[Coco] Coco4..Who needs it? I do.

[jdaggett at gate.net]

On 30 Dec 2008 at 0:30, Carlos Santiago wrote:

> I believe that it is time for a more advanced Coco. The coco3 is 
> great, but there are some features and capabilities that could be
> added.
> 
> For example, if a new Coco was developed with a newer, faster, or 
> different processor. It could still run all the legacy software in an 
> emulated mode. Furthermore, the new coco could be integrated into an 
> FPGA. This would allow some really cool features to be added. Here are
>  some ideas:
> 
> 1. Higher resolution graphics with more colors and support for VGA 
> monitors.
> 

Being worked on. My vision is to have a desktop with 800x600 resolution and 
256 colors.

RS Basic will appear in a window whose size will be determined by the values of 
the HRES and CRES bits. This keeps some backward compatabiliy.

> 2. Network connections (ethernet and Wi Fi).
> 

Best done with a peripheral processor in my opinion. Why burden the 6809 down 
with the TCP/IP work?

> 3. Additional I/O such as USB.
> 

While a USB core can be include into the FPGA SoC, it would be better again to 
be an peripheral device and memory map it. USB chips are cheap and can be 
memory mapped or serial connected to. 

> 4. A PS/2 or USB  keyboard with key mappings for the Coco.
> 
done.

> 5. New peripherals including a USB mouse, digitizer pad, USB keys and 
> hard drives.
> 

mouse, keyboard, and storage is in progress.

> 6. The multipack interface could be integrated into the system along 
> with the floppy controller.
> 

Doable. I do have a core for a WD1772 floppy controller (not mine)  that has 
been modified to allow access to HD Floppies. It does synthesize.

> 
>   I would like to see an effort similar to the work that Steve Bjork 
> is working on but with a real goal of have working hardware in the 
> next 3 to 6 months. This may be a tough goal, but with the right 
> people involved, I think it maybe possible.
> 

One year is more likely depending on how many people are involved. 

> Consider this:
> 
> A motherboard that can be installed into a PC chassis that includes 
> the following:
> 

If FPGA based board you are looking at an 8 layer board. Look for about $120 a 
piece with silk screen and solder mask in small quantities(10) and around 40 sq 
inches in size. Besides the FPGA will most likely have to be in BGA package 
with at least 256 pins. While I have experience soldering BGA packages without 
usign IR assited reflow, that is limited to 100 pin devices max. 256 or even 480 
pins is a huge task to do by other means than IR assited reflow. 
 
> A high performance Processor (hardware or emulated in an FPGA)
> 

In FPGA it would not be emmulated. 
 
> Standard memory DIMMs with SDRAM (DDR1 or DDR2).
> 

If you are doing a 6809/6309 processor in an FPGA you can get about a ten fold 
speed increase from the ASIC version. Even with the speed increase, unless you 
increase the Program Counter to say 24 bits, you have to do page memory. 8MB 
or 16MB SDRAM would be sufficient. DDR* SDRAM would almost dictate that a 
Xilinx Spartan 3 or better device be used or the Altera equivalent. My 
reccomendation is to use a FPGA development board. There are several out 
there that range in price tags. 


> An FPGA with the enhanced graphics and I/O capability.
>  -Support for all standard Coco1,2,3 graphics modes.
>  -Hires graphics (320x200, 640 x 480, 800x600, etc) with 256 or more 
> colors.
>  -USB controller and hub.
>  -IDE interface.
>  -Multipack emulation.
>  -Standard Coco 6bit sound.
>  -Enhanced Sound controller with multiple channels.
> 
> Slots for the ROM packs and/or additional I/O cards.
> VGA output .
> CoCo joystick interfaces.
> USB interfaces.
> IDE interface for a hard disk.
> User selectable setup for system to boot directly into disk basic or a
>  different OS from the hard disk. PC keyboard and mouse support with
> Coco Keyboard mappings.
> 
> Carlos Santiago
> 
Much of what you list is doable. Forming a team to achieve this is tougher. This 
is not a 3-6 month project even with a team of 8 to 10 people. 

james