Previous column ComputerAnswers    Next column ComputerAnswers

Logo

ComputerAnswers Column 2


COMPUTER ANSWERS FEBRUARY 1985



Copyright 1984-1989 Simon N Goodwin

PREHISTORIC PROCESSING?

We have just unearthed a very old computer at my school. It is a South West Technical Products CT-82 terminal, Motorola M6800 processor unit, twin disk drives and a teletype. All of these are in separate units.

About four years ago something went wrong in one of them and, after several unsuccessful atempts to repair it, the teacher in charge locked it away in a cupboard - until now, when we have been given permission to try to fix it.

The only thing we can get to work is the terminal, and then we can only manage to type in our names. All of the manuals for the system have been thrown out, and no one remembers how to use it.

We desperately need help from anyone who knows how the system works.
Tony Reeves, 5 Main St, Howsham, Lincs, LN7 6LE.

Five years may be a long time in politics - but ten years in computing is evidently a step into prehistory! The computer you describe was quite popular in 1977-79, especially among schools. The system would have cost about £3-4,000 in 1978.

We can't really diagnose your problem without looking at the hardware, but here are a few general hints: The main circuit boards in SWTP computers were generally very reliable, but many people had problems with poor wiring between them. The most common faults were in the power-supply and reset-button circuitry, often as a result of poor soldering, so you'd be well-advised to take a close look at the wiring in those areas.

You should be able to test the disk-drives and teletype by connecting them to other computers, since they use standard interfaces. It sounds as if the fault you mention is in the processor unit.

If any PCW readers can dig out manuals or circuit-diagrams for the STWP machine, please contact Tony directly.





SCREEN DAMAGE

My big brother says that using my Spectrum on a normal colour television set will damage the coating on the tube. Is this a myth or can it happen? He says graphics and games are the worst offenders.
D.E Avison, Mosely, Birmingham

It is very unlikely that your micro will damage the coating on the TV tube. Very early video games used crude block graphics (such as the border on a tele-tennis game) and some of these could become 'burnt in' to the screen if the game was used for many hours with the TV brightness turned up to maximum.

The problem can be seen on some old black and white video-games - the area where the screen border appears can become scarred, so that you can see its outline even when the display is turned off. Of course, these machines are used to display one bright picture, day after day, for months or years, so it is hardly suprising they become worn out.

In theory you could damage your TV if you played a game with bright, static graphics for several weeks without respite, but in practice I've never known it happen. Modern computer games use finely-detailed graphics in graduated colours - it is most unlikely that these would leave a permanent impression on a TV tube, even if you played the same game for several months.

With typical attention to detail, Atari build a feature called 'attract mode' into their computers - this automatically turns down the brightness of a display, and cycles through the colours, if the computer is left unattended for more than about seven minutes. This safety feature probably stems from Atari's experience producing arcade machines.

If you want to 'play safe', you should not run programs with the TV brightness turned up to maximum, and avoid leaving the computer for long periods showing a static display. Normal micro use will have no damaging effect on a TV tube.





BBC MICRO UPDATE

Are the new Acorn Micros (ABC's) entirely compatible with the BBC Micro?
Darren Johns, Weybridge, Surrey.

The latest computers from Acorn - the ABC range - are business computers designed to compete with the IBM Personal Computer and other 'heavyweight' micros. They feature a variety of processors, including the Z80, 80286 and 32016, but there is no version with the 6502 processor used in the BBC Micro and Electron. For this reason (and a host of others - the ABCs are unique designs) the new machines are not compatible with BBC Micro software.

Acorn do produce the Torch, which can run CP/M or Beeb software. This is a hybrid business system which contains a BBC Micro circuit-board and a separate Z80 processor. A lot of BBC micro software will run on the machine, but you waste the power of the Z80 if you only use the Beeb part of a Torch system.





TAKING THE QUANTUM PLUNGE

I am thinking of buying a Sinclair QL. Will all my Spectrum software be completely compatible, or is that just advertising hype?
Angela Holiday, Cambridge

I don't think Sinclair claim that you can run Spectrum software on a QL, although one software house has pretended that it can write a program to translate machine-code from one machine to the other. In fact they made that announce- ment before they'd recruited programmers to do the job; it is impossible to write such a program.

The instruction-set, hardware addresses, display format, system variables, ROM calls and so on are completely different on the QL. You can't write an 'automatic code translator' that will work in any but the most trivial of cases.

You can't load Spectrum tapes onto a QL since the new machine has no cassette interface. The QL microdrives use an improved format - this makes them more reliable than the Spectrum version, but unfortunately also means that they can't read cartridges written by a Spectrum.

You CAN get a Spectrum and a QL to communicate via Sinclair's 'Interface One' - messages can be sent either way, via the network or RS232 link. This works fine for text (so long as you remember that the QL uses CHR$ 10 to mark the end of a line, and the Spectrum uses CHR$ 13) but it isn't very useful for transfering programs, since the keywords used on the Spectrum are quite different to those on the QL. It is possible to translate programs by hand, and it should be possible to do some automatic Basic translation (I'm working on this myself at the moment), but the technique will only ever be useful for simple Basic programs. There's no reliable way of translating PEEKs and POKEs from one machine to another, since they work in quite different ways.

To be fair to Sinclair, QL SuperBasic is a lot better than the ZX Spectrum version. It wouldn't have been possible to make the two machines entirely compatible without imposing a lot of limitations on the QL design (attribute graphics, 32 character lines, and so on), and the few QL programs I have seen so far are of a far higher quality than direct Spectrum translations could be.

Sometimes software compatability is a millstone rather than a benefit. The Commodore 64 uses a Basic interpreter designed in 1978 for the PET. In a bid to remain compatible with programs for the old machine, it contains no commands to take advantage of the graphics and sound features of the 64. The newer computer would have been a much more useful machine if Commodore had sacrificed compatibility and produced a new Basic - it's not often you see a Commodore 64 owner running programs written for the old black-and- white PET, after all!



DEATH OF A ZX 81

Seeking to raise a few shekels to buy a printer for my Commodore, I decided to sell my ZX81. To improve its performance I tightened the 9 volt input, by fitting a slightly larger jack-plug. Carefully noting the positive and negative lead positions I then obeyed Murphy's law by soldering negative to positive and vice versa.

Not suprisingly, I ended up with a blank screen. Is there anything I can other than present it to the dustman? Have I also ruined the 16K RAM pack?
R Clayton, Rownhams, Southampton.

Whoops! You've probably only blown the power-supply regulator on the ZX 81 circuit-board, although you may have killed the video modulator as well. Both of these components could be replaced for a few pounds, but you may find that the effort is not worthwhile - a second-hand ZX 81 (with RAM pack) is probably only worth about #10-20.

Your first step is to re-wire the plug the correct way. I'm suprised you tried to make the system more reliable by changing the plug, since the socket is the weak link in the chain. I'll try and make the rest of the advice relevant to anyone who has accidentally damaged a computer, or acquired a machine that doesn't work.

You should never test a computer with peripherals (such as the RAM pack) connected, unless you're sure that the computer works when used on its own. In this way you minimise the chance of damaging your entire system.

The supply from a ZX 81 plug goes directly to the video modulator, which converts a digital video signal into one that can be received by a TV. Re-connect the supply the correct way round (if you've blown the machine up it is unlikely you will damage it any more by connecting the supply properly).

Now try to tune in your TV, with the sound turned up. You won't get a normal display (unless the computer is undamaged) but you might find that the TV sound is more regular (a 'burrr' sound rather than a 'hiss'), around channel 36. The display may also be less speckled. If this is the case, the video modulator is still working, even if the computer isn't giving it a signal to chew on.

The next step is to find out if anything is happening on the processor-board. Take a portable AM radio and put it on top of the computer - as close as you can get to the circuit-board. Tune the radio to a point in between two stations. You should hear a horrible electronic crackling noise - radio interference generated as circuits in the computer turn on and off, millions of times a second. This is an old diagnostic trick from the 1950s.

Compare the noises with the computer turned on and off. If there is no difference, the power-supply to the board is faulty - replace the regulator and see if that helps. Alternatively you can run the board from a 6 volt lantern battery, via a silicon diode. Unplug the mains, and connect the battery supply directly to the processor - positive to pin 20 and negative to pin 40. It may be that other components have blown, so this may not fix the problem, but it will be a step in the right direction.

If you do hear a horrid noise when the computer supply is connected, part of the machine is probably working. The radio signal comes from two main areas - the processor (and memory) and the video-generator (the ULA, in a ZX 81). Move the radio between the two and see if you get a different tone in each position. If the noise just gets louder or softer, one or other part has been destroyed and it is not really worth fixing the machine.

If the video-generator and processor both seem to be working, but the modulator didn't affect the TV display, you may be able to bring the machine back to life by fitting a new modulator.

Don't connect the 16K RAM pack until the main computer is fixed. You run a small risk that the RAM pack has failed catastrophically - check that it doesn't short out the supply lines on the edge connector before you re-connect it. If you have any foolhardy friends, try out the RAM pack on another computer which you know to be working. It should still work, but test it very briefly at first, just in case.

At the end of all this experimentation you may not have a working ZX 81, but hopefully the experience will come in useful when you decide to re-wire the plugs on your Cray supercomputer!



DECODING ORIC DISKS

Could you please help me with any information on the Oric Atmos disk drive. They use the terms 'expression' and 'data' when related to file-handling. Does one need to declare DATA in a program? What is an expression when saving to a file?
G C Kendell, Rugby, Warwickshire.

I suspect you have been confused by some rather vague terminology in the Atmos disk manual, or a review of the system, although I'm sorry to say that I've not used the Oric drive myself.

The terms 'expression' and 'data' are often used as shorthand in specifications, to avoid writing out all the possible terms which could be used in a command. This use of the word 'data' is distinct from the Basic keyword DATA, used to mark lines which contain preset string or numeric values.

The term 'expression' is used whenever you could type a value, or a variable name, or a calculation. For example, these are all expressions: A 123+Z A*123+B(I) "FILENAME" A$+"BAS" The first three are numeric expressions - things which have a numeric value when examined by the computer - and the last two are string expressions, since they have a string value.

If a computer manual says you may type: LOAD 'string expression' it means you can type any sequence which produces a string result, after the LOAD keyword. Likewise, you might read that: PRINT #'numeric expression','data' allows you to send data to different places depending upon the value of the 'numeric expression' between the PRINT # keyword and the 'data'. Normally that expression would just be a number, but the technical term 'numeric expression' has been used to show that you could type a variable name, or a calculation at that point instead. If a book just says 'expression', it usually means 'numeric expression'.

The term 'data' can stand for almost anything. In the PRINT example, you could replace 'data' by a string, or a string expression, or a number, or a calculation, or a variable-name. Here, 'data' just means 'something that has a value'. Often 'data' may consist of more than one item, so long as you put punctuation (usually a comma) between items.





JUGGLING WITH DAISYWHEELS

I have bought a 'Quen-data' daisywheel printer because it takes Qume print wheels. However one of the wheels prints garbage - it appears to be completely in the wrong order. I expected the accents and special characters to need decoding, but all the letters and numbers are wrong as well.

My Courier 10 and Gothic 15 wheels are correct, and the Prestige Elite 12 is standard ASCII, but the Boldface PS which is marked as Bilingual/WPS is completely wrong. The problem is that I want a 12 character per inch wheel with an English pound sign. I would need a special driver for each wheel unless it conforms to a standard order.

How can I tell which will work and why they are non-standard? The suppliers of the wheels and the machine do not seem to know.
Derek Trayler, Hornchurch, Essex.

The answer to your question is coded into the name of the daisywheel. The wheel which produces characters in jumbled order is a 'bilingual proportional spacing' (PS) model, which makes it something of an oddball. You are correct in saying that your printer takes Qume daisywheels, in that they fit on the machine, but some of them are designed to take advantage of features that are not available on low cost printers.

The letters on a daisywheel are always held in a jumbled order, for two reasons. The first is to 'balance' the wheel - it is important that larger letters are opposite one another so that the wheel turns smoothly. The second reason is to minimise the distance between letters that are foten used together. The printer works by stepping from one character-position to the next, spinning the wheel as it does so, so that the next character required is at the top of the wheel when the paper is correctly positioned.

Obviously some sequences of letters are more common than others - it is possible to speed up a printer by making sure that common sequences are close together on the daisywheel.

Your rogue wheel is a 'bilingual' model, with various special accents for French printout. It is also designed for a 'proportional spacing' printer - one in which letters are packed together, so that, for example, a letter 'M' takes up more space than a 'j'. Such a scheme is used on this typeset page, but not on typewriters and cheap printers, which always allocate the same width to each letter, making 'M's look rather squashed and exclamation- marks rather spaced-out.

Your WPS daisywheel uses a different sequence to the others because it is designed to work with foreign languages, and turn smoothly even though some letters are smaller than others. Some printers (such as the up-market Qumes) recognise a special sequence of characters which tell them that they are using a bilingual wheel. You print this 'escape sequence' before you use the new wheel, so that the printer can make allowances for the changed order.

A dealer tells me that your printer cannot cope with proportional spacing anyway - the Quen-Data always steps a fixed distance from one character to the next. For this reason the PS characters, which vary in width, probably wouldn't look very good even if they came out in the right order.

Luckily, there is a wheel which will give the characters you need, on the printer you've got. It is called a 'PICA 12 ENGLAND' daisywheel - you can get it from Worldwide Computers, at 11 Worple Road, Wimbledon, London SW19 4JS. They also supply the Quen-Data printer, so they should be able to tell you which other wheels will work in your system.





MAGIC WAND WAIVED...

I am interested in obtaining more information on Magic Wand, by Small Business Applications Inc, Houston, Texas. Is the package compatible with CP/M 86 or MSDOS, for use with the Sirius? Is it possible to embed ASCII or ESC codes in the programme for an Epson LQ1500 dot matrix printer? Wh are the British suppliers of Magic Wand?
Janet Heslop, Hull.

We tested Magic Wand way back in May 1981. At that time it was available from the London Computer Centre (01-387 4455). The program does not appear in the 'ACT directory' of programs which run on the Sirius, so it is unlikely that a compatible version is available.

You can embed control codes for any printer within a document by using the OUT command.





DRAGON KEY MYSTERY

I have a strange problem with my Dragon 32 computer. Sometimes it ignores keys when I am typing quickly. For instance, if I type LIST quickly, the screen shows LST and I get a syntax error when I press Enter. The problem occurs with some words and not others. Is there something mechanically wrong with the computer? Why does it work properly when I type slowly?
Graeme Newman, Crown East, Worcester.

Your problem stems from some sloppy system programming - there's nothing wrong with the keyboard of your computer.

The Basic interpreter in the Dragon, and in the Tandy Colour Computer, contains a bug which affects the way multiple key-presses are detected. The computer is meant to have 'roll-over', which means that you can press the next key of any sequence before you release the first. Of course, this is only useful for fast typists - when you type slowly you press the keys one by one.

Hold down the 'L' key on your keyboard, then press 'E'. The second letter appears, even though 'L' is still depressed. Now hold down key 'L' and press 'I'. This time the second key is ignored.

The Dragon keyboard is wired up in groups of eight keys. The computer reads each group of keys and then works out which key is pressed by examining the resultant pattern. It keeps a record of the last key pressed, so that it can work out which is the 'new key' as your fingers 'roll' from one to the next. The snag is that this is not properly done on the Dragon - it ignores the second key if two in the same group are depressed. 'I' and 'S' are in different groups, but 'L' and 'I' are in the same group.

There's no easy solution to the problem - you've got a choice of typing more slowly or writing your own keyboard routine. The bug is not confined to program entry: INPUT and INKEY$ use the same code, and therefore also suffer from the bug. It has, however, been fixed on the Dragon 64.






Link to the top of this document    Link to the main index
Previous column ComputerAnswers    Next column ComputerAnswers