Project Breaker and the 6502 Vectrex

At the beginning of Vectrex development, a 6502 microprocessor was considered for a short period of time.

I believe that the 6502 was suggested because of its very low cost and the fact that we already had a team with 6502 experience. Those with 6502 experience were working on Project Breaker ‑ an effort to reverse-engineer the Atari 2600.

The Breaker team consisted of Mark Indictor, Paul Allen Newell and myself.

Breaker consisted of looking at the binary code from a 2600 VCS cartridge, identifying what was machine code versus binary data and then disassembling the machine code ‑ all by hand.

Since we were going to create original games, it was only necessary to examine the games to the point where we understood what the hardware was doing.

We would step through the machine code to see what it did and how it affected the hardware. We would typically see a 6502 register being loaded with a value, maybe some kind of logical operation followed by a write to a hardware port. The question would become what was important about the individual bits of that write and what did the port do when written?

Mark began the Breaker effort and was joined by Paul well before I was hired. By the time I began, the two of them had accomplished much of the analysis. Given the short amount of time for Breaker, the amount of work that they had accomplished was very impressive.

Once we got the go ahead for Vectrex, we changed gears and boxed-up the Breaker stuff, literally put it on the shelf and shifted our focus to the design of Vectrex.

As the design of Vectrex progressed, it was quickly decided that the 6502 would not be fast enough. One aspect of 'speed' was the clock rate ‑ I believe a 1 Mhz clock was used for the 6502 and 1.5 Mhz for the 6809. More important than clock rate was the architecture and instruction set of the 6809. The 6502 had a relatively primitive instruction set where doing fundamental things like indexing could be quite awkward.

The 6809 was a vast improvement over the 6502, but at many times the cost. No matter the benefits of the 6809, cost could have easily have damned us to the 6502 ‑ and trust me that when writing code for the 6502, it doesn't take long to understand what being damned truly means.

I don't know how the decision was made or who the people were in deciding to switch from the 6502 to the 6809 ‑ but I really appreciate that they made a decision driven by technology rather than one were cost was the priority. Don't get the idea that I'm minimizing the issue of cost. The significantly higher cost of the 6809 meant that more money had to be raised before production could start. It had to be hard decision for everyone involved.