136 replies to this topic

[retrocon]

Did you ever see that homebrew tetris game for the channel f? That seemed pretty impressive. Channel F actually has a frame buffer doesn't it? It's incredible how bad the games were that came out for it, but maybe it was just bad programming or lack of tools or time.

In any case, Channel F will probably go down in history as the most under-utilized console ever.. but it will always be the first

[Tom]

retrocon, on Wed Aug 3, 2005 12:36 PM, said:

Channel F actually has a frame buffer doesn't it? It's incredible how bad the games were that came out for it, but maybe it was just bad programming or lack of tools or time.

No, what's really incredible is how tedious and time consuming video RAM access on the Channel F is.

[~llama]

Tom, on Wed Aug 3, 2005 12:55 PM, said:

No, what's really incredible is how tedious and time consuming video RAM access on the Channel F is.

Amen to that.

[retrocon]

~llama, on Fri Aug 5, 2005 12:07 AM, said:

Tom, on Wed Aug 3, 2005 12:55 PM, said:

No, what's really incredible is how tedious and time consuming video RAM access on the Channel F is.

Amen to that.

Is that what that Saba chess cart that is always mentioned fixed? Did tetris use that as well?

Was this a case of hardware that might have been decent in some respects but was held back by a severe design flaw that programmers couldn't overcome easily?

Can you give an example of why it is so terrible?

I'm just curious because channel f was my very first system and I have sweet memories of it even though the games were awful. I must have played Cart 10 (Maze) a thousand times with my sister.

Edited by retrocon, Thu Aug 4, 2005 11:16 PM.

[Tom]

retrocon, on Fri Aug 5, 2005 12:13 AM, said:

Is that what that Saba chess cart that is always mentioned fixed ?

Not really.
What makes the chess cart useful for homebrewing is that it has a 3853 which decodes the F8's perverted control signals to standard control/address/data bus signals. So you can remove the chess ROMs and replace them with EPROMs containing your own software. Doesn't have anything to do with VRAM access.

retrocon, on Fri Aug 5, 2005 12:13 AM, said:

Can you give an example of why it is so terrible?

See http://www.freelists.org/archives/channelf...4/msg00025.html and http://members.cox.n...ddle/lights.asm

Edited by Tom, Fri Aug 5, 2005 5:08 AM.

[~llama]

Tom, on Fri Aug 5, 2005 12:54 AM, said:

See http://www.freelists.org/archives/channelf...4/msg00025.html and http://members.cox.n...ddle/lights.asm

I struggled for a few weeks to get a little demo running in MESS and simply couldn't do it, i never really figured out the way things work in a Channel F. I gave up on it.

[mos6507]

If you want a different homebrew platform besides the well-known I would look into the Astrocade.

It's now pretty well documented. There are several disassembled games as well as the internal ROM.

It's a pretty powerful system other than the fact that it doesn't really have hardware sprites per se, just a pseudo blitter. But it can do things that framebuffer-only systems like the Apple II can not. It allows you to change the color palette on every scanline like the Atari 2600/5200/7800, but few games ever tried that. It also has a very wide 256 color palette and the sound hardware is pretty good.

It also is ripe for hardware enhancements. It has a DMA slot in the back that can be used for peripheral/memory expansion.

Those who have done any Z80 programming for the Colecovision or Gameboy would probably be able to get up and running on the Astrocade without too much hassle.

http://www.ballyalley.com/

Edited by mos6507, Wed Aug 10, 2005 12:42 PM.

[Atari Rescue Group]

Tempest, on Tue Oct 19, 2004 2:19 PM, said:

Quote

Oh yeah, there's also the Videobrain computer that's based on the F8.  

http://www.obsoletec...m.org/vidbrain/  

Unlike the Channel F, I've actually played with one of these.  

I wrote up a section for the VideoBrain for a guide that was never used (cartirdge and hardware listing). I could post it if people were interested.

Tempest

Tempest,
Can you please post that article?
I just found a VideoBrain console, unfortunately with no controllers or power supply.
Also, there were a pile of brand new sealed cartridges I got:

Two each of:

Music Teacher 1 ED01
Math Tutor 1 ED02
Video Artist ED05
Lemonade Stand ED06
Pinball EN02
Checkers EN04
Vise Versa EN06

And one each of:

Demonstration ST01
Blackjack EN05
Timeshare CM01

The timeshare box says it changes the VideoBrain into a timeshare terminal to access data banks and computer power of a large timeshare system.

Huh? What's that?

[Atari Rescue Group]

Ok, I found your article on the VideoBrain in another thread.

[e5frog]

~llama, on Fri Aug 5, 2005 9:46 PM, said:

I struggled for a few weeks to get a little demo running in MESS and simply couldn't do it, i never really figured out the way things work in a Channel F. I gave up on it.

Now there's help!


VESwiki is now UP!

http://www.bingbangboom.us/productions/ves...title=Main_Page

[LS_Dracon]

Nice!
Good to see a pacman port for VES.

[e5frog]

Veswiki has moved to http://veswiki.com




I added up the cycle time of the operations needed to plot a single point into Channel F VRAM, it takes 24 cycles, not counting the delay needed for the older system. But you could use the delay time for other things I guess, like reloading the scratchpad registers with new data perhaps...

Edited by e5frog, Wed Nov 21, 2007 4:58 PM.

[Rev. Rob]

I just wanted to go on record saying that I will happily pay for a nice Channel F homebrew. This was a great system with some really nice games and a controller that was way ahead of its time.

[e5frog]

Rev. Rob, on Mon Jan 28, 2008 7:33 AM, said:

I just wanted to go on record saying that I will happily pay for a nice Channel F homebrew. This was a great system with some really nice games and a controller that was way ahead of its time.

I have a spare cart to modify, it all depends on how much you're willing to pay for it, it's a lot of work. Send me a message and we can discuss it.

[Rev. Rob]

LS_Dracon, on Tue Aug 14, 2007 9:12 AM, said:

Nice!
Good to see a pacman port for VES.

It would be even better to get to play it on an actual VES.

[supercat]

Bruce Tomlin, on Thu Jul 21, 2005 1:02 PM, said:

The F8 data bus is strictly 8-bit. There are no "extra bits".

What about having multiple bytes of ROM associated with each address? Every type the F8 asks for a byte of memory, fetch four bytes from the EPROM. Those bytes could then hold the modified PC value or whatever else was needed.

[e5frog]

supercat, on Thu Mar 6, 2008 5:20 AM, said:

Bruce Tomlin, on Thu Jul 21, 2005 1:02 PM, said:

The F8 data bus is strictly 8-bit. There are no "extra bits".

What about having multiple bytes of ROM associated with each address? Every type the F8 asks for a byte of memory, fetch four bytes from the EPROM. Those bytes could then hold the modified PC value or whatever else was needed.

Can you elaborate on that I don't quite get what you mean.

Thanks.

[Bruce Tomlin]

I suppose if you didn't mind:

1) either using a 24-bit wide memory or somehow generating a 3x or 4x clock to fetch individual bytes

2) modifying an assembler to generate the extra 16 bits of "hint" address (don't forget conditional jumps!) Of course existing games would be much more difficult without a thorough disassembly.

then you could do it.

I think a small CPLD (but bigger than a GAL) would be a better idea.

[supercat]

Bruce Tomlin, on Sat Mar 8, 2008 5:52 PM, said:

I suppose if you didn't mind:

1) either using a 24-bit wide memory or somehow generating a 3x or 4x clock to fetch individual bytes

2) modifying an assembler to generate the extra 16 bits of "hint" address (don't forget conditional jumps!) Of course existing games would be much more difficult without a thorough disassembly.

then you could do it.

I think a small CPLD (but bigger than a GAL) would be a better idea.

I looked through the F8 documentation and I think having a CPLD do the address calculations is really the only approach that would produce compatible code, since the 8-bit offset need not be in the code itself. I did have another thought, though:

The easiest way to make a Channel-F cartridge would probably be to have a some other microcontroller (like a PIC 18F) feed instructions to the F8. Because the F8 doesn't have its own program counter or bus controller, this could actually be somewhat easier than doing such a thing on the 2600. I would think the F8 could 'idle' more easily than the 6507 while the other CPU was busy, and the code-stuffing logic could be much simpler since it wouldn't be necessary to use tricky instruction sequences to get perfect timings.

Of course, a game coded in such fashion might not seem as much like a Channel F homebrew, but it could run in a real Channel F machine.

[LS_Dracon]

Just a mockup

Attached Thumbnails

• 

[superbee]

LS_Dracon, on Tue Mar 11, 2008 8:09 PM, said:

Just a mockup

Very nice
That look awesome, can't wait.

[LS_Dracon]

I give a look on F8 processor assembly, but this realy complicate compared to 6502, and is a bit sad to learn becouse you be restrict to channel F, learning 6502 you have a wide range of possible machines to code for.
I'll not try the channel F, not for now.
But I admire people who code for it, like e5frog.
Here's another mockup:

Attached Thumbnails

• 

Edited by LS_Dracon, Thu Mar 13, 2008 8:08 PM.

[djmips]

supercat, on Sat Mar 8, 2008 4:04 PM, said:

I looked through the F8 documentation and I think having a CPLD do the address calculations is really the only approach that would produce compatible code, since the 8-bit offset need not be in the code itself. I did have another thought, though:

I'd really like to see someone (supercat) design a simplest replacement logic that allowed compatible code. This is the only satisfying solution. I'd do it myself but when I investigated this particular concept it was too foreign for my skills and poorly documented to allow me to find a solution quickly. I'd make a game for this system if I could actually make a cart for the thing. The assembler isn't the hurdle that the strange F8 cart hardware is. I don't think there are enough SABA Chess around to modify into homebrew dev systems. I've been looking for one for a very long time.

- David

[e5frog]

Very nice graphics, a note though is that the pixels are approximately 3:4, three times wide, four times high.

I'm feeling a bit of an urge to start programming...

Edited by e5frog, Thu Mar 13, 2008 3:56 PM.

[supercat]

djmips, on Thu Mar 13, 2008 3:55 AM, said:

I'd really like to see someone (supercat) design a simplest replacement logic that allowed compatible code. This is the only satisfying solution.

An F8 cart would probably require a Xilinx XC9572 and a ROM. Those two things together should suffice. As for how to program the XC9572, that shouldn't be difficult but I don't have a Channel F and am not nearly as interested in that system as the 2600. An XC9572 would be a bit under $4 in quantity 100, so that shouldn't be too expensive; it should have enough resources to allow a fairly simple VHDL implementation without having to hand-optimize things.