Fullscreen HTML5Med Res (??MB)HTML4

When the chips are down I didn't have much trouble deciding which story to relate this month. It was a case of Murphy at his best. Start at one end of the circuit and it will be at the other end; start in the middle and you'll move towards the wrong end. Either way, you can't win. The story concerns an AW AMitsubishi 63cm colour set, model 6346, one of several models based on the ML series chassis. This is a very recent model and, in fact, the particular set was still under warranty, the owner having purchased it three months earlier. He was a new customer, having recently moved into the district, and apparently had been directed to me as the nearest serviceman authorised to provide warranty service for AW A sets. The first contact was by phone, initially to confirm that I could provide the service and then to find out what was involved. I confirmed that I could provide the service, but explained that it would be his responsibility to bring the set to me. Warranty agreements do not cover •house calls, or the cost of transport to or from the service department. The · customer accepted this situation philosophically enough: "No problem - I'll borrer me mate's ute." Those points clarified I made some attempt to determine the nature of the fault; loss of picture, loss of colour, loss of sound or what? "Aw no - the picture's gone all funny; kind of collapsed, if y'know what I mean." I wasn't sure whether I did or not. The best guess was frame collapse, although the implication was that there was still some kind of a picture to be seen. Perhaps it was only a partial collapse. There was 56 SILICON CHIP obviously no point in probing further; I would just have to wait and see. And so it was that the "mate's ute" eventually turned up with the set on board, and we carried it into the workshop. I took the opportunity to connect an antenna and switch it on, just in case there were any points to clarify while the owner was still there. I suppose I can hardly blame the owner for his "all funny" description. The first 75mm at the top of the screen was blank, the next 100mm showed a reasonably linear picture, from there to within 50mm or so of the bottom was grossly expanded, and the rest of the picture compressed into that last 50mm. It was, literally, not a pretty picture. Other commitments prevented me from delving any deeper at that stage and I put the set aside for a day or so. When I did get around to it I fished out the circuit and concentrated on the vertical section (Fig.1 ). Special Notice These notes are being contributed by the author who, from 1950 until July this year, wrote 'The Serviceman" in another magazine . . We feel sure that regular readers of that series will welcome the opportunity to continue following his adventures in SILICON CHIP. LP..Si KNOWN f>\CT1JR.€: OF \V\UR.?l-1'-i, ASou,- . TO BE: HI, B\.f A FALLING Sf\rE. Wl·\ll-'ST D'ROWN\t-JG l N A FL.ASH FL.00t> O~ AC It> RAIN ON '11'\ NV\..\...A.!!.01< t>L.A.INu •.. e:. Jungle IC In fact, there wasn't a great deal to be seen. Most of the functions were performed inside a 48-pin jungle type chip (IC201, MC1310AP), including the vertical sync separator, vertical oscillator, ramp generator, and an amplifier which delivered a signal at pin 16, which was applied directly to a vertical output pair, Q451 and 452. And that seemed to be all there was to it. I concentrated on the output stage for a start. I checked the various minor components, including two diodes, D451 and 452, but found nothing suspicious. Similarly, a voltage check around the two transistors produced figures very close to those shown on the circuit. The next logical step seemed to be to get the CRO going and check appropriata waveforms. The circuit shows two vertical waveforms, one at the vertical oscillator pin of the IC (pin 20, waveform 8), and one at the vertical output pin (pin 16, waveform 7). I checked waveform 8 first, and this came up virtually identical with the photograph on the circuit. But waveform 7 was another matter. Even the photograph on the circuit suggests that there is some "mush" on the pattern and this appeared to be even more so on my CRO. But this aside, the shape was nothing like that on the circuit, although the amplitude was approximately the same. So what was happening? Was the chip at fault and delivering a faulty waveform to the output stage, or was the output stage at fault and somehow loading the chip and distorting the waveform? Since I had already checked the output stage fairly thoroughly I hesitated to blame it. On the other hand, as I have remarked before, I am less inclined these days to suspect an IC until all other possibilities have seemingly been exhausted. And there could still be something funny about one or other of those output transistors, which didn't show up on my voltage check. The transistors are each mounted on a U-shaped aluminium heatsink, the "U" being inverted so that the heatsink sils above the main board. The ends of the vertical sections are cut to provide narrow tongues which fit into slots in the board, then bent slightly to hold them in place. It was necessary to remove the heatsinks from the board in order to gain access to the transistors, but this was not particularly difficult. Unfortunately, the effort seemed to be wasted. Both transistors tested OK, as did the two diodes, which I tested a second time while the transistors were out of circuit. I also double checked the minor components (resistors and capacitors), and the height and linearity controls, until I finally convinced myself that there was nothing in that part of the circuit which I hadn't cleared. Which seemed to put suspicion squarely back on the chip. But still I hesitated, trying to think of anything I might have overlooked. I drew a blank and was rapidly reaching the conclusion that the chip would have to be replaced, if only to prove the point one way or the other. After all, I could easily spend several hours searching vainly for some external fault, which may not actually exist, and finally be forced to change the chip anyway. On the other hand, if I changed the chip now, that point would be settled once and for all. Granted, I'd be down the cost of a chip if I was wrong (but one chip up in my stock) and down by the time needed to change the chip. Now I don't pretend to like changing chips, particularly 48-pin monsters - I don't suppose anybody does - but I've developed a pretty good routine and a fair amount of skill over the years, and can usually do the job in about 15 minutes. That's no record, and I know some blokes who can do it quicker, but it's time well spent to prove a point. Sydney or the bush A more practical snag was that I didn't have such a chip in stock, so one had to be ordered, which at least gave me time to think while I was awaiting delivery. In fact, this didn't help much; I had not thought of any other possibility by the time the IC arrived so, muttering something about "Sydney or the bush", I set to with solderwick, solder sucker and a good hot iron, and pulled out the suspect chip. The whole operation went smoothly enough and I subsequently fitted the new IC, tidied everything up, and switched on hopefully. I don't suppose I need spell it out; it wouldn't have been worth writing about if it had been as easy as all that. Suffice it to say that the set behaved exactly as before. I was back to square one. So what now? Up to this pointJ had convinced myself that I had thought of everything before I changed the chip. Quite obviously I hadn't, but I was at a loss to think of some new line of attack. I went over the circuit again. I had proved, the hard way, that the chip was not at fault. I was also convinced that there were no faulty components in the output stage. So what did that leave? It was more of a growing suspicion than a sudden inspiration but I found myself thinking more and more about the scan coils and any minor associated components. Faults in scan coils are extremely rare, but I have seen the effect of a shorted turn, and it does some weird things to the picture. In any case, there was not much else left to suspect. Tracing out the circuit from the vertical output stage to the scan coils revealed the existence of several auxiliary components, apart from the scan coils, and I ..,..,.., .... JA NUA RY 1988 57 ,RHJb , Ill 15K , IW !Ha) nsv Deflection Yoke ' C9 11 33µ 160 V 12 V 1 4 02 3 < 1 V "~ Y3 2 H Fig.1 portion of the AWA 6346 circuit showing the scan coils and associated circuitry. The pincushion correction circuit is on PCB-PCC. made a mental note that these would have to be checked. But first the coils themselves. I fished out a shorted turns tester that I had not had occasion to use for a long time, unplugged the scan coils, and checked for shorted turns. Result: a clean bill of health; there was nothing wrong there. That left only a few auxiliary components, the main ones being on a small printed circuit board bolted to the side of the cabinet and connected to the scan coil circuit via flying leads and plugs and sockets. The relevant portion of the circuit is reproduced here to assist readers to follow the story. The board is designated PCB-PCC, the "PCC" standing for Pin Cushion Correction. It consists of a transformer, T571, resistors R575 and 576, and capacitor C575. Connection to the scan coils is via two plug and socket sets; PV (Plug Vertical) and PH (Plug Horizontal). The vertical scan coils are connected to the vertical output stage via pin 4 on the deflection yoke diagram and eventually find their way back to chassis from pin 3, via the pincushion network, and a pair of 8.20 resistors in parallel between "PV1" and chassis. These resistors appear to be part of a feedback network, and are not shown here by reason of their remote location. 58 SILICON CHIP I had left the set running while I visually checked out this section and related it to the circuit, which meant that the set had been running for about 15 minutes, much longer than I had run it so far, since I had previously turned it on only long enough to check a few voltages or whether the replacement chip had achieved anything. In any case, I had not been observing the circuitry associated with the scan coils. But now my attention was drawn to R569, a 2200, 1W wirewound resistor, mounted on the main board, but effectively connected between pin 3 (vertical scan coil) and the 12V rail. I don't know what function it serves, but I noticed it for the simple reason that it was running stinking hot - literally. So hot, in fact, that it had melted the solder connecting its pigtails to the board. Well, it seemed that at last I was onto something, even if exactly what was not clear. Looking at the circuit again I suspected that, for some reason, the circuit from pin 3 to chassis was incomplete, forcing the scan current to seek a path through R569 and the 12V rail. And the most likely cause of this would be failure of the two 8.20 0.25W resistors (R462 and 464) already mentioned. So these were located and check- ed - only to draw a blank; they were spot on value and completely free from any signs of stress. So much for that theory. But I still had the idea that there was a fault somewhere in this chassis return path and a check with the ohmmeter from pin 3 of the scan coil to chassis confirmed that there was a resistance of several hundred ohms; much higher than seemed logical. So began a rather laborious process of checking individual connections involving the printed board PCC and the associated plugs, sockets and leads. In detail, PV is a female plug, on a flying lead, which mates with male pins, 1 and 2, on the board. The flying leads are crimped into the plug contacts and I suspected that there could be trouble here. Checking this wasn't easy because of the difficulty of ensuring that one was making an effective connection with the contact inside the plug. But, after several attempts, I finally gave them a tentative all clear. But what about the board itself, and particularly the pins. The pins are hollow and are mounted by inserting them in a hole in the board, then expanding them with a flaring tool to make a moderately strong mechanical joint. Then, when the board goes through the solder bath, the flared end of the pin is soldered to the surrounding copper pattern. I pulled the board out and examined it closely, using a magnifying glass. As far as I could see, all four pins were soldered perfectly to the copper pattern. But the ohmmeter told a different story. Measuring from pin 1 of PCC to the surrounding copper pattern showed virtually zero resistance, but not so pin 2. Here there was a varying resistance of around several hundred ohms. Significantly, even though I knew there was a fault there, I could not pick it visually. But a few moments work with a hot iron, and a somewhat lavish application of solder and flux, removed any doubt. In fact, I treated all four pins, because I could no longer trust a visual check. TETIA CORNER Blaupunkt (Bridge Rectifier Chassis) Symptom: Fuse S1242 blows repeatedly. If the fuse is replaced often enough, the set will eventually fail completely with a clattering noise coming from the chassis. Cure: D1245 (TAG 3-400) SCA shorted or breaking down under load. An emergency repair can be effected by removing diodes D1242 and D1 244 from the bridge board. This reverts the set to its original half-wave rectifier design, in which state it seems to run quite happily. This information supplied by The Electronics Technicians' Institute of Australia (Tasmanian branch). Relief and frustration Then I switched the set on and up came a perfect picture. I viewed the end result with mixed feelings; relief that I had finally cured the fault, but also a certain amount of anger at the frustration I had experienced, and the time I had wasted, all because of one miserable dry joint. Only someone else who has been through it all will know how I felt. Still, that's what the game is all about; one has to take the rough with the smooth and, in the ultimate, measure success by the end result. On that basis this job had been a complete success, even if it had been less than satisfactory financially. And I am still puzzled as to how that joint could look so good yet be so poor. ~tb.2. ANl:> 46~ FR.ON\ AN'-f 5\G,~S Of course, I had destroyed any evidence in proving the point, and there wasn't much help for that. As I saw the joint, the solder had flowed into the hollow pin and right out to, and over, the edge of the flared end. It had also flowed quite freely over the copper pattern which ran under the flared section, and it looked as though these two runs of solder had mated into one. Quite obviously they hadn't but, short of using an electron microscope that looks around corners, I am at a loss to suggest how it might have been better checked. The only real answer, I suppose, is if in doubt, resolder it. Which I did of course, but the "doubt" was a long time a-coming. And now, for a change of pace, c..oMr>LE:'1""<=-L.'-f r~ee- Or s~~ss.. _. here is a story from one of my regular contributors, Mr J.L. of Tasmania. It is an intriguing story involving both technical problems and the ingenuity in solving them, and the frustration caused by the replacement parts problem. This is his story, for which he has nominated the following appropriate title: Sharp shooting One thing about television servicing that appeals to me is the uncertainty principle. When a dead set comes into the workshop I never know what killed it, and finding out involves a series of tests and guesses that will, hopefully, point to the faulty part. This was never more clearly shown than with a Sharp CX2020 that came in recently. Tracking down the fault took both practical and theoretical knowledge, plus experience, suitable test equipment and a pile of patience. The customer remarked that the set failed to start up when switched on at 7pm, although it had been working perfectly at 5pm. There had been no sparks or smoke, just a total refusal to show any sign of life. He tried the set on a known good power point, just in case, but it was totally dead. As is the way with so many customers, he felt sure that the picture tube was done for. He wasn't really convinced when I tried to explain that if it was the tube, he should still have sound. When I switched it on in the workshop, I heard the degaussing coils go "boing" which indicated that part of the set was working. That characteristic degaussing sound meant that the power lead, mains switch and mains fuses were all intact. I could tell this much without even taking the back off. But that's as far as I could go with the back on. Next it was inside to check the DC power supply. Specifically, there should be about 300V on the collector of the chopper transistor (Q701) and, in this set, 115V on its emitter. The 300V was there, but not the 115V. Quite obviously the chopper was not running. This set uses a self-oscillating JANUARY 1988 59 TTINO I - ® C635 10µ(16VJ I DX0048CE I V R50 3 I I ,-'II-------+--+----------+., 10 ·SIZE - ----, )Li~~ • ~:a:, • (Fl. R60I 27K R- 82K C l1,1( + C608 471,1 . 039 - (16V) (ML):; C609 SW60I R6l0 5.6K(3W) [F-"°__cJ_N.I) Fig.2: horizontal oscillator (IC501) and line output stage of the Sharp CX2020. Note the protective diode incorporated in Q602. chopper which should self-start, then come into sync with the line frequency as the line output stage gets going. The pulse from the line stage (Fig.2) is not essential to run the chopper but, as well as its synchronising function, it is used by the protection network to indicate the presence of faults on the line output. No line pulse may mean a fault and so the chopper is shut down. So where should I start to look for a fault in a roundabout network of this kind. My first line of attack is to determine if the chopper is trying to selfstart. I do this by looking at the emitter of the chopper transistor with the 'scope. If I'm lucky I will see a brief flick of the trace which indicates that the transistor is being turned on, if only for one or two cycles. The output voltage appearing on the emitter is immediately reflected to the protection circuit, and if all is not right, the base drive to the chopper is very promptly terminated. In this case, the flicker of the scope trace, when it did appear, was so brief as to be almost unnoticeable. Still, it was enough to say that the chopper was trying, and at this point I had to make an 60 SILICON CHIP educated guess as to where the troubl-e lay. It could be either a fault in the supply itself, or a fault in the set proper. It's a "heads or tails" situation, but experience helps a little. Line output stage faults are more common than power supply faults. In fact, faulty line output transistors are so often the cause of stoppages that most servicemen check that component first of all. It's usually simple to reach the collector and measure its resistance to chassis. If the meter shows zero ohms, then the transistor is like the Christmas turkey. In this case there was no such indication and the transistor (Q602) checked out perfectly. Well, sort of perfectly, because the base-emitter junction is difficult to check in circuit, normally showing as a short through the line drive transformer. Still, if collectoremitter shows no sign of leakage then it's a safe bet that the base emitter junction is OK. Or so one might think. My next move was to connect the CRO to the base of the line output transistor to see if there was any drive reaching this point. Again, this was inconclusive because the trace merely flickered, without any evidence of a drive waveform. I transferred the probe to the collector of the transistor, hoping to see a similar flicker there. I already knew that the 115V rail appeared briefly, so if the line output transistor was being driven at all, there should be some signs on the collector side. But there wasn't a sausage. Unusual failure Up to this point I hadn't switched on the soldering iron, but the time had come and in a couple of minutes I had the 2SD869 line output transistor lying naked on the bench. And now I could really detect the cause of the trouble. The base emitter junction was as short a short as any short I have ever seen. It is an unusual failure for a power transistor because the basecollector junction seemed undamaged, and there was no trace of leakage between collector and emitter. So now I had the problem of replacing a 2SD869. This transistor is one of those odd animals with a built in protection diode, used in a number of recent vintage Japanese sets. Replacing these can be quite a VOOD FOR Cl-I/PS ... WOOD FOR Cl-I/PS ... WOOD FOR Cl-I/PS ... WOOD FOR CHIPS .. . WOOD FOR CHIPS ... WOOD FOR CHIPS:.. WOOD FOR C SCOOP PURCHASE ~~;~~ !~: (s ilicon Chips of LM7912CT LM7915CT LM79L05ACZ LM79L12Acz s1.oo $1.40 SO.BO so.90 course!) This month we're listing a few of our ~~:9;~~c:EFE~;~Es LM1 13H s19_10 LM313H s11.oo Doni Forget - it's still WOOD Geoff has managed to get hold of a limited CHIPS quantity of genuine PA PST fans. They're the 75mm type for 220V operation. While they last you can have one for just$25.95 IC SPECIALS 27C64 200nS 8kx 8 CMOS EPROM 12.5.VPP. These are stock items in the ~~~~~~ linear range. !;~~ LM3290Z s1.1s prime quality, not factory seconds or floor sweepings! Whatever your requirement give Geoff a call And they're only $5.99 each CD4503/80C97 Hex tor CMOS VOLTAGE REGULATORS AVAILABLE AT LAST PORTASOL PROFESSIONAL (it's the complete kit!) It's a gas knife .. . it's a soldering iron ... it's hot blow. And a blow torch ... it's a hot it comes in a neat carry kit complete with a bit wiper. No cords or batteries yet it gives the equivalent of a 1 O to 60W iron. You can get up to 90 minutes average continuous use from a single fill. And you refill it in seconds using a standard butane gas lighter refill. · Tip temperatures as high as 400°C can be set. The kit includes one soldering tip, a hot kntte, blow torch and hot blow. The cap contains a flint lighter. The complete kit comes in a handy case (with stand for the iron) which just about fits Porta-Sol Professional is $81.00. PORTASOL STANDARD SOLDERING IRON in your pocket. Geoff has sold hundreds of 'em to servicemen and technicians. Complete and ready to use like the Professional but you only get the iron and bit $39.95. P.ORTASOL TIPS Expand the capability of your Portasol Iron with spare tips available for standard iron in 1 mm, 2.4mm, 3.2mm, 4.8mm and hot knife tip. Professional tips come in same sizes plus hot blow and blow torch. Tips are JlQl interchangeable between irons, so specify Standard or Professional when ordering. Tips are all $12.95 each. IBM and Compatible PC Users! Save a power point - Get a rewireable IEC plug from Geoff. It's so easy -on the back of your PC you'll fi_n d an IEC outlet which is controlled by the power switch on the computer. So cbop the mains plug off your monitor and connect the L2298 and hey presto you 'll never forget to switch your monitor off again! Qu_ality Belling Lee (Geoff has a full range of IEC connectors) Ask for !e t Q: 0 II.. Cl 0 i LM304H $6.05 LM305AH $5.85 LM3051-1 $1.75 LM309H $5.45 LM309K-STEEL $5.90 LM317H $7.30 LM317HVH $10.85 LM317HVK-STEEL $10.85 LM317K-STEEL $6.60 LM3 17KC $5.20 LM317MP $1.85 LM317T $1.05 LM320H-5.0 $9.65 LM320K-12 $7.70 LM320K-15 $7.70 LM320K-5.0 $7.70 LM320MP-12 $3.25 LM320MP-15 $3.25 LM320MP-5.0 $3.25 LM320T-12 $3.70 LM323K-STEEL $5.50 LM325N $10.75 LM326H $11.75 LM326N $10.75 LM330T-5.0 $1.75 LM333K-STEEL $14.35 LM333T $6. 70 LM337H $7.95 LM337HVH $13.90 LM337HVK-STEEL $17.45 LM337K -STEEL $9.15 LM337LZ $1.40 LM337T $2.55 LM338K-STEEL $13.25 LM340K -15 $5.90 LM340KC-12 $2.30 LM340KC-15 $2.50 LM340KC-5.0 $2.25 LM340T-12 SO.BO LM340T-15 S0.75 LM340T-5.0 SO.BO LM341P-12 $1.35 LM341P-15 $1.35 $1 .35 LM341P-5.0 LM342P-12 $1.20 LM342P -15 $1.20 LM342P-5.0 $1.20 LM350K-STEEL $9.65 LM376N S0.90 LM396K-STEEL $31.90 LM723CN SO.BO LM2925T $4.60 $1.75 LM29:l>T-8.0 LM2931CT $3.05 LM2935T $4.80 LM2940CT-5.0 $2.80 LM3524N $3.65 LM76601N $3.10 LM78L12ACH $2.05 LM7BL12ACZ S0.60 LM78L15ACZ so.so LM7905CK $4.15 LM7912CK $2.55 ~~~~~- 2 5 LM336Z-5.o LM368H-5.o LM385Z-1.2 LM385Z-2.5 LM399H LM3999Z LM349N :~: s1 .90 s 11 .90 $2.40 $2.40 $6.15 $4.75 $16.85 $1.85 $1.20 $2.65 $2.80 S0.60 $7.95 $7.95 $4.80 $1 .25 $1.05 $1.55 $2.40 $1 .20 AUDIO AMPLIFIERS LM380N LM380N-B LM381AN LM38 1N LM382N LM383A T LM384N LM386N-1 LM387N LM388N-1 LM389N LM390N LM1875T $1.90 $1.90 $6.40 $3.95 $3.05 $4.'30 $3.50 $1.65 $2.80 $2.40 $2.40 $1.95 $8.00 INSTRUMElfTA nON AMPLIFIERS LM3630 LM363H-10 LM363H-500 LM725CH $38.00 $24.00 $24.00 $9.20 OPERAnONAL AMPLIFIERS LM10CLH LM10CLN LM10CN LM1 1CH LM11CLH LM11CLN LM11CN LM301AH LM301AN LM307H LM308AH LM308AN LM308H LM308N LM310H LM310N LM312H LM316H LM318N LM321AH LM321H LM324AN LM324N LM344H LM346N LM348N $8.30 $5.95 $9.30 $9.75 $6.70 $3.30 $3.50 $1.30 S0.70 $1.95 $6.95 $5.10 $2.45 $1.00 $4.20 $4.25 $6.00 $11.20 $2.20 $28.05 $10.15 $4.70 SO.BO $11.20 $4.75 $1.90 $2.20 SO.SO $3.10 $1.25 $2.45 $1.05 $0.80 $0.50 $0.30 $1.60 $1.10 $1.90 $2.80 $0.55 $1.40 $0.90 $2.30 $2.25 $2.25 $2.25 TRANSISTOR ARRAYS VOLTAGE COMPARATORS LM306H LM311H LM311N LM319N LM339AN LM339N LM360N LM360N- 14 LM361 N LM392N LM393N LM710CH LM1414N LM3302N LM358N LM359N LM709CN LM733CH LM733CN LM74 1CH LM 741CN LM741CN-SGS LM747CH LM747CN LM833N LM1458H LM1458N LM:l>BON LM3900N LM4250CN LM13080N LM13600N LM13700N LM394CH LM394CN LM394H LM395T LM3046N LM:ll86N LM3146N $5.40 $5.40 $7.40 $3.85 $1 .50 $1.15 $2.15 TEMPERATURE SENSORS LM3351-1 LM335Z LM35CAH LM35CH LM 35DZ LM3911H-46 LM3911N $3.15 $2.40 $12.95 $13.95 $2.60 $5.05 $2.65 SPEOAL FUNCn0N BLOCKS LM331AN LM331H LM331N LMC669CCN LM1812N LM18:l>N LM1889M LM1893N LM2907N LM2907N-8 LM291 7N LM2917N -8 LM 3909N LM3915N LM3916N $9.85 $13.20 $7.60 $11 .05 $6.20 $4 .40 $5.60 $19.85 $2.35 $4.25 $4.40 $4.20 $ 1.85 $4 .25 $4 .65 COMMI.INICA nONS ORCUITS LM565CH LM567CN LMC567CN LM 1496H LM1496N LM 1886N LM:ll89N LM3189N LM3820N $5.25 $1 .40 $2.35 $5.05 $2.05 $8.05 $3.95 $5.15 $3.20 nMERS LM322N LM555CN LMCSSSCN LM556CN LM3905N $3.05 S0.50 $1.00 $1 .10 $2.35 EXAR PRODUCTS XR-2201 CP XR-2202CP XR-2200CP XR-2204CP XR-2206CP XR-2209CP XR -2211CP XR-2240CP XR-2243CP XR-5533AP XR -5534ACP XR-558CP XR -8038ACP $1 .65 $1 .65 $1.65 $1 .65 $9.40 $5.20 $7.50 $3.45 $4.20 $4.15 $3.90 $3.30 $7.30 L2298 at only $4.75 8.30 to 5 Monday to Friday, 8.30 to 12 Sat. Mail Orders add $5.00 to cover postal charges. GEOFF WOOD ELECTRONICS P/L All prices INCLUDE sales tax. (02) 427 16 76 Tax exemption certificates accepted if line value exceeds $10.00. p~~==~[ ;:::: e:::====; 229 BURNS BAY RD. (CORNER BEATRICE ST.) J~" ; BANKCARD, MASTERCARD, VISA , CHEQUES LANE COVE WEST N.SW. IN C IN NSW TWX71996 P.O. BOX 671 LANE COVE N.SW. 2066 8RI~ OR CASH CHEERFULLY ACCEPTED specialising in electronic components for the professional and hobbyist. pain as each manufacturer seems to have his own design and none seem to be compatible. Here in Tasmania we have a real spare parts problem as all manufacturers have withdrawn their spare parts services to centres in Melbourne or Sydney. So our Tasmanian customers are now faced with 10 to 20 day delays while we order the part, then wait for the invoice, then send the cheque, then finally get the part. (I know - some firms do have faster services, but most do not and our gripe is with the latter ones). The only alternative for independent servicemen like myself is to cultivate our friendships with the manufacturers' service agents. If they have what I require, and my face is welcome in their workshops, then I might be able to get the parts needed to put my customers' sets back into working order in a reasonable time. So my immediate need for a 2SD869 led me to the local Sharp agent. When I asked his receptionist about the availability of 2SD869's, a voice from deep within the workshop declaimed "Yer don't need wunna them! A 2SD350 is OK in that set." The voice was soon joined to the jovial face of their technician who volunteered the information that the 2SD869 was only a 1000V transistor, and the diode was included to catch any spikes over that level. On the other hand, a 2SD350 was a 1500V device and to his knowledge, none had ever failed in these sets. He wouldn't tell me the price of the 2SD869, but as the 2SD350 is about as cheap as any power transistor can be, I decided then and there to use one in this repair. So I soon had a 2SD350 in the chassis, in place of the original transistor. Pressing the power switch produced not the expected burst of sound, but quite the opposite - absolutely nothing. This was a real disappointment because I was convinced that the transistor had been the only faulty part in the set. So I had to start troubleshooting all over again. This time the CRO on the chopper base showed a quarter second burst of drive waveform. The 115V rail also showed a change 62 SILICON CHIP - the 'scope trace at the emitter leapt up the screen, then quickly dropped back to the zero line. This was a much more positive response than the earlier flicker, but it was no closer to restoring full operation. Continuing my investigations, I found that the line output transistor now showed more enthusiasm at switch-on, but couldn't be resolved into actual drive. It just flickered into some kind of action, them died. So what now? It could be (a) a faulty line oscillator chip (IC501), (b) a dud line driver transistor (Q601), (c) a bad line driver transformer (T601), etc, etc. With a chain of doubts like this, it is perhaps best to start at the beginning. My 'scope was not able to resolve any trace of line drive from the oscillator chip at switch on, but then a quarter of a second or less is not very long to resolve anything. What I had to do was power up the chip and see if it produced the right waveform. I used a 9V battery fitted with leads and alligator clips to apply Vee to the chip. The set uses a 12V rail at this point, but 9V is enough to see if things will work at all. In this case, 9V produced a solid train of square waves and proved the chip to be 100 % . While the battery was connected I was able to trace the signal up to the base of the line driver transistor. They went no further because the collector is powered from the 115V rail and this wasn't working yet. So I had to devise a way to get voltage onto this rail. One of the most useful pieces of equipment in my workshop is a Variac, a continuously variable autotransformer. I have also built up a simple DC power supply using an old TV transformer, a couple of diodes and a 350V electrolytic capacitor. When fed from the Variac, this supply can deliver from about 20V up to something over 200V. (It's rather crude, but I haven't had either the time or the money to build a better high voltage power supply). For this Sharp job, I connected the DC supply to the 115V rail and slowly cranked up the Variac. The 9V battery was still connected, and the IC was delivering drive pulses to the line driver transistor. The voltmeter monitoring the 115V rail at first indicated a rising voltage, but it stabilised at about 5V and went no higher. Even 100V AC into the DC supply could produce no more than 5V out, and the Variac was humming ominously. The story is nearly finished! I took no time at all to find that the 115V rail was shorted almost down to chassis. Examination of the schematic showed two components as the likely culprits. One was C713, a l00µF 160V electrolytic capacitor, and the other was ZD702, a zener diode , type EX0074CE, connected between the rail and chassis. Roast dinner The more I thought about that zener the more dubious it looked. Then in the parts list I found that it was a 130V type, twice as high a voltage as any zener I've ever seen. I'd bet a dud fuse to a roast dinner that it was the villain. And so it was. Not quite a dead short, about 10 ohms, but near enough to stop the set dead in its tracks. The zener is only in the set as a protection device and the set works quite happily without it. But it is there for safety reasons and had to be replaced. So now came the second spare parts trauma with this set - where do you get a 130V zener if not from the set's maker? Fortunately, the zener value is not critical and I was able to replace the faulty item with two 62V zeners in series. The set is now, if anything, a little safer than it was. Getting back to the philosophy expressed at the beginning of this story, I find television servicing an occupation that is intellectually stimulating and full of interest. It is never boring and the smile on the face 'of the customer when he learns that it wasn't the picture tube after all is all the job satisfaction that I could ask for. Alternative transistors Thank you J.L. for a most interesting story and an insight into the problems faced by our colleagues in more remote areas. I have no doubt that there may be those who would question the wisdom of substituting the unprotected 2SD350 in place of the original 2SD869, in spite of the higher rating of the former . In fact, when I showed this story to a colleague he immediately rattled of a list of protected output transistors, which he felt could at least be considered. These included the 2SD870, 871, 899, 900, 951, 952, 953 and 954. However, neither he nor I would be prepared to stick our necks out and claim that all of these would be compatible with the Sharp circuit. (Incidentally, most transistors of this general type, with built in protection diodes, also have an in-built resistor between base and emitter. This is not always shown on the circuit symbol and could be mistaken for leakage. Typical value is about 400. More to the point, of course, is the essentially practical problem faced by J.L. and others who live in places remote from the manufacturers ' distributors. And even then the expression "on back order" crops up all too frequently. Obviously, all is not cider and skittles in the Apple Isle, at least as far as spare parts are concerned. This being so, people like J.L. have to do the best they can with what is available. And at least the substitution was recommended by the manufacturer's agent who, in turn, appeared to be basing the recommendation on practical experience. And I'm sure that no one would have been happier than J.L. had he been able to fit the correct replacement but, if it's not available and a long wait is involved what should one do? What would you do? More to the point, perhaps, what would I do? See you next month? it l F ,-r 'S NOT A\JA\l.A~l.£, W~Prt' 8~0ULt> ONE 'OO? W~AT WOUl.t> '<OU 'Do? VJHAT \NOUl.t> ~EV t>O? '<~, W\o\Ai ~HOUt..t> WE t>o? 8. ~110:.~ >WAAT MlGHT' SHE. oo? ALll\OVGH, WMAT WOUL.t) IT 'Do? Su, Tl-\EN W~AT WOUl...t> Tu.JO OR MOR6 l)O~~E 10 ME POINT, W'°'~••. ~ ~-=? )) JANUARY 1988 63