The next EVOLUTION in Future TI Cartridge(s)

[Omega-TI]

It's been pointed out to me recently that the TI Guru's have been designing cartridges that are hitting the functional maximum capacity for a standard TI case. Basically, you just cannot cram any more in there. So, where do we go from here?

 

I see different paths we could take:

 

High Density SMT

While this method may not be for the average hobbyist, some of you have the technical knowledge and capacity to pull this off. This would pack more into the standard case design. Who knows where this would eventually lead or what new capacities a future TI could have. With bank switching and huge memory chips, I can see a whole new class of programs on the horizon.

 

While going SMT would functionally add more to a standard case design, a redesigned or fatter case would also allow for a multi-level or stacked SMT design as well. What could the future bring with that?

 

Case Redesign

A longer case would allow for some larger cards, this would enable the average hobbyist to add more components to a standard density PCB design. Of course someone, or a team of someones, would have get together to take the first step in cartridge design and manufacturing as well as a designing the new core format for the new PCB design to reside in the card. Once people have the layout for the board, those designers could then do their thing.

 

 

I can see this adding some excitement over the next couple of years, it would keep the hobby fresh and interesting, as well a take the TI to a whole new level. Comments, suggestions?

[Willsy]

Nah. The future is a micro-controller inside the cart case, and an SD card slot on the front of the cartridge. There'll be no ROM in the cartridge at all. Just uController RAM. The uController will load an image from the SD card and 'represent it' to the 9900 CPU by driving the cart bus via it's I/O. Probably need a fairly quick uController to do it, but I think it's possible.

 

The cart could also include a small (say 1x20) LCD display and a single button on the front. Then you could scroll through available images on the SD card, and pick a "cartridge".

[am1933]

Nah. The future is a micro-controller inside the cart case, and an SD card slot on the front of the cartridge. There'll be no ROM in the cartridge at all. Just uController RAM. The uController will load an image from the SD card and 'represent it' to the 9900 CPU by driving the cart bus via it's I/O. Probably need a fairly quick uController to do it, but I think it's possible.

 

The cart could also include a small (say 1x20) LCD display and a single button on the front. Then you could scroll through available images on the SD card, and pick a "cartridge".

So it is in effect a cartridge cartridge, "how much more cartridge can you get"?-I would have to answer-"none, none more cartridge!"

[matthew180]

You have to be careful with the expansions. If the 99/4A were implemented as an SoC, do you think people would stop using real hardware? After all, an SoC (whether done with a uC or FPGA) is not the real deal. When you are dealing with retro hardware like this, the majority of the appeal is nostalgia IMO.

 

I see the same thing in many retro forums where people are talking about making 24-bit color video boards with tons of pixels, 3D support, multi-megabytes of memory, MP3 quality sound, blah blah blah. They are describing a modern computer, and those kinds of enhancements, although possible, are grossly mismatched to the original system. Some people also finally realize that making a 3D accelerated video board is not as "easy" as they thought. Cheap and mass availability of computer hardware has lead people to believe that making this stuff is quick, cheap, and easy.

 

If we had a mega-cart, what would you do with it? You still have the problem of zero software support until people write new programs to use the new hardware features. Then you enter into the dilemma where people don't write software for the new hardware because not everyone has it, or they like the challenge of making something work on the original configuration.

 

This discussion has been had many times before. I started a thread a long time ago about the desired features for a mega-cart. Everyone has their own ideas. I don't think the community will ever reach a consensus on this because everyone has their own thing they want to do. Some people want GRAM/GROM krackers just to mess around with copying carts, for no other reason than to do it. Some people want tons of RAM just to say "I have X megabytes of RAM!", etc.

 

Personally I have thought about this a lot. Too much actually. I have started, and shelved, my mega-cart many times. If someone could come up with a reasonable set of design goals, that might help make such a cart a reality.

 

My requirements are something like this:

 

* SMD parts. Availability, cost, size, *low power*.

* USB drive support, i.e. no SD cards or special formatting.

* At least 1MB of memory as 8K banks.

* Each 8K bank configurable as ROM, RAM, GROM, or a stack.

* Possibly an on-board processor.

 

I never worked out how the configuration would be stored or if software would have to set it up each time. Also, the interaction with the USB drive was undefined, but some sort of menu/loading support would be nice to support a mega-cart with all known cartridges available.

 

The components I imagine would be a uC to provide the USB interface and FAT16/FAT32 support, an FPGA or programmable logic to support the host interface and implement the bank switching, etc., and a real memory chip (1MB SRAM) to contain the current "active" banks of memory.

 

With USB support I could also see a real-time connection back to a PC where the cartridge files could be located, or the memory could be read or written. This would make programming and debugging on the real hardware much faster and easier.

Edited April 3, 2014 by matthew180

[Omega-TI]

Nostalgia is a big part of the hobby, and Matthew makes some good points. For me, being able to do a little something other than playing video games is important. I need a little functionality to justify dumping money into this thing. A few functional enhancements would do the trick.

"Back in the day", would most of us have put a 10 megabyte RAM disk in our TI if it was affordable and possible? I think the answer is "probably". Would more of us have upgraded to VGA back then if it had been affordable? Again, "probably". Today many people use the F18A. To me even with the F18A the TI is still a TI, it's not a different machine just because it now has an awesome display.

I do not know if a 10 megabyte RAM disk is possible, but with todays prices, it could probably be 100 megabytes or more for the price, but then again, would the TI even be able to address that much memory? Like you say, someone would have to build it, how many would buy it and who would make programs for it?

You asked, "If we had a Mega Cart, what would you use it for?" For starters, I'd love to retire my widget so I could use a dust cover again. Having everything in one cart would be fantastic. We've always had cartridges that added "a little extra" to the TI, Mini Memories, Super Carts, Multiplan, etc. It's still a TI no matter how cool the cartridge.

Also, I actually see the possibility of convergence with a couple of things happening now. For instance, one guy is working on a Web Browser for the TI. A project like that could easily grow and become a super-browser over time and require a mega-cart for the program. There was also discussion in the past of where the received pages would reside, a super-RAM disk would fit that bill too. I dunno about you, but cruising the TI forum on a TI is just too damn cool!

[Gary from OPA]

Soldering one wire inside the console, to allow GROM access out of sidecar expansion slot on right, and adding an atmel controller doing the grom simulator logic to new version of NanoPEB/CF7+ allowing it to have GROM cartridges as images on the SDCard along with floppy disks,

[Davvel]

I think that a Cartridge that can save our work within itself on a regular SD card would be great. One would not need anything but this cartridge. I think that this is possible but maybe the market is small for it. I am not sure if one can have an XB 2.7 on the same cart that one can mount a virtual disk to save the source code to. Maybe it is technically impossible and that is why someone invented the NanoPeb or CF7.

 

I would appreciate comments and clarifications as I really would like to learn on these things.

[Omega-TI]

One would not need anything but this cartridge.

 

Well, they would probably need a 32K memory upgrade, but that can be accomplished for 'around' $11.03!

 

I gotta admit it would be sweet and with the lower price-point to get that kind of capability it might draw a few more people to the TI as a platform own and use.

[RickyDean]

 

Well, they would probably need a 32K memory upgrade, but that can be accomplished for 'around' $11.03!

 

I gotta admit it would be sweet and with the lower price-point to get that kind of capability it might draw a few more people to the TI as a platform own and use.

I think(my opinion) that such items as floppy emulators, hard drive emulators and even FPGA replacements for the Geneve (like the Geneve 2) should be considered first, so that others can experience the level that some of us have the capability of experiencing now, if our equipment is working. If this includes SD cards, CF cards , external usb devices, and so forth then bring them on.

[Omega-TI]

There are many different aspects to the TI hobby, some like games, others are into hardware, programming or whatever. As people grow into the hobby, many ending up spending more on new hardware, firmware and other associated things... and they become hooked. One thing I'd like to see is something to draw new or returning TI'ers into the fold. If the price point of the BAIT is low enough, we can "HOOK 'EM".

 

Some guys have 10 or more backup consoles, 'lending' one with a single compilation cartridge could get people interested too. Just a few years ago I came across Tursi's Classic 99 in a web search, which led to further searching, that led me to a user group and then to Atari age, that led to me hooking up with my first "TI-Crack-Dealer"... the rest speaks for itself. We've had a few others join up over the past couple of years too.

[+Ksarul]

Large RAM disks are theoretically possible--and could probably easily be built in sizes up to 20MB or so using old-school chips. There are 36-pin, 2MB RAM chips out there. . .so the only real drivers are demand, desire to realize the design, and the money to build test articles. . .of course, all of this takes that most precious of resources: time. So many fun ideas, and so few who have the time to actually work on them. There lies our biggest problem--we do not lack for usable ideas, we lack the time to realize them. I'm glad the current series of cartridge boards has such a high level of use, as it provides increased access to old software at the same time it allows wonderful new applications to be designed and distributed. Just look at all of the new games and applications developed in the last four years!

[Tursi]

Flash would be nicer for a big RAMdisk... would hate to restore a 20MB RAMdisk cause the DSR glitched or the battery contact was loose. Sure, writing is a bit slower, but it should be more stable. I've got some nice 128MB parallel flash chips on my desk right now.

 

A cartridge DSR is unfortunately limited to GROM, which again unfortunately, most pieces of software don't support GROM DSRs. So sadly the all-in-one solution isn't that easy.. you need SOMETHING either inside the console or off the side port.

 

But yeah, time is the big deal. Right now it feels like I'll never have time again.

[RickyDean]

Flash would be nicer for a big RAMdisk... would hate to restore a 20MB RAMdisk cause the DSR glitched or the battery contact was loose. Sure, writing is a bit slower, but it should be more stable. I've got some nice 128MB parallel flash chips on my desk right now.

 

A cartridge DSR is unfortunately limited to GROM, which again unfortunately, most pieces of software don't support GROM DSRs. So sadly the all-in-one solution isn't that easy.. you need SOMETHING either inside the console or off the side port.

 

But yeah, time is the big deal. Right now it feels like I'll never have time again.

yeh, I wondered why you are always asleep at your desk

[mister35mm]

How should a SMALL cartridge with a microcontroller ( Raspberry Pi Zero?) and a pair of USB ports, one for USB memory sticks, the other for PC connectivity.