The basic video system on both the NASCOM computers was 15 lines of 48 characters each. This was a fully memory-mapped display. In this type of display each character position on the screen corresponds to a particular address in the processor’s memory. Thus, by writing to a memory address, the processor can change the character displayed on the screen at that position. As one memory address can hold a value of up to 256 (FF hex) this is the number of possible characters. The advantage of this is that the display is cheap to implement and simple to use. The disadvantages are many! True line graphics are impossible to produce as all the available characters are held in read-only memory (ROM).
Colour was not possible on a basic NASCOM system. Not only was insufficient memory available (the NASCOM screen memory is a total of 1k – not all of it used!), but the necessary logic to produce the colour system was not fitted.
The NASCOM screen memory was mapped to commence at the second screen text line, continue down for the next 15 then back to text line 16 at the top of the screen. This allowed the monitor to scroll text lines 2 to 15 and leave column headings fixed at the top of the screen. It did, however, make life a little difficult if you wanted to scroll all 16 lines! In addition to this, the hardware generated 64 characters to a line but only displayed 48 of these, leaving 9 characters at the left and 7 at the right hand end of each line not displayed. These invisible characters were affected by clear and scroll commands so the space was effectively wasted as the contents of those addresses were subject to change without notice!As mentioned above, the character set was “hard-wired” in ROM and could not be changed. This was extremely limiting as no special graphics characters were included, indeed the normally unused characters from 00h to 1Fh were representations of the corresponding ASCII commands. The standard character set is shown below.
The memory-mapped screen was produced by switching the display memory between the display system and the processor. This switching action, although very fast, if carried out at the wrong instant caused flashes on the screen. These were quite bad if there was a lot of screen action going on. Some of the games (yes, we had some!) were almost unplayable. A simple circuit to reduce this was published in ETI and later appeared in the first INMC News magazine. This was the “Snow Plough” modification. This worked for many users but later a much more effective (but more complex) modification called the “Snow Dinger” appeared.NASCOM-2 owners had the option of installing the NAS-GRA EPROM into a second socket on the main board. This contained (at last!) a few games orientated characters together with various box characters and a set of block graphics similar (but not identical) to those available on the Tandy TRS-80. Note that the little “robot” characters actually have feet on the EPROM but the NASCOM-2 didn’t display the bottom two rows of the character cells, so they didn’t appear, the NASCOM-1 though displayed all 16 lines so the feet first appeared when the ROM was made available to them! A lot of design work was put into the NASCOM-2 to remove the white screen flashes which had plagued the NASCOM-1. Unfortunately, this resulted in black flashes! These were a bit less objectionable, but modifications appeared from various sources again.
This made life much easier, and, of course, many people tried to use the new
block graphics to draw lines and shapes on the screen. This was not easy because of the strange screen memory system. However, NAS-GRAPHPAC appeared (from CCsoft) which worked with 8k BASIC to give a bit more control.
Of course the poor NASCOM-1 owners had to get this wonderful facility on their
machines. One technique was to add a second socket “piggy-backed” onto the character generator which could accept the new EPROM. This was switched in by a little modification to the NASCOM-1 board to produce the necessary signal. This system worked well, but at the time 2k EPROMs were rather expensive, the official NAS-GRA EPROM was only available at a high price (10 UK Pounds at that time) and half of the expensive EPROM required was filled with block graphic characters.
Nascom made the “Econographics” board available (if you could find
anyone that had any). This worked in a similar manner to the above modification but used its own logic rather than modify the PCB.
Another alternative was the Line Graphics board from COMP Computer Components.
This consisted of a small pcb which fitted into the character generator ROM socket. Rather than use half of the space on an EPROM to produce the block graphics, this used hardware to produce double height characters. This effectively gave the possibility of continuous vertical lines and almost continuous horizontal lines, but not always where they were needed! Also, of course, there were no games characters available using this technique – but it was a lot cheaper than an EPROM! A major problem was that this board wasn’t even slightly compatible with anything else.
Yet another system was designed by a member of the MNUG, J.M.Stevenson. This was a slightly different idea, using a 1k EPROM to provide a (different, incompatible but often more useful) set of graphic characters and hardware to produce the block graphics. This time the block graphics were compatible with those of the NASCOM-2. As the graphic characters were stored in EPROM it was quite possible to recreate the NAS-GRA characters as an alternative set. Although I still have the circuit diagrams for this I no longer have a pcb.
The next idea was to produce a system where the user could change the graphics
characters to any pattern that could be fitted into a character cell. I’m not sure what systems were made commercially available (it was definitely included on the MNUG’s FPCS board) but I have a circuit diagram (obtained from the MNUG again). Either 2k (if used with the standard character generator) or 4k (if all characters were to be user-definable) of memory was mapped into the processor memory space. The required character set was then simply copied into this area of memory. I seem to remember that one member of the club used the 2k of monitor space. Writing to the monitor (normally impossible as the monitor was in EPROM) would change the graphics characters. Reading the same addresses would read the monitor of course.
IO SYSTEMS GRAPHICS BOARD
This was an interesting add-on for NASCOM-2 systems. It made
dot-addressable graphics available for the first time. Installation was not for the faint-hearted as there were quite a lot of connections to be made to both your NASCOM and the memory board. The card did not have its own memory, it used a 16k bank of the existing memory. The resolution was 384 dots x 224 dots – and gave a very good display. The “snow” problem was cured by stopping the Z80 processor during screen refresh. Unfortunately this also slowed down the system. A similar technique was later used on the Sinclair Spectrum.
DEDICATED VIDEO SYSTEMS
No further video development happened until the “big split” when
Nascom and Gemini went, more or less, their own ways. Gemini, because their system was designed as such from the beginning, were the first to have a seperate video board – the GM812 – also known as the IVC or Intelligent Video Controller.
GEMINI IVC BOARD GM812
Gemini had decided to produce what was basically a new computer but using a very similar bus systen to NAS-BUS. There were several fundamental changes to the NASCOM computers, the most obvious being that the minimum system used two boards rather than one. The IVC board had its own Z80 processor (for control) and MC6845 video processor complete with all the necessary RAM and logic to drive the display. It also included the keyboard input and thus made additional keyboard facilities available.
The next step was colour graphics. This was difficult and expensive at the time as colour monitors were not easy to obtain. TVs could be used, of course, but were not suitable for the required resolution (this still applies – you can’t read 80 characters accross a colour TV tube). Several boards were advertised, but all of them were extremely difficult to obtain. Nascom and Gemini both tended to advertise well in advance of their products becoming available through dealers. The other companies may have worked in a similar manner!
WINCHESTER TECHNOLOGY CARD
One of the first colour cards to appear, this used the Mullard Teletext ICs to give the colour and graphics display. Of course, this produced the familiar Ceefax/Prestel display. It certainly wasn’t high resolution but it had the advantage of requiring little screen memory and was easily programmed. Unfortunately this card disappeared just as the idea of using a computer for Prestel was catching on. The board was produced in the familiar 8in square format for the NAS-BUS/80-BUS.
IO SYSTEMS PLUTO GRAPHICS CARD
This was a truly stunning card. It was available in several versions ranging
from the Baby Pluto to the Pluto Palette. It was also the most expensive graphics card available, the standard version costing some 450 UK Pounds in 1983!
Standard 8in square format card
Dedicated 8088 16-bit processor
196k RAM fitted, 192k used for display
Basic resolution of 640×288 in two pages
“Instant” page switching
8 colour display
The Baby Pluto was simplified, having 96k RAM and single page display. The card was port-driven rather than memory mapped. An extended command ROM was available as an optional extra. This provided commands for drawing circles and complex filling amongst others. This board had its own on-board expansion bus for further additions. I think that the Palette and Baby Palette versions utilised this, but I cannot verify this. Gemini took over production and marketing of this board, but only in its standard form, as IO828.
ANIMATED GRAPHICS BOARD
This design *may* have originated in Radio & Electronics World, but I cannot verify this.
NASCOM AVC BOARD
One of the first colour boards to be advertised (and one of the last to become available), the AVC board was a clever piece of design work. I think it was also the last board to be manufactured by Lucas/Nascom. The board was not expensive (by colour board standards!), costing about 150 UK Pounds in 1983. The cost was kept down by using the main processor, rather than a dedicated unit, to control the screen.
The board was large, being 10in by 8in so it stuck out of the front of most rack systems! It used a dedicated MC6845 video display processor. The screen memory consisted of three 16k blocks, one for each of the three primary colours Red, Green and Blue. These were memory mapped to the same addresses, so can be imagined as three overlapped screens. The screen resolution was 392 dots horizontally and 256 dots vertically. If the three R, G and B outputs were fed to a RGB monitor then the three overlapping layers were mixed on the screen and a multi-colour image seen. Selection of the three screen layers was by “paging” the appropriate layer into memory, updating it and “paging” it out again.
It was also possible to place two memory pages side by side to give a very high resolution (784×256) screen overlapping a lower resolution (392×256) screen. This was nice (and gave an 80×25 character display), but text on this screen was effectively bitmapped. Scrolling the screen required that 32k of screen RAM had to be updated for each scroll – making the operation rather slow.
What must be one of the first personal CAD-type drawing packages was written for this board, LOTTI (named after the developer’s daughter I believe), or as it was later known, NAS-CAD.
CLIMAX COMPUTERS MV256
This board (produced in 1983), like the Pluto, was not really intended to replace the existing screen display, merely to add graphics display to the existing system on a separate monitor.
GEMINI SVC BOARD GM832
Back to mono displays again! This was a very much enhanced IVC board. Generally the system was much faster.
Dedicated 6MHz Z80B processor
HD46505 graphics control processor
80 characters 25 line mono main display
256 Programmable character generator
Limited character attribute support
160×75 “pixel” graphics
256×256 bitmapped graphics display
Keyboard input (with user-definable keys)
Light pen input
Combined video output (not via an RF modulator)
I think this may have been the last Gemini board produced before the company disappeared.