Atari 800 vs. Commodore 64 – The Brief Tale of Two 8-Bit Home Computers

IMHO the 8-bit Ataris were a lot more innovative than the Commodore 8-bitters (including the C64), despite being a couple years earlier to market.

For instance the Atari display chip ANTIC was an actual co-processor, running a sequence of simple instruction via DMA to produce the display (very similar to the Amiga's copper).

The C64 (and other 8-bit machines) could only reprogram the display chip mid-frame by directly writing hardware registers with the CPU at the right time (often controlled by a raster interrupt).

The 8-bit Ataris were the actual predecessor to the Commodore Amiga (not surprising, because the same people who designed the Atari later designed the Amiga).

The Atari 800 was nice I heard, but there was a war between Commodore and Atari users on the BBSES.

The Atari 800 had more colors than the Commodore 64, but the SID chip on the Commodore 64 had better sounds, which attracted musicians to it.

Oddly enough management of Atari and Commodore switched and the Atari ST was the successor of the Commodore 64, but I had the Commodore Amiga 1000 that was designed by Jay Miner and other Atari engineers.

...but during the same period Radio Shack attempted to compete with the C64 and Atari 8-bit machines with their TRS-80 Color Computer line; specifically the Color Computer 2 (the Color Computer 3 was a more worthy competitor, but since it had 128K and came into being at the tail end of the 1980s when things were transitioning to the PC and the 8-bit machines were "dying out" - it doesn't really figure here).

https://en.wikipedia.org/wiki/TRS-80_Color_Computer

The thing about the Color Computer 2 (this applied as well to a great extent with the 3), is that it was mostly a reference design by Motorola, and subsequently virtually all of the main component ICs were Motorola parts.

In this case, the CPU was the 8-bit 6809 (clocked at roughly 1 MHz - though there were ways to double this speed), and the "display chip" was the MC6847:

http://www.cocopedia.com/wiki/index.php/6847 https://en.wikipedia.org/wiki/Motorola_6847

As you can tell from the wikipedia article, it's capabilities were nowhere near that of the machines it was competing with. There was a way to get some weird modes called "semigraphics" via messing with another chip in the machine called the "SAM" chip, but they were rarely used.

In short - for everything the Color Computer 2 did, it relied almost entirely upon the speed and capabilities of the CPU itself! Graphics blitting, sound generation (there was no "sound chip" - merely a 6-bit DAC - which was shared with other parts of the system as well), keyboard/joystick input - all went through the CPU.

It has been argued that the Motorola 6809 is a far superior CPU (in design and capabilities) than the CPUs used in the Atari or Commodore 8-bit machines (or the Apple IIe); in fact, many arcade and gambling machines of the era also used the 6809 (generally in multiples) over other offerings.

Where it fell down, though, was it's price. It was easily several times more expensive than other offerings from other manufacturers, which the other machines used (mainly variants of the MOS 6502).

I've often wondered what/why Tandy/Radio Shack took this route? Now originally, the Color Computer (well, the 6809 and 6847 combo) grew out of a "videotext" style terminal that was originally marketed toward farmers, with the idea of receiving up-to-date "almanac" and planning information via phone lines and the terminal. This was during the late 1970s/early 1980s. I can't recall what this device was called, but I recall it had a blue case, but looked identical to the Color Computer otherwise; it can be found in old Radio Shack catalogs now online, if you care to look.

Did Tandy/Radio Shack have some kind of exclusive license deal with Motorola, to only use Motorola or Tandy-designed parts for the machines? I don't think the SAM was a Motorola component (?); and later with the Color Computer 3 it had it's own custom chip (the GIME) too - but everything else was Motorola...

The price point of the CPU prevented them from putting multiple CPUs in the machine (I often wonder how things might have been different had there been a second 6809 CPU in the Color Computer to handle sound and/or graphics); but what prevented them from putting in a cheaper CPU (like the 6502?) or a sound chip (there were a few cheap ones out there made by TI and others)?

Instead - for additional sound capabilities, you had to purchase cartridge expansion - either third-party or from Radio Shack (the Speech/Sound Pak, or the Orchestra-90 Pak), as well as an expansion device (the Multi-Pak interface), especially if you already had a floppy disk system and/or modem/rs-232 pak. That upped the price considerably, and as a result, few did it, and so those sound options were not well supported in games or other applications.

Instead, the built-in sound was exploited as far as things could take it at the time (and more recently, even further). Later, to improve the graphics of the system with the release of the Color Computer 3, Tandy added the GIME chip, which basically took over the duties of the SAM chip and 6847, but only partially (semigraphics modes were not available - all other modes were, though); but it was a seemingly in-house designed chip, and they also had to have a third-party update the BASIC (that was originally provided by Microsoft) to support the new resolutions and colors enabled by that chip.

The company that did this was Microware, who had long supplied a multi-user multi-tasking OS for the 6809 called "OS-9"; they later produced a version for the Color Computer 3 called "OS-9 Level II" - not that this should not be confused with OS9 for the Mac (totally different products).

Microware essentially hacked the BASIC. The original code was from Microsoft, but they had moved on and didn't want to do the updates to the code, so what Microware did was kinda ingenious - they had their ROM of "patches", and on bootup, they would copy the ROM code for the MS BASIC over into RAM, then overlay hooks to their ROM routines in the RAM copy, then run the modified RAM copy of the BASIC interpreter; this was called "all-RAM mode", and it was something that was long possible to do with the CoCo 2 (and was used in a similar manner by third-parties to extend commands in BASIC for their own products and extensions). Microware's only changes were being able to do it at boot time. Anyhow, they extended it and it worked (mostly) well enough.

But even at that late date (circa-1986) - Tandy had to go with a custom chip, and not anything third-party outside of Motorola (from what I recall, Motorola didn't have any display chips of a better spec for the 8-bit line; they had mainly moved on to their 16 bit 68K lineup). So what was preventing them? I honestly don't know...

Interestingly, some time in the early 2000s it came to light that Microware had probably the only (maybe outside of Tandy itself) "prototype" of the Color Computer 3 - it was a giant wire-wrapped board; it's most curious feature being that it lacks a GIME chip! This is especially important as there is no documentation/datasheet or internal information about how the GIME chip worked, outside of what was provided to programmers. In short, the GIME chip is an "unobtainable" component (there are two versions - one from 1986, the other from 1987, with minor "bug fixes"). It is hoped that this rare board might provide clues as to how the GIME chip was designed, but right now the board is a curiosity (I can't recall who it was given to, but it was given to that person by employees of Microware, which was still in business at the time, and may still be (?), who found it in a back closet or something like that).