MY EARLY DAYS WITH COMPUTERS
By Robin Morom (K4R)
Part 2
Those who have read Part 1 will know that in October, 1979 I had
at last obtained a computer, the UK101, in kit form. "Now," as they
say, "read on."
The single PCB was very well made, with plated-through holes,
considered the state of the art in those days. The keyboard and the
voltage regulator were on the same board, with only the mains
transformer as a separate item.
4k of RAM was included in the price and as the chips (2114) were
only 1k by 4-bit it needed 8 chips just for the minimum. There was
room for a further 4k on board but I made do for the time being. An
extra 8 chips would cost £49.00 and may I remind you of the inflation
since then. It seemed a great deal of money in those days.
I stayed up half the night and got the thing assembled. Inspect
for dry joints - switch on hopefully - no joy - switch off - check
again - find one dry joint and two solder 'bridges' - switch on again
- looks good - retune mono TV to find display - great, screen says '
D/C/W/M '. It seems absolutely nothing today but this was the first
writing I had seen on the screen that was not part of a TV programme.
A wonderful moment! Right, now read the manual. 'Press C for cold
start.' OK, done that. Screen says MEMORY SIZE? Manual says 'Press
RETURN for default' OK, done that. PANIC!! Screen says '252 BYTES
FREE'. Should say 3324 BYTES i.e. 4k minus RAM claimed by operating
system. Switch off again and think - re-examine RAM and find one chip
with a pin bent under - correct this - switch on - this time 3324
BYTES - good. Next step...... And so on and so on.
Eventually I got it working as it was meant to and was delighted
with it. There was a cassette tape supplied which had two games and
the manual had a few simple Basic programs. As the Basic was 8k
Microsoft (in four separate chips in the original version) it had no
proper graphics, colour or sound commands but it did have string
handling such as MID$, LEFT$, RIGHT$ etc. LET was optional, rather
unusual at the time and PRINT could be abbreviated as ?
Of course, like all these things, I wanted to do more. Could I
use the computer to run a model railway, and how about controlling the
organ which I mentioned in part one? Was there any limit?
Talking about the Maplin organ, I said last time that it was a
kit. This was a slip of the keyboard. I should have said that it was a
detailed design with many options and all the parts available
separately from Maplin. Also while on the subject, I might as well
admit that I must have had delusions of grandeur over that project. It
was not until I had built it, or enough for it to be heard, that it
slowly dawned on me that, despite being a fan of the cinema organ, not
only could I not read a note of music but neither could I play ANY
musical instrument! I bought several books of the 'Teach Yourself'
variety and started to learn. And then it was practice, practice,
practice. Eventually I got so good that the neighbours would come
round and break the windows so that they could hear me better. Well,
what I lacked in musical skill I made up for in VOLUME.
To get back to the computer. There is nothing like having an
uncased board in front of one to encourage hardware modifications. The
most obvious thing was the arrangement of the <RESET> keys. There were
TWO of these, wired in series, the idea being to prevent accidental
resets. Stupidly, however the two keys were alongside each other and
dangerously close to the <RETURN> key. It was very easy to press BOTH.
I soon removed one of them to the other end of the keyboard, next to
the <Q>, which meant that it needed two hands to do a reset. I
sometimes think this should have been done to the <BREAK> key on the
BBC. Next mod was automatic reset when switching on. Now it is no
hassle to do a reset manually, but how much nicer to have it done for
you by the computer. A couple of components and it's done. On studying
the circuits, it was obvious that there were several spare gates on
the TTL chips. They must be useful for something. One toggle switch
and a couple of cuts in the PCB tracks and it was possible to have
inverse video (black-on-white instead of white-on-black) as an option.
By about Easter, 1980 there were plenty of UK101s and Superboards
about and everyone was coming up with suggestions for improvements. 16
lines on the screen is not very many and since the screen was memory
mapped into 1k it seemed likely that a further 1k would provide
another 16 lines if the decoding and the software could cope. There
was actually no space on the board so the additional screen memory
chips had to be piggy-backed on to the existing ones. A little extra
decoding and some juggling with the clocks plus another toggle switch
and there it was - 16 or 32 lines as required. It had to be switched
since not all programs were suitable for both. Talking of clocks, the
6502 microprocessor was intended to run at 1Mhz but it was possible to
run many of them at 2Mhz. I was lucky, mine did. Double calculating
speed on demand! Yet another toggle switch and it was an option. It
was very important NOT to try and change this while running a program.
The computer would usually crash and need a cold reset.
Now I come to what I suspect may be the most surprising thing to
many members. I have come across very few people who have even heard
of it. The cassette loading on the UK101 was interesting because
everything was written to the screen as it loaded so that any garbage
was obvious and you could rewind a line or two and try again. The rate
was 300 baud. Having nothing to compare it with I did not realise how
slow this was but you can try it on the BBC or Master using *TAPE3.
The BBC default is of course 1200 baud. There were no options for baud
rate built into the UK101 but the clocks were very accessible so I
tried doubling the rate to 600 baud. It worked but was a little touchy
to load. The trouble with higher baud rates on cassette is that you
are recording onto an analogue device. The recorded pulses get
'rounded' and are likely to be misread on reloading. Then I read about
a new circuit available as an add-on board. This was the Cottis
Blandford interface. This was brilliant! Originally intended for the
Nascom computer, it cost £14 or thereabouts and its purpose was to
clean up the pulses when saving and then reclean them on loading. It
was then possible to use a higher baud rate. When connected it
performed beautifully. Being very cautious I first tried it at the
original rate. Perfect. Now 600 baud. No problem. 1200? Yes. Try 2400?
YES. The instructions said that higher rates 'might be possible'.
Well, no harm in trying 4800. It still worked! Up to now I had been
using a very cheap cassette recorder and the C12 cassettes which were
sold for computing use. I tried 9600 baud and was hardly surprised to
have some difficulty. However, when I used a high quality cassette
deck even this rate worked with only the very occasional problem. I
even used 19200 baud with my broadcast-quality reel-to-reel recorder
but this was more trouble than it was worth. For one thing it took
very much longer to locate the program than to load it! This was at 16
times the rate which Acorn subsequently used for the BBC! So why
didn't they use something similar? Presumably because they wanted to
eventually sell disc systems and be able to emphasise the difference.
19.2k was about the same loading rate as the then 'standard' i.e.
8-inch discs. (5-inch discs were 'mini-floppies') Perhaps I should add
for the experts reading this that I admit having greatly simplified my
explanation. (Yes, there were problems in timing and getting the right
number of nulls, etc.)
Eventually I added lots more to the UK101 such as a memory
expansion which got me up to 24K, a PIA and A/D chip so that I could use a
joystick, etc. There were colour boards available (at a price!), disc
controllers and sound boards in the course of time, but by then I was
ready to move on.
NEXT TIME: a look at a few of the many machines which began to
appear in the early eighties.