[Coco] Re: Coco Digest, Vol 31, Issue 10

[farna at att.net]

--
Frank Swygert
Publisher, "American Independent 
Magazine" (AIM)
For all AMC enthusiasts
http://farna.home.att.net/AIM.html
(free download available!)

 -------------- Original message ----------------------

> Message: 4
> Date: Sat, 4 Mar 2006 11:11:28 EST
> From: RJRTTY at aol.com
>
> In a message dated 3/4/06 9:53:08 A.M. Eastern Standard Time, farna at att.net  
> writes:
> 
> >Making an FPGA enhanced CoCo does sound feasible, but I have to  ask why? 
> That's a lot 
> >of work for a product that will end up costing more than most would  want to 
> pay. I 
> >understand the desire for a high performance CoCo, but with the cost of  PC 
> hardware and 
> >fast emulators, the "CoCo4" is already there. A decent Pentium 3  processor 
> will run an 
> >emulator much faster than a CoCo
>  
> David Kiel's emulator can break even with the coco3 on a 16 mhz 386  machine.
> It is written in X86 assembler so it is the fastest emulator by far.   It can 
> also
> access real coco double density disk drives directly from the  emulator.
> It can also access high density drives directly from the emulator.

So running it on a Pentium class or higher machine would be much faster. The VIA processors I mention below are the "Eden" chips, comparable to P3 class processors and running at 400-1500 MHz. Like the AMD processors, the VIAs perform a little better than Intels at the same clock speed, so ratchet that speed up ~20% for a direct comparison to Intels. The VIA C7 is comparable to a Pentium M (P4 core). VIA has concentrated on low power applications, and there are lots of small boards made using the VIA chips, some about the same size as a 5.25" floppy drive. A full height floppy case could house a half height 5.25" floppy drive and the computer board -- with all PC type I/O included. That's with a VGA controller built in. The VIA brand boards are Mini ITX (6.7" square) or Nano ITX (4.75" square) boards. Most have a PCI slot in them. There are other companies, including AMPRO and several other embedded system makers, with various baords. I'm sure a low cost one could be found. Then th
ere may be a market for using the system as a development tool also. 

>  
>  
> >So why not tweak an emulator to extend
> > the CoCo capabilities?  That should be
> >doable while still maintaining  compatibility.
> > It would be easy enough in NitrOS-9. 
> >DECB could  always have the option of
> > running a standard emulator. Make the  thing
> >self booting with one of the free DOS systems,
> > and  run from something like a VIA 
> >mini computer board, package in a small  case
> > with either a small laptop HD or a 
> >flash card to boot from,  and there you have it! 
> 
> Now this is the best idea I have heard so far for
> a coco4 project.

Actually, I'd think of it more as a CoCo 5 or 6 since it would be well past the capabilities of hte CoCo 3. Emulation would be used for compatibility, though I don't know how much compatibility is really needed. It really just needs to be an easy to program and interface machine, but for obvious reasons this group *would* want it to be familiar. 

> >I'd like to see DECB patched to something 
> >like the old 512K  BASIC capabilities, and 
> >most of the ADOS enhancements added.
> > A  configuration utility similar to a BIOS 
> >editing screen on the PC so that  configuration
> > of drives (at least) could be changed 
> >easily, and  as often as desired, would be nice.
> > Support for larger disk drives would  
> >be a necessity also (and that might be a problem
> > -- but maybe  not if an enhanced 
> >and standard emulator were with each machine...)  
> >Once all that was worked out and 
> >running nicely, it could then  be programmed into
> > an FPGA and a new board made if 
> >that's a real  desire. At least a standard for the 
> >extended capabilities would be  worked 
> >out. 
> 
> Do it first in the emulator THEN you have a guide to build
> the real thing.  The first step anybody should take in
> building a coco4 is to emulate it first on a PC.

Yes, that would be a logical development step! 

>  
> 
> >The only thing missing is an easy way
> > to interface to the  outside world. There's your
> >hardware project: an external interface  board
> > that plugs into a PCI slot. I always liked 
> >being able to  easily use the CoCo joystick
> > ports (and even the cassette port) for  
> >easy I/O. Maybe even emulate the CoCo cartridge port on the same card!  Some 
> >easy to program data I/O ports and a half dozen relays on a card  would be 
> realy 
> >nice. The CoCo's strong point has always been  experimentation and easy 
> >programming -- an I/O card and the standard  DECB with extended memory 
> >capabilities (long BASIC programs -- and might  want to extend variables to 
> more 
> >than two character names) would really  open things up. 

>  
> YES !!!!    Implement all the ports on the coco3 on a ISA or  PCI card.   
> THis is
> good.  And you could use the real parallel and serial  ports of the PC as 
> coco4 
> enchancements beside the standard ones.   How about adding  connectors to
> the cartridge port so more that one real coco cartridges could be plugged  
> in? 
> I think I am going to look into doing this as my next hardware  project.
>  
>  
> Roy

I'm glad I inspired someone with the time and skills to consider this! I wish I had them. I'd extend the cartridge port out, but not necessarily with the edge card connector. I'd suggest a dual row header suitable for wire wrap or standard header connectors. An adapter for a cartridge can always be made. I don't really see the need for more than one though. It would really be useful for experiments/projects, rarely used for actual cartridges, and with all the needed hardware I/O built in why would more than one be needed? Build an adapter for transferring the occasional cartridge if necessary, but that's all I can think of using it for.  Maybe testing some old CoCo hardware?? I think most cartridges can be obtained as .dsk files now. But the joystick ports, even as header connectors on the board, would be really nice! There are real serial ports on most of the boards, but the D/A conversion of the old joysticks would be nice, especially if the emulation software is modified so they
 can be read like the CoCo ports. Maybe some kind of improved I/O board that's easily programmed in the "improved DECB" of the "improved" emulation package instead of the normal joystick ports? Easy to program I/O for projects is the main thing. 

I really think the main market for such a machine would be hobbyist and experimenters who prefer to program on the machine they will be using, not using a PC then loading onto something like a PIC. The CoCo was originally targeted for home users as a step into a PC, so Tandy limited it's power. This "new CoCo" would be a limited computer in some ways, built as a learning tool for programming and for experimenters/hobbyist. It's just to difficult to program on a PC! You can't get much easier than the old line number BASIC!!