History of Project Support Engineering (PSE)

KyleF

Member
Joined
Aug 15, 2025
Messages
15
Reaction score
67
=== PART 1 ===
Several forum members asked that I post a thread with some historical knowledge I have on PSE. PSE was a small, pioneering company that developed several arcade games at the dawn of the video game industry. This is written in 2025 from my memories of that time in the mid-1970s, so my apologies if there are any inaccuracies in the following narrative.

About me
My name is Kyle and I worked as a game designer for PSE. In 1975 when I was 17 years old and a senior in high school, I signed up for a vocational work program where I could get on a bus at noon and go to NASA / Ames Research Center in Mountain View, California. Not being a fan of school, I jumped at the chance to leave school while still getting full credit for attending. At NASA I was assigned to a small design team working on a top-secret project to build one of the first unmanned, remotely piloted drone aircraft. After graduating high school in 1976, NASA offered me a job as an associate engineer working on the same project. There was a hiring freeze in place, so I was hired as a contractor. PSE provided contract services to NASA, so NASA asked PSE to put me on their payroll. This was my introduction to PSE.

The NASA project was cancelled a year later and PSE offered me a job working at their office. I had built an early solid-state video camera at NASA and taught myself to program in Fortran, so I had experience in the fledgling field of digital video design. This fit nicely with the work PSE was doing with its video game product line.

PSE background
I believe PSE was founded by John Chaudhry in the early 70s as a contract design and manufacturing company located in Sunnyvale, California. Basically, PSE would design and manufacture electronics as a subcontractor to other companies. This is the reason for their name, "Project Support Engineering". They were located a few miles away from Atari, where Nolan Bushnell's team had released Pong around the same time.

It was a family business run by John, his wife and several grown children. The family dog, a large German Shepard, was also present and patrolled the hallways and offices. The building was perhaps 8,000 square feet with about 25 employees. It was split about one-half conditioned office space with the remainder an unconditioned manufacturing warehouse.

Game history
John was a brilliant digital design engineer out of IBM. He was responsible for the design of Knights in Armor, Maneater and their other early video games. This was before I started in 1977, but I remember one of their first video games was to take Atari Pong games and modify them to use 2-axis joysticks. This upgraded the game so the video paddles could be moved horizontally and vertically, instead of the limited up/down motion found in the original Pong. I did not see this game noted on the Museum of the Game website. I'm wondering if any of them still exist.

Knights in Armor and Maneater used very complex discrete digital circuits to create their video images. They had very large printed circuit boards containing over a hundred chips. This was a time when microprocessors were just being released, but PSE never used one. RAM memory was tiny – nowhere near the size required to produce a bit-mapped game image that's so common today. As with Atari Pong, the electronics in the PSE games creates the video images directly from discrete digital circuitry. There was no "computer", no programming in these early games – just hardware design.

Bazooka
After designing the crazy, complex circuitry for Nights in Armor and Maneater, John came up with the idea of a "programmable" game hardware platform. Keep in mind that computers at that time were large and expensive machines. Small microprocessors were just coming on the scene and Jobs and Wozniack were just starting to design the first Apple computer in their garage a few miles away. It's taken for granted nowadays, but a game you could change by programming was a brilliant idea for its time.

The Bazooka architecture provided a control board (Board 1) that allowed the game play to be programmed using PROMS (Programmable Read Only Memory) chips. Instead of having to design large, complex single-use circuit boards, the Bazooka platform could host a variety of unique games by just changing the program in the PROM chips. John hoped that this flexible architecture would allow the company to quickly release new games, boosting their sales in the fledgling, and often struggling, video game industry.

John tasked me with writing the "code" for the games, which involved creating the program stored in the two PROMs on Board 1. These were one-time programmable chips. If you wanted to change the code, you threw the chips away and burned new ones. I also designed the electronic hardware upgrades found in the later PSE games. The electronic architecture was essentially a simplified microprocessor, made from discrete chips. Programming was a manual process, done on a homemade wooden box with a panel having a row of toggle switches for each bit, and pushbuttons to store entries into a small RAM memory. There was no programming language for this, no keyboard, no display monitor. I had to memorize all of the unique binary numbers (opcodes) representing different software instructions. By hand I would enter each opcode into the box, a bit at a time. With a memory capacity of 512 lines, it was a tedious process. There was no editing or insert function, you could only overwrite previous lines of code. There was no permanent storage, the program could only be copied over to a PROM chip for production. If the power went off, you would lose everything and have to start over. Because of this, I wrote the program opcodes down on sheets of paper before entering them into the RAM box.

There would typically be three circuit boards in a game, plugged into a common motherboard with power supply. Board 1 provided the programmable "brains" of the game, along with the required video synchronization signals. Board 2 contained the unique digital circuitry required to produce the actual video images. Board 3 contained the audio circuitry and interfaces for the coin switches and other game controls.

1757188899482.png

The Bazooka game play consisted of military vehicles crossing the screen from left to right. A small mock-up of a bazooka gun was mounted on the game console that allowed the player to fire a shot at the vehicles to score points. You lose points if you accidentally hit the stretcher or ambulance.

1757188922723.png

Digital scoring is displayed at the bottom of the screen, produced by Board 1 circuitry. The text above them is actually made by reversed-out stickers stuck to the face of the display monitor. Computer-generated text was not yet a reality.
The actual bazooka gun was made from PVC pipe parts found at the local hardware store. A potentiometer was attached to the gun's rotating support tube using a rather poor mechanical design. It was a frequent source of trouble for the game.

(continued...)
 

Attachments

  • 1757188947583.png
    1757188947583.png
    231.9 KB · Views: 11
  • 1757188979856.png
    1757188979856.png
    147.3 KB · Views: 11
=== PART 2 ===
Desert Patrol

Using an electronic design that was an upgrade to Bazooka, this game had a similar game display showing aircraft flying across the screen. A double-barrel machine gun was mounted on the cabinet that could be aimed both up/down and left/right. An improvement from the horizontal only aiming found in Bazooka.

1757190088422.png

John wanted a more realistic machine gun for Desert Patrol – a definite step-up from the primitive Bazooka gun. PSE did not have a mechanical engineer so he hired an outside consultant to design a machine gun with motorized barrels that would quickly oscillate in and out when the trigger was pressed. While it worked, and was quite impressive, the mechanical design turned out to be poor and the gun would eventually self-destruct from the unbalanced reciprocating motion. I believe the moving barrels were eventually dropped from manufactured games, in favor of two fixed barrels. I'm curious to know if any Desert Patrols with motorized gun barrels still exist.

In the game play, pilots can appear parachuting from destroyed aircraft. You're not supposed to shoot them, but if you do, the pilot plummets to Earth in a death spiral. I added that feature to the game play, but many in the company thought it was too violent. Being 19 at the time, I thought it was pretty cool. I decided to add a scream sound to accompany the falling pilot. I asked John's daughter, who was PSE's receptionist, to come over so I could record her screaming into a primitive digital voice recorder I built. Her digitized scream was copied into a small 8574 PROM memory chip on Board 3 and gets played back whenever a paratrooper is shot.

Game Tree
Continuing to expand on the programmable game platform, PSE released Game Tree. The game play has various animals running around, with squirrels jumping from tree to tree. A wooden rifle mounted on the cabinet allowed you to shoot them for points.

Game Tree used an improved hardware design that allowed multiple images to be moved horizontally and vertically on the screen, under program control. As with all PSE games, all video images are generated directly by electronic circuitry, rather than a display memory.

The game cabinet contained a two-way mirror that superimposed the black and white monitor image onto a printed full-color graphic forest scene reflected on the two-way mirror. The use of a mirror doubled the distance to the player, making the game targets appear farther away. The superimposed image gave the illusion of a color display. Game Tree was my favorite of the three PSE games I worked on and they are apparently quite rare these days.

1757190098964.png

Game Tree was released soon before the company closed in 1978, resulting in the low number of Game Trees produced. On a Friday, John came over and told me the company wasn't going to make it and I didn't need to come in on Monday. As there was no warning of the closure, it was quite a surprise to me at the time. I went on to found my own company at age 20 and continued to work in the development of other video games and products.

The Game Tree Operator's Manual (found on this website) has the best documentation for the three games. I wrote the text in this manual, which includes a detailed technical description of how the circuitry works, which is a bit of a mind-bender to read today, but it made sense to me at the time. You'll find my initials in the schematic title block for Board 2. Other schematics, manual images and the game cabinet graphics were created by PSE's in-house graphic artist (Sharon Davies???). Although she was not an engineer, you'll see her initials on other sheets for having created nice looking versions from John's hand-drawn schematics.

I guess that's enough rambling for now, please post if you have any questions.
 
=== PART 2 ===
Desert Patrol

Using an electronic design that was an upgrade to Bazooka, this game had a similar game display showing aircraft flying across the screen. A double-barrel machine gun was mounted on the cabinet that could be aimed both up/down and left/right. An improvement from the horizontal only aiming found in Bazooka.

View attachment 844779

John wanted a more realistic machine gun for Desert Patrol – a definite step-up from the primitive Bazooka gun. PSE did not have a mechanical engineer so he hired an outside consultant to design a machine gun with motorized barrels that would quickly oscillate in and out when the trigger was pressed. While it worked, and was quite impressive, the mechanical design turned out to be poor and the gun would eventually self-destruct from the unbalanced reciprocating motion. I believe the moving barrels were eventually dropped from manufactured games, in favor of two fixed barrels. I'm curious to know if any Desert Patrols with motorized gun barrels still exist.

In the game play, pilots can appear parachuting from destroyed aircraft. You're not supposed to shoot them, but if you do, the pilot plummets to Earth in a death spiral. I added that feature to the game play, but many in the company thought it was too violent. Being 19 at the time, I thought it was pretty cool. I decided to add a scream sound to accompany the falling pilot. I asked John's daughter, who was PSE's receptionist, to come over so I could record her screaming into a primitive digital voice recorder I built. Her digitized scream was copied into a small 8574 PROM memory chip on Board 3 and gets played back whenever a paratrooper is shot.

Game Tree
Continuing to expand on the programmable game platform, PSE released Game Tree. The game play has various animals running around, with squirrels jumping from tree to tree. A wooden rifle mounted on the cabinet allowed you to shoot them for points.

Game Tree used an improved hardware design that allowed multiple images to be moved horizontally and vertically on the screen, under program control. As with all PSE games, all video images are generated directly by electronic circuitry, rather than a display memory.

The game cabinet contained a two-way mirror that superimposed the black and white monitor image onto a printed full-color graphic forest scene reflected on the two-way mirror. The use of a mirror doubled the distance to the player, making the game targets appear farther away. The superimposed image gave the illusion of a color display. Game Tree was my favorite of the three PSE games I worked on and they are apparently quite rare these days.

View attachment 844780

Game Tree was released soon before the company closed in 1978, resulting in the low number of Game Trees produced. On a Friday, John came over and told me the company wasn't going to make it and I didn't need to come in on Monday. As there was no warning of the closure, it was quite a surprise to me at the time. I went on to found my own company at age 20 and continued to work in the development of other video games and products.

The Game Tree Operator's Manual (found on this website) has the best documentation for the three games. I wrote the text in this manual, which includes a detailed technical description of how the circuitry works, which is a bit of a mind-bender to read today, but it made sense to me at the time. You'll find my initials in the schematic title block for Board 2. Other schematics, manual images and the game cabinet graphics were created by PSE's in-house graphic artist (Sharon Davies???). Although she was not an engineer, you'll see her initials on other sheets for having created nice looking versions from John's hand-drawn schematics.

I guess that's enough rambling for now, please post if you have any questions.
Thanks for posting all this!! Can't wait to hear more.

@Dillweed get some mint Game Tree pics to share!
 
This is amazing, thanks for sharing this! Do you have any insight around the early partnerships with Atari? I have an early test rig from Atari that was manufactured by PSE.

View attachment 844792
That's an interesting find! Built before I started with PSE, so I can only speculate. It might have been a project PSE developed for Atari under contract. Contract engineering was their business at that time and Atari was located near by. The PSE Pong game conversion replaced the Pong control panel with a two joystick panel (the same joysticks as in your picture). Perhaps this contract is what lead PSE into conversion video games? Just a guess... When I started, PSE was no longer working with Atari, probably because they emerged as a competitor. They had an original Pong game in storage that I bought for $50. It was hand made from Bushnell's team, serial #45 I think. Had the modified television inside with hand-cut cardboard trim around the screen. I gave it away to a friend a few years later...

Thanks for sharing!
 
Sweet! That looks to be in great shape. Let me know if you have any technical issues bringing it up. After shipping, I'd pull the boards and check for proper voltages on the power supply first. If OK, then power down, plug the boards in and see what happens. Be careful handling the boards. PSE electronics often had lots of small wire jumpers on the back (another story...). Easy to break those old connections...

Be careful to not shoot the hunting dog! :)
 
Sweet! That looks to be in great shape. Let me know if you have any technical issues bringing it up. After shipping, I'd pull the boards and check for proper voltages on the power supply first. If OK, then power down, plug the boards in and see what happens.

Be careful to not shoot the hunting dog! :)
It's missing the power transformer that plugs into the fuse board and logic board card rack... I am debating on using a newer power supply just on the fact that i want these boards to be as safe as possible haha.
 
New po
It's missing the power transformer that plugs into the fuse board and logic board card rack... I am debating on using a newer power supply just on the fact that i want these boards to be as safe as possible haha.
A new switching power supply should be OK, just make sure the +5V is adjustable. It was normal to adjust the 5V a little to get the PSE games to work. Setting it to 5.0V might not be the sweet spot...

If you want to keep it original, the original transformer has 8.5V RMS and 32V RMS center-tapped secondary windings. It allowed for international line voltages, but you only need a 110V primary. You might have to use two transformers to get these voltages, as I believe the PSE transformer was custom. It should be the same as Bazooka or Desert Patrol if there are any spare parts out there...


1757198164149.png


Power supply schematic is on page 36 of the Game Tree manual. You'll need +5VDC at least 6A, and +/-12VDC @ 1A.

1757198986561.png

Just PM me if you have any questions.
 

Attachments

  • 1757197876082.png
    1757197876082.png
    16.6 KB · Views: 1
  • 1757198049209.png
    1757198049209.png
    113 KB · Views: 3
Enjoyed reading your write up! Especially your account of programming with toggle switches and no terminal viewer, insert, delete, etc. And I thought roughing it was entering op-codes directly into memory using the built-in Apple II terminal program. Today software development is built on layers of abstraction as deep as the sediment layers of the Grand Canyon, but there was something special about working that close to the "bare metal" of the system :)
 
Enjoyed reading your write up! Especially your account of programming with toggle switches and no terminal viewer, insert, delete, etc. And I thought roughing it was entering op-codes directly into memory using the built-in Apple II terminal program. Today software development is built on layers of abstraction as deep as the sediment layers of the Grand Canyon, but there was something special about working that close to the "bare metal" of the system :)
You're spot on. Today's programming is very abstract from the underlying hardware. I first learned to program in 1975 using IBM punch cards at NASA. One line of code per 6" paper card. You used a mechanical keypunch machine that made the proper holes in the card. Make a mistake and you had to throw the card out and start over. It was a beautifully designed mechanical device, and quite expensive.

1757204948497.png

1757204918376.png

You ended up with a stack of hundreds of cards that made up your program and carried them in a big cardboard box. To run the program, you carried your box of cards over to the computer building and gave it to the lady behind the counter to run on NASA's powerful IBM mainframe computer. You could come back the next day to get your results. As there were no display monitors yet, output was limited to a paper printout. You could make some simple graphs and even lousy images by using a grid of printed Xs, but mostly it was all text and numbers. In contrast to today, where you can modify and run a program in seconds, each single run of your program took one day. Any bug, error, or change in your code meant printing new cards and carrying your box back to the nice lady so you could get an updated printout tomorrow. I tripped on a curb once and spilled my cards into the parking lot, so I spent the rest of that day getting them all sorted back into the proper order. I learned that it was smarter to draw a diagonal line on the side of the card stack to make it easier to reorder them.

Of course this was state-of-the art computer programming at the time. Something only a big facility like NASA would have with their huge and expensive IBM mainframe. At PSE we had to improvise and design our own methods to create programs in a unique binary language we created. It started with toggle switches but a year or so later we progressed to using a teletype machine to enter programs. Teletypes were designed to send text messages between companies using phone lines, but they got adapted to programming out of necessity. This was a vast improvement, as you could now store your programs on long strips of punched paper tape instead of volatile RAM memory. The machine could read the tape at a blazing speed of 10 characters per second! There were no hard drives yet so storing a program permanently was a challenge.

1757205787761.png

1757205873293.png

It was the dawn of Silicon Valley (unknown to us at the time). Digital electronics was exploding. New chips came out almost every day. Big 8" floppy disks came on the scene. Video display monitors with keyboards showed up. And of course, the primitive, first personal computers with microprocessors came into being. PSE was gone by that time, but I think it's interesting history and probably not commonly known to those of a younger generation. Video games were adapted from televisions. But televisions at that time only displayed whatever the transmitting video camera was pointed at. There were no computer-generated graphics, text or pictures. The TV weather report consisted of an attractive woman sticking Velcro sun and rain pictures onto a felt wall. The idea that a computer could create an image on its own, without a camera, was really new and exciting. Imagine, a computer making a TV picture! A lot has changed in 50 years...
 
@KyleF Do you know if Game Tree's serial numbers started at 0 or 100? My machine lands at S/N 102. I'm unsure how many Game Trees made it out of there.
The only Space Out found was S/N 8, so by that comparison it seems at least 100 Game Tree cabs actually made it out the door?
 
@KyleF Do you know if Game Tree's serial numbers started at 0 or 100? My machine lands at S/N 102. I'm unsure how many Game Trees made it out of there.
The only Space Out found was S/N 8, so by that comparison it seems at least 100 Game Tree cabs actually made it out the door?
Sorry, I don't know if the S/N started at 1 or 100. My guess is 1. I recall seeing lots of Game Trees sitting in the assembly area. Maybe 200-400 were made? Just a guess... There used to be a date of manufacture sticker placed inside the PCB card cage, so look for that too...
 
This was a fantastic read, even more so since it happened locally for me. I sometimes can't remember what I did last week, let alone 50+ years ago! Kudos on your recollection!

I had an internship with NASA/Ames in the 2000's, however mine was off site (from Moffett field) at foothill college. I still got to go to a lot of the cool presentations with SETI, in the wind tunnel, the fire squad, etc.
 
Back
Top Bottom