Why was the original IBM PC 4.77 Megahertz?

In 1979, I had a summer job on Long Island, NY designing a circuit board for the front panel of a satellite control system. That was the time Apple II and Radio Shack were top selling computers.

We were designing a circuit board that had an embedded processor. Developing the software for it was much easier with something called an ICE: In Circuit Emulation.

Basically, an ICE is a computer that had a ribbon cable that terminated in a DIP (Dual On-line Package) connector that could plug into the prototype circuit board under development and emulate the target CPU.

This allowed the developer to have a compiler, linker, keyboard and display while prototyping. I could use a text editor to write the software (assembly code for the 8080), compile it using a cross-compiler (a compiler that runs on one architecture that compiles for another) and download it to the “virtual processor” to see if the circuit board really worked.

Because we were Intel customers, we received a preliminary specification of a new processor family that Intel was coming out with, named the 8086/8088 family. They were pretty much the same, except that the 8088 had an 8 bit data bus and the 8086 has a 16 bit data bus. Because most microcomputers in those days used 8 bit components, the 8088 was used for the IBM PC. I remember scanning that document and marveling at the possibilities and groaning at the fact that all I knew about 8080 assembly code would be changing.

At around that time, Phil (Don) Estridge was assigned the task of building a new microcomputer for IBM. Previously, IBM hardware and software were all proprietary: they made the memory, the disk drives, the processors, etc. Estridge bucked the trend and used a Shugart floppy, an Intel processor, and other off the shelf parts, and published the design and the BIOS. I still have the BIOS listing and schematics of the original IBM PC sitting on my shelf in the PC Technical Reference Manual.

He probably saw the same preliminary spec that I did.

Sadly, Estridge died in the Dallas wind shear DC-10 crash in 1985 and was unable to see the wild success of his product.

For a stable timing source, circuit boards required a crystal oscillator. Because there were many easily available crystals for TVs, they were pretty cheap.

However, TV oscillators were 14.318 Megahertz, which is 4 times 3.579 Mhz, the frequency of the color subcarrier of a baseband video signal. That means if you put a baseband video signal (SuperHeterodyne modulates between baseband (0-6 Mhz) and a particular channel (2 is 50-56 Mhz)) on an spectrum analyzer, you’ll see a spike at 3.57 Mhz where the color information is encoded. B&W TVs just ignored this.

That’s how Color and B&W TVs could use the same signal: another growing up mystery solved by understanding subcarriers.

(I visited a house in the 60’s and they had Color TV ! Wow! They must be rich! Some hotels actually had signs that said Color TV and the TV Guides used to have a ‘c’ next to shows that were broadcast in color. And the only input was from an antenna! How could that be? Another mystery solved by the understanding of SuperHeterodyne and subcarriers).

The Intel 8088 could allow a maximum clock rate of 5 Megahertz.

Thus simple circuitry to divide the 14.318 Megahertz crystal by 3 yielded 4.77 Megahertz: just under the maximum.

Nowadays computers use clocks that are 1000 times faster, measured in Gigahertz!