2600 on-a-chip update

[Carlos_Lopez]

Just as a follow up to this thread, I wanted to try by myself how a scaled 2600 frame would look on the GBA screen. IMO for a NTSC game there is not significative loss in resolution. However, I did some simple tests and I concluded that the GBA is not able to scale the image at 60 fps. Anyway, this task can be done by hardware in the same cartridge and DMA the video directly to the GBA video memory.

 

Also, I'm generating RGB video now instead of composite, and the picture looks great on a Commodore 1084S monitor

 

 

http://www.mindspring.com/~2600onachip

[MegaManFan]

As usual Carlos, your pics are not only a site to behold but make me drool for the eventual possibilities your project may lead to. Keep up the VERY good work!

[jeff_a]

you KNOW I'm happy about the GBA testing

[Albert]

The RGB output is great, will this be an option for the final product or is it something you just whipped up for your development setup? Would be amazing if this could be translated to VGA output!

 

Also, those 2600 bitmaps running on the GBA are some tease..

 

..Al

[Paul Slocum]

Great to see you're making some progress, Carlos. The RGB is really cool!

 

-Paul

[mos6507]

15khz RGB output can be accepted by monitors via a scan doubler device. A scant few monitors will sync down to 15khz. Of course, C= monitors like the 1080/1084 are designed for 15Khz-only.

 

Scan doublers are very common for things like the Amiga. It might even be a simple enough device these days that it could be included in the 2600-on-a-chip design.

 

But I don't think I'd want to pump a 2600 to a VGA monitor unless it was to get it to LCD or something.

 

At best I'd want to output to component as some high end TVs have.

[Albert]

At best I'd want to output to component as some high end TVs have.

 

Yes, component output would be nice also, I have one television that can accept that signal. Of course, I already have two devices that output component signals (DVD player and PS2), and my receiver only has inputs for two component inputs so that would be a problem. Personally as long as s-video output is available I'll be happy. The differences beween composite and s-video are quite noticeable, but you'll be harder pressed to see the difference between s-video and component.

 

..Al

[Thomas Jentzsch]

IMO for a NTSC game there is not significative loss in resolution. However, I did some simple tests and I concluded that the GBA is not able to scale the image at 60 fps.

To bad, Thrust looked very good.

 

Anyway, this task can be done by hardware in the same cartridge and DMA the video directly to the GBA video memory.

Do you think it's worth the effort? And how much would something like that cost?

[Carlos_Lopez]

The RGB output is great, will this be an option for the final product or is it something you just whipped up for your development setup?  Would be amazing if this could be translated to VGA output!  

..Al

 

I decided to use RGB video because I was having some color shifting with composite video. That's because I was using a 28.6363 MHz oscillator as the clock source, and the chroma signal was derived from dividing this clock by 16 with the proper phase. It seems that the color delays obtained from this method are different than the ones generated by the TIA analog delayer.

 

To answer your question, the 2600-on-a-chip core will have digital luma and chroma output (directly from the luma-color registers). If the particular application needs analog video output, I'm considering to interface a video DAC as the ADV7177. This chip is not very expensive (less than $10) and you can select from composite, S-Video or RGB outputs, among other features. For cheaper single-chip applications, I will use clock shifting to generate composite video as I did.

 

Would be amazing if this could be translated to VGA output!  

..Al

 

As mos6507 mentioned, the problem is the horizontal sync frequency. Maybe some VGA monitors (multisync?) can handle 15 KHz, but I'm not sure.

 

One thing that can be done is to use a video buffer, and then just scale the frames to a higher resolution, but that would require an extra RAM chip.

 

Thank you all for the good feedback

[Carlos_Lopez]

Anyway, this task can be done by hardware in the same cartridge and DMA the video directly to the GBA video memory.

Do you think it's worth the effort? And how much would something like that cost?

 

Mmmh, I don't think it would be more expensive than doing this by software, but I can be wrong. The GBA already has DMA, and using bililear resampling to scale a 320 x 200 picture to a 240 x 160 on the same FPGA wouldn't take a lot of extra resources. I need to try this I guess

[CPUWIZ]

Well ?

 

* hit me with the digits *

[christianscott]

i'll be speaking way out of my depth here in term of specs and signals but i think the rosetta stone for a portable VCS would be something that you could hook up to a game gear screen. game gears are cheap and plentiful on ebay and the screen is one of the better ones. the case is big enough i presume and there are enough buttons to support all the features of a 2600.

 

somebody please make this possible, i know enough about it to know that the stumbling block is converting the video signal.

[Carlos_Lopez]

Well ?    

 

* hit me with the digits *  

 

Well, I really couldn't tell you about the cost of this at high production volumes. Basically a GBA cart of this kind would need the 2600 on-a-chip with the required glue logic (inside the same chip maybe), a RAM for video buffering, and some kind of ROM or FLASH memory to store the rom(s).

 

If I was to produce this at a low scale I would have to use a FPGA or a big CPLD (instead of an ASIC) for the chip. This component alone would cost about $30 in small quantities, so adding the cost of the RAM, the ROM/FLASH and the PCB maybe it would be around $50 in components (not considering the case). Perhaps there is a cheaper Actel or Quicklogic antifuse-based FPGA available from some distributor, but I haven't do a lot of research on this.

 

Anyway, I may be wrong but I don't think this can be homebrew produced because of legal issues, so I will have to persuade an investor to produce this at least at a mid scale.

 

I have a lot of work to do, but I'm hoping to have a prototype either of the GBA cart or a custom portable 2600 using my chip ready for CGE.

[CPUWIZ]

Sounds great.

[jeff_a]

i would pay anything for my gba to play atari 2600 games.

[cschell]

Just thought I'd post a quick update on this thread. I went over to Carlos's apt. last night to check out the 2600 on a chip, and the RGB picture does look nice. I'd say it's on par with the output from a 2600/7800 with a reasonable s-video mod. It's probably capable of producing a better picture than most of the video mods, but I didn't really look closely for whether or not it removed some of the common artifacts found when using those circuits. (And I think Carlos's monitor has seen better days as well.)

 

And in case anyone is skeptical given the recent track record of certain Atari projects and purveyors, I can verify that the 2600 on a chip is exactly as described. No 2600's or fake screen shots in sight.

 

Chad

[cschell]

Hi,

 

Someone sent me a PM regarding this, but I hit the wrong button and deleted it before reading it. So if you want a response you'll have to send it again.

 

Ugh,

Chad

[Albert]

Chad,

 

Thanks for the update, glad to hear you had a chance to check it out in person. I can't wait to see what happens with this project when Carlos has finished it!

 

..Al

[Rob Mitchell]

Next project .. Cuttle-Cart-on-a-chip marries 2600-on-a-chip

 

Of course .. I'm eating potato chips as I type this!

 

Rob Mitchell, Atlanta, GA

[EricDeLee]

The breakthrough of a Atari2600 on a chip would be enormous! Think of the possibilities!!!!

 

Right now, you can get someone to make you a portable atari probably for about $3-400. Problem is... whoever is making it, will not be making much of a profit. I'm building a tabletop Atari 2600, with a 5" color screen at the time of this writing, and I know for a fact that it is very time consuming. Each project is a labor of love!

 

But an Atari on a chip!!! WOW! Right now, my project Atari is a 4X4 square... that's about as small as I dare to cut it up. 4X4 still takes a lot of room.

 

anyways.. great job on the chip. I'll keep an eye on that!