TV manufacturers over the years have had fleeting flirtations with making their products easier to service. In the black and white TV days, valves were easy to replace and this led to the term "valve jockeys" being used to describe technicians who just changed valves.

When transistors came along, we even had one or two sets that used transistor sockets (in the early seventies, you could purchase cheap Russian radios with every transistor socketed – like Tektronix CROs). Then, with colour, we had socketed ICs and plug-in modules. German-made sets also had easy diagnostic systems, with LEDs monitoring critical voltages and flow charts to show which module to replace.

With the advent of microprocessors, On Screen Displays (OSD) and the I²C bus, most sets now offer some sort of error code system. The only trouble is that most faults will not give an OSD, so they then have a standby LED flashing codes. The only difficulty here is actually counting the number of flashes and then comparing them with the service manual.

The nature of faults has also changed somewhat over the years. These days, faults are often found in protection circuits, or there are software problems.

Of course, ICs are no longer socketed and more and more are surface-mount devices. Some manufacturers actually expect you to replace these but an increasing number are moving to board rather than component level replacement. In fact, plasma TV manufacturers generally offer nothing smaller than board replacement and insist you do not repair boards – not that you would actually get a detailed circuit diagram.

Philips sets

Philips invented the I²C bus and has been using error codes the longest but they can be unreliable. I have had quite a few Philips models with FL1 and FL1.2 series chassis where multiple codes randomly displayed on switch-on and yet, when the set warms up, then settle down with all functions working properly. I haven’t been able to get to the bottom of that one.

The A10 series chassis often display error numbers (eg, 17, 23) which, although shown in the service manual, have no meaning and certainly don’t interfere with the working of the set.

One Philips set I looked at recently – a 32FL2881/75R (FL2-G chassis) – was dead. I had replaced the flyback transformer 18 months earlier. This set not only has OSD but also another seven red LEDs internally, none of which was giving an error code.

In this case, the horizontal output transistor (Q7610, 2SC4288A) was short circuit. Despite access being rather difficult, I managed to replace the transistor by completely removing the entire chassis.

The set now came on and gave a perfect picture, except it was a little underscanned in all modes. I checked the voltage at TP78 as +141V and it was spot on. This presented me with a dilemma – did the set need realignment or did it have an additional fault that was going to kill more transistors (eg, the tuning capacitors on the collector of Q7610)?

I had it running for over three working days and it didn’t falter and neither were there any error messages in the SDM (self-diagnostic mode). In the end, I told the client that I didn’t know what had caused the failure and that I could not guarantee it until more work was done.

However, he wasn’t prepared to wait any longer and didn’t want to spend any more money on this 11-year old TV, despite it having cost over $5000 new and this being only the second time it had ever broken down. In fact, he said that he really wanted to buy a new plasma set.

Anyway, he picked it up, paying only for the cost of the transistor and one hour’s labour, but phoned four days later to complain that it had died again. And of course, he wanted it fixed under warranty despite my earlier warning. You can’t win them all.

Another 2004 Philips I had was a 32PW4523/79R which wouldn’t turn on and was making a clicking noise inside. The label on the back said L2K3 but it actually it looked like an L01.1A chassis. Its diagnostic LED was flashing twice, or so I thought – it was hard to tell.

The clicking noise was coming from the degaussing relay (1515). I began by checking the B+ rail and it was low and pulsating. However, when I shorted out the base and emitter of the line output transistor and hung a 100W globe on the collector, it came up fully and the globe illuminated without pulsating. In addition, the relay stopped clicking.

This told me the fault was in the line output stage.

After making sure the vertical output IC was not dry jointed, I tried desoldering it to see if that would affect the protection circuit. It didn’t so I then tried disconnecting all the secondary circuits on the flyback transformer to see if there were any shorts.

This revealed nothing untoward, which threw suspicion onto the transformer itself. I ordered a new one (3128 138 21921) and fitting it fixed the problem and confirmed my diagnosis.

I checked the SDM afterwards to find it had error codes 2 (horizontal protection), 6 (I²C) and 10 (timer ID). I can only conclude that the 6 and 10 errors referred to previous faults and the errors hadn’t been cleared after repair.

Not quite the same fault

Once I had two identical 1999 Sony KV-XA29M31 (BG3S chassis) in the workshop at the same time with what appeared to be identical fault symptoms. They were both dead, with the report that the red LED was flashing twice.

The self-diagnostic function reports error 2 as being OCP (over current protection) or OVP (over voltage protection). This can be due to the horizontal output transistor Q511 being short circuit, IC701 or IC503 being faulty, or no -13V rail. In each case, the TV resets itself into standby mode.

Access to these chassis isn’t easy, so on the first set I went for the line of least resistance and decided to change the vertical output IC (IC503, TDA8172) first after checking that Q511 was OK. This turned out to be a good move, as this fixed the problem straight away. I also resoldered the pins on the line drive transformer (T501), as this sometimes causes premature failure of the horizontal output transistor.

Moving onto the second set, I figured I would tackle it the same way as the first one. However, it was soon apparent that this was an entirely different fault.

Once again Q511 was OK but there was no +135V going to it from D615. Neither was there a short on the line. In fact, it turned out to be a general failure of the main switchmode power supply.

A quick check soon showed that R632 (0.1W) was open circuit, while IC601 (STR-F6656) had gone short circuit. Further checking then revealed that D610 had also gone short circuit.

So why did the supply fail? I couldn’t find anything specific, so I decided to also change C623 100mF, which is about the only electro in the primary of the power supply – not that I could find anything wrong with it! Anyway, that fixed it but I couldn’t quite see why the self-diagnosis was flashing twice with this configuration.

Another Sony

The next Sony I had was flashing two times again. This set was a 2003 KV-AR221M30 using a BX-1S chassis.

This time, the line output transistor (Q805, 2SC5885) was short circuit and the +135V feed resistor (R827) was open circuit. I then checked transformer T800 (HDT) for dry joints but found none.

So why did the line output transistor fail? I removed the flyback transformer (T801) and checked it with a shorted turns tester. As I suspected, it had failed and replacing all three components completed the repair.

Dead Philips

A 1997 Philips 25PT448A/75R PV4 4.0AA came in "dead". Well, not quite – the power supply wasn’t dead and the front LED was flashing. V_BAT measured OK at +137V all the way to the line output transistor (7448) and its driver (7440) but there was obviously no horizontal or vertical drive from the Jungle IC (7200).

I could see that the vertical output IC (7960) was dry-jointed on several pins, so I resoldered it. Unfortunately, this made no difference on switch on but I still suspected it, figuring that it might have been destroyed because of the dry joints. Fitting a new one (TDA9302H) restored picture and sound but only in the bottom quarter of the screen.

The new IC was getting hot too, so I checked the ±13V supply from the flyback transformer with the CRO and noticed that the negative rail had a large amount of ripple on it. As a result, I replaced the 470mF electrolytic capacitor (2472) but it made no difference.

Next, I checked the voltages around the IC and found that pin 1 was very low. The twin vertical drive to the IC comes out of the jungle IC (IC7200) from pins 46 (+) and 47 (-). It is then fed through an RC network before going to pins 1 and 7 respectively of the vertical output IC (IC7960). A quick check with the CRO showed plenty of negative drive to pin 7 but virtually no positive drive to pin 1.

It didn’t take long to find that pin 1 of IC7960 was short circuit to ground. However, it wasn’t the IC that was at fault. Instead, it was C2910 – a 470pF surface-mounted chip capacitor – that had gone short circuit.

Replacing it restored the set to full working condition. Interestingly, there is no error code for this.

Storm damaged TV

An unusual white-coloured TV was recently brought in for repair. Years ago, most 43cm sets were white in colour but nowadays, they are all silver.

The set in question was a 1996 51cm Grundig T51-731 Text FFS-TG/51C using a CUC 7303 chassis (GCE5084). It had gone dead after a storm but ironically, the only thing that had failed was the mains fuse.

Once the set was running again, I had to tune in the stations, which was rather tricky. You definitely need the instruction book and the remote control (Telepilot 711, now substituted by a TP715/TP750C). In fact, there are only three controls in total on this set, including the power switch!

To do the tuning via the remote, you have to press "i" and then OK to get into a rather daunting tuning menu. However, with perseverance and the book, I eventually cracked it (in fact, auto tuning is easier and you do that by pressing "PC/AUX" for four seconds and then "OK").

This revealed further problems. I have a good antenna and reception at my workshop is good but the picture on this set was just slightly snowy. As a result, I decided to give the RF AGC (automatic gain control) a nudge to see if this would improve matters.

To get into the Service Menu, you press "i" and switch the set on with the main power switch. The AGC was set at 29 but increasing or decreasing this value made no difference to the picture and it was beginning to look like the tuner (Part No. 29504 3010100) might be faulty. As a result, I took it out, fully anticipating that a new one would have to be fitted (in the old days, you could get the tuner fixed by taking it to a tuner specialist but unfortunately, that’s no longer the case).

Anyway, I took a quick look inside the tuner to see if anything was obviously faulty and was surprised to see that the centre pin of the antenna socket to the tuner PC board had disappeared – it had vaporised completely. I fitted a small wire link and reassembled the set and that finally fixed the problem.

The dummy spit

I am still meeting Sanyo TVs that have those 120kW resistors that go high, particularly in the start-up circuitry of the power supply (eg, R620 & R621 in model CP-29ST2-01, AC2-A29 chassis). This is a fairly common fault that’s help put the bread on my plate – and do I need it!

I recently had a later Sanyo model CP-29AF2X(A)-55 using a FC3-G229 chassis come in. This had had a spectacular dummy spit for a set barely out of its egg, the fault blowing the daylights out of eleven components in the power supply: R613 330W, R614 22kW, R624 10W, R632 0.27W, R633 0.33W, Q611 & Q625 2SC2274, Q612 2SA984, Q613 2SK3102, Q432 2SD2634YB and D610 (PC123 optocoupler).

The worrying aspect of such an explosion is why did it happen? In this case, it wasn’t possible to figure this out but it concerns me when I don’t know, as it almost invites the fault to recur.

Kamikaze fly lava

A kamikaze fly lava chose to be born (and die) inside a brand new Sanyo TV, model CP14SR1-50 AC5-G chassis. Unfortunately, he, she or it selected the switchmode power supply FET and killed the set by turning the power supply into a pile of what flies like best.

This suicide terrorist only managed to blow eight components – R612 10W, R613 1W, R617 68kW, Q613 2SK2647, Q614 2SA984, Q432 2SD2634, D610 PC123 and D611 ISS133 – but at least this time we knew what the cause was. I don’t think any anti-terrorist laws will stop this one either!

When changing the horizontal output transistor (Q432) in this set, it is important to check that the jungle IC201 is a QXXAVB889- - - M. If it’s not, you should fit a CP14SR1MK modification kit.

In any event, check that C351 is a 47mF 16V electrolytic and that C661 is 10mF 16V and replace C501 (1000mF 35V), otherwise you might have more unexplained failures of the horizontal output transistor.

Shrunken widescreen

A brand new wide-screen Panasonic TX-76PW60A (EURO9S chassis) TV was brought in to the workshop under warranty. The fault was that there was very little width or pincushion correction and there were vertical shaded lines spaced regularly across the picture, which faded off completely within a minute or two.

Initially, I felt that this was going to be a simple problem in the east-west correction circuits. When I got the back off, there was a small plug-in module (DF) on the righthand side which is supported by a metal bracket. This bracket and the module were severely bent, so it looked like this was the problem. However, when I had removed the module, I found I could straighten it and no part of the PC board’s circuitry had been compromised or damaged. Besides which, it didn’t have much to do with the east-west circuitry.

Next, I located the circuit I really wanted – ie, transistor Q703 (a surface-mounted 2SK2231) and the diodes inside D580 on the D Board. These measured OK, as did R702 (1W) and there was about 40V on the drain. I also checked that there was a clean, healthy parabola on the drain and for a square-wave pulse coming into this circuit from the A Board.

While I had the oscilloscope out, I also checked the line pulses on the collector of the horizontal output transistor (Q551, 2SC5905). These were perfectly clean but despite this, Q551 was getting very hot, as were resistors R705 and R706 (150W) near Q703.

Unfortunately, by the time I got the set into the service mode (for less adjustments), the picture had faded off completely. As a result, I was unable to determine whether the digital control system on the A Board was working or not.

I have had cases where the microprocessor has caused these symptoms but as the A Board is worth around $750, we naturally we didn’t have one on hand to substitute.

Despite the fact that it would appear to be a fairly obvious fault with simple circuits to troubleshoot, I was initially clueless (most people would assert that this is indeed my normal state). Naturally, I checked almost every component in and around the circuit, including D559 and D560, but without result.

I also carefully check-ed the coils in the east-west circuit – L701 220mH and L704 1.8mH. These measured correctly on my inductance meter but I went further in case the insulation was breaking down under load. I had noticed that L582 was also a 220mH unit, so I decided to swap it and L701 over.

Interestingly, these two coils had different part numbers even though they have the same inductance. What’s more, the gauge and type of wiring differed between them. In particular, one has a solid copper wire, while the other uses multi-stranded wire. Anyway, to satisfy my curiosity, I swapped them over but it made absolutely no difference.

In the end, it was luck that solved the puzzle. After swapping the coils, I noticed one nearby capacitor (C582, 470nF 400V) with what looked like a small lump of clear varnish on it. Of course, a small deformity like this is not really unusual in a brown polyester capacitor – after all, it’s what’s on the inside that counts and looks aren’t really a concern.

Anyway, because I was out of ideas and there wasn’t anything else to do, I decided to remove this capacitor and take a closer look. And, as it turned out, it wasn’t a lump of varnish at all but was instead a lump of material that had been pushed out of a crack in the capacitor’s body.

A new one soon restored the set to normal operation.