Trying to create "new" system

[The Mysterious SirPaul]

Hello.

 

I'm trying to create a video game system which has the "Classic" feel to it, and I need a little help: What kind of hardware should it use, so that it would strike nostalgia to the old atari users, and still make it graphical enough to not bore the younger crowd? thanks in advance for your advice.

[Noble Kale]

First, you'll need to make the choice. Is this for personal use (as a 'custom machine'), or do you want to try and sell it? The reason behind this question, is things like: Production costs, parts... etc..

 

Your budget will influence the kind of hardware you will use. There was a similar thread on the newsgroup rec.games.video.classic a long time ago (when I read it about a year ago), which became very detailed and would be of interest to you.

 

I wish I could help you more, but all I can do is point you in the right direction. However, I can suggest you consider the following factors when designing this system:

- Ease of use

- Display quality (ok, so you want a 'nostalgic feel'... but from what era? Are we talking about 2600 nostalgia? 7800 nostalgia? or maybe even slightly later systems like the master system?)

- Production Cost/Ease of Production (I strongly suggest you read the earlier threads about mass producing atari carts with regard to manufacturing issues.)

 

And the big one...

-Ease of Programming - You don't want to make a system no one wants to write games for (ie: you wouldn't want to be writing at the same level as the 2600). Various chips/processors will have different levels of programming skill requirements. My first suggestion, is that you go and get your hold on all the programming notes/resources you can get your hands on, and realise *what you are up against*. The thing to remember is: if you want to produce this console for private use, you'll have to do *all* the programming yourself. If you want it to be commercially manufactured and available, you'll be able to bring in others who may wish to develop for the system, but you'll bring on more headaches with regard to manufacturing issues.

 

Basically, if you want to avoid manufacturing problems, you'll be the only one with the console, and the only developer for it. If you want to overcome programming problems, you'll have to deal with manufacturing issues.

 

Sound like a headache? It is. I've been long considering doing the same thing, but I'm a software kind of person, not hardware. You'd be wise to consider all factors

 

That said, I wish you the best of luck. Consider everything, and if you still want to do it, make sure to talk to the people around the community. I'm sure they'll have some tips/ideas on the development of the system. I hope you succeed :)

 

Kale

[The Mysterious SirPaul]

maybe I should have posted more info. It's going to be manufactured and sold, it will have 2 cartridge slots (1 for Atari 2600/7800 compatibility, the other for 5200 compatibility), and it will be comparable to the SNES in "new game" quality. Otherwise, I have no ideas on what it would have.

[Viso]

Sounds like it could be an exciting project. I'd say that if you're looking for more than compatibility with the older systems, you should think about doing the following:

 

* Providing video memory to cover the full screen and maybe some extra for buffered graphics. Makes programming easier.

* Support for the full TV resolution and some lower resolution modes. This should allow for some very nice graphics or for graphics that that have the look of an older system.

* Support for digitized sound samples to be played back as instruments, like the old mod file. The Amiga has a chip that plays mod files directly.

* A processor that can support games written in C. This goes a long way to providing the ease of programming Kale mentioned. C libraries to use the hardware will be needed at some point, too.

* Support for loading new game software, and maybe even the Atari game software, without a cartridge. Using ethernet would work well for those of us with computers, and that can be done with low-end processors.

* If using ethernet as mentioned above, allow new games access to the TCP stack, or at least the ethernet hardware (in which case the games could provide their own stack).

* Support for a generally availble controller for another modern game console rather than requiring the use of old controllers or controllers that are specific to this system. I recommend Playstation controllers -- they're pretty easy to use.

 

BTW: While Kale is a software person, I'm a software person that tinkers with electronics. I've been playing with PIC microcontrollers lately and I have one reading data out of a Playstation controller and putting the results on an alphanumeric LCD. This is part of what I've been doing for fun over the past month or so.

[The Mysterious SirPaul]

True, Playstation controllers would work, but what about games which use paddles? would I have to emulate the analog stick(s) to work like paddles, or what? Also, I've been considering the system, and I decided to make one modification: I would make it a home computer/video game system, with it's own os and everything. I do have the programming knowhow, and it's been my dream since I was 4. Do you think I should? or would it mess the system up?

[Noble Kale]

While Kale is a software person, I'm a software person that tinkers with electronics

I'm making my way through university electronics

I don't get to tinker as much as I would like, but since it's a robotics course, I'm going to need a firm knowledge, and this place is a goldmine of resources in that area

[Viso]

True, Playstation controllers would work, but what about games which use paddles? would I have to emulate the analog stick(s) to work like paddles, or what?

Sure, but I can't imagine that being horribly difficult. I'm about to start implementing this kind of emulation, except that the output is going to an Atari 2600 / 7800. The biggest problem is the timing, but I already know how I'm going to deal with that.

Also, I've been considering the system, and I decided to make one modification:  I would make it a home computer/video game system, with it's own os and everything.  I do have the programming knowhow, and it's been my dream since I was 4.  Do you think I should? or would it mess the system up?

I can't see why it would mess the system up, especially if the game software is given free reign over the system, which is still common on modern consoles. If game software cannot have free reign, then more care will be needed in the design of the OS and its API's.

 

You may wish to consider using parts of, or even a whole, Open Source OS on the system rather than make an entirely new OS. It'll still take a fair amount of work to get running on the new hardware, but the result will be a more mature OS with support for a good number of goodies, like TCP/IP and ethernet, and lots of useful tools will be a compile away. There are some versions of Linux for embedded systems that might work well. Of course, these typically don't support a GUI, but that may be prefered for your system.

 

That said, you'll need to maintain a workable balance between complexity, capability, and cost. If you run an OS with a game, the system will likely need more RAM. To run an OS at all it will help to have something a good amount faster than the CPU's in any Atari system. The faster CPU will help run software written in C, so that will also help on the ease of coding issue.

 

If you go with a nice OS, having more than 16-bits of address space for an 8-bit bus may become a nessecity. Going to a 16-bit bus doubles the amount of addressable data, but 128Kb may still be limiting. Bank switching could make it difficult for the OS to run programs, unless the OS just runs a program at a time and provides API's the rest of the time, like DOS.

 

I don't know how much complexity or cost you are willing to deal with or what ideas you have come up with, so the following might not be useful. I just checked out DigiKey. They sell 33MHz 386DX's for under $14, and 33MHz 486GX's (I don't know what a GX is) for $33. Those might make for a nice CPU in your system. There may be other processors that will work better, though, but these will provide the performance you're looking for and some more. You might need to write the 6502, 6507, TIA, POKEY, etc... emulators in assembly, though. Structure it right and you can use that code in the creation of an NES emulator.

 

I'm making my way through university electronics

I don't get to tinker as much as I would like, but since it's a robotics course, I'm going to need a firm knowledge, and this place is a goldmine of resources in that area

Cool! Sounds interesting and fun! I have to admit I haven't messed with robotics, just PIC microcontrollers and digital logic chips mostly. Still, they can do plenty, they just can't move.

[The Mysterious SirPaul]

hrm.. sounds good, though I was thinking of somehow implementing the processors within the system, if possible. Also, I kind of doubt I'd use an intel processor, as they run a little slower than Motorola processors. Also, I would think that the games would run better, since they also use motorola processors. Of course, they would use their own processors, so...

[Thelen]

sounds nice, designing a new system, but first i would forget the 2600/7800/5200 compatabiltity. i doubt if you know enough to create a new 'computer'. i'm thinking of creating a small computer too, but with a little lcd display and a 6507 processor (atari 2600) and a little ram etc.

 

would it be possible to program a atari 2600 emulator in a microcontroller ? (never done anything with a microcontroller..)

 

thelen

[The Mysterious SirPaul]

hmm.. good point. I don't really need the systems in it, and I could make clones of the games. Also, about your question, I haven't really worked with microcontrollers, so I'm not sure either.

[Viso]

hmm.. good point.  I don't really need the systems in it, and I could make clones of the games.  Also, about your question, I haven't really worked with microcontrollers, so I'm not sure either.

So, you were going to include all the 5200 & 7800 hardware just so your system could run the old games? That is silly. You should use a fast enough processor so that you can emulate the Atari consoles. There may also be some address space overlap between the 5200 and 7800, so if the various chips from those systems were used, they'd also need to be dynamically disconnected from the data bus to prevent two things from using the same address space.

 

As for using microcontrollers to emulate the 2600, it would take a high-end microcontroller to run emulation in real-time. Most microcontrollers don't have enough performance.

 

Also, the 6502 is not a Motorola invention, but it was made by ex-Motorola engeneers.

[The Mysterious SirPaul]

Yeah.. I guess I wasn't thinking straight or something. The reason I wanted to use the hardware was because I used Atari emulators before, and they didn't seem to work right. The colors and sound were all off. I guess I was using the wrong options or something.

[Thelen]

an intel 386/486 is a good point, and cheap too. but i don't like them. i think a 6502 compatible processor is better. but you wanted to make a computer that created graphics like the atari 2600 and some sort of that, have you been thinking how you will generate a display/video signal ? what about the video hardware ? i think this is the hardest part......

 

making a computer with a processor isn't so difficult,the processor will execute your program, but to show some graphics/pixels on a TV....

 

thelen

[mos6507]

When it comes to classic-style hardware architectures, the Amiga was the machine that took it to the limit. The AGA chipset was the final evolution of that style of system, I think.

 

It gave you all the flexibility to control the raster beam on every scanline or even in the middle of the scanline, while still maintianing a full framebuffer.

 

The only thing the Amiga lacked was chunky graphics modes and hardware text modes. Text scrolling in higher color modes was notoriously slow on the Amiga due to having to shift multiple bitplanes.

 

If you are going to design an all new system I'd try to target a particular base-level CPU and clock speed and use that as your challenge.

 

The Amiga started with a 68000 architecture so memory footprint was no longer a problem.

 

It would be more challenging to see how far you could go with an 8-bit CPU like a 6809 or something. There were some very powerful arcade systems in the mid 80s that were still using 6809 CPUs that looked like they were using 68000s.

[Atari Master]

If you want my opinion. Make it a 32-bit system with classic type games. Like with a smooth arcade feel.

[The Mysterious SirPaul]

good idea about the 6502 processor, though a 32-bit processor might be better. I was thinking maybe making 2 systems: one that was more like a computer, with an operating system on it, and the other one being solely a video game system. It might work, but it may be harder. What do you think?

[eggplant_casserole]

Don't mean to sound pessimistic, but there are lots of things to consider.

 

Firstly, size. A 7800 and 5400 in one is bulky. Then there's component availability - several chips were custom chips. Sources aren't infinite.

 

This involves having some incompatible hardware, so you may as well emulate.

 

But the biggy - copyright. Don't forget Atari copyrights DO still exist, and I dunno how the holder would feel about people making things like this. Royalty payments would make the project expensive.

 

Why is the paddle controller a problem? Its just a pot FFS! If you're making a system, make a pad as well. Don't go using PSX pads, the ports will cost you a fortune! Not to mention being sued by Sony