In November 1975 the third issue of BYTE carried an article headed “Son of Motorola, or, the $20 CPU Chip”. It began: “We thought that the ‘age of the affordable computer’ had arrived when you could buy a microprocessor chip for $150.” Now “the calculator firm, MOS Technology of Norristown, Pennsylvania” was selling one for $20, and for $25 you could have a version with its own clock circuit, the MCS6502. Within a few years that $25 chip, and its descendants, were running the Apple II, Atari’s computers, Commodore’s PET, VIC-20 and Commodore 64, and later the Nintendo Entertainment System.
MOS Technology is remembered for its chips rather than its own products. Commodore bought it in 1976, and for the next decade it was the company’s chip designer and factory. The 6502, the VIC video chip, the VIC-II and the SID all came from its engineers.
The 6502
MOS began as a maker of calculator chips. In 1974 Chuck Peddle, an engineer who had worked on Motorola’s 6800 microprocessor, brought a team from Motorola to MOS to design a cheaper one. BYTE later wrote that “by producing the 6502 and selling it for only $25 while other semiconductor houses were saying the price would never fall below $200, Peddle made the personal computer possible.”
MOS actually launched two processors. The 6501 fitted the same socket as Motorola’s 6800, which Motorola did not welcome; according to Tim Danton’s history it sued Peddle’s team. The 6502 had the same instructions but a different pin arrangement and a clock generator on the chip, which BYTE noted made it “probably cheaper in an original design”.
The chips were launched at the WESCON electronics show in 1975. Danton describes the show organisers banning sales on the floor, so buyers crowded into a hotel suite nearby to pay their $25. MOS also sold the KIM-1, a single-board computer with a calculator-style keypad and LED display, for $245, and it sold well to hobbyists and engineers learning the new chip.
For programmers the 6502 was small and quick. It has one accumulator, two index registers (X and Y) and a stack that always lives in page 1 of memory, addresses $0100 to $01FF, with 56 instructions and thirteen addressing modes. BYTE’s 1975 comparison found some common operations quicker than on the 6800: a subroutine call took 6 cycles against the Motorola chip’s 9. The first 256 bytes of memory, “page zero”, can be reached with shorter, faster instructions, so 6502 programmers treat it almost as a bank of extra registers. There is more on the processor in the 6502 entry.
Bought by Commodore
MOS had a hit but little money. Danton records that its main backer, Allen-Bradley, withdrew in early 1976, and that MOS was also losing its calculator business to Texas Instruments’ price cuts. Commodore had the same problem from the other side: Jack Tramiel’s calculator company had been buying its chips from TI, and TI had started selling its own calculators for less. Tramiel’s answer was to own a chip maker. “The best thing was to try to buy a semi-conductor company,” he recalled. “And I looked around and at the end I found MOS Technology.” Commodore bought it in 1976.
The purchase came with Peddle, who persuaded Tramiel that the KIM-1 board, a keyboard and a screen could make a computer. That computer was the Commodore PET, shown in 1977.
Commodore’s chip designers
Once it was part of Commodore, MOS designed the chips that made Commodore’s home computers cheap. Its LSI (large-scale integration) group, headed by Al Charpentier, designed the VIC video chip. When Tramiel wanted a colour computer for under $300, the MOS engineers built one around it, and it became the VIC-20.
In January 1981 the same group began a new video chip and a new sound chip, which became the VIC-II and the 6581 SID. Charpentier led the graphics work and Bob Yannes, a synthesiser enthusiast, designed the SID. At the end of 1981 Tramiel decided to build them into a 64K computer, and together with the 6510, a 6502 with an extra input/output port, they became the Commodore 64.
Owning the factory mattered as much as owning the designs. The first C64s cost Commodore about $135 to build. Ahoy! traced the advantage back to Tramiel in 1984: his “obsession with controlling the cost of every phase of the manufacturing process led to Commodore’s acquisition of MOS Technology and other semiconductor manufacturers”, and Commodore’s vertical integration “allowed it to earn record profits in the midst of a price war which has left competitors reeling from huge losses”. Yannes made the same point about the SID from outside the company in 1984: a custom chip was expensive, and “Commodore was obviously able to handle that since they have their own design center and processing.”
The SID’s data sheet shows the two identities side by side: it is headed “mos technology” and carries the address of MOS Technology, Inc., 950 Rittenhouse Road, Norristown, together with the name “Commodore Semiconductor Group”. Later data sheets drop the old name and describe the Commodore Semiconductor Group as “a division of Commodore Business Machines”. The division lasted until Commodore went into liquidation in 1994. Its former managers bought the plant and ran it as GMT Microelectronics until it closed in 2001.
Sources
- BYTE issue 3, November 1975, pp. 56 onwards: Daniel Fylstra’s comparison of the Motorola 6800 and the MOS Technology 6500 series, with the editor’s introduction “Son of Motorola, or, the $20 CPU Chip”; $150 as the previous low; the MCS6501 at $20 and the MCS6502 at $25 with its on-chip clock; the subroutine-call timings.
- BYTE, November 1982, p. 256: Phil Lemmons’s interview with Chuck Peddle; Peddle’s team moving from Motorola to MOS in 1974; the $25 price “while other semiconductor houses were saying the price would never fall below $200”; MOS selling 6502s to Atari and Steve Jobs before Commodore bought it.
- Tim Danton, The Computers That Made Britain (Raspberry Pi Press, 2021), chapter on the Commodore PET 2001: Peddle’s move from Motorola, the 6501 and the Motorola lawsuit, WESCON 1975 and the hotel suite, the KIM-1 at $245, Allen-Bradley’s withdrawal in early 1976 and Commodore’s purchase.
- MOS Technology, MCS6500 Microcomputer Family Programming Manual (second edition, January 1976): the index registers, zero page addressing, the stack “always stored in Page 1”; the company’s Norristown address.
- Richard Mansfield, Machine Language for Beginners (Compute! Books, 1983): the 6502’s 56 instructions and thirteen addressing modes.
- Roger Kean, The Story of the Commodore 64 in Pixels (Fusion Retro Books), pp. 17–19: Tramiel on Texas Instruments, vertical integration and finding MOS Technology; Peddle and the PET.
- Tekla S. Perry and Paul Wallich, “Design case history: the Commodore 64”, IEEE Spectrum, vol. 22 no. 3, March 1985 (read in the version republished at spectrum.ieee.org): the LSI group under Charpentier, the VIC-20 chips, the chip project from January 1981, the C64’s $135 build cost.
- Ahoy! issue 5, May 1984, p. 44: vertical integration, the acquisition of MOS Technology and “record profits in the midst of a price war”.
- Compute!’s Gazette issue 11, May 1984, p. 19: Yannes on the cost of a custom sound chip.
- MOS Technology, 6581 Sound Interface Device data sheet: the MOS Technology and Commodore Semiconductor Group names. Commodore Semiconductor Group, 6582 data sheet: “a division of Commodore Business Machines”.
- Wikipedia, “MOS Technology” (web, secondary): founding in 1969; the plant’s sale after Commodore’s liquidation and its closure in 2001. https://en.wikipedia.org/wiki/MOS_Technology