TI-99/8

[mozartpc27]

I am more of a Commodore guy but I have an interest in all old systems and the TI-99/4 is one I am currently trying to get back into (sold off a collection I had for some cash for another project, but now feeling the itch again).

 

But as a Commodore guy I keep a close watch on current projects out there in the scene and by far the most exciting is this: the Mega65, a latter-day attempt to create, essentially, the infamous never-produced Commodore 65 prototype machine for production. Details here: www.mega65.net

 

It honestly looks like it is coming together, slowly but surely, and amazingly.

 

It made me think what other unproduced machines might be take np by their corresponding communities. The TI community is so strong, I naturally thought of the TI-99/8.

 

Has there been any thought anywhere for folks to work on this?

[willbilly]

I've put some thought into it. I know that there's quite a few custom chips in the 99/8 to contend with. There is also the option to produce new build Myarc boards, possibly even putting the myarc architecture into new cases.

[+Ksarul]

The complete logic diagrams right down to the gate level are up on WHT for most of the special chips, as I put them up there a few years ago once I scanned the originals that I have. The Data for the OSO, POLLO, and MOFETTA chips is complete (as is the data for the HOMBRE chip used in later versions of the 99/4A). The AMIGO chip data isn't available at this level, but it is fully desribed in the documentation, to include all of the JEDEC equations for the portions of the chip that are just performing combinatorial logic. There is enough there to reconstruct it once you suss out how to build the logic for the described memory mapping cells as well.

[+mizapf]

Mind that I managed to build a TI-99/8 emulation from those documents on WHTech. If someone plans to rebuild a TI-99/8, the files 998board.cpp and ti99_8.cpp in MAME should deliver valuable information.

Edited December 23, 2018 by mizapf

[Opry99er]

The dumped ROMs are available, and our friend Klaus has posted several videos on YouTube on his lovely collection.

 

[+mizapf]

From my experiences with the emulation, a rebuild of the TI-99/8 should be feasible, and not too hard. The only challenges are to get a TMS9995, a TMS5200C (speech), and a TMS9118 (video). The rest (the custom chips for logical/physical space mapping, the memory map registers, and Hexbus) is more or less logic stuff. The TMS chips may require to be emulated by FPGA, for the 9118 we have a F18A. The ROM dumps are complete and verified. Any volunteer?

[fabrice montupet]

Using FPGA? It will not be me
Before launching a such project, it may be important to know how many people would be interested in it. Many people love the 99/4A, few knows the 8. The second question is: Is the 99/8 very interesting? Sadly not. If Texas Instruments had sold the 99/8, it would have been a failure because a lack of inovations.

[Omega-TI]

Many people love the 99/4A, few knows the 8. The second question is: Is the 99/8 very interesting? Sadly not. If Texas Instruments had sold the 99/8, it would have been a failure because a lack of inovations.

 

One also has to consider software and compatibility. What, if anything, was ever made for that unreleased computer that would set it apart from the 99/4A to the end user? If a 99/8 was released tomorrow, how long would it take, -for anything- to be developed for such a small audience? It's been years since a couple of items have hit the market for the 4A, in that time a few things have been written for one device, and next to nothing for the other... and that is for the computer most us here already have.

[+adamantyr]

I agree, the TI-99/8 is an unfinished product anyway. It only had a TMS9918A because they didn't have a next generation video chip designed as yet. I remember thinking it was puzzling it would have an older video chip in the prototypes. The Geneve is the spiritual successor of the TI-99/8 and we can see how much use it gets, sadly.

 

Somewhere in an alternate Earth (Earth-99 perhaps?), where Jack Tramiel never rose to prominence and Texas Instruments encouraged 3rd party development, maybe the TI-99/8 was finished and is a glorious machine indeed...

[Casey]

Just because, I'd like a TI 99/8 (I had one at one point many years ago), and we could probably make some great programs in Extended BASIC II that take advantage of the speed of the 99/8 - but I suspect precious few people would spend money on one. The 99/8 originally had a different sized expansion connector that isn't compatible with the 99/4A, so unless that was somehow rectified, very few expansion peripherals would work.

 

That being said, I wouldn't sneeze at a new 99/8 that was compatible with the TIPI.

[JB]

The 99/8 originally had a different sized expansion connector that isn't compatible with the 99/4A, so unless that was somehow rectified, very few expansion peripherals would work.

The Mega65 was mentioned earlier in the thread, and they had a similar problem. They changed the expansion bus to be compatible with the wide array of C64 accessories instead of the nonexistent C65 accessories. We could do the same. Or include the 99/8 connector on the back and the 99/4 connector on the side.

Assuming the system architecture isn't fundamentally incompatible with existing peripherals, of course.

[+adamantyr]

It's really all about the applications... The Geneve is proof of that. It was a solid machine with excellent hardware, but there's almost no software for it. Which is really a shame.

 

The TI-99/8 is a great "What could have been" machine, it's impressive what the engineers accomplished when they weren't being forced to kludge together two different designs for a game console and a home computer and try and make it work.

[Omega-TI]

It's really all about the applications... The Geneve is proof of that. It was a solid machine with excellent hardware, but there's almost no software for it. Which is really a shame.

 

Which proves the old adage, "Without software even the worlds most powerful supercomputer is nothing more than a boat anchor".

[+retroclouds]

Think that the TI-99/4a with todays extensions such as FGROM, F18A, nanopeb and the like has a lot more potential compared to the TI-99/8. And with ralphs SD99 coming...

 

That being said, really would love a TI-99/8. It is such a pretty machine.

[+InsaneMultitasker]

Keep in mind the Geneve was intended to be nearly compatible with the 4A. It was built with hardware advancements appropriate for the era such as the 9938 video and gram kracker style GRAM emulation. Yes, there isn't as much Geneve-specific software out there but that was due to its smaller hardware base, smaller programmer base, and the wide swath of software already available as written for the TI. I have mostly considered the Geneve an advancement over the 99/8 or perhaps what the 99/8 should/could have been with some better engineering and foresight.

[fabrice montupet]

I have mostly considered the Geneve an advancement over the 99/8 or perhaps what the 99/8 should/could have been with some better engineering and foresight.

 

I'm absolutely agree with you.

[LASooner]

Sounds like the Geneve suffered the same issue the Commodore 128 did, it ran C-64 software so people rarely made something specifically for the 128's capabilities.

[Omega-TI]

Sounds like the Geneve suffered the same issue the Commodore 128 did, it ran C-64 software so people rarely made something specifically for the 128's capabilities.

 

That makes total sense to me. If you were going to invest a lot of time and effort to write a program, what group would you target?

 

Group A) The largest group with the most users.

Group B) The smaller group with fewest users.

 

Answer BOTH! If you could write for BOTH groups, why limit your earning potential by locking out the larger audience? The early success of the 64 actually discouraged exploitation of the 128.

[JB]

Sounds like the Geneve suffered the same issue the Commodore 128 did, it ran C-64 software so people rarely made something specifically for the 128's capabilities.

I get the impression the Geneve also suffered another issue the C-128 had: It was more like two computers in one box than it was an upgraded version of the older computer. Quite literally in the 128's case.

You can't easily make software that runs nicely on the newer system and gracefully downgrades itself on the older one, in the way you can on an IBM-compatible or a GameBoy Color. You can't choose to make Geneve or 128 software that degrades gracefully on a 4a or a C64.

 

The 99/8 was (more or less) just a faster 4a with more RAM, so there was a lot of potential for software sharing. Just as a lazy example, a TI-Writer II with larger(or even multi-)file support on a 99/8 that ALSO ran on a 99/4a by dialing back to TI-Writer I file behaviors, while still benefiting from any other changes.

You can't really do that with a C128 or a Geneve, because the new computer behaves so differently outside of compatibility mode. .

 

 

ON THE OTHER HAND, the 99/8's graphics capabilities would have been a serious sore spot. The 9918a was looking pretty dated, and my impression is that the 9118 didn't actually bring anything new to the table.

[+adamantyr]

ON THE OTHER HAND, the 99/8's graphics capabilities would have been a serious sore spot. The 9918a was looking pretty dated, and my impression is that the 9118 didn't actually bring anything new to the table.

 

I had the same thought myself. What I heard was the prototype models they had so far were only using the 9918A because they hadn't designed the video chip intended for the system yet. So if the finished system had reached market it may have had something else by then.

[+9640News]

While folks are dreaming of the capabilities of the 99/8 and the limitations, etc.,then one should take a second thought at the 99/5C. It's a 99/4A with 1 meg of ram with 9938 capabilities.

 

In some earlier discussions with Eric, he hinted it had the capability of emulating a Geneve. If it can emulate a Geneve and all of its complexities, then a 99/8 may be possible as well.

 

The 99/5C has VGA, the 1 MB ram, IDE drive support, 9938 capability, USB mouse, USB gamepad, and USB keyboard. The pieces are there, but if the support isn't there for something like the 99/5C, then I do not see how building a 99/8 is going to be even be more supported.

 

Just saying..............

 

Beery

[Tursi]

then one should take a second thought at the 99/5C

Yeah... but didn't you say it was like $600? That's a lot to pay for a retro machine... I was pretty interested until I heard that. (If I misunderstood, please correct me!)

 

Fabrice's machine is a work of art, and might be worth that, but a ARM-based emulation platform at that price point is a hard sell, at least for me. Personally I consider the 9938 to be a limitation, rather than a boon - the community is shifting behind the F18A - there's already nearly as much software for it as the 9938 ever got on the 99/4A. (The other parts are great, though).

 

I've no interest in a 99/8 reproduction. There's no software for it /at all/, and indeed it seems to be an incomplete design. It functions as a prototype, but there was probably more to come. Back in the day I'd have done a heck of a lot for one, but today it's a curiosity (to me, at least). Without software a curiosity is all a computer can ever be. We have MAME to satisfy that.

[Omega-TI]

 

The 99/5C has VGA, the 1 MB ram, IDE drive support, 9938 capability, USB mouse, USB gamepad, and USB keyboard. The pieces are there, but if the support isn't there for something like the 99/5C, then I do not see how building a 99/8 is going to be even be more supported.

 

 

The good old 4A already has much of this available now.

 

A good portion of the community already have 1meg cards in their expansion boxes, but after three years there is little new software that exploits it.

The TIPI pretty much satisfies the storage and mouse issues along with other benefits.... at an affordable cost.

The F18A has attained traction in the community and I know people who would love to see another run of them now, instead of waiting for a newer version at 'some time in the future'.

 

There is something to be said for working with what we already have and building upon that. It's worked so far, and from a purely monetary aspect, probably a more user friendly and inviting prospect for most people.

[Opry99er]

Fabrice's machine is a work of art, and might be worth that, but a ARM-based emulation platform at that price point is a hard sell, at least for me. Personally I consider the 9938 to be a limitation, rather than a boon - the community is shifting behind the F18A - there's already nearly as much software for it as the 9938 ever got on the 99/4A. (The other parts are great, though).

 

I have to agree 100% with this. The fact that this is an ARM-based emulator implementation on hardware makes it really hard for me to buy in. I was there, just like many others, and can tell you that it felt like emulation.... now, it was still new and maybe not fully debugged, but I've played Parsec and Super Demon Attack enough to know what they feel like on real hardware.

 

Now maybe this 5C isn't designed to be a frame-perfect arcade machine, and possibly it is designed more to attract developers and those wanting 9938-based 80 column capabilities. But I want it all.

 

All that said, the 5C is a brilliant piece of hardware and software... just not for me at that price point, unfortunately.

[mozartpc27]

I get the impression the Geneve also suffered another issue the C-128 had: It was more like two computers in one box than it was an upgraded version of the older computer. Quite literally in the 128's case.

You can't easily make software that runs nicely on the newer system and gracefully downgrades itself on the older one, in the way you can on an IBM-compatible or a GameBoy Color. You can't choose to make Geneve or 128 software that degrades gracefully on a 4a or a C64.

 

The 99/8 was (more or less) just a faster 4a with more RAM, so there was a lot of potential for software sharing. Just as a lazy example, a TI-Writer II with larger(or even multi-)file support on a 99/8 that ALSO ran on a 99/4a by dialing back to TI-Writer I file behaviors, while still benefiting from any other changes.

You can't really do that with a C128 or a Geneve, because the new computer behaves so differently outside of compatibility mode. .

 

 

ON THE OTHER HAND, the 99/8's graphics capabilities would have been a serious sore spot. The 9918a was looking pretty dated, and my impression is that the 9118 didn't actually bring anything new to the table.

 

 

I started this thread with a comparison to the C65 and the project to basically re-create and re-invent that machine. Commodore's last 8 bit got nearly to production, but not quite, and came very late in the game - 1991. It was a wholly different concept from the earlier 128, although it was to have, like the 128, a 64 mode. But whereas, as you say, the 128's 64 "mode" was in fact just a second 64 computer inside the 128 box, with a dedicated processor and everything, producing 99.9% compatibility or something like it, the "Commodore 65" would only have emulated the 64 in software, resulting in a compatibility of 75% or so (which I have never really understood what that means - it seems like it could mean that like all of the software would be 75% compatible, or, to put that another way, not compatible at all!).

 

The issue about compatible software will exist for this new machine just like what you are talking about with the TI-99/8; since the REAL C65 was never released, there is NO software for its native mode. Yet these folks at mega65 are doing it anyway. I must say the Commodore 64 software/hardware scene is rather robust for a machine dead 25 years now, and I guess that is not maybe the case with the TI-99 scene, but it seemed like it was from a (mostly) outsider's perspective.

 

The 65 was to be, in other words, quite a different machine, and of course it has long since reached mythical status amongst Commodore fans. I was so impressed that there were actually people developing a latter-day 65 - the time involved seems just enormous - but it made me wonder if any other retro communities had similar project developing - or if other retro communities would even be aware of just how far these Commodore guys seem to think they can take it.