A tricky little Panasonic TC-51S90A (MX5A chassis) waltzed its way onto my workshop bench with an irritating fault – it would turn itself off intermittently, when hot.

I started by checking all the voltage rails but could find nothing suspicious. I then checked and changed all the small electrolytics in the power supply. That also failed to fix the problem and after checking for dry joints throughout the set (especially on the line driver transformer and the IC regulators), I had a go with the freezer. But even this refused to reveal the little blighter’s secrets!

I subsequently spent an indecent amount of time with this set (more than it was worth, in fact) before realising that it might be the microprocessor. I checked the clock oscillator and the 5V reset line but became really suspicious when I was lucky enough to notice a slight drop in the 5V rail to pin 39 of IC1101 (MN1871681TFH). In addition, the supply IC regulator (IC850) was getting rather hot.

However, before ordering a new microprocessor, I changed the IC regulator and the electrolytics on the supply rail. I then discovered that the impedance across this rail – which essentially feeds just the micro – was just 50W, even after desoldering the reset IC and pin 39.

It didn’t take long to discover that the real culprit was 5.6V zener diode D1160, which was intermittently leaky. I replaced this and its sister D871 for good measure.

Items Covered This Month Panasonic TC-51S90A TV set (MX5A chassis) LG RT-54NA61RB rear projection TV set (chassis MP-03AB). Philips 29SX8761/75R TV set (G8 chassis) Toshiba Active Vision CT-90101 TV set (model 36ZP48P(B)) Hitachi C29F300B TV set (V3AL1 chassis)

LG rear projection TV

A while back, we had an LG RT-54NA61RB (chassis MP-03AB) rear projection TV which had three coloured dots at the top of the screen.

This came in as a warranty claim and we soon discovered that the fault had occurred after the removalists had moved the set to a new location. At first, we thought this was due to some kind of screen burn but after removing the screen and displaying the picture on the ceiling, it was obvious that it was nothing to do with the tubes or electronics. Instead, it was soon traced to the mirror which had three tiny dimples in it.

A new one was ordered and fitted, which fixed the problem.

After the set had gone out, we examined the mirror to try and find out how the dimples had got there. Our suspicions were that the removalists had dropped something on the back and deformed the mirror.

The mirror consists of a strong welded steel frame with pre-tensioned silver foil stretched out across it. Behind this is a polystyrene block packing filler. The interesting thing was that when we lifted the mirror so that it was face up, the dimples disappeared – they were only there, in exactly the same place, when it was pointing down! With some difficulty, we managed to prise out the polystyrene block but it was faultless. However, with the polystyrene removed, the dimples in the mirror had disappeared, regardless of its inclination.

We then reinstalled the polystyrene block and the dimples were still not there. Our conclusion was that three specks of dirt or dust had embedded themselves in the polystyrene when the set was moved. When we removed the block, we displaced them and now it was all OK.

The fountain of knowledge

I used to supply technical advice and support to colleagues in this dwindling profession. I guess I found it flattering that someone out there doesn’t consider me to be total dill and generally speaking, what goes around comes around. Hopefully, when I have lost more marbles than I have left, someone might return the favour – that is, of course, if there are any technicians left in our increasingly "throw-away" society!

Thomas, a sem-retired technician, wrote in from interstate with obvious frustration. He had a Philips TV in which he suspected a faulty P83CO54 microcontroller, which was no longer obtainable. He needed to know a little more about the I²C data bus so he could perhaps understand and maybe fix his problem (as yet undefined).

I guessed that he was fixing a Philips 29SX8761/75R (G8 chassis) and suggested that before he went down the expensive and difficult road of getting a replacement IC7222 (Part No: 4822 00932 953) that he might try replacing the EEPROM (IC7223, 24C04), as it is much cheaper and often gives the same symptoms as the microcontroller.

I briefly explained that the "Inter IC" (I²C) Bus is a twin bidirectional 7/8-bit (10 fast mode) digital TV control communications system developed by Philips in the early 90s and now used in almost every TV set. One line is the Serial Data (SDA) and the other is the Serial CLock (SCL) and these lines connect all the digital ICs to the microprocessor and EEPROM.

The data is transmitted in this byte/word format: Start condition, Slave address (to select the IC), Acknowledge, Sub Address (to select the function), Acknowledge, Data change (varies the function level), Acknowledge and P (= stop condition). The eighth bit determines the data direction R/W. The remainder are for I²C extensions (eg, RS232 Access bus protocols) and fast mode.

I also suggested he check the following around this circuit: +5V, reset 5V and the clock crystal/resonator frequency. And he was advised to use a CRO to check for digital noise on the +5V, SDA and SCL lines and to check the error (clear first and then reread) and option codes (for this particular model). If something is wrong with these, you disconnect the data lines to each IC until the basic conditions are restored.

Other checks included disconnecting the remote control receiver in case it was transmitting noise – the control keyboard is scanned and faulty switches can cause funny effects. Memory problems indicate problems in the EEPROM, while some sets have backup batteries that fail.

Most microcontroller faults will not let the set power up. The way around this is to override the standby/on IC pin and then carry out the above checks.

I suggested that he initially try the Service Mode switch (SW1907) and read the error codes. It was at this point he explained he had no picture or sound and was not able to read any OSD error codes in the Service Mode, nor was he able to discern any significant flashing of the Standby LED except to note that it was glowing yellow or orange. However, he did manage to conclude that there was activity on the SDA and SCL lines and they were normal at +5V DC.

He also turned up the G2 control on the flyback transformer and found he had a fully-scanned raster with flyback lines but absolutely nothing else. IC7551 (TDA4681) was getting extremely hot so he replaced it but it made no difference.

As we both had service manuals, we discussed various voltages and waveforms like sync pulses and sandcastle, etc and basically nothing was coming out of IC7551 on the Small Signal Board (SSB) to the CRT D Board. Thomas then discovered Q7186, R3181 and R3182 on the CRT board were also faulty but replacing them didn’t alter the symptoms.

Eventually, I suggested he send me his CRT board and some of the ICs that he wasn’t sure about and I would install them in one of my own sets. This he did and apart from a small error on the CRT board, I found everything was OK except for IC7466 (TDA4651) which gave the same symptoms on my set as he had. And ironically, his EEPROM gave no sound.

I sent the parts back, hoping these would fix all his problems. Unfortunately they didn’t. He was unable to get a TDA4651 but a TDA4650 didn’t work either.

Finally, he sent the entire chassis to try in my set. I had to socket all the ICs in both sets in order to check them. Also, I found there were significant differences between his model and mine, which meant I couldn’t directly connect the Large Signal Panel into my set (and vice versa) without changes to plug L20 (pins 8 and 9), which gave no vertical deflection.

In the end, I found that the fault lay on the SSB and I swapped all the ICs (except for the microcontroller), the tuner and the Sound H Board (which is soldered to the SSB. This made no difference, so it looked like the fault really was in the microcontroller.

The problem was that the 42-pin high-density microcontroller IC could not sustain repeated desoldering and soldering. As a result, Thomas invested in some expensive 64-pin IC sockets from Radio Spares. I cut them up to fit the ICs and swapped them over.

And that was it – his microcontroller IC, as he had predicted, was faulty, as mine worked fine in his chassis. Mine is a PC87C055 but the P87C054 is also still available from Philips.

I sent the chassis back by post and Thomas refitted it to find everything was now OK. I also advised him that multiple failures – such as in this set – are due to things like lightning strikes, power surges or EHT flashover within the picture tube to either the CRT socket or the yoke.

As a result, I suggested he clean the dust and dirt off the CRT, especially around the EHT final anode, and check the earthing to the yoke clamps. I also told him to fit two back-to-back series 24V zener diodes to the set, with their anodes connected to the P5 Aquadag and Link 9199 on the CRT D Board, respectively.

Subsequently, he told me that he had another problem involving an audible "plop" or "thump" to all three loudspeakers when the set was switched on or off. although I never noticed this symptom when I had the chassis.

The circuits in the service manuals show a variety of muting circuits that keep the set quiet when the power is switched off. His set had yet another variation and he changed a couple of electrolytics (C2676 and C2677) in this mute circuit from 22mF to 100mF, which reduced the problem to a level which was acceptable to him.

Not being one to leave loose ends, I was intrigued with the problem and would like to have resolved it. However, I respected his decision to leave it alone!

Apparently, the master on/off power switch doesn’t produce any noises, so presumably the problem only occurs in the remote control mode. This set is fitted with no less than six audio amplifiers, so was his problem confined to just three of them or all of them? The mute circuit monitors the 12V rail and mutes the set until this rail was either fully charged or discharged. The audio amplifier IC (IC7621, TA8218AH) that provides the left, right and sub-woofer channels uses a 23V rail instead.

I don’t quite understand the circuit changes between the models but I would have changed C2620 to 330mF to give a longer time constant in the PLOP MUTE circuit.

Coming the raw prawn

I had a "recall" demanded by a client on a Grundig TV I had fixed three months earlier.

Despite always pointing out that I can only guarantee the work I do and the parts used for a particular fault, I still get clients expecting me to guarantee the whole set after it has been repaired. Basically, they expect me to take responsibility for every aspect of their set’s subsequent performance – including self-inflicted faults.

Of course, that’s impossible – it’s like asking a mechanic to guarantee all subsequent mechanical faults in your car after he’s repaired the brakes.

This clown’s set previously had several dry joints in the deflection circuitry, which caused all sorts of sync and horizontal scanning problems. This time, however, he wanted a house call because he had a chrominance problem.

When I arrived, he triumphantly demonstrated that there was a chroma delay problem, resulting in monochrome leading edges and colour-smeared trailing edges. This definitely was nothing to do with my previous work and had not been present when I had re-delivered the set 90 days earlier – as I was quick to point out.

Mr "know-it-all" had the set connected to a brand-new set-top box, so I began by checking the set’s own OSD menu. This gave crisp, clear colour registration. Next, I connected an indoor aerial and tuned in a terrestrial station. That too came in without the colour delay.

Finally, I demonstrated it with a video cassette and DVD. And again, the pictures on the set were perfect.

By now, it was clear to me that the problem lay with the set-top box – which was still under warranty – and had nothing to do with the TV. It wasn’t so clear to the client, who felt that it was his God-given right to call me out on this for free.

The British Toshiba

George Barrington-Smythe was fresh off the boat with his brand new Toshiba "telly" from the Old Country. This was a Toshiba Active Vision CT-
90101 model 36ZP48P(B) with a Mitsu-
bishi W86LXK003X30 picture tube.

What poor George hadn’t realised is that Toshiba TV sets in the UK are entirely different from the models imported from Japan by Toshiba’s agency in Australia (Castel). Nor did he realise that the Earth’s magnetic field is significantly different Down Under – at least, in terms of its orientation.

Still, he tackled the inevitable col-our purity problem he was getting with true traditional British stiff upper lip.

He had already had another technician around to his flat, who degaussed the tube. Unfortunately, this had only helped to reduce the symptoms slightly. Furthermore, this technician had now gone home to China as prospects are better there for TV technicians! How times have changed!

Anyway, he phoned me and asked what could be done, "my good man". I explained that the Earth’s magnetic field, being different in the southern hemisphere, was the basis for his problems. I also explained that problems could be caused by other magnetic sources, like the steel RSJs in his unit or even hifi loudspeakers and little kids with high-power magnets.

Most of this stuff he dismissed but I did suggest two practical plans of action to prove my claim and possibly fix the problem. First, I told him to rotate the set through 90° relative to its current position and see if the colour patches moved. And second, I told him to find the geometric adjustment in the set’s on-screen menu, as this can also affect the colour purity. He duly followed both suggestions but neither gave any real improvement to the colour purity.

In the end, I said that there was nothing for it but for him to bring this 76kg British set into my true-blue workshop. There, I could check the set’s own degaussing circuit to see if that was OK. If that failed to turn anything up, the picture tube would have to be rotated through 180° – ie, mounted upside down. I explained that the latter would be expensive in terms of labour but our man was made of sterner stuff and brought the set in.

The problem was quite evident when it was on my bench, the set displaying prominent colour patches. I unplugged the degaussing coils, switched the set on and degaussed the screen manually with my wand. I also tried the geometric menu myself but nothing would shift the patches.

I was now faced with a dilemma – should I first try to realign the deflection yoke and purity magnets or just dive straight in and rotate the picture tube? I settled on the latter. If it didn’t work, I could then redo the set’s purity.

The logistics of turning the tube upside down are considerable and require military precision. First, I needed at least four telephone books and an assistant to help with the lifting. Next, the back and chassis were removed, after which we carefully lowered the screen so that it sat face down on top of the telephone books.

That done, the 10mm coach bolts were removed and the front of the cabinet lowered gently to the bench around the books, leaving the CRT. We then rotated the tube 180° and lifted the cabinet back up and refitted the bolts.

The next problem I had was extending the degaussing and geomagnetic coils to reach their respective sockets. I also cut and resoldered the deflection yoke wires, so that the picture would be the correct way around, before refitting the chassis.

And that was all that was required to resolve the purity – no further adjustments were needed. All that remained was to extract the money from George Barrington-Smythe’s wallet.

Dead Hitachi

I was called out to a dead 3-year old Hitachi on the third floor of a home unit. It was an Hitachi C29F300B (V3AL1 chassis) but Mrs Tracy had it in the middle of a huge wooden cabinet in her lounge. The set was chosen because it almost exactly fitted the cabinet hole and while it looked great, it was a nightmare to fix because of the logistics.

For a start, the power point was unreachable without pulling out the entire unit. The TV set then had to be partly eased out of the cabinet and swung around, with the added constraint of all the AV leads that were connected to the rear.

Eventually, by half-balancing it on the edge of the wall unit, I managed to get it into a position where I could remove the back. It was at this moment that the battery in my trusty Panasonic electric screwdriver chose to go flat and I didn’t have a spare. This meant that I had to manually remove all the screws (and there’s lots of them), while balancing the set with my knee and other arm. And to make it even harder, the screws are black and are recessed into invisible deep black holes.

Once the back was finally off, it didn’t take long to find that the main fuse (F901, 4A) had blown violently. I unplugged the AC power before replacing the fuse and then reconnected it before switching the set on. Much to my surprise, the set fired up and everything seemed OK.

That bothered me because this fuse didn’t blow without reason. Of course, it could have been a power surge but nothing else had been damaged in the unit. Anyway, the set was working now so there was nothing could I do but replace the back and push the set back into position.

This process took some time but eventually the job was all done. I then switched the set on to check that all was well and the fuse immediately blew again! You get days like this!

There was nothing for it but to start over and track down the real fault. There were several possibilities but the most likely suspect was a faulty degaussing coil PTC.

After removing the back again, I unplugged the degaussing coils and fitted a new fuse before switching on and off several times. So far, so good – the set was still working OK (or was it a bit like the guy who fell off the Empire State Building and had fallen 20 floors – so far, so good!). I then reconnected the degaussing coils and switched on again – the fuse blew immediately.

The only way I could repair this set now was to take the chassis to the workshop and order and fit the correct PTC thermistor (TH901, Part No. 2341325) and then bring it back. And this time it all worked.

Of course, the diagnosis and repair was relatively simple compared to the logistics of the exercise. As far as I am concerned, I mainly earned my money for patience and persistence, rather than the repair itself.