Whatever became of Turbo-816?

[+Larry]

In 1988 or so, DataQue (Chuck Steinman) developed and released the Turbo-816. It was real, and it did run in a demo at our user's group meeting. There were a series of articles published about it and DataQue's plans for add-ons for it. I think these were originally in MAM (Michigan Atari Magazine) or AIM (Atari Interface Magazine) and maybe a few installments later in Antic.

 

Does anyone actually have one of these working? Was the Turbo-816 patched OS useful? (I am aware that Drac030 has developed his own 65816 OS.) Did the Turbo-816 have any relation to the later Sweet-16 that I believe was primarily developed by Bob Woolley? Whatever became of DataQue? I last saw him in 1992(?) at the Chicago Atarifest. *If possible*, would this device be worth "bringing back" since no other alternatives seem available?

 

I know that several folks have worked hard on 65C816 accelerator boards, but I am not aware of a released product for the Atari community -- e.g. send in your money and in a week or so you get a package in the mail. I don't believe that the "Sweet-16" was ever released in any significant numbers, and had timing issues in some computers. (?) Would going back to "square one" be useful?

 

-Larry

Edited October 1, 2007 by Larry

[Rybags]

I've read of a few upgrades around the place. F7 was another, IIRC (F7 being the "turbo" key for A800Win+).

 

The problem with such upgrades is that probably so few people would get them anyway.

 

IMO, something along the lines of a co-processor + shared RAM addon which could reside in a cartridge would be an ideal upgrade.

 

Even if it was just something that could do stuff like generate software sprites and screens in the background, it could be a worthy addition to the system.

Edited October 1, 2007 by Rybags

[+Allan]

In addition to the hardware, someone would have to write at least one piece of software for it (most likely a good game) in order for it to sell. It's a lot of knowledge, hard-work, and money to put into it. It's probably why it's never happened. You need all three and every time someone has attempted to do it, they only had one or two of the qualifications. Hopefully it will happen one day.

 

Allan

[Curt Vendel]

I spoke with Chuck frequently while he was working on the Tubro-816 while I was testing and doing the icon designs for Diamond for Alan Reeve. Chuck had sent me some documents on his hardware and he was also talking with Alan about a version of Diamond on the 816.

 

I had heard many years ago that 2-3 people had actual 816s but I have never seen an actual working unit to date.

 

 

Curt

 

 

 

In 1988 or so, DataQue (Chuck Steinman) developed and released the Turbo-816. It was real, and it did run in a demo at our user's group meeting. There were a series of articles published about it and DataQue's plans for add-ons for it. I think these were originally in MAM (Michigan Atari Magazine) or AIM (Atari Interface Magazine) and maybe a few installments later in Antic.

 

Does anyone actually have one of these working? Was the Turbo-816 patched OS useful? (I am aware that Drac030 has developed his own 65816 OS.) Did the Turbo-816 have any relation to the later Sweet-16 that I believe was primarily developed by Bob Woolley? Whatever became of DataQue? I last saw him in 1992(?) at the Chicago Atarifest. *If possible*, would this device be worth "bringing back" since no other alternatives seem available?

 

I know that several folks have worked hard on 65C816 accelerator boards, but I am not aware of a released product for the Atari community -- e.g. send in your money and in a week or so you get a package in the mail. I don't believe that the "Sweet-16" was ever released in any significant numbers, and had timing issues in some computers. (?) Would going back to "square one" be useful?

 

-Larry

[mos6507]

IMHO, from an end-user perspective the problem with the Turbo-816 was that it wasn't really Turbo. It still ran at 1.79mhz. So all you get is the new 65816 opcodes and the flat memory model. That really was not worth the money. Consider the fact that by then most had extended bankswitched memory and too few programs were even using that for anything more than a RAMdisk.

[+Larry]

IMHO, from an end-user perspective the problem with the Turbo-816 was that it wasn't really Turbo. It still ran at 1.79mhz. So all you get is the new 65816 opcodes and the flat memory model. That really was not worth the money. Consider the fact that by then most had extended bankswitched memory and too few programs were even using that for anything more than a RAMdisk.

 

Yes, that was my opinion when I watched the demo. All I could tell was that the normal SIO sounded a little faster. I think they ran a couple of BASIC programs that timed about 10% quicker, also. Certainly not earth-shaking.

 

But it would still be very interesting to investigate further. If the OS had pretty good "hooks" so that the 65816 floating point could be used, that alone would be worth something.

 

BTW, I found 7 installments from DataQue in my old AIM clippings about the T816. Does anyone know if there was an eighth?

 

-Larry

[danwinslow]

Couldn't we scale up the clock in multiples of 1.79 or something? Or would we need a whole seperate buss for that?

[ndary]

Few years a go in an ATARI party in Poland i met Michal Pasiecznik who showed me an upgraded 130xe with:

2meg simm upgrade, SpartaDos X Built in, IDE interface and a TURBO-816 PCB

 

 

In the ZIP file below you will find all the information i have on the TURBO 816 upgrade

 

 

Turbo_816.zip

 

Ndary

[mos6507]

Couldn't we scale up the clock in multiples of 1.79 or something? Or would we need a whole seperate buss for that?

 

What you'd want to do is get John Harris to release the design of his 816 system which he used for his CG software. That one supported a faster clock speed as well as other cool enhancements. We're kind of at the point where if you want to upgrade your 8-bit it really makes more sense to make a new board design and just transplant the core chips over than to keep hacking away at the old one. I would seriously purchase a 1200XL-sized board with all the bells and whistles on it.

Edited October 1, 2007 by mos6507

[Shawn Jefferson]

Probably the best bet for a 65816 based system is this:

http://drac030.krap.pl/en-main.php

 

He seems to have made good headway with the OS and SpartaDOS as well as the hardware.

[peteym5]

Doing a 65816 internal upgrade would be cool. I do agree with Allen that you need a few software products like games to demonstrate what a 16 processor can do. I know the F7 upgrade did run from the expansion slot on back of a 130XE and I think they had theirs running at 16mhz. I know WDC has a 65816s out and not sure what the speed limit is. Would be interesting if it ran a few GHZs. You will not need to do bank switching and be able to use 16 megabytes of RAM. Now you will have faster manipulation of screen graphics and player/missile graphics. You can write huge VBI routines that can do major scrolling and quick moves onto the screen. A DLI can change all the PMG and Color registers on a single pixel line, even pull data from a table in this case. If you add Video Board XE, Duel Pokeys, along with a 3rd sound chip, then these upgrades look even more attractive.

 

I have mentioned the ideal of building a new 65816 based MB with the Atari 8-bit ICs and a 16bit bus, maybe add the Video Board XE with a few other ICs. I think with all that, you will have one hell of a game machine that is backward compatible with the Atari 8-bit. Do you think that would be really cool to see if someone build a few of these machines.

 

I know people say this machine will not be a true Atari, they act like it wont be backward compatible or something. Almost like saying my AMD 3 GHZ based computer is not a true IBM PC because it is no longer using a 8088 4.77 MHZ processor. One point is to make it run the old programs, but also run a new set of programs made that take advantage of the new hardware. Did people really object when IBM decided to go from a 4.77mhz 8088 PC to a 16mhz 80286 and kept going from there? What if Atari decided to take the same route and kept upgrading their 8-bit machine upto today? or Commodore with the 64, or the Apple with the Apple II. I know Apple had the IIgs, but they released it after the MAC was popular. Each company and machine has a good size software library with some very popular games with people who like playing them.

[JamesD]

Doing a 65816 internal upgrade would be cool. I do agree with Allen that you need a few software products like games to demonstrate what a 16 processor can do. I know the F7 upgrade did run from the expansion slot on back of a 130XE and I think they had theirs running at 16mhz. I know WDC has a 65816s out and not sure what the speed limit is. Would be interesting if it ran a few GHZs.

FWIW, Western Design Center created the 65816.

They currently mention 400MHz but don't sell a 400MHz part. I'm guessing it's a licensed core in an ASIC that achieved that speed.

The fastest they sell last I checked was 16MHz... but many of those parts have been overclocked at up to 24MHz without problems.

 

(A 400MHz Atari... Yikes!)

 

You will not need to do bank switching and be able to use 16 megabytes of RAM. Now you will have faster manipulation of screen graphics and player/missile graphics.

That's not entirely true. The CPU has to slow down whenever it accesses regular system memory and the custom chips don't have access anything but standard RAM so some things would be the same speed. Address lines are also multiplexed on the 65816 so it can't access existing memory expansions without slowing down since they wouldn't be it it's memory map unless you redesigned the custom chips.

 

You can write huge VBI routines that can do major scrolling and quick moves onto the screen. A DLI can change all the PMG and Color registers on a single pixel line, even pull data from a table in this case. If you add Video Board XE, Duel Pokeys, along with a 3rd sound chip, then these upgrades look even more attractive.

Any access to old RAM/hardware would run at standard speeds. You could manipulate DATA in fast memory and then dump the result to the screen and that would make it faster.

 

I have mentioned the ideal of building a new 65816 based MB with the Atari 8-bit ICs and a 16bit bus, maybe add the Video Board XE with a few other ICs. I think with all that, you will have one hell of a game machine that is backward compatible with the Atari 8-bit. Do you think that would be really cool to see if someone build a few of these machines.

The buss is already 16 bits on the 6502. If it were 8 you'd have access to 256 bytes of RAM... but I guess it would be fast since everything would be in Page 0.

FWIW the 65816 extends it to 24 bits and real 16 bit CPUs typically have 24 to 32 bits but the instruction sets often handle 32 bit addresses. The 65816 may deal with some 16 bit numbers but other than memory handling and the combined A-B register it's still pretty much an 8 bit CPU.

It is a big improvement on the 6502 but don't expect miracles. Try out an Apple IIgs. Still slower than similar clocked 68000s.

 

The easiest thing to do would be to isolate the old 16 bit buss and the buss of the 65816. Run it at a multiple of the normal system clock (8x or 16x as fast) and keep any areas not accessed by the custom chips mapped to the fast RAM. Page 0, the stack... everything runs at full speed until you have to access old RAM or hardware. It would also need a software selectable high/normal speed select to insure full compatibility. It wouldn't be quite as fast as a total redo but a redesign of the custom chips isn't required.

 

I know people say this machine will not be a true Atari, they act like it wont be backward compatible or something. Almost like saying my AMD 3 GHZ based computer is not a true IBM PC because it is no longer using a 8088 4.77 MHZ processor. One point is to make it run the old programs, but also run a new set of programs made that take advantage of the new hardware. Did people really object when IBM decided to go from a 4.77mhz 8088 PC to a 16mhz 80286 and kept going from there? What if Atari decided to take the same route and kept upgrading their 8-bit machine upto today? or Commodore with the 64, or the Apple with the Apple II. I know Apple had the IIgs, but they released it after the MAC was popular. Each company and machine has a good size software library with some very popular games with people who like playing them.

Well, it wouldn't be a "true Atari" but only because it's not built by Atari.

[ndary]

Here are some pictures i got of the F7 - 65C816 accelerator by PasiuSSG

 

 

NdARY

[carmel_andrews]

If memory recalls, Steinman was trying to make the t816 upgrade totally comaptible with the existing A8 o/s, apparently he tried offering Atari the upgrade (i.e. allowing Atari to do their own version) and in return allowing steinmann to making the upgrade atari o/s compatible...unfortunately Atari nixed the idea (from what i heard)

 

Steinmann later tried to design an upgraded A8 system designed around the t8/16 upgrade and using inexpensive pc based expansion cards (mainly low end sound and graphics cards) inside a tower casing...called the XLE computers

 

You can see articles about this on the TGR disk mags (TGR= the grim reaper) i beleive Mr Bacardi has most/all TGR disk mags archived on his w/s

[+Larry]

What you'd want to do is get John Harris to release the design of his 816 system which he used for his CG software. That one supported a faster clock speed as well as other cool enhancements. We're kind of at the point where if you want to upgrade your 8-bit it really makes more sense to make a new board design and just transplant the core chips over than to keep hacking away at the old one. I would seriously purchase a 1200XL-sized board with all the bells and whistles on it.

 

At the World of Atari show in Las Vegas (1998) John Harris showed me what I believe is the system that you mentioned. It ran at regular clock speed inside the main 64K, but ran at something like 8 MHz outside of the base ram. This is, I believe, similar in concept to the Warp 4 accelerator being developed in Europe. He gave me the name of the company that sold these, which I wrote down as "Multiplex Technology" in Brea, CA. I tried to contact this company but never made contact. I certainly wanted one of those units. Dunno much else about them (perhaps you know more?). Haven't head from John Harris in a long time.

 

At the current time, it still looks like the Warp 4 is our best hope.

 

-Larry

[ndary]

I found WEB page on Pasiu F7 Accelerator: http://pasiu.krap.pl/

 

Ndary

[+Larry]

Few years a go in an ATARI party in Poland i met Michal Pasiecznik who showed me an upgraded 130xe with:

2meg simm upgrade, SpartaDos X Built in, IDE interface and a TURBO-816 PCB

 

In the ZIP file below you will find all the information i have on the TURBO 816 upgrade

 

Ndary

 

Hi Nir-

 

Thanks for the pics and info. I am slightly confused though, because (from memory of 1988 or so) that board in the picture is quite different than the one I saw. Way more chips on this one, and of course, the original used a DIP-40 IC. My recollection is of the plug-in board with just a few logic chips in addition to the cpu. I wonder if this is a later/enhanced design?

 

-Larry

[peteym5]

If you built a new motherboard with a 65816, you probably wont need to hack an old Atari to get the ICs out. You can order these chips from http://www.best-electronics-ca.com. (ANTIC/GTIA/POKEY/FREDDIE/PIA) Why would you need to slow down your main 64k ram anyway? A good custom design probably will have new memory SIMMS or DIMMS. Could you keep the CPU at high speed (14 or 16 mhz) and put the Atari ICs on a seperate board and run those at 1.79mhz? The ANTIC chip cannot go any faster and needs halt the main CPU to access the main bus for a short time. When you write to those chip registers, it would be going to the 1.79 mhz board(s). That Duel Antic/GTIA upgrade used a seperate board for the 2nd chipset and is given its own memory and bus that ran intendant from the main bus/cpu. Maybe do the Duel Antic/GTIA thing on a custom system. The plan for Videoboard XE is running at 7+MHz, independent from the main system. The operating system itself is all software on an EPROM and the F7 builders were making a custom operating system at high speed. I would keep the PIA/FREDDIE chipset in there to emulate 130XE style bank switching and keep it backward compatible.

 

I should have said make a new system with a 32bit bus before instead 8/16.

Edited October 2, 2007 by peteym5

[JamesD]

If you built a new motherboard with a 65816, you probably wont need to hack an old Atari to get the ICs out. You can order these chips from http://www.best-electronics-ca.com. (ANTIC/GTIA/POKEY/FREDDIE/PIA) Why would you need to slow down your main 64k ram anyway? A good custom design probably will have new memory SIMMS or DIMMS. Could you keep the CPU at high speed (14 or 16 mhz) and put the Atari ICs on a seperate board and run those at 1.79mhz? The ANTIC chip cannot go any faster and needs halt the main CPU to access the main bus for a short time. When you write to those chip registers, it would be going to the 1.79 mhz board(s). That Duel Antic/GTIA upgrade used a seperate board for the 2nd chipset and is given its own memory and bus that ran intendant from the main bus/cpu. Maybe do the Duel Antic/GTIA thing on a custom system. The plan for Videoboard XE is running at 7+MHz, independent from the main system. The operating system itself is all software on an EPROM and the F7 builders were making a custom operating system at high speed. I would keep the PIA/FREDDIE chipset in there to emulate 130XE style bank switching and keep it backward compatible.

 

I should have said make a new system with a 32bit bus before instead 8/16.

If you want to interface the 65816 to the slow system or the slow custom chips to fast RAM you need to design an interface that isolates the two.

It's not difficult to do either way.

For the slow custom chips on the fast buss:

READS - latch the data from the fast RAM at the rising edge of the slow clock and hold it till the falling edge of the slow clock. Takes 1 fast clock cycle.

WRITES - latch the date from the slow chip and hold it till the falling edge of the slow clock, then write it to fast RAM. Takes 1 fast clock. It *could* be on the rising edge or anywhere in the slow clock cycle if needed but I think making it appear to take place on the last clock insures the chips are done.

Handshaking with the CPU would need to be modified accordingly and it may have some undesirable effects. The chipset has to appear to be busy for all those clocks where access is pending... something that may not have existed before. However all CPU accesses to RAM are fast.

The interface would also translate memory addresses from the slow memory to the 24 bit memory model so expanded RAM is the same on both sides of the interface.

The clock speed would have to be a multiple of the standard clock speed so the rising/falling edges of the slow and fast buss would match. *IF* the fast buss was fast enough that the next fast clock would take place before custom chips expect the data to be valid, then unmatched speeds would be ok for the custom chips.

The CPU will experience wait states if it ever reads from the custom chips so replacing the chipset with a compatible one that interfaces at a higher clock speed would be important for maximum speed.

 

At least I think that's how it would work.

[potatohead]

IMHO, an attractive option would be to put the A8 on one of those little single board designs, like the C64.

 

Put the enhancements in there, 65816, etc... then get them distributed as toys. All that would be required for us retro fans is the right signals brought out somewhere cheap. From there, put this thing into an XL case, or something. Then you've got an Atari, looks like an Atari, plays like one, but does more.

 

The nice thing about the C64 design was the extensions to the system. It's still really C64 like, and that's cool. IMHO, the same kinds of things could be done for the Atari, preserving the look and many of the limitations, allowing for the art of it to come through.

 

That's key right there. Enhancements that preserve the art give us new content challenges, and that's where the fun is --at least it's there for me.

 

If the design were out early, people could actually write for it, thus adding some value right out of the gate. Maybe tweak some great content to polish it up for the toy, etc...

 

Would much rather see that kind of thing done, over an expansion board only a few will own. The board will be damn cool --in fact, the coolest! But it won't drive any authoring efforts and that's really more important for the hobby as a whole.

 

***And I've always wanted to program on the 65816 too! Still have the data book somewhere. As a kid, my good friend had a CoCo, with the excellent 6809. At that time, we often wondered what building a CPU board for the 400 / 800 series computers would be like. Clearly for the 6809, a new OS would be needed, but the 65816 could potentially use the existing one with some minor changes.

Edited October 2, 2007 by potatohead

[mos6507]

IMHO, an attractive option would be to put the A8 on one of those little single board designs, like the C64.

 

Put the enhancements in there, 65816, etc... then get them distributed as toys. All that would be required for us retro fans is the right signals brought out somewhere cheap. From there, put this thing into an XL case, or something. Then you've got an Atari, looks like an Atari, plays like one, but does more.

 

That's a good idea. The Flashback3 was going to be an 8-bit system. If that ever gets dusted off, then the enhancements would be a must, just like the C=64 stick has extra graphics modes and such.

[potatohead]

IMHO, an attractive option would be to put the A8 on one of those little single board designs, like the C64.

 

Put the enhancements in there, 65816, etc... then get them distributed as toys. All that would be required for us retro fans is the right signals brought out somewhere cheap. From there, put this thing into an XL case, or something. Then you've got an Atari, looks like an Atari, plays like one, but does more.

 

That's a good idea. The Flashback3 was going to be an 8-bit system. If that ever gets dusted off, then the enhancements would be a must, just like the C=64 stick has extra graphics modes and such.

 

Totally.

 

A lot of people really like those C64 all in one things. There have been programs, mods, etc... done for them on the hobby side too. IMHO, the 8 bit is more than deserving of this treatment. Lots of potential on the table --even with just a faster CPU, and a tweak or two to the GFX system. Heck, do the '816, stereo POKEY (maybe fix the damn high pass filter this time around), open up the GFX for a few more bits per pixel in a fashion similar to what was done for the VICII high color mode. With SIO connections made available, devices new and old could be attached in the same way they are being done on that C64 toy. All really cool stuff.

 

The TV toy path is a good one. Gets the hardware out there and paid for, which is all the hobby really needs to do it's thing on a very longer term basis. I don't think I've seen a C64 case brain transplant yet, but if I were a C64 fan, I would have done it in a second! (Ideally with that nice portable unit --damn those things are cool.) I'm quite sure there is more than enough info available for this to happen on the Atari side.

 

In fact, the only thing I've not seen anywhere is the communication between ANTIC and GTIA. What passes between those two chips anyway, and can that be leveraged some how? I'm almost thinking if that were made available to a faster CPU, driving the GTIA directly would make for Atari like GFX, and allow for software driven graphics modes, sprites, etc... in a fashion not unlike what I'm experiencing on the Propeller right now. (and that's proven to be fun stuff, where I'm concerned anyway.)

 

At the least, having some of the better titles on a portable, new bit of hardware would just be nice to have, just as that flashback II is. And that thing has been popular. The kids play it regularly, and it gets passed around plenty. To me, that has more than demonstrated that good game experiences are just that --timeless things anyone can enjoy.

 

Guess I'm rambling. (home sick today...) It all comes down to the art. We are only gonna get new experiences of the kind we all know and love, if:

 

-hardware makes it into peoples hands

 

-said hardware comes with well planned limitations aimed at extending the art in a meaningful way

 

-provisions are made for home brew authors to do their thing.

 

That's just not gonna happen with hardware expansion boards. Maybe your Chimera might cross that line, but I suspect not. It's likely to be an AtariVOX thing. Make no mistake, that's really cool and I want one, along with supercat's 4A50 magic write board. But what I really want is more content to enjoy and to watch people get into the art of it. That's the big return.

 

I find it amazing we are here so many years later still squeezing so much more out of the older hardware. The smarter we become, and the harder we work, the more the stuff does! Of all the new toys that come out, only the Propeller seems to have that quality. So, there is a vacuum for new gear that has those attributes. We are seeing younger people get into this stuff. Heck, Ian has a class running every year, doing just that! Ongoing efforts to get hardware toys paid for really should be the focus, for these reasons, IMHO.

 

***Glad to see you progressing on that project too Glenn. I was feeling for you when it all broke down. Don't know how many here realize the personal commitment you put into these things, but I do. Owner #97 of the Stella CD, first run.****

 

Maybe you could consider a toy design, project effort. You are not bad at these things. (Stuff actually gets done, and that's more than a lot of people can say.) Why not?

 

If it goes, you and whoever else is contributing, hits it outta the park. If not, well it was damn fun right?

[JamesD]

IMHO, an attractive option would be to put the A8 on one of those little single board designs, like the C64.

 

Put the enhancements in there, 65816, etc... then get them distributed as toys. All that would be required for us retro fans is the right signals brought out somewhere cheap. From there, put this thing into an XL case, or something. Then you've got an Atari, looks like an Atari, plays like one, but does more.

Well, doesn't the 65816 have illegal instruction traps in place of undocumented ones? That could be bad for running some games.

Unless it's only when in native mode anyway.

 

The nice thing about the C64 design was the extensions to the system. It's still really C64 like, and that's cool. IMHO, the same kinds of things could be done for the Atari, preserving the look and many of the limitations, allowing for the art of it to come through.

Too bad the extensions differ from one version to the next and you can't count on them to be on all the versions.

 

Would much rather see that kind of thing done, over an expansion board only a few will own. The board will be damn cool --in fact, the coolest! But it won't drive any authoring efforts and that's really more important for the hobby as a whole.

I have to agree with you there.

 

***And I've always wanted to program on the 65816 too! Still have the data book somewhere. As a kid, my good friend had a CoCo, with the excellent 6809. At that time, we often wondered what building a CPU board for the 400 / 800 series computers would be like. Clearly for the 6809, a new OS would be needed, but the 65816 could potentially use the existing one with some minor changes.

It think it should use the existing OS as is. To use the new features you'd just need to load software that switches to native mode. If the standard ROMs could be banked out for FLASH a custom OS could be put in it's place. If it's like some of the DTV (joystick C64) models then the OS was flashable.

 

The 65816 borrows a few tricks from the 6809 but it still has a few more limitations... not that the 6809 doesn't have it's own.

The 6309 was even nicer than the 6809.

[potatohead]

Yep. 6309 is 8 bit nirvana. Very fun chip. I really enjoyed programming on that design. Moto hit it outta the park with the 6809. Just didn't get enough traction. The 6309 follow on just polished what was already the best 8 bit micro ever, IMHO. At least we ended up with the CoCo 3.

 

Those version differences are a major bummer. If this ended up happening with the Atari 8bit design, it would be terrible. Well, better than nothing, but not really by much. A strong software driven enhancement design would mitigate this. Make room for code, and let people do their thing. Keep actual hardware changes to a minimum. That seems the best.

 

It would have to be running in native mode for sure. That's kind of a given, just to keep overhead to launch something even remotely sane. The key is to be able to get something up and running that people would enjoy. It sells on that basis. Packing in the goodies just can't get in the way of that at all. From there, let the hobby people, and maybe interested others, leverage the extras. One great example, I saw the other day, was this Texas Hold'em TV game. There was nothing to it graphically that could not have been done with the extended C64 TV game design. (version issues aside) I think it was $30 or so. This is just the kind of thing that drives hardware out there, and that could easily leverage enhancements in the real world, as well as in hobby land.

 

Of course, that means having to deal with limitations and not be tempted to just feature creep the things all over the place to fit various niches. I think it can be done, but the planning of those enhancements would have to be solid, and as software driven as is possible, for it to really be useful enough to power hardware distribution.

 

 

IMHO, an attractive option would be to put the A8 on one of those little single board designs, like the C64.

 

Put the enhancements in there, 65816, etc... then get them distributed as toys. All that would be required for us retro fans is the right signals brought out somewhere cheap. From there, put this thing into an XL case, or something. Then you've got an Atari, looks like an Atari, plays like one, but does more.

Well, doesn't the 65816 have illegal instruction traps in place of undocumented ones? That could be bad for running some games.

Unless it's only when in native mode anyway.

 

The nice thing about the C64 design was the extensions to the system. It's still really C64 like, and that's cool. IMHO, the same kinds of things could be done for the Atari, preserving the look and many of the limitations, allowing for the art of it to come through.

Too bad the extensions differ from one version to the next and you can't count on them to be on all the versions.

 

Would much rather see that kind of thing done, over an expansion board only a few will own. The board will be damn cool --in fact, the coolest! But it won't drive any authoring efforts and that's really more important for the hobby as a whole.

I have to agree with you there.

 

***And I've always wanted to program on the 65816 too! Still have the data book somewhere. As a kid, my good friend had a CoCo, with the excellent 6809. At that time, we often wondered what building a CPU board for the 400 / 800 series computers would be like. Clearly for the 6809, a new OS would be needed, but the 65816 could potentially use the existing one with some minor changes.

It think it should use the existing OS as is. To use the new features you'd just need to load software that switches to native mode. If the standard ROMs could be banked out for FLASH a custom OS could be put in it's place. If it's like some of the DTV (joystick C64) models then the OS was flashable.

 

The 65816 borrows a few tricks from the 6809 but it still has a few more limitations... not that the 6809 doesn't have it's own.

The 6309 was even nicer than the 6809.

[peteym5]

If you want to interface the 65816 to the slow system or the slow custom chips to fast RAM you need to design an interface that isolates the two.

It's not difficult to do either way.

For the slow custom chips on the fast buss:

READS - latch the data from the fast RAM at the rising edge of the slow clock and hold it till the falling edge of the slow clock. Takes 1 fast clock cycle.

WRITES - latch the date from the slow chip and hold it till the falling edge of the slow clock, then write it to fast RAM. Takes 1 fast clock. It *could* be on the rising edge or anywhere in the slow clock cycle if needed but I think making it appear to take place on the last clock insures the chips are done.

Handshaking with the CPU would need to be modified accordingly and it may have some undesirable effects. The chipset has to appear to be busy for all those clocks where access is pending... something that may not have existed before. However all CPU accesses to RAM are fast.

The interface would also translate memory addresses from the slow memory to the 24 bit memory model so expanded RAM is the same on both sides of the interface.

The clock speed would have to be a multiple of the standard clock speed so the rising/falling edges of the slow and fast buss would match. *IF* the fast buss was fast enough that the next fast clock would take place before custom chips expect the data to be valid, then unmatched speeds would be ok for the custom chips.

The CPU will experience wait states if it ever reads from the custom chips so replacing the chipset with a compatible one that interfaces at a higher clock speed would be important for maximum speed.

 

At least I think that's how it would work.

 

If I recall correctly, the Antic/GTIA have their own internal clock speed on the chip. However, the read/writes can probably only be done at 1.79 MHZ. I am not sure if it is possible to write to these chips at higher speeds. Of course another option is make ICs that are backward compatible with Antic/GTIA. Maybe have optional enhancements like increasing the HOZ resolution by a multiple for the current Antic Mode. Double the Vertical resolution by writing to the interlace lines. Already we have someone making VideoBoard XE for a new sprite layer. Aren't there 2 open memory locations in the Antic area? You have 2 more registers, along with the other 240 bytes past the Antic area. GTIA has 224 more bytes beyond its memory area. Extra registers can be added without loosing backward compatibility. Maybe combine the two ICs into one chip like Atari was proposing with the CGIA.

 

My Ideal about increasing the HOZ res, would take like Antic Mode 2,3 4,5, 13, 14, and 15 which are 40 bytes wide. I would add a register in the Antic area, if you leave it at 0, those modes will stay the same. Set it to 1 to go to 64 bytes wide, 2 would be 2x res giving us 80 columns, a 3 would be 3x+ res to an even 128 bytes. Maybe go beyond that to 256 bytes wide. Going 128 bytes wide or more probably not be practical for 2 text modes, would useful for 4 and 16 color GTIA modes. I probably would have one of the bits in this register to toggle to one byte (8bits) per pixel/color clock (256 color mode!). I think you get the picture of where I am going with this. The color clocks will increase proportional to the resolution and the chip probably need to run at 16mhz, the same as the proposed speed 65816.

Edited October 2, 2007 by peteym5