Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN'S LOG What’s happened to service backup? I have a rather mixed bag this month, ranging from an el cheapo, downmarket TV set to an allsinging, all-dancing, state-of-the-art model with more bells and whistles than are ever likely to be used. Neither set was an easy exercise. The first story concerns a 34cm portable colour TV set made in China and marketed under the Vision brandname (model VIS-146R). It is not only a story of technical problems, although it has its share, but also one of woefully ineffective backup and support for imports from this part of Asia. More on that later. The technical problem, as described by the owner, was simply a form of frame collapse or, as he put it, “after the set’s been running for a while, a black band comes down from the top and up from the bottom.” Well, there’s nothing new about a complaint like that. And, while I had no data of any kind for this set – it was the first one I had seen – I didn’t imagine that it would a particularly difficult fault to fix. On the bench, the set behaved exactly as the owner had said. It performed normally for about 10 or 15 minutes then suddenly went into the fault condition. Fortunately, even without a circuit, the vertical system appeared to be fairly conventional and based on a 7-pin IC (IC205, a TA8403K). After taking some voltage measurements around this IC, in both the normal and fault conditions, and checking associated components, I concluded that the IC was faulty. Unfortunately, my regular spare parts supplier did not list this IC, so I approached another organisation who were supposed to be the agents for this set. And this was the first hurdle. When I identified the set and nominated the IC type number, I was promptly informed that that IC was not fitted in that set. 34  Silicon Chip And in vain did I emphasise that I had the set on the bench and the IC in my hand. “Aw no, it can’t be”, was the only re­sponse I could get. There followed quite an argument, which culminated in a grudging offer to investigate and ring me back. Of course they didn’t and I had to follow up with several more phone calls, only to be shoved around from technician to storeman and back again. The upshot was that, while tacitly admitting that I had correctly quoted the number in the set on my bench, they suggested a substitute unit, LA7830. And they quoted a price of something over $20. I said, “thank you very much; I’ll let you know”. And there was a very good reason why I didn’t immediately place an order. I was familiar with the LA7830 and I knew that my regular supplier stocked it. More importantly, he listed it at about half the quoted price. So the replacement IC was duly acquired and fitted. And that fixed the fault; the set came good immediately. I gave it a thorough workout for the next week or so (the owner had stressed that he was not in a hurry) and, in view of subsequent events, I was very glad I did. During the test week, however, it didn’t miss a beat, in spite of all the abuse I could heap upon it. And so it was duly collected by the owner. That was the last I heard of it for the next couple of months. Then the owner was on the phone with the bad news; the set had failed again with the same fault. He was quite reasonable about it – almost apologetic – but it was a bit of a shock to have the set bounce. I asked him to bring it in immediately. On the bench, the set did appear to be exhibiting the same fault – at least superficially. And it even fooled me at first. But it wasn’t exactly the same. I recalled that the original fault had been quite predictable in its onset; it would appear every time within 10 or 15 minutes of switch-on. Not so now. Sometimes the set would run perfectly for hours, then suddenly go into fault condition. At other times it was in fault condition at switch-on. My suspicions aroused, I made a check around the previous culprit, IC205. There were none of the voltage changes which had lead me to this component be­fore. But there was one change, not evident on the previous occasion. The supply rail to the IC was normally 25V but now, in fault condition, it dropped by about 3V. Another IC fault? Not very likely I felt, since the IC gave no sign of overheating. So was it a fault in the power supply, or somewhere along the line to the IC? That was probably the explanation but it was time to call a halt. I wasn’t prepared to go any further without some service data – a circuit diagram at the very least. A sorry tale And that brings me to hurdle number two – and a sorry tale it is. I approached the aforementioned agents and asked what they could provide. The best they could offer was a circuit diagram. I was prepared to settle for this and they promised to send one. When it didn’t arrived in a reasonable time, I rang them to see what the problem was. They made some vague excuses about a shortage but my impression was that they had forgotten all about it. They promised to chase it up but it still didn’t arrive and further phone calls over several weeks pro- duced further vague excuses. Finally, after some fairly straight talking on my part, a circuit arrived. But – you’ve guessed it – it was the wrong circuit. In fact, it was nothing like the set on the bench. It took several more phone calls (and related excuses) before the correct circuit turned up – for what it was worth. My guess is that it was about a 10th generation photocopy, made via a couple of pretty grotty copying machines along the way. It was almost completely unreadable with component values, type numbers, identification symbols, voltages (such as there were) IC pin numbers and any lettering all just blobs. The best I could do was try to relate actual component type numbers or values with the blobs on the circuit and see whether they seemed to match. In most cases, it was guess work more than anything. And that was all I had to work with. Granted, it was better than nothing, in the sense that I could at least follow the general circuit trend which all seemed fairly conventional. Well, that was something. Power supply checks So back to the fault. Following up the lowered 25V rail clue, I checked the main HT rail. With the set running normally, it sat at about 120V but in fault condition, this dropped to around 93V, although this figure varied somewhat. OK, so we had a main power supply fault. And this fitted in with another earlier observation. When the vertical deflection decreased, so did the horizontal scan (though less obviously) – something that suggested a common fault. The power supply appeared to be quite conventional and very similar to some used in Samsung sets. And it used the same common IC, an STR5404. Well, that was a small plus. I turned the set on and waited for the first sign of the fault. When it appeared, I began prodding around the board, hoping to get some kind of a lead. And I did; the board was extremely sensitive, particularly around the IC. In fact, the IC itself appeared to be the most sensitive. My next trick was to try a spot of freezer on the IC while the set was in fault condition, taking great care to keep the freezer off any other components. Result – an instant cure. And it worked every time. I carry this IC in stock and, with only five pins involved, it took only a few minutes to fit a new one. My self-confidence was shattered immediately at switch – there was absolutely no improvement. In fact, I soon established that I could cure the fault by spraying the new IC or any component on the board. I was back to square one. Well, not quite. All these observations added up to a strong suggestion of a dry joint or a hairline crack in a copper track. I pulled the board out again and went over the copper side with a magnifying glass. And in spite of a careful inspection, I could find nothing even vaguely like a faulty joint. But there had to be a fault in there somewhere, so I decided on a brute force approach – go right over the board and resolder every joint. Yes, I know, it takes time and may also dent one’s ego a little but it is often the most effective approach. And it certainly was in this case. It took me about 20 minutes to do the job – and I could have spent more time than that just prodding and pondering – and it cured the fault. And when all is said and done, that was the purpose of the exercise. Another intermittent Naturally, I gave the set a good workout over the next few days and nothing I could do would produce the fault. But it wasn’t the end of the story. During this procedure, I became aware of another fault – also intermittent – which the owner had apparently not September 1995  35 of that set or any of it brethren. In fact, another one turned up a few weeks later but I had to say “sorry, I can’t service it”. And I went on to explain the problems of obtaining data and other technical back­up. And unless those concerned can get their act together, I suspect other servicemen will be forced to adopt the same atti­tude. The snack Fig.1: the IF and stereo sound decoding circuitry in the Sony KV2764EC. IC102, which contains the FM detector, is at top left, with L111 between pins 9 & 10 and L105 and C124 between pins 7 & 8. noticed and which I hadn’t noticed either in the confusion associated with the original fault. Now that I could watch the screen in a more relaxed manner, I suddenly became aware that there was a loss of blue in the pic­ture from time to time. The effect could be somewhat subtle at times, depending on the overall colour content, but it was defi­nitely happening. This part of the circuit is quite conventional. Three drive transistors – red, blue and green – on the neck board feed the three picture tube cathodes and, in turn, are fed from the PAL decoder IC on the main board. My first reaction was to suspect the blue drive transistor. This is a common type, a BF422, and it was easy enough to replace it with one from stock. But no joy, the fault remained. Next, I checked the voltages, particularly the base voltag­ es, around all three transistors, first while the set was normal, then when the fault appeared. Under normal conditions the reading on all three bases was about 4V but, in the fault condition, 36  Silicon Chip this was something less the 2V on the blue base. The base of this transistor connects to a 2.8kΩ trimpot, used for colour balance adjustment, and from there to the main board and the PAL decoder blue output. However, there are two more components in the line: a small RF type choke and a 100Ω resistor. Tracing this line showed the low voltage at the connection to the main board but the correct voltage at the PAL decoder. By then retracing the line from the decoder, I found the correct voltage up to and beyond the 100Ω resistor but not beyond the choke. I checked the choke’s soldered joints very carefully but they appeared to be perfect. Nor could I find anything wrong with the choke when I pulled it out and tested it. As a result, I refitted the choke, this time taking particular care with the soldering. And that was it. It took me several days of constant monitor­ ing to be quite sure but the fault was fixed. And it has remained fixed for several months now. So I hope that is the last I shall see Well, after all that, a change of pace is called for. Here are some shorter stories; stories I have been holding back for some time, due to space problems. And the first one is something of a reversal of the usual theme. One of my common themes is the fault which looks like a snack but turns out to be a real stinker. In this case, it was a strange new fault which I thought would be a stinker but which turned to be much easier than I expected, if only by good luck. It concerns a Sony KV2764EC colour TV set (PE-3 chassis) and the owner complained that there was very little or no sound when operating the VCR, although there was plenty from the chan­nels on the TV set. It had started by being intermittent but was now permanent. The set was too heavy for the owner to carry it in by himself, so we brought it in together, placed it straight on the bench, and connected a VCR to it in the usual manner. In some perverse way, and in spite of my apprehension, I was actually looking forward to seeing this fault because it sounded like a challenge. When the owner first mentioned the trouble I had thought he may have connected the VCR incorrectly to the audio and video connections of the 21-pin (SCART) socket on the rear. Or perhaps he suffered from finger trouble and had pressed the wrong buttons on the TV set. But he turned out to be an intelligent bloke, in­trigued as much as I was by the fault. More importantly, he was fortunate enough to have two VCRs. So, when he first noticed the fault, he swapped them around but to no avail. Then he tried swapping the video RF output from Ch1 to Ch0 and retuning the TV set to it. Still no luck. Not only was the sound weak when playing tapes but also when selecting the stations on the VCR’s own tuner (sometimes called the EE mode). And, alter the fault again. I even tried changing the IC, a TDA2546A, but no joy. I noticed the freezer seemed to have more effect around the end near L111 and L105, where there also happened to be a 470pF styro capacitor C124. I pounced on this because this type of capacitor has a habit of changing its characteristics. I gently squeezed and moved C124 whilst the set was on and was rewarded by a momentary change in sound level. That was enough; I changed it for a new one and the sound was restored. I phoned Sony technical support on another matter and in the course of the conversation told them about this fault, only to find that this is a common one and the styro is now replaced by the higher grade gold version! I guess I should have phoned them first. Anyway, the customer was as delighted as I was and departed as happy as a sand-boy. But I have to thank lady luck. The super Mitsubishi as I soon discovered, the situation was no better with my VCR which used an RF output on UHF Ch37. Next I tried using the video-in/ audio-in connections via the SCART socket. Ah-ha, plenty of sound, so it must be some sort of RF/IF problem. But why on earth was it discriminating against VCRs? Didn’t this set like them or something? More realistically, what was the difference between an off-air TV station and a VCR? The shape of the sync pulses? No – it took a while but the penny finally dropped. All our TV chan­ nels transmit in stereo, using the German Zweiton or two-tone system. But a VCR always transmits in monaural sound, probably because it is too expensive to provide stereo modulation. So the set was discriminating between a stereo transmission and a monaural one and was muting the latter for some reason. I confirmed this to some extent with an RF signal generator but, not having a stereo generator, I tried attenuating a TV channel until the stereo light went out, which dropped the sound simulta­neously. But where was the fault? Was it in the tuner, IF, stereo decoder, or audio logic circuits? Fortunately, all these areas, except the last, are located on board “A”. I decided to start working from the decoder and move out from there. The two inter­carrier signals at 5.5MHz and 5.74MHz are fed into IC151 and IC152 where they are decoded. I began by measuring the voltages around these ICs and found them close to those shown on the circuit. But I was getting a gut feeling that I should concentrate on the 5.5MHz signal which is the usual monaural signal (L+R) rather than the 5.74MHz signal (2R), used for stereo only. This focused my suspicion to IC102 which contains the FM detector. I decided to hit the components in this area with the hot and cold treatment, on the off-chance of a clue. I was lucky – the freezer momentarily brought the sound up but it quickly disappeared and heating or freezing the components around the IC didn’t It wasn’t much later when I encountered a similar fault to this involving a Mitsubishi CT-2553EST, fitted with an E4-Z chas­sis (very similar to the AS2M chassis). This is one of the newer generation, all bells and whistles, super-deluxe models, boasting just about every feature you can think of, including a multi-standards facility (PAL, NTSC and SECAM). It’s is all very clever but I do wonder how often these features are used, or even understood, by the average person. Of course, it makes for a beaut sales pitch but it means a higher price tag and, more to the point, much greater complexity when trying to service the monster. And so it was in this case. A major problem in this set is access; trying to take measurements on the main board and the modules while the set is on is a very dodgy procedure. The leads are very tightly dressed in a wiring loom, preventing the board from being easily inverted, while the plug-in modules are just too close together for examination. And to unplug them is a major operation! Anyway, the complaint amounted to intermittent loss of luminance; ie, a ghostly, over-saturated, dark picture. It was clearly a job for the CRO and we started from pin 10, video out, of the VIF pack, IP1A1, on the main board, September 1995  37 Fig.2: a corroded track in the 12V supply line to transistor Q254 caused the original loss of sound in the Mitsubishi CT2553EST. Note that the 12V rail comes in on pin 3 of socket VC2 at lower right. PCB-MAIN. There was nothing wrong here, the CRO showing a very clean composite wave­form. OK, but which way next? What follows from here is a bewildering maze of video switching circuits but I managed to trace it, eventually, to PCB-Y/C-SW, via plug/socket VS3, thence through transistors Q2P1 and Q2P2, IC2P1 and Q2P3. From here, despite many blind alleys, I traced the signal to PCB-VC/RGB-CTI via plug/ socket YS1 and thence to the emitter of Q254, where the trail became cold, or rather distorted. The collector voltage of Q254, shown as 11.9V, was very low. This is derived from a 12V rail coming in on plug/socket VC2-3 and which was correct at that point. I was getting close. To cut a long story short, after much bad temper, due to the awkward access to this module, I traced the 12V copper track around the edge of the board. Near the top, it disappeared under that horrible brown glue which I have mentioned in these notes in the past and which the Japanese are so fond of using to anchor parts. Hasn’t anyone ever told them – or don’t they listen? I’m still finding this glue in near new sets. Anyway, as usual, the glue had hardened, changed colour (darkened) and corroded the copper track, changing it, effective­ly, into a high value resistor. I linked the break and 12V was restored, along with the luminance. It all sounds so simple but it meant taking the module out, measuring and working on it, and replacing it a number of times. This was a very slow and weari­some process . And that should have been the end of the story. Unfortunate­ly, after reinstallation at its home with its VCR, it bounced. Yes, you’ve guessed it; intermittent no sound, or rather very little sound, from the VCR. Unfortunately, I sensed that the owner wasn’t very happy. He didn’t say much; it was rather what he didn’t say that alerted me. It was obviously a different fault, for which I was not to blame, but it was too difficult to argue at that stage. Fig.3: the TU1A1 U/V TUNER and IP1A1 VIF-PACK are only available as a combined pair in the Mitsubishi CT2553EST. The VIF-PACK is prone to dry joints around the filters, coils and the SAW filter. 38  Silicon Chip I confirmed it was OK through the audio/video sockets and I was half-tempted just to supply suitable leads and leave it at that (no, not seriously)! But my reputation was at stake. Since I was – apparently – the one who had wrecked a beloved set, I had to be sure I “fixed it proper”! As with the Sony, I started at the decoder, located on the PCB-SOUND board, and went straight to IC3001, a TDA3803A. I have had a lot of trouble with the TDA3800G which was used widely on earlier models, especially JVC. But changing this made no dif­ ference and I moved to IC3000 and its associated circuitry. I examined this very carefully, particularly capacitor C3002, which plays a similar role to C124 in the Sony circuit. But everything checked out OK and wouldn’t succumb to the hot and cold treat­ment. The next major suspect was the previously mentioned VIF-PACK IP1A1. I have had a lot of trouble with this module over the years, including a variety of intermittent faults, mainly video, and everything from no picture to patterning, or snow on UHF only. And you cannot purchase a new module by itself. Instead, it is part of a (very expensive) matched pair consisting of Tuner TU1A1 and VIFPACK IP1A1. I unsoldered the module and examined it. Normally, it is very prone to dry joints everywhere but more so around the filters, coils and the SAW filter. But this set had already been worked over so I reinstalled it with the covers off. I switched the set on and tried freezing the ceramic filters, something that is not usually recommended as it can sometimes damage them. In this case it made no difference. I then tried a trick I learned from a colleague some time ago; running a moist finger over the circuit. (Don’t try this trick unless you’re very sure you know what you’re doing; par­ ticularly on a live chassis!) It was his extension of the tradi­tional bash and prod, wobble and twist, hot and cold techniques, often needed as a last resort to pinpoint a difficult fault – particularly one involving a change in a characteristic, rather than a clean cut failure. I had tried it a couple of times before without any luck. But this time it worked. I noticed a change in the sound level when touching around the 5.5MHz ceramic filters X3 & X4. The module was removed again and two new filters fitted. And that was it; sweet success – the sound level was back to normal. Apparently, one of the filters was faulty and off frequen­ cy. My finger on the board pattern was re-tuning it slightly. Once again I was lucky. And, hopefully, my reputation had SC been restored. Especially For Model Railway Enthusiasts Includes 14 projects for model railway layouts, including throttle controllers, sound simulators (diesel & steam) & a level crossing detector. Price: $7.95 plus $3 for postage. Order today by phoning (02) 9979 5644 & quoting your credit card number; or fax the details to (02) 9979 6503; or send cheque, money order or credit card details to: Silicon Chip Publications, PO Box 139, Collaroy, NSW 2097. September 1995  39