Date : Thu, 18 Sep 2008 22:38:01 +0100
From : mfirth@... (Michael Firth)
Subject: BBC Micro on an FPGA?
----- Original Message -----
From: "Mark McDougall" <msmcdoug@...>
To: <bbc-micro@...>
Sent: Thursday, September 18, 2008 9:45 PM
Subject: Re: [BBC-Micro] BBC Micro on an FPGA?
> michael.firth@... wrote:
>
>> Have you got any more information on what hardware is needed for the
>> 'PACE' project, and where it is available from?
>
> <http://pacedev.net/forums/showthread.php?t=3>
>
I did have a look there, but it seems that a lot of the projects are
designed to work on either the 'PACE-P2 ' or 'PACE-P3' boards, which seem to
be proprietary boards you've developed, for which there doesn't seem to be
much details of the hardware spec (beyond the core FPGA).
>> Your sites seem to have a lot of information about the software /
>> firmware side of the project, but don't have much on what hardware is
>> needed to use it - e.g. what format the CVBS interface would take.
>
> Currently the composite output is designed for an ADV724, so it's strictly
> RGB out with composite video timing inside the FPGA.
>
Useful info, I'll look into what the ADV724 requires.
> A problem I'd like to solve is getting 50Hz composite video displayed on
> modern VGA monitors. 60Hz is easy using scan doubling, but 50Hz is not so
> straight-forward. On a lot of platforms, the video timing is central to
> the
> whole system timing, so the refresh rate isn't something that is easily
> changed.
>
For info, there was a version of the BBC designed for the US market with a
60Hz output.
I don't know how much standard 50Hz BBC software would run on it though.
> The idea of PACE is to be as target-independent as possible. There's a
> *lot*
> of FPGA dev boards out there - both Xilinx and Altera - so I'm trying to
> make it as accessible as possible. However, it *will* require some minimal
> porting effort (and possibly some hardware add-ons) to get it running on a
> new platform.
>
>> As we've a lot of old Xilinx evaluation boards lying around at work, I'd
>> hope that there would be something I could push into service.
>
> The current designs utilise internal RAM to varying extents atm - it's
> something I'm trying to move away from where possible. It should be
> possible to run the BBC micro on a relatively modest FPGA/dev board by
> moving the system roms into flash or external SRAM, for example.
>
>> Have you done any investigation on creating a 'Tube' system in an FPGA? -
>> presumably if you can emulate one 6502 then doing two should be easy if
>> the FPGA resources are there.
>
> I'm not overly familiar with the BBC but the Tube is something I've been
> interested in - it would be almost trivial in an FPGA! But not having any
> software or experience in using it, I wouldn't be able to do much in the
> way
> of testing etc so I haven't attempted to implement anything yet...
>
If you have a desire to gain more familiarity with the BBC platform in
general, it might be worth spending some time with the 'BeebEm' PC emulator.
It emulates most of the BBC hardware, including Econet, Hard Drives and Tube
second processors, and is open source.
The Windows version is available from:
http://www.mikebuk.dsl.pipex.com/beebem/
Specifically for the Tube, Robert Sprowson has developed a new second
processor for the BBC in the 21st century
(http://www.sprow.co.uk/bbc/armcopro.htm), and uses an Altera device to
perform the Tube processing. As such, it may be possible that the logic
required to glue two CPUs together could be made available to add to the
PACE BBC project.
One hope that I have is that an FPGA implemented BBC would allow the
possibility of a similar flexibility of system configuration to that
available in BeebEm, but, because it is a deterministic hardware solution,
it should be possible for it to match the performance of a real BBC system a
lot better (e.g. the sound on BeebEm isn't perfect, and if you run more than
2 instances on even a very modern PC they get rather sluggish)
Michael