Mail Archives: djgpp-workers/2003/05/30/19:55:31.1
Eric Rudd wrote:
>
... snip ...
>
> If you are concerned about small frequency errors, you may be interested
> in the origin of the 1.19-MHz rate that is used in the PC timer chip.
> In the early days of personal computing, it was considered important to
> maintain compatibility with NTSC, since people used NTSC monitors on
> their computers. According to the EIA RS 170 A spec, the chroma
> subcarrier is at 3.579545 MHz +/- 10 Hz, and the PC clock was chosen to
> be 1/3 of that, or 1.19 MHz approximately. The specified 3.579545 MHz
> rate itself is an approximation; it is exactly 455/2 times the
> horizontal sweep rate, which was originally 15750 Hz in the early
> black-and-white days, but was changed to 15750/1.001 when NTSC was
> devised. Thus, the chroma subcarrier works out to 39.375/11 MHz, and
> the PC timer chip runs at 13.125/11 MHz. As I recall, the system clock
> on the original IBM PC was 4 times that, or 4.77 MHz.
The fundamental reason for the selection was that 3.579545 MHz
crystals were cheap and plentiful- they existed in every color TV
receiver. About 6 years earlier I selected a frequency for my
systems based on exact generation of baud rates from 110 baud
through 19200, and the resultant 8080 clock was slightly below 2
Mhz. Don't remember the exact result any more.
--
Chuck F (cbfalconer AT yahoo DOT com) (cbfalconer AT worldnet DOT att DOT net)
Available for consulting/temporary embedded and systems.
<http://cbfalconer.home.att.net> USE worldnet address!
- Raw text -