VDU-K Interak 1 VDU Interface Section 3: Assembly and Testing
As this card is supplied as a "bare board" as well as a kit, many users will be assembling it in a way which suits their own particular needs and wishes. Certain items (such as the UHF modulator, selection switches, front panels, video and UHF connectors) are omitted from the parts kit to follow the policy that only the basic most essential parts are supplied, to give individuals the chance to build their system as cheaply or expensively as they themselves wish. However to avoid permeating these constructional notes too much with mention of these options it will be tacitly assumed that the majority of all the options are to be fitted. Therefore if you find yourself being instructed to drill a hole for say a switch or video connector, which you haven't got, and don't want, then remember to ignore that section of these notes.
emitter e collector c Top View (leads away from you) base b
Note that this diagram does not follow the convention given on transistor data sheets, i.e. giving an under-view. Instead it follows the convention given on IC data sheets which generally show top views.
This concludes the constructional notes. The next section of the Manual covers testing and setting up.
ZYBASIC VDUK Setting-up Program.
10 SCROLL: PRINT "Hello": GOTO 10
Machine-code Setting up Program.
0800 3A00F0 START: LD A,F000 ; Access the VDU 0803 18FB JR START ; And again and again
Once the program (whichever you use, preferably try both) is running adjust RV1 until a snow-free display is achieved, if this occurs over a range then set RV1 in the middle of that range. If an oscilloscope is available confirm that the monostable pulse (measured at U3 pin 8) is approximately 640ns at the optimum setting.
For Amusement Only
The Z80A-CPU timing is most stringent when it is executing an op-code fetch ("M1 Cycle"). Even with the 200ns type of RAM used, the access time of the video RAM is too long to guarantee operation for an op-code fetch in the video RAM, but if you have a fair wind behind you may just be able to use it to run an executable program in the video RAM itself, e.g.
F200 18FE MAYBE: JR MAYBE; Thats it looks easy but it's not!
It takes some ingenuity to get a program into the video RAM (Suggestion: use ZYMON to fill F000 onwards with 00 (NOP), then just copy in the two bytes of the program from somewhere else use ZYMON's COPY command. You can then execute at F000 and just trickle through the NOPs until you hit 18FE). Adjust RV1 as before. If you succeed in running this program, and if you have the facilities, try adding some wait states, or dropping the clock frequency to 2.0 MHz, and see what a mess there would have been on the screen without the anti-snow feature!
10 FOR J = #F000 TO #F1FF 20 POKE J,J 30 NEXT J
This concludes testing and setting up. The next section of the manual covers fault finding, and return for service.
As a fault will be the result of any one or more of hundreds of possible things going wrong (he said cheerfully) it is of course not practicable to give any more than a guide to the location and correction of faults, but the procedures below will at least give some idea of how an attempt at fault-finding can be made.
Experience with other cards has shown that the major cause of faults in cards assembled from kits is due to soldering errors (e.g. "dry" or unsoldered joints, and bridged tracks). The next most likely cause is the setting up of the DIL switch options, then comes misplaced components and even obvious mistakes like missing ICs or ones plugged in upside-down. In the case of "bare boards", other common causes of faults are incorrect component substitutions (the user can't see any reason why he cant use some other component to that specified, but the VDU-K card can! ), and mis-insertion of ICs into their sockets; we shall have a great deal more to say on this latter subject later.
It is very rare for an IC to be supplied faulty (each one has to be tested and characterised by the manufacturer as part of the manufacturing process), and this is borne out by the fact that "ready made" Interak Cards very rarely suffer defective chips; faults found on test are invariably due to some other cause.
However, all the time a board is being tested it is "at risk" (ICs are being pulled in and out, voltmeter and test probes are being prodded about and may short pins of ICs and PCB tracks together). It is frighteningly true that if you test a good board often enough you will eventually cause a fault!
When reading the following procedures remember that they are only suggestions, and there are plenty of other faults which often cause misleading symptoms.
If an oscilloscope is available some waveforms can be measured on the VDU-K card if is suspected this is at fault. Some places to look are given later in this section.
Many professional users will be in a position to find any faults without undue difficulty, and will therefore want to do their own fault-finding should the need arise. Other less experienced users will wish to do the same, but this time on the grounds that there is no better method of gaining a real understanding of computer hardware than tracing faults and fixing them themselves.
However laudable as these two aims may be, the company who supplied the kit will be prepared to help any of their customers who are in difficulties. The charge will probably be very little more than the cost of return postage and any parts which have been damaged.
There is no shame in admitting defeat some tricky faults can baffle even the most experienced expert, and in many cases advanced test equipment will be needed which will just not be available to the average user.
If it is necessary to send a circuit board through the post, make sure the recipient knows to expect it. When making arrangements for the return of a board, be ready to quote the number of the invoice on which it was supplied, or approximate date of purchase. (A supplier will understandably not be so keen to help if the kit was purchased from someone else!).
For transit through the post, pack the board(s) well, wrapped in e.g. aluminium cooking foil for anti-static protection, and remember to include your name and address, and payment e.g. by means of an ACCESS or VISA card. A suitable postal service should be used both ways, e.g. Registered Post, Recorded Delivery, or Compensation Fee Parcel Post, even though this does cost more.
There is a great art to replacing components and working in general on a plated-through hole board, and high quality, expensive special tools are needed, and the skill and experience of people who have spent many long years learning their craft.
Having seen the results of some peoples efforts we urge you to do anything other than try to unsolder an IC socket from a through-hole plated card. Cut tracks, plug the IC in with one lead bent out of the socket, and wire-wrap the missing connection onto the leg of the IC, anything! If you do have to remove a socket then don't try to salvage it. Smash it to pieces, and using tweezers remove the pins one by one, applying the minimum of heat and force to avoid pulling them out complete with the plated through holes. (The possibility of this activity being enjoined explains why it is not recommended that the very expensive turned pin in solid plastic type of socket be used. If the socket cannot be broken into pieces for removal then there is a severe risk of damage to the card when trying to desolder all leads of the IC at once.)
The same goes for discrete components. To remove a resistor, chop it in two; a new one is only a few pence and it will cost several hundred times that figure to buy a replacement for a damaged board.
The great difficulty in suggesting fault-finding procedures, is that the methods vary so much according to what equipment you have and your skill, knowledge and experience.
As the VDU-K is often one of the "essential" cards in a system, i.e. without it the system cannot be used, it is very difficult to be sure that any fault in the system is in fact located in the VDU-K card. Almost anything going wrong with any part of the system can cause something to happen to the display.
In many ways however the VDU-K is one of the easier cards to test. In a manner of speaking it can stand alone. Although it will then be able only to display a random assortment of unchanging characters it takes a good deal of the circuit on the card to perform even this simple task, and so if you see anything at all there isn't likely to be much wrong.
In the event of any trouble do switch off and re-examine the card, checking quickly for any overheating. If an integrated circuit is inserted wrong way up it will often cause a heavy burden on the power supply (which fortunately is well protected internally if it is one of the recommended switch-mode types, and so should not suffer any damage). If the load is so great that it lowers the voltage on any of the supply rails then this could easily stop the computer working. It is almost certain that the heavy currents would cause the integrated circuit to overheat, in an extreme case causing smoke to come from it, and perhaps a minor explosion. Incredibly, integrated circuits subjected to such abuse can often still be fully functional when they are re-inserted correctly, but you are advised to replace them as soon as you can, rather than leave a potential source of weakness in the computer.
Experience shows that there is quite a good chance that there will be some visible physical fault, e.g. solder splashes on the tracks, unsoldered or "dry" joints, incorrectly inserted ICs, and so on, (e.g. ICs in the wrong sockets, or inserted badly, so that a pin is bent and makes a poor contact).
The integrated circuit sockets used in the kits are a high reliability type which has been most carefully selected after considerable experience. They have a tenacious grip on the ICs; it is also easy to enter the IC leads into the type of socket supplied, because of the lead-in ramp, and the generous and visible contact area.
Bare board users who have used some of the inferior types of socket widely available, must take special care to insert the ICs correctly, and should remember the sockets can be the cause of faults. (The "inferior" type of socket has a very flimsy contact and weak grip on the IC leads, and has the contacts hidden behind small "windows" through which the IC leads must pass. It is vital with such a socket to ensure that the IC is inserted so that the leads are true and straight so that they do not slip to the wrong side of the contact, or bend the contacts out of line an IC insertion tool should really be used.)
The main requirement in successful fault finding is a thorough knowledge of the whole card and the way it operates. Once you have this (and it is hoped you will have if you have studied this Manual so far) you will be able to localise the fault to a particular area of the circuit, and then "home in" on the particular source of the trouble. For example if a perfectly stabilised and steady display was obtained, but the fault was that no reverse video characters could be displayed it would be fairly pointless to start looking at the timing generator and counting chains, because they must be working fairly well to display any kind of stable picture. In such a case a sensible place to start looking would be the circuit which controls the selection of reverse video characters, the most significant bit from the RAM, the Alt/Rev selection switch and so on.
On the other hand, if the fault was a very weak, grey picture, it would be foolish to start looking for trouble in say the data bus transceiver or the address decoder, whether the data are good bad or indifferent they should still be displayed perfectly.
The circuit has already been described at length, in various degrees of detail, in Section 2 of this Manual, but it is proposed to take another pass through the diagrams, this time from the point of view of looking for various things which can be considered when fault finding.
With the exception of printing mistakes there is not much to go wrong here, but it is recommended that this be studied first when fault finding so that a "battle plan" can be formulated. Ask yourself where could the fault be to explain the symptoms you are suffering. For example if you are getting characters displayed in the margins, where they should be invisible, could it be the crystal oscillator at fault? Not really if that wasn't working there wouldn't be a picture at all. Could it be the UHF Modulator? Again, not really, if it was being incorrectly driven it wouldn't display anything, certainly it is inconceivable that it could be generating characters of its own in the margin. By asking questions like this it should be possible to deduce from the block diagram alone, the general area of the fault, and you can then move on to study the chosen part of the circuit in depth.
Is the crystal oscillator oscillating at 12 MHz? (Probe at U8 pin 6 rather than the crystal circuit itself to avoid disturbing its operating conditions and possibly stopping it yourself.) Are there the various binary divisions in the outputs from U14? If not, are the reset pins (U14/2,12) jammed high? Pin 2 should be always low, pin 12 should have the waveform shown in the appropriate diagram in Section 5 of this Manual. Is the Z80A-CPU receiving a full swing clock signal of the correct frequency? If not is the active pull up, Q1, working correctly? Is it connected right way round? Has some one substituted a different transistor? Is it PNP? With the correct pinouts (b,c,e)? Have the right value resistors been used? 22R not mixed up with 220R?
If the clock is OK is the Z80A-CPU working? Look at U1 pin 27, the M1 line. There should be plenty of activity here if it is executing instructions. If not make sure it isn't permanently being held reset. Ul/26 should be high. If not, why not? Is C11 correct polarity, not holding the line down low? Are U1/16,17,24,25 all high?
If the trouble is to do with the sync. pulses (i.e. you are getting a picture but it is not stable, or it is "torn"), is the "S" signal (U13 pin 2) correct, see timing diagram? If not, is it the timing program?
Check for read and write pulses on U1/21,22. Is there activity on the TA0-TA8 address lines? Do any signals look so similar that they could be short circuited? Ditto the TD0-TD7 data lines? Are the 0V and +5V connections to U2 (and the rest) all present? Probe the actual pins of the IC, remember it could be badly inserted into its socket. See if closing S2h has any effect.
This is most difficult to test in a non working system as it is impossible to set up and run any test programs. If you are really desperate and you are sure your fault is here you can set up some static tests, e.g. hardwire the AB10-AB15 lines to some chosen address, NRFSH high, and NlfREQ low, to see if you can get a match with the DIL switch S2 setting. Are you doing something silly, like setting the switches on when they should be off and vice versa? Are you setting them back to front, i.e. most significant digit confused with least significant? Is SIL1 plugged in right way round, are you sure pin 1 is pin 1? Fortunately, there is one part of the circuit which will hardly affect the operation of the card if it is faulty, and that is the monostable, U3. This is used purely for the anti-snow feature, which is a cosmetic improvement, the board will otherwise work perfectly well without it, assuming nothing drastic like the output stuck permanently high.
It is quite difficult to test this part of the circuit under the high speed conditions under which a computer operates, especially when as is likely to be the case if you are having trouble with this part of the circuit, the computer is not operating. Any small error will throw the whole logic out of operation, and without great experience and/or sophisticated equipment it is really only a matter of luck to be able to find a fault here. To give luck a chance, you can probe the various elements of the circuit although the various waveforms encountered generally will be meaningless, you can look for lines which are "stuck" at "0 or "1", or which dont appear to be reaching a low enough "0", or a high enough "1", inverters which dont seem to be inverting, and so on.
It may be helpful to read through the Detailed Circuit Description earlier in this Ilanual, and to study the waveforms both on paper, and physically on the Card. As you gain an understanding of how the circuit operates you will be able to direct your own test procedures so as to find which piece of the circuit is malfunctioning, which is the first step to repairing it.
It is stressed again that if you do not have success in finding a fault that is present, you can always hand it over to someone else. The after sales service which is available to purchasers of the kits is one of the many things which mark out the Interak 1 System as being a unique and worth-while product.
In contrast, this part of the circuit is very easy to test, in that it is a very simple counter chain. For all the counters on this sheet of the diagram it is a simple matter of checking if there is an input, if so, is there an output, and so on for all the outputs? If there isn't an output, or if it isn't the correct frequency, then find out if the supply voltages are connected (to the actual pins of the ICs, not just the sockets). Are the reset pins jammed to the wrong logic state? Are they shorting to some other line, causing premature resets?
This is another of the more difficult parts to test when the computer is not working. If the fault is in a lack of picture of steady well defined (but random) characters, then check pin 1 of the multiplexers U16 to U18. Mostly (in fact all the time if nothing is reading or writing to the video RAM) pin 1 should be low. This selects the cycle of counts on the "A" inputs (pins 2, 5, 11, 14), and transfers them unaltered to the "7" outputs (pins 4, 7,9, 12). Check that it does. Be sure the !CS line on the RAMs (U26/8 and U27/8) is low, and if it is check the ouputs (pins 11, 12, 13, 14) for some sort of activity. If they seem to be very low in amplitude see if they are in some sort of contention with U19, which could be caused if U19/1,19 were both low.
When nothing is accessing the card the outputs of the latch U25 should follow the inputs. If they dont, see that U25/11 is high, and U25/1 is low. If the latch was jammed to some particular character that character would be displayed on the whole screen, but often this symptom is the result of a runaway program writing the same data throughout the RAM, including the video RAM.
If a display of random characters is being obtained but it is not possible to change it then (assuming the rest of the computer system is free from faults) the fault could quite likely be in the multiplexers or buffer, or of course in the signal which control them. A very obvious cause of this difficulty is the lack of a write pulse for one reason or another, but you also have to be sure that the control pin 1 of the multiplexers is being activated, and the control pins of the octal tranceiver i.e. U19/1,19.
This is perhaps in the "easier" category. Problems associated with bit LD7 (reverse video which wont reverse, or an alternate character set which can't be obtained) should be fairly easy to find by tracing the circuit from LD7 through the links and/or switches. Is inverter U4/3,4 working? Are the character generators plugged in correctly, correct type selected on P5?
If recognisable characters are not being displayed are LD0-LD6 active? Are PL0-PL3 cycling round a count of 10 correctly? Are the outputs CD0-CD7 jammed for some reason? Or perhaps just one bit? (Pretty obvious from the display.)
Is the shift register U22 working, clocking out dots? Are the waveforms in the circuit just above the shift register basically correct (see Timing Diagram in Section 5.2 of this Manual)?
Another fairly easy circuit to search for faults. For lack of picture content on a well stabilised display trace back from the line marked "VID". Just follow back through the inputs of the gates, until you see where the signal has come to grief. Checking the two flip-flops (contained in U20) is easy enough. If an output is wrong, then is there a clock input? If there is, are there data ("D") inputs of the correct polarity? If not, why not?
If the fault is a picture of sorts, which cannot be sysnchronised then check the line marked "S". Are the combining circuits which bring together the video and sync. information working, for either or both the video monitor output and the drive to the UHF modulator. Is CR2 the correct way round? Has it been snapped by clumsy bending of the leads? Are the resistor values correct?
Beginners, dont make the mistake of assuming that the UHF modulator is faulty if you cant see any activity on an oscilloscope connected to its TV aerial output youre not likely to unless you've got a very special oscilloscope! Is the case of the modulator earthed, via its earth tags, and is this connected to the front panel TV aerial socket?
There is not much on this sheet to go wrong, and it is a bit late in the day now to be checking if the power is connected, but if you have the world's most classic fault, a short circuited supply rail, then you will have to look at at all of the various decoupling capacitors shown on this diagram to see if theyve got anything to do with it. Probably it is best to pull out all of the ICs just to be sure, but then it is a matter of cutting tracks, drilling out plated through holes and so on until you can find the faulty section of track. Better still send it back to the firm who supplied it, they were daft enough to volunteer to fix any fault you can put into a board, call their bluff!
This concludes the section on fault finding so the authors of this manual hope the card is fixed by now, because they've run out of ideas!