No Software Required: How Atari Made Its First Video Games Without A Computer

Paul Monopoli examines the early history of a future gaming giant

In the early 1970s, computers were expensive! Like, really expensive, man! So if you were a young Nolan Bushnell entranced by Spacewar and you wanted to bring it to the masses, how could you do it?

Throughout his career Nolan Bushnell has demonstrated the ability to take an idea and turn it into a success. Nolan is a fan of science fiction and as a child he wanted nothing more than to live in the futuristic worlds he could only read about. His endeavours in entertainment show that he enjoys the fun side of science fiction, with gadgets, gizmos and items that make people’s lives enjoyable. During the 1970s and early 1980s Bushnell began his work to bring elements of sci-fi culture into the real world. His pursuits included Atari, which popularised coin operated and home video games and Chuck E. Cheese restaurants, which entertained diners in its pizza bar theatres.

Ever the tinkerer, at the tender age of 6 Nolan created a space ship control panel out of a crate and convinced his neighbours that they were under attack from a UFO of his own creation. When he finished school and made his way to college, one of the majors he chose was engineering. He had seemingly done enough of it in his youth, so it was time to see if he could make a career out of it.

While he was studying he claims to have come across the first ever video game, Spacewar. He spent many hours in the computer lab, playing the game as often as he could. Bushnell was able to see the hold that the game had on himself and others, and considered the possibility of releasing it as a coin operated machine. After graduation Nolan found a job as an engineer at Ampex where he met future collaborator, Ted Dabney.

Bushnell introduced Dabney to Spacewar, and while the young engineer wasn’t as impressed as his friend, he was never one to turn away from a challenge. Bushnell wanted to create a simplified version of Spacewar that could be mass produced. The duo started by looking into parts that could be easily accessible and cheap.

The project took some time to get moving and as the days turned into months Dabney had started to give up on the idea. On the other hand, Bushnell was as enthusiastic as ever and started to toy with image manipulation on his television. He went to Dabney and posed the question of why the vertical sync on a TV set can make the image move and up down. A follow up question of whether this can be controlled led to Dabney looking into whether it was possible. The men continued their work in the bedroom of Ted’s daughter, though some sources claim it was Nolan’s daughter who was kicked out of her room. Either way, not even the men’s children could get in the way of their masterplan, and the girl was made to sleep on the couch.

Unknown to him, Nolan’s discovery had also been made by another engineer known as Ralph Baer. Baer and Bushnell would cross paths numerous times in the following years, each taking inspiration from the other. For now, while Baer was focused on home entertainment Bushnell was looking to develop the first ever coin operated video game.

While Bushnell appeared to drive the project and had the original idea, it was Dabney who worked on the primary electronics and engineering. His big hurdle was that the analog signal he thought could generate the image would not be linear, something that could cause a problem with the timing. This presented new challenges as the engineer had to replicate the process using digital technology. Eventually Dabney’s original circuitry was able to display a pixel that could be moved around the screen by external control switches. The basic building blocks were now in place and Bushnell took over the project. Dabney was left in the dark when Nolan, by chance, discussed the project with his dentist. The dentist suggested that Nolan meet with another patient of his, Dave Ralston. Ralston worked for arcade manufacturer Nutting Associates, and after the meeting felt that it would be the perfect home for Nolan’s project.

In 1970 Nutting’s claim to fame was a game called Computer Quiz, a game that was released in a variety of flavours, including Sports Computer Quiz. The company pooled all of its resources into the project and while the game was popular at college campuses, it failed to pull in the general public. Founder Bill Nutting also had a problem when his chief engineer left the company, leaving them short handed. When the opportunity to manufacture Computer Space came up the company was in the throes of financial trouble. The situation meant that Bushnell didn’t have to try too hard to charm his way into a job, attempting to take Dabney along for the ride.

Initially Dabney didn’t have as much faith in the project as Bushnell, so he decided to stick with his secure, full time job at Ampex. Charlie Steinberg, the Ampex president, was concerned about Nolan and told him that it was a mistake leaving such a secure and highly sought after job to work with video games. Undeterred, Bushnell continued with his plans to leave. During negotiations with Nutting Nolan managed to retain the rights to Computer Space and received a 5% royalty on sales of the cabinet. Nutting would cover the costs of manufacture and Bushnell became their one man engineering department. He would work on Nutting’s existing projects during the day and continue to develop Computer Space after hours. In spite of his initial reluctance, Nolan managed to convince Ted Dabney to leave Ampex though he was still working on the game with Nolan after hours.

The end result, completed in 1971, was a fibreglass cabinet that contained a raised TV screen and a control panel with buttons. The game is reminiscent of Atari’s future classic Asteroids, but the ship has to destroy two UFOs instead of space rubble. Dotted outlines represent the ships on the screen, an expansion on the single dot technology that Dabney had developed. He was now able to use the controls to move multiple dots around the screen. The dots were placed and the circuitry on the board told the controls which dots to move. It was Bushnell who had an idea to make the motion of the ship as smooth as possible. He played the dots in an array that would show the ship at four different angles. These dots would be flipped vertically or horizontally if the ship was pointing in another direction.

The AI of the rival ships was extremely simple. The ships would not shoot at the player, but rather in the quadrant of the screen that the player was on at the time. It had to be basic as the game does not contain a CPU and uses no software. In 1970 this would have been prohibitively expensive. Instead the game uses TTL or Transistor to Transistor logic chips. These were connected to a 16 inch monitor that contains five vacuum tubes. The simplicity behind their design meant that this new ‘computer’ was only capable of playing Computer Space.

When the design was complete Bushnell could see a Computer Space machine standing along side a pinball table at every club, arcade or his own line of amusement restaurants. Chuck E. Cheese was already a germ in Bushnell’s mind in 1971 but the decision was made to focus on the new game for now.

After testing at a few locations the game was showcased at the Music Operators of America expo in October 1971. The cabinets that Bushnell and Dabney prepared for the show were found to have been damaged in transit. The monitors had fallen out of the cabinets. The duo managed to repair three of the cabinets but the fourth was unable to be fixed. Bushnell had an idea. If he was going to show the game off to the buying public then maybe they should see how it worked. The fourth damaged unit was opened up so that potential buyers could see how simple the inner workings of the game actually were. Bushnell was reported to be extremely enthusiastic at the expo, talking the ears off anyone who would pass by their booth, whether or not they were actually interested.

While Bushnell claims that no sales were made during the expo, Nutting’s accounts tell a different story. Bushnell was suddenly the company golden boy and he even earned a ride in Bill Nutting’s personal vintage biplane. Reports of Computer Space’s success vary, with critics raving about the game but Nutting management being less than impressed over sales. Bill Nutting’s increasing disinterest in the company is one of the suggested reasons for the lack of sales. While exact numbers vary depending on the source you read, the game was considered a failure by Nutting Associates.

Reports vary, but most outlets report that 1500 cabinets were sold after the game’s release. A paltry number, but in 1972 newspapers claimed that Nutting was increasing production due to demand. Though records state that production did increase, it might not have been on as large a scale as was reported at the time. Regardless, the game had a hard core fan base and Bushnell was given a lucrative job offer which Dabney managed to talk him out of.

The main feedback from the public was that the game was too complicated. In a world of easy to play pinball machines the average gamer didn’t want to spend a couple of minutes reading instructions. They just wanted to insert a coin and start their game. Bushnell took all of the feedback on board and approached Bill Nutting for another shot at it. If he could simplify the game then he knew he would have a hit on his hands. It sounded great, but Bushnell wanted something extra for his trouble. He would create a simplified Computer Space for a one third owner ship in the company and a larger role in managing it.

Nolan Bushnell is a personable and charming man, but his status with Bill Nutting was dropping fast. His counter offer was 5% of the company and his continued services as an engineer. While he was no longer the golden boy with the boss he still had his partner, Ted Dabney. The duo met and discussed the glass ceiling they had hit at Nutting Associates and decided it was time to leave and go it alone. They would take their business and make it a greater success than Nutting Associates ever was.

The duo had already started their business in 1969, but there was a problem with the name they chose. Bushnell and Danbey weren’t aware that the name Syzygy, the nearly straight line of 3 celestial bodies in a gravitational system, was already taken by another company. Nolan and Ted were keen chess players and this interest led them to play other competitive games, one of which was the Japanese board game Go. Like Chess, one of the positions in Go is ‘check’ where the opponent is forced to make a move that will save themselves. In Japanese the word is ‘Atari’, and this is the name that was chosen.

Bushnell understood that the long winded instructions and controls of Computer Space were its downfall, so the next game needed to be immediately accessible. Nolan imagined people playing the game with one hand and holding a glass of beer in the other. Meanwhile, future Atari employee, Steve Bristow, was working with Nutting on a two player version of Computer Space. There are conflicting reports over Bushnell’s involvement with the project, but as he claimed no ownership over it (something that would have been unusual for him at the time) it’s more than likely he was too busy focusing on his new company and his next project.

In a 2012 interview with Retrogamer Magazine, Ralph Baer claimed that Nolan Bushnell had seen his Odyssey console at the Magnavox dealership in California in March of 1972. He had taken an interest in the Table Tennis game that was built into the system. Several witnesses have also gone on record claiming to have seen Bushnell at the show. Taking inspiration from this experience Bushnell considered his next move as he hired a young engineer he had worked with at Ampex to help him. The man’s name was Allan Alcorn.

At the time Bushnell left Ampex, Alcorn, like many others, was concerned for him. No one ever left Ampex, it was a job for life. Later, when Alcorn received the call to join Nolan he accepted, though in the back of his mind he always assumed he could return to Ampex if he had to. In the meantime this might be a fun little project to work on. Nolan was an interesting guy and the two men had gotten along very well.

Alcorn was initially bought in to work on a driving game, though he thinks this project was only used to entice him to work at Atari and didn’t actually exist. When he started work Bushnell had something else for Alcorn to do. He told Allan that Atari had signed a deal with General Electric to create the bat and ball game, though this was also a ruse. The exercise was actually intended to help Alcorn become familiar with the process of creating video games while Bushnell himself worked on a more complex idea. Presumably once Alcorn had completed his side project then Nolan would be ready with the next big game, by which time the young engineer would have some experience in game design. It was a perfect plan, but often the best plans can take on a life of their own.

Alcorn had access to Bushnell’s original designs for Computer Space and he was familiar with the workings of the TTL circuitry that was used. Bushnell’s documentation of Computer Space probably made complete sense to him, but for everyone else it was practically illegible. Alcorn tried to decipher the mess he was presented with but in the end decided to use his own knowledge of the circuitry and his understanding of Computer Space to work on the project.

Pong did not use a CPU or software to run the game. Instead Alcorn improved on the existing technology used for Computer Space. Like the former game Alcorn started with working on the horizontal and vertical sync pulses. Digital timing circuits were developed that turn objects off and on, depending where they are on the screen. Some elements had to be combined, such as the counters only being allowed to tick over when the ball is displayed

While Alcorn worked on the project for three months, the basic building blocks of the game were developed within a week and a half. He tested the game and found that it was boring. It was certainly simpler than Computer Space, but probably a little too much. He thought about ways to punch up the gameplay and started to work on improving the basic bat and ball concept.

Alcorn split the paddle into 8 sections. When hit, the three sections on either side would send the ball back at a different angle with the centre two knocking the ball straight ahead. The ball itself has three basic horizontal speeds, with the ball getting faster the longer each round lasts. Restarting a round resets the speed. The paddles are unable to make their way to the top of the screen as Alcorn discovered limitations in the circuitry. In the end he decided to make it a feature, as a talented player could angle the ball to go over their opponent’s paddle. This also ensured that rounds would not last forever.

Alcorn admits that he kept making improvements as he was not getting positive feedback from Bushnell, who loved the game but wasn’t forthcoming with his praise. When Bushnell and Dabney did give Alcorn feedback they asked him to put in some sound effects. Nolan wanted the sound of a crowd cheering when a point was scored while Dabney asked for booing when a point was lost. Alcorn was running out of space on the circuit board, so he had to use the existing on board sync generator to create these effects. He claims that this took him half a day to complete.

Unlike Computer Space the only instructions contained on the original cabinet read, ‘avoid missing ball for high score.’ The control mechanism was a simple circle, which when spun to allow the on screen paddle to move up and down. According to Alcorn, initially the Pong prototype was installed with cheap potentiometers to control the paddles. After they broke he calculated the average amount of turns a person would make per game and how many turns the cheap device could withstand in the real world. Eventually the decision was made to upgrade to a more expensive potentiometer, supplied by a military contractor.

Prior to its release in 1972 Bushnell approached his former employer at Nutting Associates. He proposed a deal whereby Nutting would develop Pong, Atari would retain the rights and receive a 10% royalty on each machine produced. Bill Nutting turned his former golden boy down, a decision he would soon regret. Five years later Nutting was sold, but not before they developed their own line of unlicensed Pong clones.

Famously, the test unit for Pong was installed in Andy Capp’s Tavern. Apparently not only named after the famous British comic, the tavern was allegedly known for its ever flowing beer taps and smoke filled interior. Alcorn was unsure about this move as he felt that the cabinet wasn’t robust enough to withstand much punishment. A simple knock on the side could damage the machine or the silicon chips could burn out.

Urban legend tells that the Pong machine was so popular that it broke on the second day. The bar owner was reportedly quite angry at this, but Alcorn claims that the whole story was just a myth. From the money that Nolan had made from Computer Space he was able to buy some existing machines and install them at local arcades and bars. Atari already had machines at Andy Capp’s Tavern and the owner was well known to them, hence their decision to test the machine there. The famous phone call came about a week or ten days later. Surrounded by upset patrons who wanted to play a game Alcorn opened up the machine and discovered the overflowing coin box. There was roughly $100 in quarters, meaning that around 400 games had been played during the week. When told about this Nolan simply responded by saying, ‘that’s interesting…’

Andy Capp’s Tavern had regular customers who would pile in at 9AM after breakfast and start drinking. However, one morning some new faces entered the bar along with the regulars. They didn’t order any drinks, but instead headed straight for the Pong machine. They inspected the machine as much as they possibly could without opening it up, then they left. From their observations they developed a clone of the game, an achievement Alcorn says was admirable considering they hadn’t even seen the circuit board.

Nolan was enthusiastic about the game, and in typical Nolan style he spoke to everybody who would listen. Bally and Midway both expressed interest in the game, but Bushnell was considering his options. He manipulated both companies into thinking that the other was no longer interested and they both backed away.

With Nutting, Bally and Midway all out of the picture there was no one else who could develop the cabinets. He decided that he would develop them himself and a line of credit was granted by Wells Fargo. Nolan expanded Atari to include a new warehouse and assembly line for the machines. In late 1972 the game was released across pubs and clubs around the USA, and eventually other parts of the world. It was a resounding success as Bushnell knew it would be.

Nolan’s original plan was for Alcorn to develop games after creating his ping pong game, and this hadn’t changed. Alcorn worked with Dabney on Atari’s second game, Space Race. Originally called Asteroid, Space Race was conceived by Nolan Bushnell and Ted Dabney, though Alcorn engineered the final release. Both he and Dabney claim to have designed the game. The graphical techniques used are similar to those found in Computer Space. The whole game was designed to be a simpler version of Nolan’s original vision, but with a race between spaceships rather than a shooting match. Both Dabney and Alcorn claim that Bushnell had some input into the project, though how much is up for debate.

The game bombed as players demanded more Pong. Shortly after release the market was flooded with clones, though in all of the excitement Atari themselves seemed to forget that they had made a clone of the Magnavox tennis game, a fact that Magnavox themselves hadn’t forgotten. Certain parties claim that at this point Magnavox was already attempting to approach Atari over this fact, but it would be a bit longer before the lawyers were called in.

In late 1973 Atari suffered a few failures with Space Race and its successor, Gotcha. Atari recalled many of the Space Race cabinets and outfitted them with their new game, Pong Doubles. Like the real game, now four players could partake in video table tennis. 1973 also saw Ted Dabney leave the company he co-founded. According to Alcorn, Atari was expanding at such a rate that everybody had to perform different jobs, including manufacturing and shipping. In the end Dabney simply wasn’t interested in the business side of Atari, so he left. Alcorn claims to have been heartbroken over his decision.

Dabney was also upset over the plethora of Pong clones that had swamped arcades. Alcorn claims that of the twenty-odd thousand that were made only three thousand belonged to Atari. Nolan Bushnell claims that Atari sold many more, but history shows that there were more clones than genuine Pong cabinets in pubs and clubs around the world. Worse yet, almost all of these featured clones of Atari’s board inside the cabinets, circuit for circuit.

Even though the world was still obsessed with Pong, Allan Alcorn was ready to move on. He continued working for Atari but others oversaw development of the Pong expansions. Even Bushnell was wary of Atari being a one trick pony and he inspired his staff to work on new ideas and innovations. It was Atari engineer Harold Lee who came up with the idea of a home Pong system. In 1974 Atari started work on the project. Alcorn already had the game running on a standard black and white TV, so the big problem was taking his design and shrinking it down. He assisted with the project while working on other games.

1974 was also the year that Magnavox decided enough was enough. They had attempted to contact Bushnell on numerous occasions, and it was time to bring in the lawyers. They argued that Ralph Baer had come up with the original TV tennis game, and Atari was infringing on their copyright. Unlike the well organised Baer, Bushnell did think too much about patents and licensing and in 1974 Atari was still considered to be a new company. It didn’t have the financial clout to withstand Magnavox and its legal team.

Ever the charmer, Nolan was able to hammer out a deal with Magnavox which was beneficial to both parties. Atari would pay a one off licensing fee of $700,000 and Magnavox would get the right to any Atari product that was released for the next 365 days, starting in June 1976. During the year Magnavox lawyers made many visits to Atari HQ, though everybody kept quiet about any new developments. Nolan’s negotiations saw that the Magnavox deal would end a week before the next CES, meaning that Atari could announce new products without sacrificing the rights to Magnavox.

In 1974 Bushnell decided to focus on running Atari, leaving development of new games to his engineers. Bushnell was the ideas man, and he had been burned by the Magnavox settlement. He decided that he needed to stop being so blasé about the business end of Atari and focus on understanding the industry. He started to apply for patents that would protect existing Atari products. He also pushed his team to work on new games, releasing several new titles in 1974. He didn’t want another Pong clone situation, so to keep the market fresh and beat the imitators he wanted a continual string of hits. Unfortunately for Atari it was the Pong clones that were continuing to receive all the attention, and Atari themselves released several upgrades to the existing game to meet demand.

Work was continuing on the home system, something that Bushnell hoped would put a stop to the clones. People wold be able to play Pong at home, meaning that when they went out they were more likely to look for a fresh experience. Atari had recently achieved success with a new game, Gran Trak 10, an overhead racing game. Bushnell had wanted to work on a driving game since Atari’s inception, but it was an idea he kept pushing aside in favour of something else. The game was starting to outsell Pong but there was a problem. Poor accounting records meant that the game was sold to vendors for $995 when it should have been $1095. Atari suffered a loss on what was supposed to be their next big hit. But there was a light at the end of the tunnel; the Pong home unit was completed.

Bushnell took the unit to toy shows, only to find attendees disinterested in the product. Bushnell started to feel a bit lost. Atari could sell the console themselves, but they were already stretching their resources with the arcade games they were developing. Eventually the retailer Sears approached Bushnell and offered to purchase every Pong TV game that Atari could make. Sears put up the money for a major marketing campaign and with the extra cash Bushnell was able to increase production. In 1975 Atari sales increased by $40 million, a turnaround from the disaster of 1974.

Rather than invest the money Bushnell decided to throw wild parties and buy expensive cars. The profits kept rolling in and Atari was going from strength to strength, but the Pong problem hadn’t gone away. Instead of clones appearing in arcades and pubs they were popping up on store shelves. While many of them were from small toy companies, some of them were from bigger names. At this time future gaming juggernaut Nintendo were starting to explore the video game market, a move which saw them releasing five of their own Pong clones between 1977 and 1980.

Bushnell was an innovator and releasing Pong systems on the TV wasn’t enough for him. Atari now had a few successful arcade releases that he wanted to get into homes. He needed a system that could play all of these games, and anything else that the company developed. In early 1976 he put Steve Mayer and Ron Milner on the project with Al Alcorn assisting them. They developed a prototype of their interchangeable cartridge system within three months which they codenamed Stella. The problem was the Magnavox agreement had just started. Al Alcorn claims that keeping the Video Computer System (VCS) a secret during the remainder of the allocated year was very difficult. After the agreement had ended Atari debuted the VCS at the 1977 CES to a surprised audience but Atari had already been beaten to the finish line.

Fairchild Semiconductor engineer Jerry Lawson had begun work on a home entertainment system, known as the Channel F. This would be the first home console to use a central processing unit and programmable ROM cartridges. In the latter half of the 1970s prices of semiconductors had significantly dropped in price, justifying their use for entertainment purposes. Atari themselves had already started using them in their arcade games in 1976. Without the Magnavox agreement it’s possible that Atari may have beaten Fairchild to market, though Atari did have the superior hardware and an established name. In the end only 27 games were released for the Fairchild Channel F, while the Atari VCS (later named the 2600) remained in production until 1992.

A new era of gaming had begun but while there were still rocky roads ahead, gamers were now able to have a taste of the arcades while sitting on their favourite couch. Home video games were here to stay.

Pong designer Al Alcorn says a few words about his creation.

Hi Allan, thanks for taking some time out to talk with us about Pong. Did you have any idea when you were developing it that Pong could ever be as big as it was?

I was only 24 years old and had no expectations for Atari and Pong other than a paycheck.  But the work was both interesting and challenging as raster scan displays, like a TV, were analog devices and the Pong logic board was a digital state machine. 

A “state machine”? Could you elaborate on that?

To understand how Pong was designed you need to understand how an NTSC display works.  The beam on the display scans horizontally at 15,750 Hz and the vertical scans at 60 Hz,  First I had to build a sync generator that used high speed counters that counted the 14.31 MHz clock down to the 15,750 Hz signal for horizontal.  It was a simple binary counter that allowed me to access all the intermediate counts so I could define any vertical slice by simply using logic gates to map them out.  Similar thing was done to the vertical that counted the horizontal scan lines.  By gating the correct horizontal and vertical counts I could map any region on the screen like where the paddles appeared or the net.

So what about the ball? It moved both horizontally and vertically…

To generate the ball and allow it to move anywhere on the screen I built another sync generator similar to the first sync generator but this sync generator was built so that I could change the count by +/- one, two or three.  These counters were gated so that it created a small dot or ball.  If I set the counters to have the same counts at the first sync generator then the ball would be stationary.  If I set the vertical part of the counter to be one count less than the first sync counter then the ball would move up the screen one scan line per frame.

How did the “gap” at the top of the screen come about?

The start of the paddle is set by a 555 timer controlled by the potentiometer.  The gap at the top is an error with my control of the 555 and I left it in as it allowed a way for the player to lose so they would not play forever.

How does the “english” on the paddle work?

The vertical speed of the ball is set by where the ball hit the paddle.  The paddle is 8 scan lines high and I capture the count of where the ball hit to set the vertical speed.  Three speeds up, three speeds down and one with no vertical speed.

Thanks Al!

…and Nolan Bushnell says fewer words…

When did you first come across Spacewar? What was it like? What did you think of it?

Graphics lab at University of Utah in 1965 or 6.  I was mesmerized.

Was it then that you had the idea to turn it into an arcade machine? Or later?

Because of my knowledge of the arcade business I immediately thought the game had arcade potential if the cost could be brought inline with the current coin op game cost.

Why wasn’t it possible to follow through on your idea?

That computer had a cost of several hundred thousand dollars.

How did you hit on the idea to use hardware circuitry instead of a computer?

Several years later I was working on my project and the computers were too slow so I kept having to build small circuits that would offload certain functions to free up computer time.  Once I offloaded enough I realized I could offload everything onto small digital circuits.

How were the graphics drawn / rasterised to the CRT?

We scanned out pixel by pixel and created dot matrices using diodes.

What was your relationship with Nutting Associates? How did Computer Space do and why?

I licensed Computer Space to Nutting and then they needed me to come on as chief engineer.  CS sold about $3M which seemed good at the time but in retrospect was disappointing.

And that led to setting up Atari?

Yes, I left Nutting and focused on Atari.

Thanks Nolan!

Be the first to comment

Leave a Reply