problems with radio transmissions. So it is useful to better understand them. Solar flare and sunspot activity peak every eleven years, and NASA launched SMM in 1980 during such a peak to study these and other solar phenomena.
Knowing a satellite’s attitude is needed to do scientific experiments. Using it we can accurately point the satellite and its scientific instruments toward the object we are studying. For example, the telescope on the Hubble satellite needs to be pointed at a star or galaxy astronomers want to observe. In our case, we needed to point SMM’s scientific instruments at our own star, the sun. SMM, like the human body, was equipped with various sensors to help us know how it was oriented with respect to its surroundings. Similar to our imaginary spacesuit, SMM had sensors to help determine and refine its attitude. It had two sun sensors to coarsely detect the location of the sun, another to more finely detect the location of the sun, two star sensors that could identify background stars by sensing star patterns, and a magnetometer that could sense the magnetic field of the earth. The SMM mission support team used these sensors to determine SMM’s attitude. The procedure was to first get a course determination of the attitude using the coarse sun sensors and magnetometer data and refine it with star sensor data. The refined attitude was needed to operate the fine sun sensor, which can pinpoint a solar flare.
Knowing the attitude also helps manage a satellite’s power needs. If the satellite had solar paddles (like SMM), we would like to be able to make sure those panels were oriented so that the maximum amount of sunlight fell on them. Also, knowing the attitude can help keep the satellite healthy by re-orienting it to protect any sensitive instruments from the effects of solar storms.
Back in 1977, it took me awhile to be able visualize and mentally manipulate a satellite’s attitude in three dimensions. However, today’s children (and adults) who are good at assembling complex Lego kits would be great at these mental maneuvers because Lego instructions are just wordless pictures you need to visualize in three dimensions to be able to assemble the kit.
Each of the two star sensors on SMM looked at a small patch of the sky and matched a pattern of stars in its view with a pattern in its small on-board star catalog. The goal was for each star sensor to pick a known “guide star” and use it for a while until it needed to pick another one when the first moved out of the star sensor’s limited field of view. If each star sensor could anchor itself onto one bright star we could nail down and refine SMM attitude with help from the other sensors. The problem after launch was that the star sensors could not identify any stars.
A Stellar Solution
Luckily, our immediate NASA supervisor told the executive to leave us alone while we huddled in a small room to uncover the problem and propose a solution. A short time later we did just that.
The problem came from how the star images in the onboard star catalog were associated with images impinging on the star sensor’s field of view. It was as if the star sensor was looking at a mirror image of the stars that were further rotated by 90 degrees. To illustrate the problem take a square plate of glass (or waxed paper) and place it on a table. Then, draw the pattern illustrated by the figure on the left.
Lift the glass from the table, hold the lower left corner of the glass, turn the glass over, and bring the corner you are holding to the top right corner. Or, alternately write the image on a piece of paper, hold the image up to a mirror, and rotate it by 90 degrees. The figure illustrates the result.
To solve the problem, we modified our star sensor algorithm by taking a mirror image of the star sensor data and rotating it by 90 degrees.
To this day, I am not sure how the incorrect algorithm was implemented. We conducted tests on the algorithm before launch, and the government conducted acceptance tests. Was there a miscommunication between us and the star sensor manufacturer? Luckily, the correction was simple to implement, upload to the satellite, and test. SMM went on to become a very successful scientific mission.
After I wrote the above description of SMM’s problem and its solution as I remembered it, I searched on the internet for a historical description. Given these events occurred over four decades ago, I had long since discarded any SMM documents that I worked on. However in 1995, NASA added a few of these 1980 SMM documents to its on-line archive. One document “Attitude Ground Support System for the Solar Maximum Mission Spacecraft” was written by Gopi Nair, one of the people on our team. He described the incident in few paragraphs. He used a bit of jargon, but the gist of the issue and its resolution is clear (I added some explanations within brackets):
“Several attempts to acquire fine roll attitude using the primary procedure [using the star sensors] were unsuccessful. Analysis showed that this was due to a misunderstanding in the definition of the star tracker coordinate system [as I illustrated with the image of “STAR” above]. This was verified by a careful analysis of the star motion in the camera fields of view during small slew maneuvers [moving the satellite]. After some investigation, the appropriate FHST [star sensor] documentation [from the manufacturer] was received and the correct tracker coordinate definition was established.”
The spacecraft roll attitude was established immediately thereafter.” It is nice to have an official sounding description for a situation that caused a little panic.
The Map is not the Territory
SMM had a distorted map of the stars; people can have a distorted map of reality. Consciousness operates at different levels of abstraction. People express the surface level by deleting, distorting and generalizing the deeper levels. The distorted perceptual view of SMM’s star sensors reminds me of experiments conducted with people wearing glasses that made the world appear upside down. After a few days people eventually were able to function almost normally. Here is a description of the results of such an experiment done with three people who wore the glasses all day for 7 to 10 days and were blindfolded at night (Linden, et al, Perception, 1999):
“All subjects showed a rapid adaptation of motor skills. On the third day of the experiment, they were capable of walking freely without a stick. Subjects performed all tasks of everyday life with none or minimal aid. During the second half of the experiment, they were able to find their way in a crowded department store and to ride a bicycle.”
When their glasses were removed, the participants took a while to re-adjust to their normal vision. This is a situation, unlike SMM, where people were able to deal and function with a distorted view of the world because their brains were eventually able to adjust to the situation.
Sometimes our distortions get us into trouble when we incorrectly assume things are true, and we behave accordingly. We create frames of reference that don’t serve us well. One frame is the blame frame, where we continual seek a primary cause for an unhappy situation. This behavior is evident in political tweets I see every week. In truth, there may be multiple causes, and the emotional states we use when blaming are not necessarily conducive to clearly determining the problem or finding a solution.
During our SMM crisis in February 1980, I don’t recall anyone blaming anyone else, before the crisis or after it. There was no time; we all wanted the mission to succeed. “Blaming” is an example of a cognitive distortion, which is a pattern of behavior that convinces us of something that is not true, usually reinforcing negative thinking or emotions (see my resource map of cognitive distortions). When I said earlier that the NASA executive was angry and frustrated, I was exhibiting another cognitive distortion called mind-reading. We mind-read someone when we project our beliefs and thoughts outside ourselves onto that person without checking to see if they are true. In fact, I had no idea what the executive was feeling, and I was not about to ask him: “Excuse me sir, what are you really feeling now?”
When coaching clients, it is important to avoid mind-reading to understand the client’s map of the world and minimize the effect of our own maps. One way we do that is by accessing a “know- nothing state”, where we assume we know nothing about the client’s world. Then, we ask the client enough questions to flesh out his or her map.
By labeling us worse than high school students, the NASA executive was apparently expressing his frustration about the problem, perhaps hoping his negative assessment of our progress would motivate us to succeed faster. I felt more encouraged by our immediate NASA supervisor who suggested we go off to a room by ourselves and think about the problem. He had more of an outcome-oriented frame: stating what he wanted in the positive and providing the means for us to achieve it. This is the function of the well-formed outcome (WFO) questions (see the associated WFO resource map) using during a coaching session. They provide a tool to help understand what a person really wants, and if what they want is consistent with their values and highest intentions.
Postscript on SMM
There was another bump in the road for SMM in December 1980, when a fuse blew within the satellite’s attitude control system. This caused SMM to wobble making it impossible for it to continue its scientific mission. At that time I remember people joking: “Why would any sane person put a fuse on a satellite? Was someone going up there to change it?” Actually, SMM was designed to be serviceable and was placed in a sleep (“standby”) mode for three years until the Challenger Space Shuttle was launched and astronauts moved SMM into its shuttle’s payload bay to repair it.
Appendix: SMM Documents
I was surprised to find the following SMM documents, published in October 1980, at a NASA site because they were created long before the internet we know was established. They were also written with those ancient “typewriters.” NASA uploaded them to the site in 1995. They are of historical importance and also helped me jar my memory.
“Algorithms for autonomous star identification” Primary Contributor: P. J. Gambardella; Other Contributors: R. Byrne, M. Shuster, G. Lefferts, and R. Werking
“In-flight calibration and performance evaluation of the fixed head star trackers for the solar maximum mission”, R.H. Thompson, P.J. Gambardella
“In-flight calibration of the fine pointing Sun sensor on the solar maximum mission” P.J. Gambardella and R.H. Thompson