Items Covered This Month Akai CT-2868AT NEC FS-6831S Sharp CX-51E3 NEC Cromavision FS-6807S Mitsubishi DNA CT 29AX1(A) Philips 25GR6771 G111S chassis

So what is this so-called "self-destructing" TV set? I had heard about this phenomenon from several different sources but I had never actually come across one until now.

As it turned out, the term "self-destructing" is a gross exaggeration and is much too dramatic - "intermittent failure of a number of power supply components" is a much more accurate description of the fault.

Anyway, this particular problem applies mostly to the NEC FS-6831S series employing PWC-3517 chassis. These are generally very reliable TV sets but several rental companies have complained of a few sets arriving in the workshop dead, with the power supply "destroyed", along with IC601 (STR41090) and up to about 10 surrounding components. All the parts are replaced and the set works again for another month or so before it fails again - mostly on start-up with the front panel power switch. Or, at least, that's the story.

Mr Spicer owned one of these sets and it had been fixed several times, by different technicians, until he became fed up and gave it away to another technician who happened to be a colleague.

Well, he was determined to fix it and, over the next six months or so, changed just about every part in the power supply. I obtained a scrapped chassis for him so that he could even replace the chopper and feedback transformers (T601 and T602) without cost. Wherever possible, he used genuine new NEC spares but in the end he lost patience with its repeated failures and offered it to me for spare parts.

Though a few technicians have complained about this particular symptom, NEC Technical Support were really unaware of the problem. This chassis has been in production for over 10 years and as previously stated, has been very reliable.

Bath time

I decided I would give it a go and treat it with great respect. Firstly, I noticed it was a pretty grubby chassis with lots of dust and grime on it, so I decided to give it a bath! No kidding! The secret to washing a chassis is being quick and allowing it a very long time to dry. Note, however, that mains transformers and the like should be removed first as, invariably, they will break down later if allowed to get wet.

I squirted the chassis with a commercial household cleaner and brushed the particularly dirty areas before using a high pressure fresh water spray and then very hot water. Pure water is in fact an insulator, not a conductor - but how often does one encounter pure water in real life?

I let it dry for a week, even though it looked dry enough within 24 hours, and it indeed looked very clean - like new. Any remaining grubby spots were then cleaned with more conventional PC board cleaning aerosols.

Now that it was clean, I could see that some of the electrolytic capacitors had been leaking and had bad-
ly corroded the surrounding board
areas, especially around C406, C411, C501, C611 and C625. These were all replaced after giving the area a very thorough scrubbing and finally applying a fine spray of CRC 2-26.

I then soldered every faulty joint I could find on the chassis, paying special attention to areas around the power diodes before replacing all the blown components.

I noticed when checking the component values against the circuit diagram that some were not quite the same. I dug up my file on the set and discovered that, over the life of this chassis, two distinct power supplies were made available - one for the 63cm set and one for the 68cm version. This set used a 63cm picture tube and I replaced seven components that a previous technician had incorrectly substituted from the 68cm version (Q601, R417A, R602, R603, R606, R608 and R610).

I also applied some of the modifications that have been recommended over the years, especially changing C603 and C604 across the bridge rectifier from 220pF to 4700pF. I also fitted a new insulating wafer for IC601.

I didn't attribute any significance to a failure when using the front panel switch, as this controls the CPU (IC1001) and the 19V/15V supply transistors (Q652), both of which were totally isolated from the power supply. The remote control also operates the CPU IC.

When all was ready, I reassembled the chassis and fitted it into the cabinet. I gingerly switched on and everything worked properly. I checked the main power rails, noting that the HT was slightly low at 113.7V at TP91. However, the picture size and blooming were steady with variations of beam current and everything was working properly.

It is now a month down the track and the set is still switching on and off OK and running fine.

If I am right, I would attribute the problem to previous technicians not replacing all the leaky capacitors and failing to clean away the leaked electrolyte. In addition, a previous technician had used incorrect components in this particular set and there were also possible faulty solder joints.

Persistence doesn't pay

As I hinted earlier, this story poses the question as to why we sometimes persist with a problem, long after it has become commercially non-viable. In this case, my colleague hung on to the set for about six months and I lost count of how much time I spent on it. Granted, this was spare time (short periods between other jobs, etc) which would otherwise be wasted but how does one justify it anyway?

The simple answer is "plain stubbornness". But there is more to it than that. The solution to the problem has a real value and is not just a cause of satisfaction.

Every solution becomes another item in one's stock-in-trade; something to call upon the next time a similar fault is encountered. Every time something is learned, it makes it just that much easier to earn real money the next time around.

Not so easy after all!

I imagined that Mr Klein's Akai CT-2868AT TV set was going to be easy when he described the fault as no colour - that is, until he told me he had tried to get it fixed elsewhere and they had given up on it. Usually, most colour faults are fairly easy to fix if one is familiar with the CCIR B/G PAL D system. However, this is gradually becoming more difficult as digital techniques are being added.

Anyway, I still felt confident that I could make short work of it. However, when I switched it on and connected a colour bar generator to the AV input, I was amazed at the appearance of the fault which, I have to confess, is hard to describe. It was as though the chrominance reference oscillator hadn't locked but instead of different colours, the picture showed alternating bands of monochrome and colour. And at the same time, there was a large shading band (like a shadow) moving slowly from left to right - most bizarre!

I also found that the previous technician had changed the jungle section, including chroma decoder IC302, an AN5601K (the same as is used in Panasonic TV sets). He had also replaced about 10 electrolytic capacitors on the motherboard, mainly in the chroma circuits.

This was a bit of a worry, as it told me two things. One was that the previous technician was no slouch and two, he had already done most of the things I had planned to do. It didn't leave me much room to manoeuvre.

Voltage checks

I started by checking the voltage rails and all the voltages on IC302. They were all spot on, so I checked the reference oscillator and that was also spot on. I also checked all the work done by the previous technician but everything was totally correct. Yet, to produce such a dramatic effect, something had to be off - really off.

I connected a monitor to the video output socket and saw that the picture was perfect up to there. Using the CRO, I then followed the luminance and chrominance signals through Q103 (for TV reception) to IC204 LA7016 (AV/TV switch), and then on through the Teletext module CN501, pins 2 and 7 (note: the Teletext module can be removed provided a link is inserted between these pins). From there, I went to the luminance delay line where the chrominance and luminance are separated - the former goes to pin 5 of IC302, while the latter goes through Q306 to pins 15 & 16.

I thought I could detect a slight imperfection of the chroma input on pin 5 and spent a long time investigating this, especially as it seemed to come from the base of the luminance transistor Q306. I tried replacing C312 but in the end I had no explanation and assumed it was a red herring. What I did find was distortion on pin 2 (reference oscillator) and on the chrominance output to the delay line (pin 7) and onwards to the CRT.

By now I had almost exhausted all my ideas. The fault seemed to involve two different areas: the alternate couple of lines going black (horizontal frequency) and then the slow moving shadow across the screen (vertical frequency). These could be due to some distortion in the horizontal and vertical pulses but these all where correct on the oscilloscope. The 12V and 5V rails (pins 29 and 10 respectively) had no ripple on them.

In the end, I decided I needed to scope every pin of IC302 and see if anything significant would show up. Fortunately, I only reached pin 3, which is the chrominance level control from pin 37 of the microprocessor, before finding a real clue - this pin should have only a pure DC voltage present but there were all sorts of other signals present as well

A check with the multimeter showed that a DC voltage was present but that it varied with the colour control. It was now obvious that decoupling capacitor C317 (2.2μF) was leaky and replacing it fixed everything.

I left it on test for a week before Mr Klein picked it up. However, just as I was demonstrating it, I suddenly noticed a minor new fault symptom on the screen. It was giving an effect like flagwaving at the very top. Even though I noticed it, Mr Klein didn't, and it was just too late to do anything about it - so out it went.

I am now waiting by the phone, biting my nails, waiting for the recall!

Dead Sharp

Mr Tanundra's Sharp CX-51E3 TV set (H Chassis) was dead and I was immensely pleased that he had brought it into the workshop instead of trying to get me to fix it in his house.

I don't have a service manual for this model but I do have one for the CX51B2 which is a 14B chassis and not all that different - or so I thought.

Although the set was reported as "dead", it wasn't really totally dead - just no sound or picture. For starters, there was 115V on TP702 and to the line output transistor Q602 which was good. There was also voltage on the collector of the horizontal driver transistor (Q601) but it was a bit high at about 60V.

The oscilloscope showed that the line drive was present from pin 37 of the jungle IC (801, TDA8362) all the way to the horizontal output transformer (T602). However, its frequency was far too high. Consequently, the EHT was down to about 15kV. When I turned up the screen control, there was just a horizontal line, indicating no vertical timebase either.

I checked the voltage to IC801 pin 10 and found it was low but it was the horizontal and vertical oscillators I was most concerned about. There was only a small amount of signal on the 8MHz crystal (X802) on pin 35 and the voltage on it was almost non-existent, especially on pin 36.

I followed this back, expecting 9V to appear via D633 and R672 but there was none. I thought D623, an 8.2V zener, might be faulty but it was fine. The 9V rail was missing completely.

On the circuit I was using (14B), the 9V is delivered via Q604 and D733. However, in the H chassis, there is no Q604 but instead there is IC602, a 7809 3-terminal regulator. And there was voltage going in to this device but none coming out.

Fitting a new regulator fixed the problem completely, though I am surprised at how hot it ran. I have soak tested the set for a week and Mr Tanundra is happy.

A crook NEC

Mrs Richardson had arranged for her husband to deliver their 1999 NEC Cromavision FS-6807S TV set to the workshop with the complaint that there was no picture.

Because it is such a recent model, I didn't have a circuit for it and I told him that I would have to order it. His response was that he was under some pressure to get it fixed as quickly as possible although he didn't say who from.

Well, I duly put the order in straight away but in the meantime, I thought that I would take a quick squiz at it. The fault turned out to be a black raster, no screen display and no sound.

The chassis was a Thai built PWC-2477 which I removed to check for faulty joints. Well, as luck would have it, there were a couple of enormous cracks in the solder around the connections to IC501 and IC502, both 9V 3-terminal regulators.

This is just the sort of break a technician needs, being so much easier than the complex microprocessor and/or signal problems I had been was envisaging. Indeed, it was exactly what the doctor ordered and repairing the cracks fixed the problem

Mrs Richardson was delighted to have her set fixed so quickly.

A couple of house calls

It was a fairly quiet Thursday in what had been a quiet week and so when I received two requests for house calls just outside my service area, I reluctantly agreed - mostly because they were close to each other.

The first was Mrs Mason's 1996 (1994) Mitsubishi DIVA TV set, CT 29AX1(A) (A1 chassis), which she complained had a "green picture". This sounded relatively easy to fix and was probably a faulty joint on the CRT socket.

However, when I arrived, I found the 68cm set to be in a tight, dark entertainment centre connected to all sorts of accessories. What's more, the fault was actually a small dark green raster and there was no sound. To cap it all off, there were two totally undisciplined, yapping dogs running around the room, determined to trip me. It was going to be one of those days!

I stupidly thought there might be something simple I could do and, anyway, I had to look as though I had done something - even though I suspected that this was going to be an expensive workshop job.

I hadn't worked on one of these sets before and it was one with all the bells and whistles, including a wretched motorised swivel base to turn the set with the remote control. I couldn't help feeling that the user would only ever be likely to use this control once but now it was giving me a real headache. After I had managed to pull the set out far enough to unplug a bunch of AV leads, turn the set around and undo umpteen screws, in the dark, I was beginning to lose it. Trouble was the back was jammed at the bottom by the swivel base and the whole set was precariously balancing on the edge of the pull-out swivel platform of the entertainment centre.

Finally, when I lifted it off, I was presented with a large flat chassis with no access under the PC board. Also, I couldn't work out how to remove it from the case. The plastic frame looked as though it had some sort of catch system on either side near the tube but nothing I could do would free the front control panels which were wedged under the tube. So I tried to regain some self-composure by doing some meter measurements from on top of the chassis.

I was trying to identify the main HT rail from the component side of a complex set I had never seen before. And in any case, I had absolutely no idea of the voltage to expect on this rail, even if I did identify it. I looked for large high voltage rating electros and any diodes nearby which might indicate a chopper transformer but even when I thought I could identify it, there was absolutely no access. Similarly with the horizontal output transistor, which I nearly blew up.

Mrs Mason had been watching my activities closely and every so often had been making useful comments - such as, "there's not much in them these days" and also "I would have thought it would have lasted a lot longer considering how much it cost" (about $2000 four years ago).

I kept thinking, "why didn't you buy a sewing machine instead" but bit my lip and stuck to bland mutterings along the lines of "isn't the weather good?" and "the cricket is fantastic".

In the end, I gave an Oscar-winning performance of measuring the resistance of one piece of screening metal to another to convince Mrs Mason that I knew something, before pronouncing that the power supply was faulty and that the set definitely had to go to the workshop - where it should have gone an hour beforehand!

She wanted a cast iron quote and an ETA for the return of the set. I was past caring at this point and plucked a figure of $300 and three weeks out of the air. Darkly she said "just wait a minute while I phone my husband".

Surprisingly, five minutes later, she reappeared, smiled, and said that would be OK. Oh well, "them's the breaks", I suppose.

The next drama I had was getting it downstairs and past the dogs and into the car - without letting the dogs out. The car was now full and I still had the other TV set to fix. I prayed I could do something with this in the home, because I couldn't take two back to the shop.

This set was a 1990 Philips 25GR6771, G111S chassis, that was dead. It had had a line across the screen before it died.

I arrived to find the set was situated on the second floor of Mr Paton's home which had a small rickety spiral staircase. I knew that this set wasn't going to leave the house that day and started the conversation with "...and how old is this set?" "Eleven years old, eh? Well, you have had a good run out of it . . . but it is probably past its 'use-by' date. Have you thought about getting a new one?"

In the event, the "power" problem turned out to be a faulty on/off switch which I didn't have with me, so I shorted one side of the switch and told them to use the remote control or the wall socket switch to turn it on and off. The white line across the screen was due to a faulty joint around the pins of the vertical deflection socket, which connects to the vertical deflection yoke. I pointed out that this was a temporary repair that would do until they bought a new set. Mr Paton was quite happy about this and I drove off back to the workshop.

Back at the ranch

I ordered a service manual for the Mitsubishi before examining it more carefully in more congenial surroundings. The reason the chassis wouldn't come out was because of two screws in the front panel holding it in. I removed it completely and used an air compressor to blow the dust out. I then removed the PC boards from the plastic frame to get a good look at them.

When I removed the power deflection board, I could see that the whole thing had faulty joints everywhere. I spent several hours resoldering everything on all boards before reassembling them.

Eventually, I managed to get it all back together and switched the set back on. Amazingly, it all worked perfectly, so I set about retuning the stations and setting up the various menus. And the picture was truly excellent.

I found and checked the main power rail at test point TP91A. It checked out at 130V, so I replaced the back and put the set aside to soak test for a week.

In the meantime, a few days later, I received a call from Mr Paton that I didn't really need. He said that the set was dead again and it was making a burning smell just like when I was there. What smell? There had been no burning smell, save possibly the soldering iron. Reluctantly, I said I would call back but that I couldn't until I could deliver the other TV set, as they were both in the same district.

A few days later, I manoeuvred the Mitsubishi past the Mason's dogs and back into the entertainment centre. Apart from retuning the VCR, DVD and Foxtel, it was plain sailing. And Mrs Mason was happy that it cost less than I had quoted.

Mr Paton's set, on the other hand, I had to treat very carefully in case I had a fire on my hands. Gingerly, I removed the back and put it into a position where I could see what was going on when I reconnected the power. However, it was all an anticlimax; the set came on perfectly without any bother.

I checked for bad connections and any sign of smouldering but just couldn't fault it. I put it back together again and it still switched on and
off perfectly via the wall switch!

Mr Paton's credibility had suddenly taken a huge dive - so much for it being dead and burning! If it breaks down again, he can (1) buy a new set (2), take it elsewhere, or (3) bring it into the workshop himself.