Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN'S LOG I don’t like house calls I don’t like making house calls but sometimes they are inevitable. If it’s a large TV set, one usually doesn’t have much choice but if the patient is an old 286 computer, the cost of the call doesn’t make much sense. Some weeks ago, I was called to a house to attend to a Sharp SX-68A7 stereo TV. I was reluctant to go out at the time for three reasons: (1) I was unfamiliar with the set; (2) it had an intermittent sound fault; and (3) someone else had already had a look at it. However, as the owners were semi-retired and didn’t have a car, it was going to be too difficult for them to bring it into the workshop. Besides, they did ask nicely and really I am a great big softy at heart! When I arrived, Mrs Jones made some tea while I extricated the monster from its dark hole in the “entertainment” cabinet. This set is initially a little confusing to operate, even with the instruction book (I often think they ought to have an in­struction book for the instruction book). There are four LED displays on the front, the first three being marked MONO, S/VIDEO and SURROUND. The fourth is unmarked but apparently is the power ON indicator. When I turned it on, only the first and last LEDs were illuminated. However, no matter how hard I tried, I couldn’t get any sound. There was only noise from the loudspeakers, even when the volume was turned fully up. I put a small screwdriver in the RCA audio input socket at the rear, selected AV (audio-visual) and was rewarded with a buzzing noise. Obviously, the output amplifi­er was OK and the fault had to be between that and the IF stage, as the picture was excellent. My guess was that the problem lay in either the muting, AV switching or stereo decoder circuits. I have had a few cases in the past where a TV set muted in the mono mode, so I decided to test this area first. Now most stereo TVs have LED displays that illuminate on stereo or bilingual broadcasts but this TV illumi­nated a LED for mono transmissions, which was rather confusing. Perhaps Sharp thought that as most broadcasts are in stereo, the consumer may need to know when reception was in mono rather than vice versa. In any event, the sound didn’t work in either mode. I removed the large plastic rear shell of the set and tried to get my bearings on the unfamiliar chassis. It is a large flat chassis with three vertically mounted modules on the righthand side. Two of these modules are encased in metal screening which is soldered to the motherboard. Without the service manual, it was very difficult to determine their functions and I couldn’t even find a marked control or test point that might give a clue. I tried tapping them gently with the butt end of a screwdriver but it made no difference. Mr Jones didn’t have much praise for the hapless serviceman who called previously because he “only made it worse”. Before he came he could at least occasionally get some sound and now he didn’t get anything at all. And besides, the previous serviceman “was far too keen to take it to the workshop”. Well, he wasn’t the only one. If it hadn’t been for the fact that it was a 68cm TV with lots of stairs between it and the van, I would have insisted that it go straight to my workshop. What’s more, it was beginning to look as though this was going to be the inevitable course of action. But first, there were a few more things to try. I tuned in to the VCR first but this made no difference. I then I got some AV leads and connected them between the VCR and the TV set. Mrs Jones, who was watching me like a hawk, thought that I was a genius when the sound miraculously appeared but I had to gently deflate her enthusiasm by informing her that it wasn’t really fixed. This would have to be a temporary arrangement while I obtained a service manual so that it could be fixed properly. Even then, there was a good chance that it would have to go the workshop. When the manual eventually came, I was able to work out that module PWB-E to the far right of the chassis was the stereo decoder. This circuit consisted of two ICs (IC351 & IC352) and a transistor (Q351). I enquired as to the trade cost of these parts which came to $55 including freight and tax. The question was, would it be better in the long run to get these parts in now or risk possibly yet another service call and/or a trip to the workshop (ie, would the cost of my labour exceed these parts if it was later found that either was faulty). Unfortunately, the chances of using these chips for another repair if they were bought and kept in stock would be remote. I decided to put the options to the Jones’ and let them make the choice. They decided on ordering in the parts immediate­ly which was just as well because they had to come from Japan. Eventually the parts arrived and, armed with the service manual and an audio probe (a battery powered amplifier to detect audio), I felt reasonably confident I could knock this one off in one hit. Unfortunately, gaining access to the underside of the main chassis PWBA June 1997  57 it rather difficult. As usual, the manufacturer had decided to save vast quantities of money by making sure that the interconnecting leads were as short as possible, thereby making it exceedingly difficult to get the chassis into some sort of serviceable position. This problem is compounded by the light­weight plastic cabinets and chassis used in modern sets – one has to be careful to ensure that the set doesn’t roll onto its face due to the weight of the tube and the remaining front half of the cabinet. Anyway, I finally managed to remove the screen covers from the stereo decoder module and unplug it from the main board. At least I could now work on it in comfort on a table and in good light. A careful examination of the module didn’t reveal any problems so I fitted one of the ICs and plugged it back in. There was still no sound but I quickly realised that I had forgotten to reconnect the flying lead to socket (YA). This time there was some intermittent sound so I got the freezer out and progressive­ly sprayed small areas on the copper side of the board. 58  Silicon Chip It didn’t take long to discover that the sound changed significantly when I hit the area around Q351. I removed the board again and concentrated my search around this component. What I could barely discern was a very faint hairline fracture around the collector pin of the transistor. I resoldered it and plugged the board back in. Success at last – the sound was always there no matter what I did to the board. Before refitting the screen covers I replaced the other IC as well. This wasn’t really necessary of course but was done at Mrs Jones’ insistence, seeing that the new IC “had already been paid for”. This wasn’t as easy as it sounds as it was a 42-pin high-density IC. Anyway, the rest of the reassembly was straightforward and amazingly it all still worked when the set it was snuggled back into its enclosure. Mr and Mrs Jones were both pleased that their pride and joy had been restored and that the bill was less than they had been expecting. The old 286 computer My next house call involved an old 286 computer that would­n’t boot up. Normally, I wouldn’t consider making a house call on this type of equipment as it’s just not cost effective but the customer was very insistent. Mr Smith was a retired engineer in his late sixties and the old 286 computer had been given to him by his son. I tried hard to point out that though his computer was only eight years old it was well and truly obsolete and would probably not be worth the service cost – after all, some people are tossing out their 486s these days! His response was that he only used the machine for letter writing and didn’t really need anything better. Eventually, I agreed to have a look at the machine if only because Mr Smith had been a regular customer of mine for some years. What’s more, he readily agreed to pay for the service call and so I asked for the symptoms. Basically, he had added another lithium back-up battery to the mother­ board when the CMOS settings had been lost but he didn’t know how to reset it. On the surface, it seemed that this should be a simple job, especially as he assured me that you didn’t need to use a back-up disc and he had the original in­ s truction booklets and software. So why couldn’t he do the job himself? This he couldn’t really explain except to say that he just couldn’t do it. When I arrived, Mr Smith showed me into his little “comput­er room” to examine his ailing 286. When we switched it on, his EGA monitor displayed 512KB of RAM. The first error message simply said “keyboard” and then came about six other lines with details of incorrect disc and memory sizes. Finally, it said “press F1” to enter the start-up menu before trying to boot from either drive A or C. The first important thing I noticed was that the machine didn’t respond to the keyboard, except for making a slight noise in the speaker whenever a key was depressed. This, together with the obviously dried-up coffee stains on some of the keys, sug­gested that the keyboard may be faulty. I also noticed that the Num Lock, Caps Lock and Scroll Lock keys didn’t illuminated their respective LEDs. I checked the AT/XT switch which was parked correctly and, as an experiment, put it in the XT position and reset the comput­er (Ctl, Alt, Del didn’t work). Ironically, the three keylocks now worked correctly but that was all. Mr Smith confessed that he had “looked at” the keyboard and had had it apart, which only deepened my suspicions. I still thought that the problem was relatively straight­forward – the keyboard had been ruined by coffee and hence Mr Smith couldn’t enter the CMOS values so that it would boot from the hard drive. It would, however, boot from a floppy disc in the A: drive (despite the error messages) but it still wouldn’t accept commands from the keyboard. Because, this was my last house call for the day and be­cause I was feeling exceptionally charitable, I decided to take his computer and keyboard and test them out at home with my machine. I would then return it on my way to work the next day. The first thing I did at home (after finding an old EGA monitor I had in the garage) was to connect my own keyboard to the 286 and fire it up. You can imagine my amazement when the same “keyboard” error as before came up on screen, along with all the other error messages. I double checked the keyboard by con­ necting his to my computer and it worked perfectly! Well, if it wasn’t the keyboard it had to be the mother­ board inside the computer. Removing the cover, I found that the keyboard DIN socket was located directly under the power supply, so that too had to be removed. Once the power supply was out of the way, I could see that the old nicad battery was still on the motherboard and had leaked acid onto some of the PC tracks. I quickly snipped out the soldered battery and wiped the affected area with CRC-26 to stop further corrosion. The keyboard socket was quite close to the corroded area but I now had to ask myself whether or not I should continue with what could turn out to be a lost cause. The first scenario was to declare the repair uneconomic and return the computer to its owner. However, the owner is an old-age pensioner and obviously wouldn’t be too happy about paying for my service call without a positive result. The alternative scen­ ario was to remove the motherboard, locate the broken track by continuity checks and solder in a jumper – a piece of cake and there was nothing on telly that night anyway. The hardest part was removing the motherboard which had no less than five other boards plugged into it. Having done this, it didn’t take too long to find the offending track from the 5-pin keyboard DIN socket, the only difficulty being that it was a very thin track. I bridged the track with some fine wire, reassembled the computer and anxiously switched it on. Naturally, I was relieved to see it boot up without the keyboard error. In fact, the keyboard was now working and I punched in the correct CMOS values into the setup menus and rebooted. The computer now booted normally, processing the config.sys and autoexec.bat files to finally rest at Mr Smith’s personal menu. Unfortunately, that wasn’t quite the end of the story. This particular 286 came with 1Mb of hardwired RAM chips (640Kb base and 384Kb extended = 1024Kb) but now it could only see 512Kb of base memory and no extended memory. Obviously, it had lost a couple of 256Kb memory banks but by now I had reached the TRANSFORMERS •  TOROIDAL •  CONVENTIONAL •  POWER •  OUTPUT •  CURRENT •  INVERTER •  PLUGPACKS •  CHOKES STOCK RANGE TOROIDALS BEST PRICES APPROVED TO AS 3108-1994 SPECIALS DESIGNED & MADE 15VA to 7.5kVA Tortech Pty Ltd 24/31 Wentworth St, Greenacre 2190 Phone (02) 642 6003 Fax (02) 642 6127 P.C.B. Makers ! If you need: •  P.C.B. High Speed Drill •  P.C.B. Guillotine •  P.C.B. Material – Negative or Positive acting •  Light Box – Single or Double Sided – Large or Small •  Etch Tank – Bubble or Circulating – Large or Small •  U.V. Sensitive film for Negatives •  Electronic Components and •  •  Equipment for TAFEs, Colleges and Schools FREE ADVICE ON ANY OF OUR PRODUCTS FROM DEDICATED PEOPLE WITH HANDS-ON EXPERIENCE Prompt and Economical Delivery KALEX 40 Wallis Ave E. Ivanhoe 3079 Ph (03) 9497 3422 FAX (03) 9499 2381 •  ALL MAJOR CREDIT CARDS ACCEPTED June 1997  59 Serviceman’s Log – continued conclu­sion that enough was definitely enough. Mr Smith was delighted to got his 286 back in working order; the missing memory made no difference as he could still run his word processor. In the meantime, I’ve sworn not to even look at a 286 again no matter how simple the problem seems. The work piles up Meanwhile, back at the shop, the work had been piling up with at least four jobs that were relatively urgent. I wondered whether I could polish them all off in one day but it was not to be. Some of the problems were caused by intermittent faults and these are always time consuming. The first set off the rank was an Orion 20J that was com­pletely dead. This is another one of those sets where it is difficult to remove the chassis, mainly because no-one tells you about the concealed clips that lock into the case. It is also impossible to gain access to the PC board without unplugging the loudspeaker lead. Anyway, I measured slightly more 60  Silicon Chip than 273V on the main filter capacitor (C506) and traced this all the way to the power chopper (IC501, pin 3). However, no voltage was found on TP501 which is the main B+ rail (103V). Instead, this registered a dead short to ground. The most likely culprits were the line output transistor Q402 and diode D521 (across the B+ rail), though access to these parts is appalling. I replaced them both and then spent some time patching up the generally poor soldering around the power supply. When I put everything back together again, the set fired up OK and so I put it aside to soak test while I got on with the next job. Who was it that said that pride cometh before a fall? The customer called by the next day on the offchance that the set had been fixed and, when he saw it working, insisted on taking it home. I advised him that I would like to test it a bit longer but, as it was a Friday, he said that he would rather take it home for the weekend and reluctantly I agreed. Guess who was waiting for me at 8.30am on Monday morning with his Orion? Apparently, it had only lasted for half an hour before it died again. Such is life. Stripping it down again, I found that the same zener diode (D521) had gone short circuit. For this to die, the B+ has to rise beyond 130V, so I went back to the power supply and hoicked out all the electros for new 105°C ones. I then reassembled it again without the zener diode but with a meter monitoring the B+ rail and switched it on. The B+ still measured 103V five minutes later so I put the zener back in and left the set on, hoping that it would now stay on. It was not to be –just as my first well-deserved coffee was kicking in, the set died yet again and I was too slow to read the meter before it did. I repeated the whole procedure again and left my coffee to get cold. This time, after only five minutes, the B+ began to rise sharply. I switched the set off and touched the components around the power supply. IC501, an STR58401, was quite hot and because I didn’t have any other clues, I decided to order this in and try again later. The next job was a Sharp CX2168 that had come in with the complaint that it intermittently cut off after a few minutes. Despite having a huge range of service manuals, I didn’t have this one and besides, one cannot afford to purchase a manual for every set that you fix. However, I did have a manual for a CX2048 which didn’t look that different. At least, it used the same power chopper IC which, though not marked on the circuit, was an STR­ 41090 (IC701). The main difference was that the CX2168 was a Teletext set and it also used a relay (RY701) to switch the power on and off. And this was basically what was happening – the relay clicked off after about 10 minutes. I also discovered that the circuit has a miniature slider switch (S1101) which controls the relay driver and leaves the set switched permanently on when the power switch is on. Overriding the relay meant that I could monitor the B+ rail (115V) on the cathode of D732 in the fault condition. But, as happened previously, I wasn’t watching when the fault appeared. In this case, the phone rang after the set had been on for 15 minutes. When I returned, the set was completely dead and like the Orion, zener diode D731 was short circuit. As before, I replaced all the small electros in the power supply and soldered the many dry joints there too. This time, when I switched it on, I doggedly monitored the B+ rail and switched it off when the voltage suddenly began rising after about five minutes. IC701 was hot and as it was the logical suspect, I placed an order for it as well. So far, I hadn’t cleared any of the four sets I had planned to do. The two different ICs arrived a few days later and my hunches proved correct – both sets were still working three days later. Sounding out an NEC The third set I tackled was an NEC N4830 with intermittent no sound. Having fixed a lot of these sets, I felt confident that this would surely be an easy one, especially as wiggling either the aerial or AV socket would make the sound come and go. This set uses a Daewoo C500 chassis and I initially decided on a sweeping rework of the main chassis to eliminate any possi­ble dry joints. Initially, I thought that I had got it in one but after half an hour of soak testing, the same problem re-occurred. I dived back into the set and went for the tuning micropro­cessor, concentrating on sound related functions to resolder. Still no luck. I had saved the IF module until last because – as you’ve probably guessed – it was the most inaccessible! Not only did it and its screen have to be unsoldered from the motherboard, but the screening can also had to be unsoldered before I could get at the PC board. This board appeared to have quite a few suspect joints, so leaving nothing to chance, I reworked all the connections. Unfor­tunately, this model set is difficult to reassemble and by the time it was all back together, I was fairly fuming on how badly the day was going. On the other hand, the set now behaved per­fectly and I pronounced it fixed after three days of thorough soak testing. Another success And so to set No.4, a Hitachi Fujian HFC2125B that was dead. I measured the 278V B+ from the bridge rectifier all the way to the collector of V901 (the supply chopper) but there was nothing on either its base or emitter. Resistors R903 and R904 (82kΩ 1W) in series between the collector and base of V901 were the logical suspects. I whipped in two new ones, after first making sure I had discharged the main filter capacitor (C905) and the set came good. While I was at it, I replaced C920, a 1µF 250V capacitor which often causes the set to lose its memory (due to the -28V rail dropping to about -10V). The set was then put aside to soak test by which time I’d had enough SC for one day. June 1997  61