[Coco] [coco] the old grey ghost lives!!

[RJRTTY at aol.com]

In a message dated 5/10/2008 8:16:04 A.M.  Eastern Daylight Time, 
brucewcalkins at charter.net writes:

>> a  custom made 1 meg  expanded
>> memory card I made in  1984,

>I would like a how to on that project, including how the  software accesses 
>it.  

Well I don't think you would want to duplicate it the same
way I did it in this day and age.   I used 32 memory
chips of the 256K X 1 bit variety.   I can't remember the
exact ID number but they cost a fortune at the time.
You could do it much better with simms nowadays.
 
I used digitial comparators to open a transparent 256
byte window mapped into the main 64k space of the
processor.   This single page window could be
addressed anywhere in the main 64k space and
point anywhere in the 1 meg of card memory.  You
didn't have to physically move the data from the
1 meg card memory unless you wanted to.   The reason
this window had a 256 byte size is that DECB used
256 byte buffers and I thought that size would be a good
fit when the card was used as a ram disk.
 
Which brings me to the software.   All I did with it was  make
a ram disk.    I hacked DECB slightly so  that when a special
drive number was used to call for a disk operation it diverted
from the usual disk access code and filled the drive buffer
in use with data from the ram disk.   It worked very well
and was FAST.  It was capable of doing much more tho
if you had the time to put into it which I didn't back then.
 
I topped it off with status LEDs that signaled when data
was received or transmitted much like the ones on the
disk drives themselves.   They acted much like the LEDs on
an external modem.
 
If you want I will send you a schematic and I will dig the
software out and send it to you.   The darn thing still 
works too.    I even painted it gray with a black face
to resemble the drives.   It consisted of three piggybacked
PC boards with a central buss between them and connected
to the Multi pak interface with a ribbon cable.   The multi  pak
was necessary to act as a protective buffer between the unit and the 
main board.
 
The trickiest part of the whole thing was figuring out the refresh
timing for the DRAMs and suppressing the noise generated
by the device. It took lots of scope time and perseverance but
I was able to get it to work reliably.
 
You could probably do it better with a modern simm unit today
and a touch of programmable logic.
I wonder if I developed a modern kit out of it if people would want
to buy one?
 
Roy
 



**************Wondering what's for Dinner Tonight? Get new twists on family 
favorites at AOL Food.      
(http://food.aol.com/dinner-tonight?NCID=aolfod00030000000001)