Fullscreen HTML5Med Res (??MB)HTML4

Brand what? Never 'eard of it! Appliances which are not well known on the local market can pose problems for both the customer and the serviceman. For the customer they present the problem of finding someone to service them, while for any servicman who takes on the task there is the problem of spare parts supply and technical backup if needed. My main story this month involves such problems, along with the technical problem which started the whole sequence in the first place. In fact, the problem deserves a place in history in its own right. The offending device was a 50cm colour TV set made by Thomson. And if that brand doesn't ring a bell you're in good company. Thomson Grand Public (to give it its full name) is a French company and, by all accounts, one of the biggest electrical and engineering firms in Europe - some say second only to Philips in size. Be that as it may, they are certainly a very large organisation with interests extending far beyond mere TV sets. (What am I saying!) Anyway, the story started when a lady rang me and identified herself as a former customer; one I had not heard from for several years. Then she went on to explain that she had a Thomson TS5108 TV set which needed service and wanted to know if I could help. Fortunately, I have a nodding acquaintance with this set - and a manual for it - so I said, yes, I'd have a look at it. At this she heaved a sigh of relief and confessed that she had tried several other servicemen in the area and that none of them would even consider it. To be honest, I also had some reservations. The only reason I knew anything about the set was because a local dealer had sold a number of these about five years ago and I was then doing warranty service work for him. In fact, I performed a couple of minor service jobs on those sets at the time; enough to provide the nodding acquaintance, but nothing more. Spare parts Another possible reason why other servicemen may have shied clear of the set was concern about the availability of spare parts. Noone wants to spend several hours tracking down a difficult fault only to find that it involves a component for which no replacement is available. (That problem can be r ; / ; ' ' --r.c ·.ov,,r12,uv ,u1.,100,oj " IJ , N ~ ~~ j:~. I N 1 :.2 = .l -I 15n50V l - CF06 ~ 1'CF08 .,, I Df'04 J.. !'"""'' .____.__---.I a.-:,o c~ ~ - - -JFOl -- - - - - - i - - - - - - - ~ ,....- - :::F05 V2 • 2'tV _J HEATER 2 '·; ... J • :oov Fig.1: relevant portion of the Thomson TS5108 circuit. The vertical oscillator stage is on the left (TF03 & TF04), the output SCR to the right of centre (CC02), the horizontal output transformer below it, and the vertical scan coil (BF0l) to the left of this. 32 SILICON CHIP bad enough with well known brands). Fortunately I knew - or thought I knew - who handled spare parts for this set so I was not particularly worried. In the event I was wrong but only insofar as the particular firm was concerned. More about that later. In the meantime, back to the set. The lady turned up with it the next day and the first thing I did was try to ascertain the nature of the fault. The lady's description was that "most of the screen is blank". As it turned out, that was a very accurate description. But I needed to see it before I realised exactly what she meant, so I plugged the set in while she was there and fed it with an off-air test pattern. The result was like nothing I had ever seen before. All that appeared was the top quarter of the pattern but this was located at the bottom of the screen. Above it, the screen was black and, rather surprisingly, there was no evidence of the blanking period. The only other point that struck me was that what was visible of the pattern was perfect in all respects: colour, brightness and linearity in both axes. So I'd saddled myself with a set about which I knew virtually nothing and which was exhibiting a fault like nothing I had seen before. Why do I get myself into these situations? Down to business Anyway, down to business. After studying the circuit, particularly the section dealing with the vertical deflection, I realised that it contained a number of unusual features. In fact, some of these features were quite puzzling but more of that in a moment. The relevant portion of the circuit is reproduced here as Fig.1. I started with the vertical oscillator section consisting of transistors TF03 and TF04 at the extreme left of the diagram. Note that immediately above these transistors, the circuit shows the number "15" in a diamond. This is a waveform number and refers to one of the CRO patterns set out along one edge of the circuit. Unfortunately, these patterns are very small, measuring a mere 8mm wide by 6mm high, so it's hard to pick out the finer points of the trace. Nevertheless, it appeared that the oscillator waveform was as it was supposed to be, so I moved on to waveform 16, skipping transistors TF05, TF09 and TF06. Once again the waveform appeared to be OK, so I moved on to waveform 17, at the gate of the vertical output stage [SCR type CC02). The use of an SCR in this stage was just one of the circuit surprises as was the strange way in which the device is shown. But it was waveform 17 which really caught my attention, since it bore little resemblance to the published pattern, miniaturised though this was. This prompted a voltage check of the output stage and preceding driver stage, TF07. The latter seemed to be within tolerance but the anode of the SCR was well down. It's shown on the circuit as running at 24.4V but it only measured about half this figure. This not only threw suspicion on the SCR but also drew my attention to the quite strange circuitry around it. For a start, I was surprised at the low voltage indicated for the anode, considering the job it had to do. And, in fact, the waveform (21) at the anode is shown as having a peak-to-peak value of 190V. Even allowing for the inductive nature of the load, this seemed to be asking a bit much. Back tracking from the anode revealed that this supply line passes through a winding on the horizontal [yes, horizontal) output transformer [pins 3 and 6) and thence to a 21 V rail via the vertical MAY1989 33 SERVICEMAN'S LOG -CTD scan coils, BFOl. The 21 V rail (extreme right of the circuit, about half way up) is also derived from the horizontal output transformer. Anyway, back to the SCR. I suspected it was faulty and pulled it out for a test. Judging by the symbol used, I assumed that I should be able to measure continuity between cathode and anode, at least in one direction, of the internal diode. But no; there wasn't the slightest indication of this component. Then I measured between gate and cathode, expecting an open circuit indication. But again I was wrong; I found I was looking at around 400, measured at either polarity. All this suggested that the SCR was shot and that the obvious thing to do was to fit a new one. However, closer inspection revealed that the type number on the component was not the same as that on the circuit. The circuit shows CC02, whereas the component was marked ESM740. I didn't have either beast in stock, of course, so I had to contact the spare parts supplier for these sets. As I recalled, spare parts had been available, along with service, from Hills Industries so this was who I called. Unfortunately, this didn't help much because they didn't stock any spare parts for these sets. Service, yes; spare parts, no. For these I was referred to the importing agents for Thomson sets, Lemair Australia Pty Ltd. [Further details about these firms are given in the accompanying panel). And so to Lemair. Yes, they knew what I wanted, the type ESM7 40 was the current replacement, and they had ample stocks. The price of the part was around $10 and all I had to do was place an order. On an impulse I ordered two; a very wise decision, as it turned out. Fitting the SCR The two units came to hand in a couple of days and I promptly fitted one to the set. Alas for my expectations; the set behaved exactly as it had before. I hastily unhooked the SCR and measured it. It measured almost the same as the original, except that the resistance between gate and cathode was about 100 higher. My immediate reaction was to suspect that there was a fault in the set which had destroyed the original SCR and had now destroyed this one. It was a nasty thought considering what it would mean in terms of tracking the problem down. But this was where my impulse to buy two units paid off; I grabbed the second one and measured it. And it was exactly the same, except that the gate/cathode resistance was somewhat higher again, nudging 600. All of which meant, of course, that I'd been chasing a furphy; there was nothing wrong with the original SCR, only my interpretation of the symbol. [With hindsight I would have saved myself a few bad moments had I bothered to check the new SCR before I fitted it). Square one But while I heaved a sigh of relief over this little incident, I was less 34 SILICON CHIP happy about the overall picture. The simple fact was that I was back to square one. All I had learned was that it wasn't the SCR. And so I began a systematic check of voltages and waveforms from the SCR back towards the oscillator. I checked or replaced the components around TF07, TF06 and TF09, as well as the transistors themselves. In fact, I checked most of the components and all the voltages right back to the oscillator. This seemed to yield nothing worthwhile; all voltages were within a whisker of those marked on the circuit and as far as I could determine from the miniature waveforms, there was nothing suspicious here either. But there was one diversion. While I was doing this I suddenly became aware that the picture on the screen had changed. The top of the test pattern, previously about one quarter of the way up the screen, had moved to the middle of the screen, but only at the expense of linearity. Fortunately, the cause was soon tracked down. I checked the anode of the SCR and found zero volts. The reason for this was quickly traced to a 150 safety resistor [RL52, not shown here) in the 21 V supply line, which had gone open circuit. I assumed from this that, whatever the fault was, it was putting an additional load on the 21 V rail. A conventional 150 unit provided a temporary repair. But the main fault remained as elusive as ever. While pondering on the circuit I noticed a couple of preset controls: PFOl, a 2.2k0 pot in the base circuit of TF04, and PF04, also 2.2k0, in the base circuit of TF06. PFOl is a vertical frequency control while PF04 is the vertical centering control. So what would happen if I adjusted the latter? Suiting the action to the thought, I found that this would move the picture up the screen, with more of the pattern appearing at the bottom as this happened. In fact, shifting the control to its limit produced almost half the picture. Thus encouraged, I reset the pot to its original position and shorted out RF23, an 8200 resistor in series with it. L-E:N\A\R HAD AMPLe:. S,OCl(S 0~ -r-<PE:- G.SW\ 740.. • This also moved the picture up the screen, though not quite as far. But by adjusting PF04 again I was able to bring the picture right up the screen and display the whole test pattern. So to all intents and purposes, I had cured the fault. It would have been easy to leave it like that, give it back to the customer, grab the money and run. Fortunately, my conscience - and common sense - prevailed. But, interesting though this little exercise was, it didn't really tell me much. In fact, I was very little closer to solving the mystery than I had been when I started. I decided it was time to ask for help. I put in a call to the Hills Industries service department, where I eventually contacted the technician who normally handles these sets. I outlined the nature of the fault, described what I had tried so far, and raised the question as to how some of the less conventional parts of this section worked. I hoped that, with a few clues here, I might be able work something out. The technician was most attentive, asked whether I had tried this and that - which I had - and then confessed that, at that distance, there was little else that he could suggest. As for the strange circuit configuration, I gained the impression that he had puzzled over it just as I had. The upshot of the discussion was that there was only one component about which I could not be sure - a 180kn resistor (RF02) below TF05. One end of this resistor is connected to the 200V rail while the other end goes to the oscillator preset control, PFOl, via RF06 (15k0). I had probably tested it but couldn't be sure. And according to the Hills technician it sometimes goes high and "causes all sorts of problems". I thanked him for his time and went back to the bench. As soon as Thomson TV Receivers • Australian agents for TV receivers and spare parts: Lemair Australia Pty, Ltd, 23 Amax Avenue, Girraween, NSW 2145 . Their postal adress is PO Box 336, Wentworthville, 2145 . Telephone (02) 636 3033. • Service only: Hills Industries Ltd, 12 Wiggs Rd, Riverwood, NSW 2210 . Telephone (02) 533 4855. I fronted up to the set I was sure I had checked that resistor but I checked it again anyway. It was spot on value so that blew that theory. But now that my attention had been drawn to the 200V rail I realised that there was a resistor I hadn't checked: RF15, an 820k0 unit from the 200V rail to the base of TF09 and the collector of TF05. The voltage at this point is shown on the circuit as 22.8V and I had measured it at about 21 V, which seemed close enough after taking into account normal component tolerances. This is probably why I had neglected to check the resistor more thoroughly. It took virtually no time at all to lift one end of the resistor and check it with the meter. And that was it - RF15 was open circuit! I fitted a new resistor, removed the short across RF23, reset PF04 and switched on. The result was a perfect picture - a fair dinkum one this time. That was the end of the story as far as the job itself was concerned. I ran the set for a couple days, then gave it a routine touch up and passed it back to the customer. Post mortem But, of course, a post mortem was inevitable. Where had I gone wrong? Now that I knew what had happened, it was easier to analyse that part of the circuit. As shown in Fig.1, bias for the base of TF09 is derived from the 200V rail via a voltage divider consisting of RF15 (the 820k0 resistor which I replaced), RF14 (4.7k0) and RF28 (3 .3k0). The junction of the bottom two resistors in this divider chain is clamped to about 20.4V by diode DF02 which is in series with the 21 V rail. I'm not sure why this rather strange arrangement has been used but the end result is that the base of TF09 normally sits at about 22.8V (ie, about 2.4V above the voltage at the cathode of DF02). In fact, the nominal 21V rail was running a little high in this set, making the difference somewhat less. But that's not all. Further analysis revealed that the emitter of TF09 is also connected to the 21 V rail - one path being via resistors MAY1989 35 perhaps saves a set from being written off. They are a nice set, a good performer and easy to work on. And I understand they are still available on the Australian market. So if you strike one, don't write it off without a try. A novel intermittent I ... ,..._15u, NOW-n\A-r MY Aire:.NrlO~ HAD 'B(;.E..N PRAWN ,-o 11-\E: z.oov RA\ RF19 and RF20 (1.2kn and 1000) and another via RF21 and diode DF05. This means that, without the 820k0 resistor to the 200V rail, the base and emitter voltages of TF09 were virtually the same. Or, putting it another way, there was no forward bias on TF09's base. Yes, it was as simple as that; simple when you've found it, that is. So where had I gone wrong? Accepting the voltage on the base of TF09 as being close enough was a mistake. What I should have done, of course, was to measure the base-emitter voltage of this stage. Instead, I made the mistake of simply measur36 SILICON CHIP l--000 ing those voltages shown on the circuit and assuming that, if they seemed to be within tolerance, all was well. So there's a lesson to be learned there. And why didn't the fault show up in waveform 16? It probably did but too subtly to be appreciated by comparison with the miniature circuit reproduction. It was only at the next stage, waveform 17, that it became obvious. So, there it is; all (well, something) about the Thomson TS5108. Let's hope that it helps at both technical and backup level and And now, for a change of pace, here is an interesting story from a reader, A. M., of Turramurra, NSW. After some very nice remarks about these notes, and servicemen in general who conquer intermittents, he goes on to relate the story of a novel intermittent which he encountered and - eventually - cured. This is how he tells it. My interest in electronics is purely amateur but I have exprienced at least my share of intermittents over the years. When silicon power transistors first became available in the late 1960s, I was persuaded by an engineer friend to build an audio amplifier described in "Radiotronics" from May 1967 to February 1968. I can remember being very impressed by the clarity which this amplifier afforded over my old valve model. This was in the days before printed boards were readily available, so construction was on Veroboard with plug-in tags to hold components and provide connections. The cases were folded from sheet aluminium in much the same way as we had made chassis for valve equipment. The amplifier was eventually installed in a cabinet, with the preamp in a drawer. All went well for some years, apart from a few blown output and driver transistors. I eventually replaced the original transistors with 2N3055s and BD139/140s which proved to be much more reliable. Then one day I noticed tha t the left channel was down in volume but the next time I used the amplifier it was back to normal. It remained normal for some weeks and I was tempted to believe it had been imagination. But the fault now entered phase two; noise, very like that generated by a faulty coupling capacitor in a valve set, became evident before the left channel level dropped. It was obvious tha t I would have to do some servicing. Slammed drawers The first clue indicated that the fault was in the preamp, but after the volume control; slamming shut the drawer which held the preamp would stop the noise and restore the level. Unfortunately, I would not have won any Brownie points from a serviceman for my design. It required considerable effort to get at the circuit, first to extract the preamp from the cabinet and then to take it out of it's box. And, naturally, when I had it all set up to test, the fault had disappeared and nothing I could do would make it return. So I put it all back together again, once more cursing my poor design. There was no sign of the problem again for some months and I was beginning to believe that it might have gone away for good - while knowing that such things never happen in real life. Then, almost a year later, there was a drop in volume (still the left channel), a rustling noise and then, very gently at first, a motor-boat at about 2Hz. Although this increased in amplitude, it did not get out of hand or threaten to damage the speaker. Once again I found that slamming the drawer shut would cure the fault and that remained my treatment for about another year, as the problem still occurred only infrequently. But several times during that period I took the preamp out and tried to find the fault. Sometimes the whole unit would be sensitive to the slightest touch but, ultimately, my probing and tapping would cause the fault to vanish and that would be the end of that session. The only bright side was that I took advantage of these occasions to improve the accessibility, so that it was no longer such a chore to remove and replace the preamp. Eventually the motor-boating came on and did not respond to the drawer slamming treatment so, for the umpteenth time, out it came. And this time it remained on, so there was no excuse for not finding it. The supply line to the preamp was taken from the main supply rail via the usual resistor-capacitor TETIA TV TIP Philips KT2A-3 Symptom: Very dark picture . Voltages around picture tube base close to correct values. Voltage on pin 5 of colour difference modules (U178, 179 & 180) is +o. 7V instead of -0.2V. Cure: C430 (1 00µF 50V electro) open circuit. This capacitor is the main bypass on the negative rails and, in particular, on -20a which supplies the bias on the colour difference modules. When the bypass fails, the luminance path through the modules is cut off. TETIA TV Tip is supplied by the Tasmanian branch of The Electronic Technicians' Institute of Australia. Contact Jim Lawler, 16 Adina St, Geilston Bay, Tasmania 7015 . decoupling network. A meter across this line clearly showed the supply line gently oscillating about the 25V level. It was only varying by about 100mV, but there was no mistaking its reality. Extra bypassing of this line merely lowered the frequency and tapping, heating, or freezing the components made not a scrap of difference. This part of the preamp had an emitter follower from the volume control feeding an active tone control, using two transistors in common emitter mode. Varying the bass control had only the smallest effect on the oscillation. I had mounted the various controls on the front of the box and run short lengths of shielded wire from the Varoboard to them. I was about to give up when, for want of something better to do, I measured the resistance between the volume control and the input capacitor; ie, 10cm of inner conductor of the shielded wire. It was open circuit! After unsoldering the wire I removed the shielding and examined the inner wire. Copper went into each end of the PVC but a pull removed each piece, which was only about 1cm long inside the PVC. Carefully opening the PVC, all I found inside was a green deposit; the wire had simply corroded away. This explained the fading signal and the noise, but what about the motor-boating? Then the penny dropped. Bias for the emitter follower was provided by a divider of two 100k0 resistors from the 25V line to chassis, the base being fed from the junction. The signal was fed from a 20k0 pot through a lOµF capacitor which held the base "earthy" to AC. When this connection was broken, there was sufficient AC coupling from the final stage of the tone control, through the supply line, to the base to allow oscillation at a low frequency, these two points being in phase. But what caused the wire to corrode? I have not seen or heard of anything like this and assume that there must have been some corrosive agent in the PVC, although the outer shielding was unaffected. Perhaps I would have found the fault more quickly had I been able to use a CRO or signal tracer but I never thought of a low level signal wire disintegrating! ~ RCS Radio Pty Ltd is the only company which manufactures and sells every PCB & front panel published in SILICON CHIP, ETI and EA . 651 Forest Road, Bexley, NSW 2207 Phone (02) 587 3491 for instant prices MAY 1989 37