Fullscreen HTML5Med Res (??MB)HTML4

The line of most resistance My main story this month comes from my Tasma~an colleague, J. L. As he freely admits, it is a comedy of errors - though I doubt whether there was much to laugh about at the time. But out of his trauma comes some good advice regarding the selection of substitute transistors. Customers are the strangest people. Often they will tolerate the most objectionable faults in their TV or radio sets and do nothing about it until some other relatively minor fault prompts them to seek help. I was reminded of this when a customer brought in a small Sony colour TV set, a KV1300-AS. The complaint was "a white line across the middle of the screen". When I fired it up, I saw the white line all right but I also saw the reason for it - the top quarter of the picture was folded over itself and Teletext lines were flickering across the newsreader's face. Not only that but the bottom of the picture was up about 25mm from the bottom of the screen. It seemed to be undistorted but was probably squashed a bit more than it should have been. When I pointed out these extra faults, the customer exclaimed: "They don't worry me at all. It's the line across the centre that is such a nuisance". All he wanted me to do was cure the white line, though I couldn't imagine how I could do that without curing the other faults at the same time. I began by looking at all the electrolytic capacitors in the vertical oscillator, drive and output stages. Dried out electros are far and away the most common cause of linearity troubles in vertical stages. Unfortunately, replacing all these made not the slightest difference. Voltage checks Next, I fired up the set with the vertical board exposed and set about measuring the various voltages for comparison with those given in the manual. The vertical 05082SC926A V. OR1VE 23 46.sw· 12 I ~I 200v 1 Fig.1: vertical deflection circuitry of the Sony KV-1300-AS. The fault was tracked down by carefully measuring the voltages around Q901, Q509 and Q508 and comparing them with those marked on the circuit. 56 SILICON CHIP output stage in this set is rather unusual. It uses dissimilar output transistors, Q901 and Q509, in a kind of asymmetrical push-pull circuit. One of the transistors is mounted on a heatsink on the vertical board and the other on a small tab attached to the main chassis. The collector of Q901 is fed from a 110V rail and the yoke drive is taken from its emitter. The circuit shows the emitter at 47V and the base at 46.5V. The collector of Q509 is shown as 47V, the same value as the emitter of Q901 to which it is directly connected. This level, in turn, is set by the bias on Q509, which is shown as 0.75V, while the emitter should be at 0.6V None of the measured voltages agreed with the circuit values and in one case the difference was quite marked; the emitter of Q901 was at 53V and the base at 42V, suggesting that there was a whopping 11 V reverse bias on this stage. But this meant very little, because the vertical oscillator was still driving these stages and the voltages I was observing were almost certainly being generated by rectification in the base-emitter diode junctions of the transistors. All they indicated was that something was very wrong with this part of the circuit - but I knew that already! I would have enjoyed the chance to sit down and work out just what was going on from the voltage readings but hard economics demanded a faster solution to the problem. In fact , the voltage measurements had yielded one important clue; the emitter of Q509 was at 0.4V relative to chassis instead of the 0.6V shown on the circuit. Since this voltage was being generated across a bypassed emitter resistor, it suggested that the transistor was not drawing as much current as it ooo-"fHex J:)ON'"'\ W~'< ME:1 ,,-....., ,, 'S 11-1-e:.. L-1 N ~ AC.ROSS -n-u~: c~e: -'l1;ig ,,,..____ _ _ _ _ _~('~•89-should and thus it did not have sufficient bias on its base. The bias for this transistor is derived from the divider chain R556, VR508, R560 and R559/ TH502. VR508 is labelled vertical bias and is used with the linearity control to optimise the vertical scan. When I went to measure the voltages down this chain, it turned out that R556 was dropping 109.4V and the rest of the chain only 0.6V. I guessed that R556 was open circuit. Having made that guess, I should have turned the set off and removed R556 before checking its resistance. But I tried to be smart and bypass all that rigmarole. Using my decade resistance box, I attempted to dab a 270kQ resistor across the suspect one. The only trouble was that I miscued and shorted the 110V rail to chassis. There was a small "pop" from the power supply board and the set stopped. The chopper transistor was circuit. short Transistor substitutes This transistor, Q903, is a 2SC1454 - a type which I didn't have in stock. But I do have many other power trnnsistors so I began a search for a suitable replacement. Although there are many thousands of transistor type numbers, there are only a thousand or so transistor types. So it stands to reason that many transistors are the same as, or very similar to, others with different type numbers. As a result, one of the most useful books in any service workshop is a transistor substitution manual. Or it would be if (1) it contained all type numbers and (2) it could be relied on for accuracy. Over the years,,I have collected a dozen different substitution manuals but only one has proved to be both comprehensive and accurate. Unfortunately, in the electronics field, every book ever published has been out of date before it leaves the press. As a result, we are forever buying new books and hoping that the latest one will be more accurate than the last. So it was that I turned to my 6-week-old copy of "Up-to-Date World's Transistor Comparison Table". This has 784 close printed pages and comes close to listing all transistor types ever made. The entry for 2SC1454 showed it to be a 300V 4A 50W NPN silicon transistor, intended for use as a power switch. The recommended alternatives were BU109, BUl 10, BU210, BUY21, BUY77 and 2N6306. As bad luck would have it, there were none of these in stock so I had to look further, to find substitutes for the substitutes. Those for the BU109 and 110 were not very helpful but when I came to the BU210 things looked better. The BU526 and BU626 were listed and these are higher voltage versions of the BU326 which I did have in stock. These are all listed as TV switches and this is the characteristic required for a chopper transistor. I have used the BU326 as a replacement chopper in many sets so I felt it would be satisfactory in this one. But you live and learn. The BU326 fired up all right and worked perfectly - for about two hours. During this time I went back to the vertical problem and found that R556 was indeed open circuit. Replacing this cured the vertical problems and it looked as though the set could go home next day. As a final check on the vertical scan, I switched the set over to SBS which I knew to be broadcasting their test pattern at that time. Just as the switch clicked over, so did the power supply and everything went dead. The BU326 was shorted. What gives? At first, I thought that the VHF/UHF switch-over had blown the chopper because both events happened simultaneously. Later, I came to the conclusion that the chopper must have been about to fail. The momentary loss of line sync as the bands changed might MARCH 1989 57 have created enough extra load to finish it off. But why did the BU326 fail? Rather than replace the chopper and risk another loss, I fired up the set using an external 110V supply. This showed that the set was running OK on VHF channels but that the UHF channel was showing only a white screen with a slight hiss from the speaker. This was turning into one of those jobs that goes from bad to worse. I began to wonder if I would ever get it finished. I could find nothing wrong with either the power supply or the line output stage. Apart from the UHF channel, everything seemed to be working properly. The only conclusion was that I should have used a BU426 (or better) and that the 326 just didn't have enough grunt for this set. Faced with the problem of finding yet another substitute for a substitute, I turned to "Towers International Transistor Selector". And against 2SC1454 I found two other recommended substitutes BDY94 and 2N5157. Looking further, I found that there were no substitutes for the BDY94 but a BU105 would substitute for a 2N5157. There was one significant difference between the characteristics given in the two books. While the "Up-to-Date World's Transistor Comparison Table" shows the 2SC1454 as merely a power switch, Towers rates the transistor as a "television line output extra high voltage". In fact, all the substitutes shown in Towers were labelled "extra high voltage". Clearly, this characteristic was one which the publishers considered to be particularly important. In the light of my experience, they were probably justified. What I wanted was a transistor of sufficient power handling capability, together with an extra-high voltage rating. At first I thought of using a 2SD350 line output transistor, which at 1500 volts can be considered to be extra high voltage, at least as far as a 1 lOV chopper supply is concerned. However, its power rating is only 22W and I needed something nearer 50W. otl)plf'q e<ti<; "be i • . ff l\ -. ou\N E:L-e:c-..-rRON\~ ~VE.R'-( :l3(:X)K E.VE:1<.. 'PlJe>L.\GM~P IS OU\ OF ~ ~~e: 1--r"' L-E.A\JES 11-\E'- ~SSooo 58 SILICON CHIP Fortunately, there is one common line transistor with these ratings, the 2SD380A, and I had several of these in stock. It took little time to fit one and soon the Sony was going like a charm - except that the UHF section still wasn't working. The possibilities My first thought was that the chopper failure had spiked the UHF tuner and knocked out the RF transistor. But a replacement tuner proved to be just as inactive as the first one. So if the tuner was OK and was receiving the proper rail voltage, what else could stop it from working? The only possibilities were the AFC and AGC voltages. Unfortunately, neither of these voltages is shown at the tuner on the schematic diagram. I had to work them out from figures given on other parts of the drawing. The AFC voltage was the lesser suspect because the same line goes to both the VHF and UHF tuners and the VHF tuner was working normally. This left the AGC line. In this set, there are no less than four amplifiers in the UHF AGC line. There are two, AGC amp 1 & 2 (Q781 & 782), on the UV board and two more, UIF AGC 1 & 2 (Q211 & Q212), on the S board. Also Q751, on the UIF board, is an AGC amp/buffer so there were plenty of locations for trouble. I injected an IF signal from a test tuner into the input of the UIF board and obtained a perfect picture. However, when the output of the Sony tuner was patched across to another TV set, there was no trace of any signal. not even snow. The tuner was completely inactive. Tracking back along the AGC line, I came to the moving arm of pot (VR201), designated UIF AGC, between the 18V rail and a tapping in the third video IF transformer, T208. From here the line goes to Q211, the first AGC amplifier. The base and emitter voltages of this transistor were correct but the collector voltage was way off. Q211 is directly coupled to Q212, the second AGC amp, and here I hit the jackpot. All three pins on Q212 were at the same voltage - give or take a few millivolts. It also allowed me to complete one of the most frustrating jobs I've had in a long while. The main point to emerge from the foregoing comedy of errors is the need to buy, keep and use every transistor comparison manual you can lay hands on. Although this story mentions only two volumes, I have in fact about a dozen books, all of which get occasional use. "Towers International Transistor Selector" has proved to be the most useful over a long period and it certainly gave the right information for this story. However, the new " Up-to-Date World's Transistor Comparison Table" is very comprehensive and if its overall accuracy proves to be better than it was in this story, then it too will become a reference in constant use in my workshop. And in future, I will check more than one reference before selecting a chopper or line output transistor. YOU NEED THIS T TEST I ENT SHORTED TURNS TESTER Including meter for checking EHT trans, yoke windings, drive trans. + $3.00 P&P HI-VOLTAGE PROBE $78.00 Including meter Reads positive or negative 0-50kV For TVs, microwave ovens and high voltage equipment $84.00 & $5.00 P&P TV TUNERS (PHILIPS) ECL 2060 Repair or exchange UV461 Repair only $17 .00 ea + $2.00 P&P CHEQUE, MONEY ORDER, BANKCARD OR MASTERCARD TO TUNERS 216 Canterbury Road Revesby 2212, Sydney, Australia Phone: (02) 77 4 1154 Back at the ranch l Fig.3: relevant section of the S board showing the two AGC amplifiers, Q211 and Q212. AGC voltage from T208 (not shown) is applied to the base of Q211 and taken off the emitter of Q212. Yet when I checked this transistor with a multimeter, all the junctions appeared to be perfect. They each had the normal forward voltage drop and no sign of leakage when reverse biased. So I pulled it out for a more accurate check. On my transistor tester, the thing showed its true colours. Although there was still no sign of leakage, when tested for gain the meter slammed against the right hand stop, as though the transistor had an extremely high beta . It was just as though it was short circuited. In fact, it was behaving more like an SCR, going short circuit as soon as any bias appeared on the base. And being in a DC circuit, once shorted it stayed shorted. A new 2SC1364 restored the correct AGC level to the UHF tuner and the whole set to normal operation. Well, having digested J. L.'s trauma, I think something a little lighter is called for. This story really started with one of my own sets; a General Electric model GE482 which I acquired several years ago and which serves as a very useful second set. The GE482 is about 7 years old and apart from the fault I am about to relate has been virtually trouble free. But before delving into the story, I should point out that this chassis has appeared under at least three brands: the GE482 as above, the Rank model C2020 and the General model GC205. So the fault and its cure could apply to any one of these. The fault showed up some time ago but it was intermittent and occurred only rarely. It took the form of loss of blue or, at other times, an excess of blue. Like the cobbler whose children were the last to wear shoes, I did nothing about it for a long while. There were a number of reasons for the duck-shoving. For one thing, it seemed like a potential nasty and it's bad enough having to solve nasties when one is being paid for it let alone doing it for nothing. Then there was the intermittent nature of D.DAUNER ELECTRONIC COMP ENTS WE STOCK A WIDE RANGE OF ELECTRONIC PARTS for • Development • Repair • Radio Amateur • Industrial Electronic • Analog and Digital WHILE STOCKS LAST CFE 455 RESONATOR $2.80 ea • 7552 COAXIAL CABLE (4 SHIELDS) CG7 $0.90/MTR • 50 ONLY MET AL PROJECT INSTRUMENT CASES 370mm(L) x 90mm(H) x 180mm (D) $4.50 ea • Come and see. Showroom: 51 Georges Crescent, Georges Hall, NSW 2198 (Behind Caltex Service Station In Blrdwood Read) Phone 724 982 TRADING HOURS: Monday to Friday 9.00 a.m. to 4.00 p.m. Saturday from 9.00 a.m. to 12.00 noon. MARCH 1989 59 SERVICEMAN'S LOG the fault plus its rare appearances. Of course, Mrs Serviceman made jibes from time to time but I could always find a good excuse. In fact, I did pull the back off the set on a couple of occasions and prodded around the red, green and blue drives, looking for possible dry joints or intermittent components but without success. So I more or less gave up. Then several months ago, one of my customers brought in the Rank version of this set and complained about exactly the same fault. I ran that set for about three weeks and it displayed the fault only once during that time and then so briefly that I had no chance to get to grips with it. Finally, I was forced to return it to him with the advice to keep it until the fault developed to a more predictable condition. And did Murphy have a field day on that one. The customer was back in a couple of days with the news· that the set had "gone blue" almost as soon as he took it home and had been like that almost continuously ever since. Yet when I set it up in the workshop it worked perfectly for several more weeks - not even a hint of trouble. So once again, all I could do was return the set with suitable explanations and advice. Having heard nothing since I assumed that the fault had not returned or if it had, only briefly. I know the customer well enough to know that he would be back immediately if he felt that the set's behaviour justified it (more about this later). Another Rank C2020 And so the problem lay dormant until a few weeks ago when a different customer fronted up with a Rank C2020 having the same fault. (/ " .-v~ ~~­ ,i\_~il~ -----------=-~.~ /2-- ~ 60 SILICON CHIP (;;# Once again, it showed a marked disinclination to misbehave when the customer tried to demonstrate the fault. By now, with three examples of the same fault hanging over my head, I realised that I had to make a real effort to come to grips with it. The best I could hope for was that this latest chassis might respond more readily to prodding, tapping, freezing, etc than had its predecessors and thus provide a clue. Neck board checks In this set the red, green and blue drive transistors are on the neck board (PWC677). I went over this board with a fine tooth comb and tried every trick I knew to bring on the fault. It was all a wasted effort; the set never missed a beat. Well, perhaps it wasn't wasted. Something, probably the sheer frustration of the problem, triggered a long dormant memory concerning a problem with this chassis when it first appeared. At that time the problem was more drastic; intermittent loss of all chroma. This problem was eventually traced to the chroma processing IC, IC701, a uPC1365C . More precisely, it was the socket used to mount the IC that caused all the trouble. The first time I struck the fault I assumed it was faulty contacts and went through the ritual of spraying both the IC and the socket contacts with a contact cleaner pressure pack. This worked for a while, then the set bounced. This time I changed the IC and that worked too, for a while. Then it bounced again. By this time I'd heard rumours of the same fault in other sets and a recommendation - I think by the manufacturer - to remove the socket and solder the IC directly to the board. And that fixed it. So was this a variation of that old gremlin'? It seemed that there was a good chance that it was but in any case, there was nothing to lose and everything to gain by dispensing with the potentially troublesome socket. It is a 28-pin device, so I used up a few centimetres of solder braid getting it out. I then I fitted the IC directly on the board. That was sim- ple enough but I realised I had made a minor goof. Mounted on the copper side of the board was a small capacitor with one end connected to the earthy copper pattern and the other end to one of the IC pins. But which one? In my haste to remove the socket I had neglected to note this vital point and the situation was not helped by the fact that the markings on the capacitor were no longer readable. I probably could have worked it out from the circuit but the situation was enough to prompt me to bring my own set into the workshop the next morning. Not only would it clarify this point beyond doubt but it would be a good opportunity to modify my own set at the same time. Well, it all worked out fine. The capacitor turned out to be a .OlµF unit which connects to pin 10 and modifying my own set presented no problems. Both sets fired up at switch-on but significantly, both were in obvious need of grey scaling. My own set, in particular, was now excessively blue. So that was attended to after which I had two very nice sets running on the bench. But had I cured the fault? Only time would answer that question. I ran the customer's set for several days with no sign of trouble, which was encouraging but not conclusive. And of course, I gave my own set a thorough workout both on the bench and then at home. It performed perfectly also and thus encouraged, I gave the customer back his set with strict instructions to contact me if there was any sign of trouble. That was several weeks ago as I write and all is well so far. So here's hoping. What about the first customer? And what about that first Rank set that Murphy had had so much fun with? Quite frankly, I was puzzled that I had not heard from the customer. I felt sure he would have contacted me in the event of trouble but at the same time, I found it hard to imagine that the set had run for nearly six months without trouble. So I rang him. And would you believe it - the set had not missed t -RE:AL.\5£.D t At> MAt>E. A MINOR GOOF9.... TETIA TV TIPS AW A-Thorn N Chassis Symptom: Small picture. 118V rail down to about 90V . R907 (2200 9 watt) spring resistor may open, cutting off the set altogether. Cure: L901, a 1 OµH miniature inductor, goes open circuit. This inductor feeds drive to the series regulator and, when it goes open circuit, the regulator is cut off. The set may still operate on current fed through R907 but this will get very hot and may spring open as a result. GEC 2213-A Symptom: Loud squealing from power supply. Squeal changes pitch or loudness as the load on the supply changes but even disconnecting the set entirely doesn't stop the noise. Cure: C510 (1 µF). C512 (1 OµF) a beat since he took it home all those months ago. Among other things, it just shows how easily one can get caught. Had I made any attempt to cure the fault, I would have blissfully imagined that I had fixed it. In fact, I know that it is still there waiting to happen. and/or C521 (4 .7µF) gone low in value. The first two capacitors modify the drive waveform to the chopper transistor while the third one is in the feedback network to the line oscillator in IC501. (All three were found faulty in one set and the noise from the chopper was deafening although the set would still work). Sanyo CTP5601, 6602 etc. Symptom: Very weak sound but no apparent distortion. A signal tracer will show sound to be normal at test point 1 G, the output of the sound IF chip , but not at the volume control. Cure: C178, a 1µF 1 OV electro open circuit. This capacitor and C 1 7 7 serve only to keep DC out of the volume control but failure of either will cause loss of sound. So I have suggested that he bring the set in immediately it gives trouble - or sooner if he likes - and I will modify it. But I doubt whether he will do anything until it does misbehave; I'm sure he's convinced that I've fixed it. Some people are funny like that. ~ MARCH 1989 61