Procep SECAM Re: C16/Plus4

From: Richard Atkinson <rga24_at_cantab.net>
Date: Wed, 2 Nov 2011 13:12:01 -0000
Message-ID: <C20A94CAB54E4A34BBA8AD6EAAA63306@abion>
You needed glass delay lines to decode PAL or SECAM, not to encode them. The
two Procep decoders contain PAL delay lines because they first have to
decode PAL before they can output RVB or encode SECAM. The glass delay line
is the light green coloured box near the TDA3510 chip in the Procep boards.
This is not something that could be integrated, it's a large analogue part.

BTW I found pictures of the Procep PS 2000 external PAL to SECAM transcoder
box. They're on forum64.de, which you have to register for to see them
(title C64 externer Secam Decoder von Procep). One of the pictures shows the
circuit board inside the PS 2000. It has a 220V power supply, a TDA PAL
decoder chip and a glass delay line, a SCART socket and - curiously - what
looks like two ASTEC RF modulator boxes. It's difficult to see for sure, but
it looks like both RF modulators are coupled onto the same RF output.

France's SECAM-L system for RF channels is rather unusual, using
positive modulation for the AM vision carrier (instead of negative
modulation) and AM for the sound carrier (instead of FM). I wonder if there
wasn't a native System L RF modulator available at the time, and what we
have here instead is two separate modulators being used for the vision and
sound aspects of SECAM-L.

The big problem with how you encode SECAM is not delay lines but how you
generate accurate enough FM chrominance carriers. By contrast, both NTSC and
PAL use QAM (quadrature amplitude modulation) where both the phase and the
amplitude of a carrier carries information. The key thing here is that the
frequency never changes, therefore you can have crystal generated 3.58MHz
and 4.43MHz carriers. There's a Commodore patent on how the VIC-II chip
achieves NTSC and PAL chrominance signals; basically it takes 4x subcarrier
frequencies (the 14.31818MHz and 17.73447MHz crystals we know and love) and
divides them by four, using both rising and falling edge triggered flip
flops, to generate four phases of subcarrier (0, 90, 180 and 270 degrees)
which can then be added in pairs using different weights to produce the
VIC-II colours. (Incidentally as yet another aside, I wonder how the
6560/6561 VIC-I chrominance encoding works, as it predates this patent, and
in the case of the 6561 it only receives PAL subcarrier frequencies at 1x
subcarrier frequency)

Back to SECAM. To generate SECAM chrominance subcarriers, you have to
generate two FM carriers, each on alternate lines. The frequency of these
carriers varies according to the information they are encoding, therefore
you can't use a crystal with a fixed frequency. There is a SECAM version of
the Atari 800 colour chip GTIA, called FGTIA, which shows how Atari did it
using 1984 technology. A block diagram is on page 23 of the PDF and the
explanatory text starts on page 18.

http://ftp.pigwa.net/stuff/collections/nir_dary_cds/Tech%2520Info/FGTIA.PDF

The FGTIA chip has seven (!) external outputs to control an external voltage
controlled oscillator (VCO) as part of a phase locked loop (PLL) circuit.
Three of them are a colour value output, to produce different frequencies
from the VCO, and four are control signals. One of the luminance outputs of
the GTIA had to be removed in order to free up pins for the VCO interface,
so the SECAM Atari 800/XL/XE only has eight luminance levels (128 colours)
rather than the sixteen (256 colours) of the NTSC and PAL GTIAs. Other
inputs on the original GTIAs were multiplexed to free up pins (diagram on
page 30, explanatory text on page 29). It clearly required a lot of
re-engineering!

Richard


--------------------------------------------------
From: "Segher Boessenkool" <segher@kernel.crashing.org>
Sent: Wednesday, November 02, 2011 2:16 AM
To: <cbm-hackers@musoftware.de>
Subject: Re: C16/Plus4

>> Apparently after what I've heard, integrating the delay lines was a  huge 
>> problem
>> Perhaps that at that time it was not something that could easy done.
>
> A digital delay line would add 50% to the die area, and that's a
> low estimate.  Not going to happen :-)
>
>
> Segher
>
>
>       Message was sent through the cbm-hackers mailing list 


       Message was sent through the cbm-hackers mailing list
Received on 2011-11-02 14:00:03

Archive generated by hypermail 2.2.0.