I don't really know what FPGAs are capable of, so this may be a dumb question.

I think that what made the C64 great was the hardware support for video (sprites) and sound (nobody else had anything comparable). Can an FPGA do this? Can an FPGA generate a signal for a television?

The 6510 processor in the C64 was cool, but a Z80 could have been used instead and the machine would have been just as popular --- a TMS9900 or MC6809 would have likely made it more popular --- I would stick with the 6510 so it could run legacy programs, but would upgrade the 6502 instruction set somewhat so new programs could be smaller and faster, while still keeping with the 6502 style of programming.

I wonder if a C64 could be built today inexpensively. Only the board would be distributed, not the keyboard like in the original. You would hook up to a desktop-computer for developing programs that would be stored in some non-volatile memory on board.

In the old days, writing games for the C64 was as much or more fun than playing commercially sold games. I wonder if modern-day kids would think that this was fun too. They might complain that the game-machine is under-powered and lacks a big catalog of commercially-written games to buy --- in that case they might as well buy an XBOX.

I have no interest in the XBOX myself --- all of those games look boring to me --- there are really better ways to waste $300 or $400 dollars!

The Super Nintendo was pretty cool. It was a closed system though. Nobody could write games for it except Nintendo and the very few authorized game developers. It seems pointless to buy a game-machine and not be able to write games for it. I bought a Super Nintendo back when it was the popular game-machine, but when I found out that I couldn't write games for it myself I lost interest and gave it away.

There's half a dozen FPGA implementations (probably more) of the C64 currently in existence, the closest product match to what you describe is the Ultimate-64 from 1541ultimate.
This is an FPGA based replacement board for the C64.
http://1541ultimate.net/content/index.p ... 74&catid=9

Sprites, video modes etc, are easily accomplished on FPGA. C64 sound is a bit of a different story, the SID sound chip for the most part is straight forward to replicate, until you get to the filters, which it what lets the current designs down a bit.

All the video signal formats of yesteryear can be reproduced on FPGA, sometimes involving a few additional passive components (resistors/capacitors).

There are multiple Amiga and and Atari ST implementations based on FPGA, along with most of the systems that came before them
You mentioned Super Nintendo in your post, it too was implemented on FPGA a couple years back.
http://www.retrocollect.com/News/super- ... nsole.html

(There's also Beeb on a chip, Electron on a chip, and Atom on a chip. And the Spectrum Next is a Spectrum on a chip)

As Cray Ze says, the SID is the hard part, not because it can't be done adequately, but because any kind of compromise will get negative attention from one or other sector of the fanbase. As there were different SIDs with different quirks, it's impossible not to compromise.

Here's a post about SIDs:



See also http://mega65.org/

The first comment under the SwinSID article linked in BigEd's post suggests "a really nice FPGA based Sid replacement coming soon".
I went looking to see what was new.

FPGASID is new to me and probably what was referred to in that comment.
http://www.fpgasid.de/

Project Details.
http://www.fpgasid.de/project-definition

The project-updated page has a bunch of videos.
http://www.fpgasid.de/project-updates

That C64 FPGA system looked pretty cool! The estimated price was $260 though, which is pretty close to an XBOX.

For the most part, I think those retro machines from the late 1980s were just a lot cooler than what we have today.

On a related note, I also think that music from that era was a lot better than what we have today.

The C64 was cool, but it did have some pretty serious limitations, such as the 8-bit processor and the 64KB limit, that would make people uninterested now.

A modern-day Amiga seems like a good idea. By most accounts, that was the best computer of the 1980s. I didn't own one, so I don't know from experience. The OS was said to be quite good! The Jforth system was also said to be quite good.

I think what killed the Amiga was that it had a low-resolution screen. It had a lot of support for games (something similar to sprites, called "blobs" iirc). What people wanted though, was a high-resolution screen, because they wanted to edit digital photographs. They didn't really care about speed, or about animation support, because games are for teenagers.

The Atari-ST was primarily for music, because it had MIDI support. I'm surprised that it didn't become popular. Maybe MIDI support is fairly easy to do, and so the x86 machines were able to offer that as an option at a low cost. I don't really know anything about music production, so I can't comment on the subject.

I agree, the Ultimate-64, while nice looking, is a bit out of my acceptable price range for what it is. I'd prefer to drop a C64 core onto a cheap FPGA board and save some dollars.
The most advanced C64 style system I've seen on FPGA is the MEGA65, a C64/C65 compatible computer. Commodore never released the 65 to market though some prototypes eventually found the right hands.
http://c65gs.blogspot.com.au/

While the C64's limitations would not attract it's original target audience (the game players) in this day and age, the advent of the internet has allowed communities of a different type of audience to form, that being of electronics hobbyists, where even more limited devices have a cult following.

Screen resolution certainly contributed to the demise of the Amiga, though I think one of the main killers was the PC clone market and the sheer amount of business software that evolved around it. People would buy the machines for business, and often couldn't justify a second machine in the household just for games.

The Amiga had 8 hardware sprites though they were quite limited compared to using the blitter. Blitter objects, Bobs, were not specific hardware objects like the sprite generator, but were really just a fancy name for a programmatic series of instructions telling the blitter hardware to copy some blocks of RAM via DMA.

Yes, MIDI was the ST's niche, though a D.I.Y. MIDI box could be built very cheaply for the C64 of Amiga.
While not as popular as the Amiga, the ST did take moderate market share here in Australia.

I had never heard of the C65 or the plus/4 --- pretty interesting history! --- I remember that I wanted a C128, but they were too expensive, then the whole home-computer market collapsed anyway because everybody went to MS-DOS for both business and home.

I think that where I was going with this thread, was that I think a game-machine would be interesting if the user could upload his own games into it --- game-machines such as the Super Nintendo or Xbox are not interesting because all you can do is buy commercial games.

Resurrecting a retro computer is a good idea, because there are a lot of legacy games available that people could play right away.
This isn't crucial though, because (as I said above) the point of the exercise is that the user can upload his own games that he wrote himself --- playing legacy commercial games is no better than playing modern commercial games, and it is arguably worse because they are obsolete --- the point is not to attract gray-haired game-players with nostalgia for their teenage years, but to attract actual teenagers.

These would have to be teenagers who don't own their own desktop computer, but who do have money to buy a game-machine if it is less expensive.
The problem with this plan is that desktop computers are pretty inexpensive, and they are going to be as good or better for games than a 16-bit FPGA system (even if it does have sprites and audio, etc.).

Somebody mentioned that there is an FPGA version of the commodore Amiga. What happened to a PowerPC version of the Amiga getting sold? I read that this continued in Germany for a while. Switching to the PowerPC seems like a bad idea, because porting legacy programs would be difficult. The ColdFire seems like a much better idea because it is quite powerful, but it allows porting legacy programs easily (so long as you have the source-code so you just reassemble them).

The Canon Cat is another interesting computer. The source-code is in the public-domain now.
I doubt that the source-code for the Commodore-64, Amiga or Atari ST is in the public-domain.

I already had a C64 when the C128 hit the market, and a few people I knew bought them, but the lack of software for the 128 mode meant they were being used as C64's anyway.

I recently started on an FPGA design for something that matches a few of your ideas. If you imagine something similar to a C64, but with a much faster 6502, and graphics capabilities that exceed anything that was possible in the day, you'll have an idea of where I'm headed.

"Attract actual teenagers" - First you have to get them to release the death grip they hold over their smart(dumb)phones.

There were many PowerPC accelerator cards for the Amiga, though I don't think any came direct from Commodore.

Coldfire had been attempted by third parties, but as the source code of legacy applications was generally not available, the plan was to emulate the missing/different 68000 instructions. The end result was no better than current accelerators of the time. A summary of the differences at the following link.
https://www.sintonen.fi/coldfire-v4-m68k.txt

There was an Amiga Coldfire project.
http://www.osnews.com/story/966/The_Col ... ga_Project

Some discussion of the problems here.
http://bbrv.blogspot.com.au/2008/03/col ... egins.html

The Atari ST guys had better luck with Coldfire though.
http://firebee.org/fb-bin/index

I don't think I ever came across the Canon Cat, I had to look it up. It has some interesting ties to Apple.
There was an unrelated computer in Australia called the Cat, it was actually a rebadged Vtech Laser 3000.
http://oscilloclock.com/history/dick-sm ... e-ii-clone

The typical conspiracy theory as to why the Canon Cat died, is that Apple told Canon to stop competing against their Mac, or Apple would stop buying Canon printers --- Canon makes most of its money on printers --- the Canon Cat was an experiment for Canon, but they didn't have any real expectation that it would be a big money-maker, so they killed it.

I wanted to get a Canon Cat, because I'm a Forth programmer --- they were quite expensive though! --- the high-price may have been enough to kill it too, without any conspiracy-theory involving backroom dealings.

The most interesting thing about the Canon Cat today, is that the source-code is public-domain. As you said above --- the ColdFire is no good without source-code, because the machine-code is incompatible.