by Brian Dear
One day in July while they were waiting for their wives to pick them up at CSL they got into a discussion. They understood why Lear Siegler was using resistors, but it was expensive and difficult to go that route. In the early 1960s it was hard to find very high-resistance resistors that would do the job anyway. For the team to ultimately build a 512 x 512–pixel display, it would require 262,144 resistors embedded in the glass. It did not seem like the right path to take. “Clearly Engelbart looked at this,” says Roger Johnson, who would eventually join the plasma display team as a grad student, but Engelbart “did not quite grok the fact that he could use this charge buildup. He saw this charge buildup as a problem. ’Cuz he was still thinking DC [direct current]. So he went down that path…and that turned out to be possible to demonstrate, virtually impossible to manufacture.”
Gene Slottow working on plasma panel prototype, c. 1967 Credit 16
“Then we had a much more novel idea,” says Slottow of his afternoon discussion with Bitzer that day in July. “If we simply moved the electrodes from the inner surfaces of the glass to the outer surfaces, we would isolate each cell from the external circuits by a capacitive impedance instead of a resistance.” By moving the electrodes outside the sandwich instead of inside, they would not come into contact with the gas at all, which in theory might reduce the undesirable sputtering and arcing effects. They could skip resistors entirely, and instead use the physical sheets of glass as a capacitor. Plus, instead of using DC voltage as Lear Siegler had done, they’d use alternating current (AC). Not ordinary household AC, in which the current alternates sixty times per second. No, they had to crank up the AC for the plasma to nearly 100,000 cycles per second. When the gas glowed, it was actually flickering 100,000 times per second: so much faster than the human eye could notice that the glow seemed stable.
They tasked Willson with creating a new sandwich of glass sheets and electrodes, sealing the sheets of glass once again with Torr-Seal, a powerfully strong epoxy glue commonly used for high vacuum seals, but this time they had Willson place the electrodes on the outer surfaces of the glass sandwich. The gas would not come into contact with them, as they would be on the other side of the glass. Then they backfilled the tiny hole in the center sheet of glass with neon, and fired it up.
It glowed blue.
They had a one-pixel cell that seemed to exhibit inherent memory. If you sent a sustaining voltage through the electrodes, then spiked it higher for a brief moment, that spike of electricity excited the ions in the gas enough for them to light up. You could then throttle back to the sustaining voltage and the gas would still glow. This was the memory effect they had long sought. You could then dip the voltage down a bit, then back up, and that was enough to calm the furiously excited ions down enough for them to stop emitting photons, so the glass stopped glowing.
On and off. One and zero. Memory. They’d done it.
But: it glowed blue. It should have been orange. The blue indicated contamination. Contamination almost invariably meant one thing: a leak, something they, particularly Willson, were terribly used to. When they had backfilled the neon into the little glass sandwich, something else must have snuck in as well. Most likely it was a tiny amount of ordinary air from the room they were working in.
Knowing that air contains about 78 percent nitrogen, 21 percent oxygen, a tiny bit of argon and water vapor, and tinier bits of other elements, including carbon dioxide (more in today’s air than that of 1964, thanks to humanity’s impact on the global climate), they assumed that most of the contamination would be nitrogen, since that’s what the earth’s air is mostly made up of. They confirmed that there was contamination and set out to prevent it. “So we fixed that leak,” says Bitzer, then backfilled the glass sandwich with neon again, “and boy, now watch it work.” And it didn’t work. “The memory disappeared!” he says.
A number of experiments and measurements ensued, revealing hints as to why the memory effect came about when there was contamination. There was something happening because of some small amount of gas mixing in with the pure neon. The scientific term for the effect is the “Penning gas mixture,” in which a noble gas like neon is mixed with an extremely small amount of another gas such as argon or nitrogen. The presence of the second gas changed the behavior of the ionization, which in turn changed the color of the glow, but also created a nice “margin” between the sustain voltage and the spike voltage, which in this case helped reveal the memory effect. They didn’t have all the answers yet, but by the fall of 1964 they knew enough that it was time to file a provisional patent claim.
Plasma Display Panel patent illustration, 1966 Credit 17
It would take another year of Willson making endlessly more prototype one-pixel gas and glass sandwiches, but by the fall of 1965 they were able to confidently build the prototypes and explain why things were happening the way they were. Much of it had to do with the fact that during each momentary half cycle of AC voltage ping-ponging across the gas many thousands of times per second, there would be a buildup of charge on the surface of the glass, a so-called wall charge. That tiny bit of buildup helped keep the gas cell “on” and glowing just long enough until the next full AC cycle would come around a tiny fraction of a second later.
With the Los Angeles Dodgers and Minnesota Twins in the 1965 World Series playing on Bitzer’s basement TV in the background, Slottow, Bitzer, and Nate Scarpelli, the attorney hired by the university, worked through two big patent applications. One for PLATO, and one for a fuller continuance version of the 1964 provisional patent application. That plasma patent, Scarpelli would say years later, was the longest, most complete, and most detailed patent application he ever filed in his long legal career. The U.S. Patent and Trademark Office took their time reviewing it. It would finally issue in 1971, and subsequent industry licenses would bring millions of dollars into the university, with a fraction doled out to Bitzer, Slottow, and Willson.
—
In the fall of 1965, Bitzer got a surprise visit, just when the new semester was beginning. Suddenly on his front doorstep at the laboratory appeared Brij Arora, one of the young IIT students he’d met in India in 1964. “He had flown himself over from India,” Bitzer says, “and was there looking for a job. He said, ‘I want to work on this stuff you’ve been talking about.’ ”
Unbeknownst to Bitzer, Arora had applied to the University of Illinois with the goal of enrolling in an engineering master’s or PhD program. “I had known Illinois to be a great university to continue further studies,” says Arora. Bitzer did not send him away. Instead, he made him a research assistant, and over the next several years Arora received his master’s and PhD in electrical engineering with Bitzer as his advisor. “When I joined in fall of 1965, I was not yet assured of a research assistant position. I had some idea that there was something possible in the computer science department. However, someone guided me to the Coordinated Science Laboratory. Yes, Don must have been surprised seeing me suddenly. I must admit it was providential. That moment was a turning point for me.”
Willson got to work writing his long-delayed dissertation, and Arora was tasked with attempting to build multiple-pixel plasma display prototypes. Another engineering student, Roger Johnson, had wrapped up his bachelor’s degree in 1965 and Bitzer talked him into pursuing graduate school to work on PLATO and the plasma display project, about which Johnson up to that time knew nothing. Bitzer had a way of finding bright talent, and when he spotted it he didn’t want to lose it. “He starts throwing me at the laboratory,” Johnson says, “and clearly had picked me, in some subtle sense, to join this team that was inventing the plasma panel. He had this Einstein lab up on the fourth floor. He hauled me up there one day—this is how he does stuff, it wasn’t like oh, I had to make an appointment to see Don, he was just like, ‘Come with me’—he hauled me up there and he’s pointing at this stuff which looked like it was right out of Frankenstein and I’m trying to figure out what the hell I’m looking at, I was trying to fi
gure out what the hell he was talking about.”
Then Johnson saw it. A little orange dot.
—
They’d mastered one pixel. The next step was a primitive matrix of 2 x 2 pixels. If they could figure out 2 x 2, then they could to go to 4 x 4. At each step of the way, they ran into new challenges. Roger Johnson eventually crafted a 16 x 16 display, a major step forward on the way to 512 x 512.
For the first few years, blue was the predominant color emitted by the charged plasma gas in the cells. Over time they tinkered with the color by adjusting the mixtures and types of gases they used in the cells. Indeed, the idea had not escaped them that if you tinker with the gases and introduce some phosphors into the cells, and then counted one pixel as a trio of cells, one red, one blue, and one green, you could in theory build a display consisting of arrays of those trios, and what you would have is a full-color display. (Luckily, Nate Scarpelli made brief mention of the idea in the plasma display patent.) For PLATO’s purposes, a monochrome display was going to be more practical and affordable. The question was, what color should that mono “chrome” be? They were used to seeing blue, but there was something about the orange color that they kept coming back to, that “sight to dwell upon and never forget.”
In time they got so good at understanding what they were doing with the prototypes and the intricate wiring of electrodes and circuitry to drive them and interfaces to computing devices to tell them what to display, it became time to think about how they were going to manufacture these things in quantity and at the crazy-ambitious resolution of 512 x 512 pixels.
By 1967 Alpert and Bitzer had chosen the Owens-Illinois (OI) company, wizards of glassmaking, to be the manufacturer of the displays. OI was a major U.S. manufacturer of so-called glass bottles, which were literally the “tubes” in cathode ray tubes. OI supplied the glass bottles to a number of the major TV manufacturing firms, including Sylvania. The company saw a profitable future in flat panel displays and televisions using the plasma technology. OI were primarily glassmakers, not electrical engineers, but they knew how to etch glass as well as print decorative lines and patterns on glass, so they would approach adding the 512 horizontal electrodes on the one sheet and the 512 vertical electrodes on the other sheet by “printing” tiny strands of gold onto the glass.
One of the problems with the prototype panels that Bitzer, Slottow, and company had built is that they all relied on the “glass sandwich” idea: two outer slices, the outsides of which were lined with the tiny electrode rows and columns, and an inner slice into which had been drilled tiny holes backfilled with neon and other gases. That was all well and good when you were building 1-, 2-, 4-, 8-, and 16-pixel grids of tiny dots. But to build a panel with 512 x 512 dots would mean that the center sheet of glass would have 262,144 holes on it. The chances that some of those holes would be faulty in some way would be high, not to mention how fragile the hole-ridden sheet of glass would be. OI decided to approach the problem in a novel way: get rid of the inner slice altogether, and fill the entire area between the outer sheets of glass with gas. Amazingly, it worked. When a horizontal electrode and a vertical electrode were charged up, the gas at the intersection point glowed, but no adjacent dots glowed. For this novel approach to the design of a plasma display, OI filed a patent, called the “Baker patent.” “I always considered it one of the major inventions of the plasma display,” says Larry Weber, then one of Bitzer’s graduate students, “but it went to court, and you never know what the court’s going to do, especially in regards to patents. It’s almost, you flip a coin to determine what the outcome’s going to be, it never seems to rely on logic.” In this case, the coin flipped in favor of the Bitzer-Slottow-Willson patent, deep in the text of which was a casual mention that the center sheet of holes is not necessary—you could take the whole center sheet of glass out and fill the space with gas and it would still work—which made the manufacturing process not only simpler and more reliable, but also economically viable.
By removing the center sheet of glass, OI engineers realized that there still had to be something inside that newly empty area in order for the outer glass sheets not to bend inward due not only to the low pressure of the gas relative to outside atmospheric pressure, but also from the fact that humans would be interacting with these screens by touching them, perhaps not always gently. They needed to be durable. Using microscopes, OI painstakingly attached tiny vertical “spacers” here and there within the glass, so when the two sheets were mounted together, the tiny gap remained uniform across the entire area and now had some supporting structure. Naturally, there was a price to pay with this approach: a user sitting in front of the terminal could see the spacers as tiny lines that faintly reflected orange light behind the glass. It was an engineering compromise that everyone would have to tolerate.
At first, OI produced a prototype four-inch panel, and secretly showed it to Bitzer and Slottow in 1967. It would take four more years to get to quantity manufacturing of the eight-inch 512 x 512 panels. OI made a strong case for Bitzer or Slottow to move to Toledo to work on-site on the project. Gene Slottow agreed to go. During those four years an enormous amount of work was done. One of the first urgencies was the discovery that the early OI panels were not very reliable, burning out in under one thousand hours of use (less than PLATO III’s Raytheon storage tubes, which they were meant to replace). That led to improvements in the type of glass used—Bitzer would call OI’s glass magic a “black art”—that led to longer-lived, more reliable panels. Another discovery was that the panels tended to need some coaxing to get going once they were turned on. Oddly, they needed light from whatever room the early prototype terminals were in. The physics explanation was that the presence of photons bouncing around through the gas helped excite it enough that the pixels would light up and glow. This led to the decision to add a border of pixels out at the edges of the panel—beyond the 512 x 512 one would see when using a terminal. Behind the frame, some additional rows and columns of electrodes were printed onto the glass, and those pixels were left on all the time. The photons from the light of those pixels sufficiently excited the rest of the gas that the panel would function correctly. That unseen lit border, behind the bezel when the panel was mounted in a terminal, contributed to the orange glow across the whole screen, particularly in a dark classroom.
PLATO users themselves called the effect “the Friendly Orange Glow.” Perhaps the best way to approach the notion is to think of a campfire. It’s not the fire itself. It’s what the fire does to the smoke and air surrounding it or to the faces of the people gathered near it. Anyone who has ever sat around a summer evening campfire knows that familiar glow—it’s in our DNA going back thousands of years. Something similar was going on with the PLATO plasma panel. It wasn’t simply the glow emanating from the text and drawings on the screen—that was only part of it. It was also the eerie, ghostlike effect that one saw in the black background of the screen. Because of that hidden border of always-on pixels on the four extreme edges of the screen, even the black background of the screen was not quite truly black. It had a faint glow of a campfire. No other computer screen ever had it. It was something that made PLATO special. And friendly.
After much discussion as to the merits of light pens versus touch screens, they made the decision to add touch sensitivity to the plasma screens in the upcoming PLATO IV terminals. Bitzer assigned a team of engineers—Roger Johnson, Fred Ebeling, Richard Goldhor (another of Bitzer’s Boy Scouts), and Jim Parry—to invent and build a touch panel. They came up with an array of light-emitting diodes that gave the terminal an invisible grid of 16 x 16 infrared light beams spaced just above the surface of the plasma panel display glass. If a finger were placed somewhere on the grid, the firmware in the terminal could pass back the proper coordinates to the central computer, which would consider it just another form of input from the student. To make room for the grid of beams, they had to set back the plasma panel by about a half inch from the front surface of t
he terminal, and include little holes around the resulting frame, where the beams of infrared light would be sent out. They created a mold into which they melted plastic pellets so that they would take on the exact shape they needed and fit snugly around the plasma display. Only—the frugality finally catching up to them this time—they used the oven in Maryann Bitzer’s kitchen to melt the plastic, stinking up the house and nearly ruining the oven.
7
Two’s a Crowd
Larry Stolurow, the UI professor of psychology who had in 1960 participated in the CSL committee considering an automatic teaching project that turned into PLATO, had turned down Alpert’s offer to lead the project once it was green-lighted. Nevertheless, he kept an eye on the system as it evolved from PLATO I to II to III. Stolurow liked that PLATO collected data on students, but disliked how it was collected. PLATO associated data with the terminal’s ID number rather than the name of the student. If a student stopped halfway through a lesson, then came back later to finish, but on a different terminal, the fledgling system could not yet handle that: there would now be two separate recordings of a single student’s work, and merging them was difficult. “That led me to a lot of frustration,” Stolurow says. “And the other aspect of my frustration was in the fact that the materials that they were working with on the system were very, oh, should I say, very elementary.” Stolurow wanted to collect richer data on student responses and response times, in order to help prove or disprove various theories about student performance. “We didn’t have answers to those questions,” he says, “and I felt we needed them in order to really design a system that would be effective and do the job that was intended.”
