supplied built in either 64K or 256K versions. The RAM block is socketed to allow
easy upgrading from 64K through 128K, 192K to 256K. The quality of the pcb is of
the standard expected of NASBUS/80BUS cards, being double sided through hole
plated and coated in a solder resist lacquer. A slight departure from the norm
here, as the solder resist is blue in instead of the usual green. Dare I say it,
adding a touch of colour to the system. The pcb edge connector is gold flashed as
is to be expected. Overall quality of construction is good (the early sample I
had was hand soldered, I do not know if production boards will be flow soldered).
My only complaint from the hardware point of view is the manual supplied.
This excels in the trend started by the early Nascom manuals and continued by the
current DRI CP/M manuals in its total incomprehensibilty. (I hasten to add that
there is not a lot wrong with the current Nascom manuals.) It is the moat
difficult document to understand. I wasted two evenings trying to access more
than 64K, not because it didn’t work, but because I couldn’t understand the
words. No examples of use were given, and the description of the mapping system
was inadequate and referred to IC numbers, which in the absence of a circuit
diagram, was less than helpful.
The MAP RAM is well built and well engineered, and arguably worth the money
if you can find something to do with it (read on), let down badly by totally
inadequate documentation. Now software has been written for it, it is of much
... continued by Richard Real:
The MAP 256K card, which I shall refer to as MAP RAM, is the first memory
card for the 80-BUS which offers more than 64K. The history of the RAM cards is
First came the Nascom RAM A, which had 32K RAM and 4K ROM, which seemed
marvellous at the time. This card was plagued with problems, some not perhaps its
own fault, and it took a long time before a definitive set of modifications to
make it work perfectly was produced (see INMC News Issue 7). To be fair, the
original specification never included operation at 4MHz, which everyone wanted.
This was followed by the Nascom 48K RAM card, which worked perfectly, and
had the apparently useless new feature of a page select system allowing any of
four 64K pares to be selected. But this gave the first hints of what was to
Gemini then produced a 64K RAM card, which was useful to disk system users,
who needed the full RAM memory. It too supported the four page system. Then came
the Gemini GM813 CPU and 64K RAM card which puzzled everyone by having memory
mapping. This uses a high speed RAM between the CPU address lines and the memory
chips, which alters the addresses selected so that many different physical blocks
of 4K are mapped onto the 16 4K logical memory areas. This translation process is
set up by output of values to port FE.
When MAP decided to produce a 256K card they therefore had a challenge,
which was to produce a card which would operate with all the existing systems.
They have put a great deal of thought into their product, and have achieved this
aim excellently. MAP RAM works with the Nascom 1, Nascom 2 and Gemini GM811
computers using a new 32K paging system, which is very easy to control. It also
works with the Gemini GM813 computer, exactly as Gemini intended, by adding more
physical 4K pages which are addressed by the memory mapping system, as a logical
extension of the GM813 design.
The next important question is, “Does it work?”. The answer is a definite
YES. There is no difficulty in getting the card going. If the card is plugged in
to a Nascom or GM811 system, it immediately acts as a normal 64K RAM card. The
appropriate software can then be used to activate the other memory pages. If the
card is to be used with the GM813, then the two header plugs on the MAP board
have to be changed, and one link altered. It would be better if this link had
been made easier to change. MAP offer to help if you are not confident of making
any of these changes yourself. You can have up to four MAP RAM cards in a system,