Jeff Frederiksen Interview (Part 1) By Paul Thacker Sept. 29, 2011 I've been in contact via e-mail with Jeff Frederiksen, who was the chief engineer behind designing the Bally Professional Arcade hardware. He's now working at Apple. It's been awhile since I originally sent some questions. He indicated that he still wanted to put some more of his thoughts down, but he has sent me some "interim" answers to a couple of questions, which are actually pretty detailed, and he said it would be fine to go ahead and publish these. So, let this whet your appetite for hopefully more ahead. Paul: Could you give us a little information about your background, and how you came to work with Bally and Dave Nutting Associates (DNA)? It sounds like you were really at the forefront of using microprocessors in arcade games with solid state pinball games and Gunfight. I read about an attempt to replicate the "Bally Brain" for the Flicker pinball machine at www.tzannes.net/pinball/flicker.pdf. Do you know if that project was ever completed? Jeff: I left St. Thomas College in St. Paul, MN, in my junior year and joined the Air Force as an electronic technician. Although my majors were math and physics, the AF is where I learned electronics. At the AF base in southern Turkey, they had a Burroughs 36-bit mainframe computer system primarily fed with a card reader. I was in maintenance and had no access to the computer, but one of the computer techs gave me a programming book and I started to program error detection for the maintenance cards I was inputting. I couldn't get access to the compiler, so I programmed in binary machine code. Before using my checker program, the error rate on maintenance records was about 10%. After implementation, it went to 0% except for 1 blank card at the beginning of a deck in the 1st month. Shortly thereafter, AF headquarters sent a team to investigate what had happened. After explaining what I did, they didn't know whether to reward or court-martial me. They finally left recommending I submit the entire package as a suggestion to headquarters in Washington. Shortly thereafter I finished my tour with the AF and went back to Milwaukee, WI, to finish school. Some time Later, I started working for Milwaukee Coin Industry which was owned by Dave Nutting. During my employment, I created a dial for a safe game using a geared disk and a pair of optical sensors which we patented. This mechanism later became the basis of the mouse. After the safe game, and several other electro-mechanical games, Dave Nutting asked me to attempt an electronic pinball machine. At the time the choice was either a Fairchild calculator chip or Intel 4004 system - also primarily a calculator chip. I created the concept of multiplexing wires to drive lamps and switches in a large matrix. The 4004 was replaced with the 4040 before production and the electronics was encased in an aluminum box called the Bally Brain first used in the Flicker. In the demo to Bally management, we received 2 electro-mechanical Bally Flickers, one which we gutted and replaced with our system, and the other as a side-by-side control. The pile of wires and relays we removed were stacked in a formidable pile along side the machine which was now practically empty. The management team found the response excellent, but couldn't believe it was all electronic. They looked inside the now empty machine and got down on the floor and looked under the machine to convince themselves what we had done. Thus was born the electronic pinball era. Paul: How did the Bally Professional Arcade project come about, and how did you get involved in it? What were the goals of the system, and who set them? Did DNA have a lot of freedom, or was everything dictated by Bally? Jeff: Shortly after the pinball game, I wanted to find an alternative to the extremely complicated and massive dedicated Transistor-Transistor-Logic (TTL) video games that were evolving. I thought of using dram as a single-bit per pixel imaging buffer with a processor to write to it using the basic motion algorithm of XOR-on/XOR-off to move images. At first I considered using the same Intel 4040 processor used in the pinball machine, but soon realized it didn't have enough compute power. I then substituted the recently available Intel 8080 and built my first machine. After finally getting it to wake up - barely - Dave and I took the prototype to Midway/Bally to show what we had and to pitch the first game - Gunfight. The prototype only had some moving noise on the screen - barely enough to demonstrate that it was alive. Regardless Hank Ross and Iggy Wolverton, the owners of Midway before selling to Bally, chose to go ahead with the project. The result was the first microprocessor controlled video game - Gunfight. Midway's sales jumped for the year from a year-to-date total in October of $3M to $11M in December - a wild success. When I asked them a few years later why they originally went ahead with the project, they responded that they had nothing better to run with at the time. In spite of the simplification from using a processor, there was still significant display and i/o control electronics. My thought was to simplify this further by using a custom chip set. This was when I designed the Bally Arcade chip set using AMI. The main designer at AMI that converted the TTL design to silicon was Brian Taylor. We used 5um technology - large enough that we could debug silicon with a microscope and cat whisker probes. Although the chip set was primarily designed for commercial coin-operated games, I further designed the Bally Arcade as a possible home market. A cute gimmick added to this design was the expansion connector to allow an external keyboard, etc. for home computing. Bally went ahead with the Arcade, but when asked to go ahead with the computer add-under using the ZGrass graphics language invented by Dr. Tom DeFanti, the Bally execs said there was no future in home computing - probably their biggest blunder.