Re: Fully Functional IMSAI-8080 boot CP/M 80 for sale on e-bay

On Jun 6, 10:29 am, Alli...@xxxxxxxxxxxxx wrote:
On Fri, 5 Jun 2009 12:43:58 -0700 (PDT), "Mr Emmanuel Roche, France"





<roche...@xxxxxxxxxxx> wrote:
Hello, Andrew!

[snip]

hardware influences the software.  Further doing it using the upd7220
with 1981 designs is a awkward task as back then the 64Kx1 Dram was
barely available and the large arrays of memory  to exploit the 7220
were both physically large and costly.  Revisiting the design with
moden memory and logic parts would be interesting and also yeild a
better result.   Modern monitors can accept wider operating bandwidths
then even 10 years ago, never minding the limitations  25 years ago.

As to the software, yes, it was there and there is a limitited amount
of useful stuff.  I have the full suite for the Visual 1050.

Do you want, too, to re-invent business graphics and CAD?

He doesn't have to.  He needs the graphics card and the BIOS that
talks to it so he can use that software.  





Upon looking at the NEC 7220 datasheet, I think it is rather
reminiscent of the NEC 765A with its DRQ/DACK interface.  However,
there is no "chip select" pin that I can see -- how strange.

If I understand well, your remark: The Z-80 and the 7220 each have
their own memory (bigger for the 7220 if it has a color screen. By the
way, in France, there were only *two* known Epson QX-10 with color
screens... And I have one! I had heard, in the Journal of the CP/M
User Group (UK) how good the Epson QX-10 was, so, when I heard that
someone was selling one, I rushed to Paris, and asked to see it
running. The owner inserted a diagnostic disk, and I was in a state of
schock (remember: I started with a NorthStar Horizon, with an ASCII
terminal and a daisy-wheel printer) when I saw those incredibly sharp
letters in 8 colors, then bold, italics, underlined, blinking, etc,
etc! Incredible! If only I had a video of this demo!) and the Z-80
communicates to/from the 7220 using I/O ports. 2 ports are needed
(only). One day, I patched a screen driver (I think it was for the NCR
DecisionMate V) and ran it on my Epson QX-10 after just patching the 2
bytes corresponding to the I/O ports! This was the only change between
2 screen drivers running on different computers! (But both using the
7220) Incredible but true!

The adoption of color here was far faster as a primary market.
Nothing incredible to us here those machines where fairly common.
Thank PC archetecture for the portability of the code you only changed
two bytes in.  Then again NEC and Epson had a very close business
relationship back then.

 I wonder
if the existing Disk IO "pseudo-DMA" design in the Disk IO board
could be modified for it to work.  It would be an interesting project.

It can help you with the larger amounts of data that the system has to
move into the chips path to do even reasonable screeen updates.

Everything could be an interesting project, to an hardware type.

However, if your goal is to run plain CP/M 2.2, Digital Research did
it before you, and extended it with CP/NET (network) and GSX
(graphics).

No DRI didn't.  DRI never sold a graphics card or a complete system.
They implmented the OS, the graphics layers and a lot of applications
and a sample bios that was machine specific.  So the task that has to
be done yet again is to finsish the task that DRI started.  The is the
code often known as CBIOS or simplified to BIOS.

Unless you know more about CP/M than DRI, I suggest re-using what they
did. And they made a lot! (I have not mentioned, for example, Access
Manager and Display Manager, "production application programs" needed
to develop programs for CP/M.) (The GINSTALL of GSX-86 Version 1.4 was
programmed with Display Manager: both programs (like CBASIC Compiler
and PL/I) run under 8-bit and 16-bit versions of CP/M.)

In this day and age with and with sources, the likely case is we do
know as much or maybe more than DRI.  However DUI Access and Display
manager still does not talk directly to any hardware.   So the user of
that software Either A) buys a working old system to run it or B)
builds the hardware and low level software to utilize it.  Old
examples may be useful assuming complete sources are available
and even then it will assume a specific implementation of the upD7220.

Whew! All those useful programs existing in the CP/M world, and I seem
to be the last one who use it. Sometime, I wonder why some people are
writing in the comp.os.cpm Newsgroup. Most of the American seem
blocked on CP/M 2.2 with ASM and SID, while Digital Research sold
later CP/M Plus with MAC and SID, for example.

Because it gives someone something to feel superior about.  Sometimes
it is the tool of choice as the task may be sufficiently complex to
make it a requirement.   Then again when that was true there were
better tools than that.

Since there is  a diversity of systems out there and a responsible
programmer in the public domain  will use tools that may be less then
optimum but otherwise  widely available and understood.  There were
and are tools far better than MAC and SID that are Z80 aware or
designed.  MAC and SID are good tools I have and use them, but they
are just the state of the art in 1981and even in the Z80 world thing
have gotten better.  I still use Zmac and Zsid and DDTZ.  Those work
well paired with SDCC.

Conclusion: Why re-invent the wheel?

DRI gave us only the tire and tube, the rim, spokes and all the rest
are user or (vendor supplied).  DRI didn't create the wheel they
made it ride better.

Allison


[snip]


Hi Allison! Interesting discussion. I've done some basic internet
searching on the NEC 7220 / Intel 82720 graphics chip and it looks
great but I can find very few documented examples of implementations
using the chip. In order for even the possibility of a new
implementation of a graphics board using the chip, I'd need some
examples to study and determine how to implement as an ECB or S-100
board.

I cannot find any examples of S-100 boards using the NEC 7200 / Intel
82720. No ECB or ISA boards either. There is a STD implementation by
DY4 at Dave Dunfield's site:

http://www.classiccmp.org/dunfield/dy4/dstd777.pdf

However, after some review its clear that implementation is dependent
on three critical PALs for which there are no logic equations that I
can find. That's a "show stopper" as crucial information is lost.
This is the main reason I dislike programmable logic devices. Yes,
they have their applications but they drive the long term sustainment
community nuts when their information is proprietary or lost. Reverse
engineering a PAL is possible but its a PITA.

Would it be possible to scan the NEC 7220 or Intel 82720 application
notes and/or samples implementations for the Z80? Having some kind of
starting point would make it all lot easier to build a board. You
could post them or just mail them to me.

My observation is the first part would be pretty easy if it just uses
regular PIO. Just use chip select gated /RD and /WR signals to access
the chip. The hard part is what to do with the V and H output
signals. Say an implementation using a straight 512K x 8 SRAM (really
256K) and 8 bits per pixel for 256 gray shades. I think an 8 bit
flash D to A circuit (8 resistors) would probably work. Is a shift
register/counter circuit needed to get the sync and timing right?
What about the dot clock? I am guessing the shared address and data
lines means the 7220 <-> SRAM interface has latches?

I've briefly looked over the Intel 82720 datasheet.

http://www.df.lth.se/~pi/compis/datablad/Intel82720.pdf

Probably the answers to all my questions are in there but an example
circuit would sure help make it more understandable.

Thanks and have a nice day!

Andrew Lynch
.

Relevant Pages

  • Re: Forth Evangalists
    ... Apparently Apple has exclusive license to their most advanced chip design. ... > will need a killer app (like realtime raytracing) or a giant ...  I was wanting to implement my own, but the reality is simpler graphics, even if raytraced until the hardware changes. ...
    (comp.lang.forth)
  • Re: Fully Functional IMSAI-8080 boot CP/M 80 for sale on e-bay
    ... using the chip. ... implementation of a graphics board using the chip, ... I need the design information behind it sufficient to build ... or some documentation that includes sample implementations to evaluate ...
    (comp.os.cpm)
  • Re: Fully Functional IMSAI-8080 boot CP/M 80 for sale on e-bay
    ... searching on the NEC 7220 / Intel 82720 graphics chip and it looks ... using the chip. ... implementation of a graphics board using the chip, ... I cannot find any examples of S-100 boards using the NEC 7200 / ...
    (comp.os.cpm)
  • Re: Fully Functional IMSAI-8080 boot CP/M 80 for sale on e-bay
    ... thanks to the "NEC uPD7220 Graphics Display ... 512K SRAM chip would be an incredible device. ... if the existing Disk IO "pseudo-DMA" design in the Disk IO board could ... The best design using it was the DEC Rainbow which was z80 and 8088 ...
    (comp.os.cpm)
  • Re: Controller for Custom LCD Glass
    ... In the meantime, we found another chip of the µPD family made by NEC that could be used - but we are currently trying to get info about long term availability from them, the chip also seems to be a pretty old design. ...
    (comp.arch.embedded)