Re: IEEE-488 bus

Hi Jim,


On Jan 15, 7:30 pm, Jim Brain <br...@xxxxxxxxxx> wrote:
christianlott1 wrote:
I'm just trying to find out what can/can not be done. I'd like to be
able to expand my C64 with midi and ethernet and keep my rs232 and add
more ram (heck, i have 512mb ram sitting here for a year since my pc
is max'd at 2 slots/gb), another SID or three, Z80 or 6809 (super
pet!), and then......

Anything *can* be done. :-)


Thank you for the considerate response.

I think the main difficulty with this project (from my small
knowledge) would come from setting up the bus and memory management.
The bus would take into account not only what we just discussed (dual
port ram?/multiple bus channel layers?) but also handle something like
monitoring from an external source (for tracing and interrupt from a
PC).

The memory management (PLA/banking system) would need to be able to
map chips to ANY address, not just the standard IO space. This allows
for alternate architectures being implemented within the same system.


Of course. Words from someone who CAN do are more weighty than from
some innocent speculator like me :)

My advice for those who want these hardware projects done is this:

o Develop the idea into more than just a thought. Think through the
practical aspects of it. Projects with broader appeal and lower price
points get more traction.


o If your idea can be developed in phases, consider the minimum amount
that would make the project useful. Essentially, if the idea can be
realized in it's barest form with a minimum of effort, the windfall from
the initial introduction can sometimes carry the later phase efforts.


o Chat with those of us that play around with this stuff. But, plan
this. For example, I would not be a good choice for a SuperSID FPGA
implementation. That's way more analog than I care to research.
Besides, I don't have a passion for the SID. ONce you've identified the
appropriate folks, pitch the idea. The more complete the idea, the
better the reception.


Actually, I like that idea (SuperSID FPGA).
:)

Since you won't be working on THAT, designing a robust, scalable and
reconfigurable bus system would enable such devices to be readily
plugged in.

A few of the benefits of 'ONLY' worrying about the bus and memory
management, keeping it expandable:

A) Real chips used or FPGA, it all interfaces. Each section can be
built, rebuilt and plugged in.

B) Minimum system is the C64 architecture (sans chips) and glue logic.

C) Hobbyist hardware (eg backplane). Most people can't program VHDL,
but setting a few switches to reconfigure the chip and memory
architecture is conceivable. Custom architecture styles can be set up
as a profile by those who know how and loaded by anyone interested in
a ie. Z80 @ 30mhz with dual VIC2s.

Probably not too many people..

Can the bus handle multiple chip/cpu/ram speeds? How many lines are
feasible on the bus?



o Volunteer to handle the non-technical stuff. Keeping the
forums/USENET updated with new info, writing manuals, maybe fronting
some cash for parts, consider these. Forwarding or answering quick
questions helps as well. Rallying support for the idea to spur on the
developer is invaluable.


If the system can handle a C65 clone you have a win.

You are right. If you and I are working on a project we need to know
each others interest level at all times. Waning interest keeps either
you from working on hardware or me from working on publicity and
documentation.

The project needs to be realistic, yet not so simple as to be
uninteresting. Most of all, it needs to be GREAT and have the
appearance of being _just_ possible.



o Patience. While I won't say my projects are the norm, it could easily
take a few months to see anything. Early attempts are many times dead ends.


I've been thinking about this project on and off for an embarrassingly
long time. The reason why it's so embarrassing is because you'd have
thought by now I'd know how to implement it :)

I'm all for wasting YEARS in planning ;)



Essentially what I'm leading up to is a physical implementation of
something like MESS but being able to mix and match chips, then being
able to plug in the PC's bus.

Well, I can lay out some "boardlets" for such an item, but I think you
need to rally some interest and plan out what would be on each board.


Jim, I don't want to make promises to people about capabilities that
are not possible. Could you please explain how the architecture you
have in mind could meet specifics of my want list?

Specifically:

1. You mentioned dual port ram as allowing the CPU/VIC2 operation with
no wait states. I assume this will be the 64k on board (not some swath
of ram map'd from a remote 512mb ram module). I also assume this can
be reconfigured to perform WITH the original wait states.

2. Speaking of wait states - Is it possible to switch from NTSC to PAL
on the same machine? This leads into supplying digital video output.
If we can get a good digital output from the VIC2 (a REAL VIC2), are
effects like VICE's scale2x possible?

3. Is it possible to slow down and speed up the machine, set
breakpoints at addresses from a remote machine hooked up to the bus?
(I may be mixing two thoughts, one from timing multiplier, one from
cpu program counter monitoring)

4. 1541 module will need to play G64s :)



btw, I have one of the rs232 carts you designed. I use a custom loader
to download one filers from my pc. Thank you!

No problem. That was a quick design, and those are easy to do. While I
wish Doppleganger only the best in selling them, I am hoping that
someday I'll have some time to make up a batch and recoup some of my
costs in designing these things.


The board could be improved with a programmable boot rom and higher
max speed.


In conclusion - this project attempts to offer REAL hardware running
in REAL time, unlike emulators. Unlike FPGA, this project does not
require knowledge (nor emulation) of the main chip internals, simply
knowledge of the bus and glue.

Are there enough spare chips out there for people to have a desire to
both buy a vanilla board and populate it with those vintage chips? If
not, are there work-arounds/substitutes for the more difficult to
obtain chips?

Since the project revolves around the expansion and interface
capabilities, what is the other machine most conveniently
approachable? My bias is the Atari VCS. Once you say go I will pitch
it to that crowd for a board design (unless you have other plans..).


Christian
.

Relevant Pages

  • Re: Cube G4 ram into eMac 700
    ... If you have PC100 Chips ... function at PC133 speeds, as well as ANY Slower Speed. ... RAM does not care about "Megahertz" ... Step 1) Push 00FFAA23BAD to the address bus handling such. ...
    (comp.sys.mac.hardware.misc)
  • Re: IEEE-488 bus
    ... map chips to ANY address, not just the standard IO space. ... Custom architecture styles can be set up ... A latter day S-100 bus, ... An all-in-one C64 board may limit how it can be expanded? ...
    (comp.sys.cbm)
  • Re: A7N8X-E deluxe : new ram problem
    ... rock stable for 2 years in Dual channel 200mhz mode and since I've ... The Nforce2 chipset has three memory slots, ... Two slots fit on one data bus. ... is the use of 128Mx4 chips. ...
    (alt.comp.periphs.mainboard.asus)
  • Re: 1977 original apple II restoration (continued)...
    ... Firstly I swapped out all the RAM for some that I recently purchased, ... Pushing reset does nothing. ... *two* bad DRAM chips. ... You should also take a *very* close look at the board traces to ...
    (comp.sys.apple2)
  • Re: Asus A7V8X RAM :[
    ... > no other RAM in the motherboard and it was error free. ... > Also I talked to someone in my comp sci class, he told me that Micron ... A double sided memory has two ... As for what channels the memory chips are taking, ...
    (alt.comp.periphs.mainboard.asus)