1200XL Mysteries (Continued)

[+DrVenkman]

Inspired by my experiences today trying to diagnose an odd issue, and by a recent thread started about the W1 motherboard jumper discussed in a 1200XL document, let's talk about some other, shall we say, "questionable" choices made during the design of the 1200XL.

 

Oddity #1: the OS sockets. As discussed in a few places around the interwebs, the 1200XL has a two-chip OS. Both of these chips are 24-pin 8K ROMs, yet the board was laid out and vias exist on all production machines for 28-pin sockets. Hell, on two of my three examples, they were built with 28-pin OS sockets to begin with! What was the extra address space intended for or at least foreseen? Did Atari plan at some point to do a single-chip 28-pin OS ROM and eliminate the second chip? Was the idea of building-in BASIC already percolating for possible use of the second socket? If anyone can point to internal Atari documents addressing this, I'd love to read them!

 

Oddity #2: the angled controller ports. They look great, and I guess this is simply Aesthetics vs. Practicality, with aesthetics winning out. I will say that it can lead to problems. One of my machines had multiple bent pins in the second controller port. From the way they were bent, I determined that someone (probably a child) could have tried to force in a joystick at the more typical perpendicular way and done just about the exact type of damage shown. Fortunately that was a fairly easy fix when I finally got around to it.

 

Oddity #3 (and what inspired me to post today): the freaking power supply circuitry! It's unlike any other XL machine and the only rationale I can come up with for designing it this way doesn't make a whole hell of a lot of sense to me. So for those not familiar, an annotated photo:

 

 

The red circle is a giant AC-DC rectifier. It takes the 9VAC power supplied to the computer from a jack clear over on the back left corner of the board and converts it to approximately 12VDC. The two (2! Not 1 like a sane design!) green circles are the pair of 7805 voltage regulators that convert the 12VDC supplied by the rectifier to 5VDC used by the computer. This two step conversion process is very energy inefficient and generates a bunch of waste heat. Note the huge black metal ribbed heat sink bolted to the top of the rectifier and running around the entire right rear corner of the machine. The two 7805's are also screwed into that heat sink as well as being soldered to the circuit board below; I changed those two regulators today because my machine was being flaky, the thermal paste was drying out and I happened to have a bunch of spare 7805's and was bored.

 

So this whole design makes absolutely no sense to me. Atari used (and had used, for years) bog-standard 9VDC adapters for the 2600. Yes, the 1200XL uses more current than a 2600 but scaling up a standard 9VDC converter wouldn't be a major task at all, especially compared to all the engineering effort and the parts necessary to do the power conversion inside the computer like this. The 800XL and 600XL, which were rushed out in the wake of the 1200XL's market failure in ridiculously short order, used power brick adapters to do the conversion from wall current down to 5VDC used by the board, and kept all that weight and heat out of the computer itself. Without the rectifier, dual 7805's, giant heat sink and that enormous 10,000 uF capacitor back there, the systems were smaller, lighter and ran much cooler, as well as being more efficient and less expensive to build.

 

What was that electrical engineer thinking? I have to wonder if some manager-type came to the guy and said, "Well, we've got a few hundred thousand of these 9VAC power bricks already out there in the wild, plus thousands more in inventory for spares and production. You need to make this new box run on the same plug." And of course, the 1050 drive of the same approximate era ALSO uses the same 9VAC power supply. So yeah, commonality *might* be the answer. Yet the 600XL and 800XL continued to be sold for years afterward with their 5VDC bricks alongside the 9VAC 1050 drive. So if commonality was the reason, it sure didn't last long. The whole thing just doesn't make a lot of sense to me.

 

So, anyone else wonder about some of the design oddities of the 1200XL, or is it just me?

 

Still, the most gorgeous machine of the XL era forward.

[_The Doctor__]

the joystick cords don't pull or fall out as easy and the wires break less using the 45 degree angle placement and it looks nice....

the tunnel for the cartridge had to be done for three reasons, 1 because the cartridge would be out of the way and could not be knock around making 2 it look nice and clean, because 3 there were already difficulties with reliable cartridge reads where it was and extending the distance made it worse and no one bothered to find out why that was.... (everyone should Know the fixes for the 1200XL cart port today, posted at different places on AtariAge)

the 1200XL was being built with more than one direction, configuration and platform in mind and was going to supply a decent 5 and 12v volt rail for all kinds of goodies...when that wasn't going to happen it was just tied off quick and dirty

the different addressing, jumpers/sockets were multifaceted as well. I vaguely remember other discussions covering some of this here on AtariAge in other threads as well..

Maybe it's time to scour AtariAge and have link backs put in all the threads to traverse the mountain of information contained within. It would be nice to have it all at our finger tips.

 

BOB1200XL should have almost all the answers in his head, on paper, and sprinkled about. He's an encyclopaedia twelvehundrica ... Albert and Curt probably have some notes on parallel development as well

Edited January 13, 2018 by _The Doctor__

[ACML]

Nice write up! The number one wacky design issue on the 1200XL for me is the missing chroma signal on the monitor port. Pin 5 is already there and plenty of folks used the superior separate video (Y/C) output on the original 800. So why not connect it??? I blows my mind that the two possible explanations are either it was deliberate or it didn't get caught in testing. You'd have to be pretty incompetent for the latter, so why intentionally would you leave it out when the signal was always there? My #2 "what were you thinking" issue is not allowing the SIO power for 3rd party devices. In know why they did it, but notch designing a deficiency in your new product to ace out 3rd party hardware manufacturers was a bad call. Only served to piss off the user base.

 

Don't get me wrong, it's my favorite computer, but Atari kept shooting themselves in the foot. Was this marketing or engineering making these calls?

[_The Doctor__]

I've considered all the super blatant obvious 'mistakes' -sabotage.... either on purpose or as a result of drink and drug.... party in hall way... mad s*x in the hallway.... keystone kaper kops in the video output... you can't have pure genius all the way till final production and bungle it at the last moment this many times for it to be coincidence..

Edited January 13, 2018 by _The Doctor__

[+DrVenkman]

Yes, the "lose the Chroma signal" video jack was also silly, and it persisted through the 800XL. Why? It just makes zero sense to take an excellent feature of the 800 and just drop it for no good reason.

 

That power system though - what a PITA. But at least I figured out how to remove/replace that giant heat sink and the 7805's if I ever need to do it again. If the rectifier goes south ... screw it, I'll do a 5VDC power mod and then probably add a PBI interface with that metal monster out of the way.

[flashjazzcat]

I might be (and probably am) wrong, but is it possible chroma was left off the monitor jack to ensure compatibility with the older C64 monitor jack (and thereby cables)?

 

 

As to the other oddities, I guess Atari initially wanted to keep the 9V AC PSU which powered the 800, disk drives, etc. Joystick port angle actually seems perfectly reasonably to me and saves a little (very little) horizontal space. As for the multiple ROMs: I think splitting the OS across more than one chip was fairly common at the time. The fact is, the dual sockets neatly provide for an internal BASIC modification, so it's a positive in that respect.

[Gunstar]

Yes, the "lose the Chroma signal" video jack was also silly, and it persisted through the 800XL. Why? It just makes zero sense to take an excellent feature of the 800 and just drop it for no good reason.

 

That power system though - what a PITA. But at least I figured out how to remove/replace that giant heat sink and the 7805's if I ever need to do it again. If the rectifier goes south ... screw it, I'll do a 5VDC power mod and then probably add a PBI interface with that metal monster out of the way.

I didn't let the metal power monster stop me from installing a PBI (link to page mod is on): http://atari.boards.net/thread/947/phoenix-rising-1200xl-pbi-mod?page=3

 

I also like all the extra amperage of the 1200XL's "universal" PSU with 3.4A instead of the later XL/XE 1.5A power. Much better for upgrading, IMHO.

[+kheller2]

I know a lot of discussion usually comes up with these topics and that the 600XL/800XL were rushed to market after the failure of the 1200XL.... but looking at historical facts, that doesn't always seem to pan out as THE reason. From what I recall, the 600/1200 were already in development (and originally called the 1000) and the 1400XL and PBI specs were started in February 83. So unless the Winter CES in January was where Atari knew they had a failure on their hands I find it hard sometimes to believe that the 1200XL "caused" the birth of the other XLs.

 

The first 1200XL's, according to the serial tracker, show they came out in February 83. (Bob's is the only one with a a January model, and the 250 PILOT units were done in October 82)

 

 

As further proof against this theory of "bad 1200xl = new XL" I show this document date chart I've been creating to give me a sense of what Atari was doing and when:

 

 

Granted this isn't complete and I don't have 600XL and 800XL design docs in their yet...

But they were designing the PBI on 2/3/83 and a 1400XL on 4/8/83. Again, unless CES was so bad.... I'm sure it contributed to the rush of further systems. So did the price and the software compatibility issues that folks make a large deal about.... you know those same incompatibilities that exist in EVERY XL/XE.

[Gunstar]

Personally, I think Atari chose the power circuit to match tons of the "universal" 9V 3.4A supplies they had left in warehouse stock from the 400/800/810 and so on, so used them for the 1200XL and 1050 and so on.

[Mathy]

Hello guys

 

I seem to remember that I read an interview with one of the guys working at Atari in either ANALOG Computing or ANTIC, that they left off the chroma signal to free up a pin for a second sound signal, which should have given us stereo.

 

Erhard developed an upgrade for the 1050, where he replaced the 78-series voltage regulators with more modern hardware, that uses less energy and (therefore) produces less heat. Maybe that'll work on the 1200XL too.

 

Sincerely

 

Mathy

[_The Doctor__]

please also remember Atari had a habit of writing the spec's after the fact... it's made and then 6,8 months later the paper comes out with engineering and document writer arguing about clarification and what's correct or incorrect....

Edited January 14, 2018 by _The Doctor__

[jmccorm]

Oddity #1: the OS sockets. As discussed in a few places around the interwebs, the 1200XL has a two-chip OS. Both of these chips are 24-pin 8K ROMs, yet the board was laid out and vias exist on all production machines for 28-pin sockets. Hell, on two of my three examples, they were built with 28-pin OS sockets to begin with! What was the extra address space intended for or at least foreseen? Did Atari plan at some point to do a single-chip 28-pin OS ROM and eliminate the second chip? Was the idea of building-in BASIC already percolating for possible use of the second socket? If anyone can point to internal Atari documents addressing this, I'd love to read them!

Do you suspect that Atari originally had a legacy ROM-select feature, or they had new functionality that was cut (perhaps for fear of being able to market the new features or fragmenting their installed base) or something entirely different?

 

I did start to wonder if the 8-bit hardware could pull off a good enough 2600 emulation via a new OS ROM. 80 columns and fast math came up in your other thread. I wonder who might still authoritatively know what the intent might have been.

 

This really does seem to suggest an untold story that wouldn't have been told in the day.

[+bob1200xl]

Some 1200XLs were built with 28 pin EPROMs and some (most) were built with 24 pin ROMs. A ROM does not have (need) a -WRT pin, so you can save a few cents on board space and sockets if you use ROMs. Of course, if you need to use either ROMs or EPROMs, it might end up costing you more.

 

The angled controller ports would put less strain on the cable at the connector. May be look better?

 

The power supply looks like a WIP. There is a third mounting point for a 7805 (or, ?) on the heat sink. Atari was transitioning to an external power supply - 600XL, 800XL, 1400XL - all used external power packs.

 

I think Atari only cared much for the TV signal from the modulator. That is what the vast majority of owners would use.

 

Bob

[+DrVenkman]

I know a lot of discussion usually comes up with these topics and that the 600XL/800XL were rushed to market after the failure of the 1200XL.... but looking at historical facts, that doesn't always seem to pan out as THE reason. From what I recall, the 600/1200 were already in development (and originally called the 1000) and the 1400XL and PBI specs were started in February 83. So unless the Winter CES in January was where Atari knew they had a failure on their hands I find it hard sometimes to believe that the 1200XL "caused" the birth of the other XLs.

 

The first 1200XL's, according to the serial tracker, show they came out in February 83. (Bob's is the only one with a a January model, and the 250 PILOT units were done in October 82)

 

 

As further proof against this theory of "bad 1200xl = new XL" I show this document date chart I've been creating to give me a sense of what Atari was doing and when:

 

history copy.jpg

 

Granted this isn't complete and I don't have 600XL and 800XL design docs in their yet...

But they were designing the PBI on 2/3/83 and a 1400XL on 4/8/83. Again, unless CES was so bad.... I'm sure it contributed to the rush of further systems. So did the price and the software compatibility issues that folks make a large deal about.... you know those same incompatibilities that exist in EVERY XL/XE.

 

I would really, *REALLY* love to read some accounts from insiders from the Home Computer division of Atari, Inc. in those last 18 months or so before the Warner panic-sale to Tramiel's holding company. The 1200XL hit the market in early '83. Based on docs and info from Curt's Atari History site, it sure seems like the 1200XL/600XL were intended to basically replace the 800/400 duo as the high-end/low-end models of the refreshed 8-bit line. That much makes perfect sense.

 

​However, to that end, the 600XL we *GOT* is a hell of a lot more like the 800XL we also got to replace the 1200XL, than the 1200XL it was presumably to pair up with. The 1200XL has the angled controller ports on the left side of the case, side-mounted horizontal cartridge port, no PBI connector, and uses the legacy 9VAC transformer of the earlier systems. The 600XL (like the eventual 800XL) has a top-mounted cartridge port with spring-loaded dust cover doors, perpendicularly-mounted controller jacks on the right side case (rather than left), a new keyboard connector design, and the PBI connector, all running on an external 5VDC power supply. Its eventual "big brother" 64K machine - replacing the 1200XL in the lineup - is all of the above plus more RAM in a larger case.

 

So somewhere along the line between conceptualizing a new 1200XL/600XL lineup to replace the aging 800/400 line, and the actual line up of machines Atari, Inc. produced between very late 1982 and 1984, things changed quite a bit. I just would love to know more - how, why, etc. - from the people involved.

[+kheller2]

 

The power supply looks like a WIP. There is a third mounting point for a 7805 (or, ?) on the heat sink. Atari was transitioning to an external power supply - 600XL, 800XL, 1400XL - all used external power packs.

 

 

 

It's interesting that you mentioned that. The production sink from ATMC has that third spot and a circular hole for something else. However... the Pilot units do not have them at all...

mystery....

[+DrVenkman]

 

 

It's interesting that you mentioned that. The production sink from ATMC has that third spot and a circular hole for something else. However... the Pilot units do not have them at all...

mystery....

 

That third hole in the heat sink (between the rectifier and SIO port) is really big though, at least compared to the screw holes for the two 7508's. Maybe there was some consideration given originally - or at least when the contract was let to have the heat sinks fabricated - to mounting the 9VAC jack closer to the rectifier?

[+bob1200xl]

I don't know about the large hole over by the modulator - look on the heat sink over where it has been bent back on the right side of the m/b. Around the corner...

 

Bob

[jmccorm]

What caught my attention was one of the early code names for the project: "Sweet 16".

You'll see it mentioned in an image that is embedded in kheller2's reply (and many other places).

 

The implication seems obvious, but has anyone adequately explained the selection of that particular code name?

 

What caught my attention at first was that "Sweet 16" is the name given to code which was written by Steve Wozniak (of Apple fame) and published in 1977 as a 16-bit (software) extension to the 6502. It used zero page memory to create sixteen virtual registers/pointers of 16 bits each, alongside some new opcodes to take advantage of them. So by that time, the name Sweet 16 already had a very specific meaning in the 8-bit world: an extension which gave 16-bit functionality to an 8-bit machine.

 

Related: Someone once ported the extension to the Atari 8-bit [archive.org].

 

So, back to the implication, few if anyone seems to specific state it. Atari may not have intended to immediate create a 16 bit machine. It seems like what became our 1200XL wasn't originally a machine with a 16 bit CPU, but an 8-bit machine with a Parallel Bus Interface, perhaps with the intention of an upgrade path to an external upgrade with a future CPU that had 16 bit registers and could access 24 bits of memory? A CPU like the 65816?

 

That particular CPU was well into design by 1983, and Atari is reported as receiving samples (along with Apple) when it was finished in 1984. Apple used it to create the Apple IIGS and Atari.... was in chaos about that time, followed by cost cutting.

 

Admittedly, you have to jump through a number of hoops, but it just seems to make sense that if "old Atari" went through an effort to create a cost-cutting machine, and they added a parallel bus interface for expandability, when they give that project the code name "Sweet 16", they had some very specific ideas in mind for what they'd do with that interface.

 

Too many assumptions? Too obvious? Both?

 

TL;DNR: An 8-bit machine which would later be capable of a 16-bit hardware extension. They designed a lower-cost 8-bit machine, but with a PBI that would let a16-bit CPU control it, on a larger external 24-bit address space. They could deliver the 8-bit machine today, but the 16-bit side was going to have to wait on next year's CPU from a different manufacturer than MOS.

Edited January 14, 2018 by jmccorm

[jmccorm]

I need to avoid writing late at night. My coherency suffers.

 

I did some more reading. It looks like WDC (creator of the 65C02) had consulted with Apple on a new CPU which would effectively be a speedy hardware implementation of Wozniak's Sweet 16 virtual CPU (while still backwards-compatible with the 6502). That CPU also would have been an excellent candidate for an external CPU upgrade in Atari's new hardware platform... which was also being developed under the Sweet 16 name. (Aside from trying to hit a lower product cost, the PBI adapter gave Atari some opportunity to breathe new life into their 8-bit platform when that CPU became available.)

 

In my mind, the 1200XL and the Sweet-16 design makes today's VBXE feel even more legitimate. Atari seemed to have had a plan for a 16-bit CPU to play puppet inside the body of a refreshed Atari 8-bit line, and for expansion card slots to introduce some degree of new capabilities... graphical to be certain.

 

TL;DNR FOR REAL: The 1200XL missed its true calling as an intended base for a 16-bit computer upgrade. The expansions were never completed by Atari, Inc and it was never continued by Atari Corporation.

Edited January 14, 2018 by jmccorm

[+DrVenkman]

I need to avoid writing late at night. My coherency suffers.

 

I did some more reading. It looks like WDC (creator of the 65C02) had consulted with Apple on a new CPU which would effectively be a speedy hardware implementation of Wozniak's Sweet 16 virtual CPU (while still backwards-compatible with the 6502). That CPU also would have been an excellent candidate for an external CPU upgrade in Atari's new hardware platform... which was also being developed under the Sweet 16 name. (Aside from trying to hit a lower product cost, the PBI adapter gave Atari some opportunity to breathe new life into their 8-bit platform when that CPU became available.)

 

In my mind, the 1200XL and the Sweet-16 design makes today's VBXE feel even more legitimate. Atari seemed to have had a plan for a 16-bit CPU to play puppet inside the body of a refreshed Atari 8-bit line, and for expansion card slots to introduce some degree of new capabilities... graphical to be certain.

 

TL;DNR FOR REAL: The 1200XL missed its true calling as an intended base for a 16-bit computer upgrade. The expansions were never completed by Atari, Inc and it was never continued by Atari Corporation.

 

Except for two things: the 1200XL does not have a parallel bus interface (those didn't appear until the 800XL/600XL pair almost a year later); and the modern VBXE is more like an extension to ANTIC, not a CPU upgrade. The CPU upgrade you're probably thinking of is RAPIDUS.

[Gunstar]

Sorry, but I don't buy any of that. I've researched the "sweet-16" project myself and in all of that research NOTHING 16-bit was ever mentioned. Why do you think a project code name with the number 16 means 16-bit?!?

[jmccorm]

Sorry, but I don't buy any of that. I've researched the "sweet-16" project myself and in all of that research NOTHING 16-bit was ever mentioned. Why do you think a project code name with the number 16 means 16-bit?!?

 

Well, when they were developing the external expansion slots and cards that the documents reference, if they weren't able to jump beyond the hard specs on the page to say, "Hey, you know, a 16-bit processor could plug into our new 8-bit bus and control it", they would have had to of been willfully blind, right?

 

I like to believe I'm an honest researcher, so I can concede that Sweet-16 doesn't have to reference a 16-bit CPU or Wozniak's virtual 16-bit CPU on 8-bit hardware. To the side, I'll see if I can't poke a few people with a stick and get us something that confirms, denies, or outright explain the name's origin.

[+DrVenkman]

 

Well, when they were developing the external expansion slots and cards that the documents reference, if they weren't able to jump beyond the hard specs on the page to say, "Hey, you know, a 16-bit processor could plug into our new 8-bit bus and control it", they would have had to of been willfully blind, right?

 

I like to believe I'm an honest researcher, so I can concede that Sweet-16 doesn't have to reference a 16-bit CPU or Wozniak's virtual 16-bit CPU on 8-bit hardware. To the side, I'll see if I can't poke a few people with a stick and get us something that confirms, denies, or outright explain the name's origin.

 

More importantly, as I've pointed out more than once, the 1200XL has no PBI interface. So the idea that a possible, sometime-down-the-road-maybe 16-bit processor card for a future expansion box (ultimately the 1090) doesn't hold much credence to me as a source of the Sweet 16 code name, barring the discovery of concept design documents or personal testimony from the hardware designers.

[+kheller2]

SWEET-16 was originally called Z800. And the documentation for S16 specifically states it was not designed for professional or business use but home and beginner use only. Atari wanted nothing to do with a new computer architecture after the success of the 8bit. The Amiga tech did eventually interest them.

[jmccorm]

 

More importantly, as I've pointed out more than once, the 1200XL has no PBI interface. So the idea that a possible, sometime-down-the-road-maybe 16-bit processor card for a future expansion box (ultimately the 1090) doesn't hold much credence to me as a source of the Sweet 16 code name, barring the discovery of concept design documents or personal testimony from the hardware designers.

 

Your message was next on the list... give me a chance to agree with several good points you've made before you rule my entire train of thought as invalid!

 

You're right when you pointed out that I conflated the Rapidus with the VBXE. You also nailed it when you pointed out that although the Sweet-16 Project outlines the PBI, and although the Sweet-16 product resulted in the Atari 1200XL, the 1200XL didn't end up with the Sweet-16's Parallel Bus Interface. Actually, I really liked this take which basically described the 1200XL as an attempt to walk away from much of the Sweet-16 Project and create a consumer product instead of a hobbyist product.

 

So, although in all the years that have followed, even though nobody manages to explain what Sweet 16 really means, you were looking to me to explain it, right? Actually, that's totally fair. I'm the one selling the project's name as an a reference to 16-bit hardware. As I mentioned just now to Gunstar, I'll start poking a few people to see if we can get ourselves a smoking gun. Outside of that, I think I have been able to offer some new support that nobody has ever seen before with via the novel connection to Steve Wozniak. At least, I haven't seen someone else make that connection and integrate it into a selling point? (If they have, please, show me! I'd like to see if I missed anything!)

 

1. I've shown that "Sweet-16" was a term that had already existed in the industry when the project was named, and that same term references an enhancement for 8-bit microcomputers.

2. I've made the argument, hopefully with some success, that it is reasonable to believe that Atari's development engineers were knowledgeable in their field (if not subject matter experts), and the meaning of Sweet-16 would be known and shared among them. (These are the guys who I'd expect to know something about Wozniak's Sweet-16 enhancement.)

3. The enhancement that Wozniak's Sweet-16 refers to is for a piece of hardware that is at the center of Atari's personal computer business (the 6502 8-bit processor). This isn't just a mild tangent.

4. At the same time that they were having Sweet-16 project and design meetings at Atari, Wozniak's Sweet-16 enhancement was being developed into new CPU hardware at Western Design Center.

5. Western Design Center was reported to be working directly with Apple on the hardware implementation of Sweet-16. (This demonstrates that a physical Sweet-16 was significant.)

6. The company which was developing/manufacturing a hardware Sweet-16 would unquestionably have sought out Atari as a customer. (Even if Atari hadn't of initiated the idea, the idea would have come to Atari.)

7. The Sweet-16 CPU would have been an ideal CPU for Atari to chose as a backwards-compatible upgrade path. [Granted, that one selling point was the one thing that was obvious and known to many before exploring down this path.]

 

Why wouldn't the documents outright state all of this? If Atari was developing a 16-bit migration path for a replacement line of personal computers, for internal politics or external competitiveness, I can see them actually going out of their way NOT to document the long-term goal. (EXAMPLE: Customer says, "Oh my God! Atari is working to replace 8-bit with entirely different hardware.") Beyond that, keep in mind that Sweet-16's goal wouldn't have been to actually build a 16-bit computer. It would have been to create an 8-bit computer with the interface (that's our friend the PBI) which could communicate with a 16-bit CPU (and other technological upgrades). The CPU, expansion cards, software development, all of that would be a a different project, later down the road.

 

To those who would say that clinching proof just isn't there, I'd have to say that they're right! Sweet-16 could be nothing more than a coming-of-age story for Colleen's daughter. I just hope that I'm able to make a slightly more convincing story that the underlying project (and its name) is based more upon the connection between Wozniak's virtual enhancement to the 8-bit architecture and Atari's and WDC's making a physical path to a 16-bit architecture possible.

Edited January 14, 2018 by jmccorm