I recently received a call from our local RetraVision store, asking me if I would have a word with one of their customers who had a faulty TV. Naturally, I said yes (I need the money) and after recording her details, I asked what sort of set she had. She wasn't really sure but thought it was a 2½ - year old Sony!

After that response, there was no point in asking what the model number was, so I asked what the problem was.

According to the customer, the set's picture would "intermittently go white" and then the set would switch to standby. I asked her to explain what she meant by "intermittently" and was told that sometimes the fault would appear after about one hour while on other occasions the set would last for most of the day.

Unfortunately, this type of intermittent fault cannot readily be fixed in the home. That's because the fault might not show when I called and even if it did, I couldn't possibly guarantee that the problem had been fixed without prolonged soak testing. So in a nutshell, it had to go to the workshop to be repaired.

The customer accepted this and so we arranged that I would call and pick the set up the next afternoon. And because it was an 80cm set, I asked if there was anyone there who could help me load it into my station wagon or would I have to bring someone with me. To my surprise, she said that she would help me with the set and wasn't even the slightest bit fazed even when I told her that it could weigh up to 80kg!

As far as she was concerned, it wasn't a problem. She worked as a nurse and was able to lift patients, so she could easily manage this!

Items Covered This Month Sony KV-29X5K TV set (FE-1 chassis). Philips 32PW9631/05 Widescreen TV (GFL2.3E chassis).

A unique cure

The next afternoon, I dutifully arrived and reversed up her steep drive and opened the tailgate of the wagon in anticipation - mentally noting the number of steps up to the verandah. However, when I knocked on her door, she said "you are going to be disappointed - my husband has decided to try fixing it himself first. If it's still giving problems after two weeks, then you can pick it up".

Naturally, I was rather peeved by this. At the very least, she could have rung and told me this before I drove out to pick the set up - service calls are expensive.

Anyway, I was intrigued as to how her husband intended to fix the set and she showed me into her living room to see what they were doing. The set was a fully imported model they had bought with them from Czechoslovakia and it was a Sony KV-29X5K - a model I had never seen before (possibly a BE3D chassis?)

Anyway, their approach to solving this fault was to operate the TV completely upside down on their lounge room floor! I wished them good luck and left, grateful in the end that I had got out of the job.

Later, I speculated on just what they were trying to achieve by turning the set upside down. In the end, I could only surmise that because the set was originally from the Northern Hemisphere, they thought that rotating it in the Earth's magnetic field might fix the problem. However, this is for a colour purity problem and would not affect this particular fault. And in any case, a 2½ - year set would probably have a "Terrestrial Magnetism Correction System" (TMCS) fitted (available since February 1998) to compensate for any geomagnetic effects on the tube's shadow mask.

In greater detail, TMCS is a relatively new technology that was developed to overcome the effects of regional (horizontal) magnetic field variations on the picture tube. In Europe and the USA, for example, the average field strength is about 35mT (milli-Teslas) compared to 30mT in Japan and 40mT in Malaysia, Singapore and Hong Kong. However, large regional variations can exist (eg, due to geological factors) and even the presence of a lot of steel in a building can have a localised effect.

In fact, if the set is located in an area with a strong local magnetic field, you can see the effect change the on-screen colours as you move or rotate the set around. Because of this, all TV sets are fitted with an automatic degaussing circuit to demagnetise the shadow mask in the picture tube every time it is switched on from cold (it takes about 20 minutes for the dual posistor to cool down again).

The TMCS system takes this a step further. It relies on a geomagnetic sensor (MIU-102-N) which detects the horizontal magnetic field and varies the DC current through a beam-loading correction coil (LCC). This coil is placed near the tube and corrects the purity using Fleming's lefthand rule. In addition, a separate "Rotation Coil" is fitted and this corrects image rotation and horizontal trapezoidal distortion.

The TMCS in some models is automatic but in others you can adjust it with an on-screen menu until the colour purity is correct.

Anyway, predictably, the Sony TV didn't respond to the upside down treatment and eventually arrived at my workshop. This time, there was an additional rider on the fault symptoms. Previously, it would go white then fold to a horizontal line before cutting off. However, it was going to a vertical line before cutting off inside one hour of being turned on (these being the symptoms for vertical and horizontal timebase failure respectively).

The first problem I had was to try and get some service information on this set, so I surfed the net for answers. Eventually I found out that SCC-QO3A-A translates into a 1998 68cm FE-1 chassis made in Spain.

Service mode

Next, I discovered how to get into the "Service Mode" - ie, go to the Standby Mode and press the i+, 5, vol+ and TV buttons on the RM-883 remote control. This is slightly different from most Sonys, which substitute the "TV" button in this sequence for the "Power" button. You would have thought the set would switch on as soon as you pressed "5" but it doesn't in this sequence.

From the on-screen data menu, I discovered the software to be v5.54-00, the factory data FFh6Ch, the destination K and the text language as EAST (TV Status), as well as Status Reg 1 = 11001101 and Status Reg 2 = 10000001 (Technical).

The instructions said that there is a Test Mode 2, which is available by pressing "TEST" button 13 twice on the remote to produce "TT On Screen Display". The only problem is that there isn't a "TEST" button, nor "13". I tried 1 and 3 together and in sequence but it made no difference.

When you enter the Service Mode, the following OSD message appears:

1-AY-C-TT- -
- - - - - - - -K E

and if you key in any two numbers, they will appear after the TT accompanied by a change in the picture (eg, 13 makes the picture darker). "MENU" enables a Test menu and to get into the diagnostic mode, you press i+, 5, vol- and TV. This produces a table of errors and their frequency, plus the set's on-time duration.

In this case, I had 69 x error 2 (protection circuit trip), 24 x error 5 (AKB), 1 x error 8 (jungle controller, no I²C acknowledge at Power On), 3 x error 9 (Tuner, no acknowledge at Power On), 3 x error 10 (sound processor, no acknowledge at Power On). The set had been running for 008905h 49m since 1998.

I cleared the error buffer with 80 and then switched the set on and waited for results. Naturally, I wasn't able to sit there for an hour, waiting for the fault to appear, so I left it in a prominent place in the hope that I would observe something when it did happen. Of course, I was somewhere else when the fault finally did occur. I returned to find that the set was dead, with the front-panel LED flashing twice every second.

I turned it off and on again and it switched straight back onto a perfect sound and picture, the Service Mode now showing an "Error 2".

Next, I removed the chassis and examined it carefully for anything obvious before thoroughly resoldering it from "stem to stern". That done, the set was reassembled and placed on a prolonged soak test.

The set then worked perfectly for a whole week. But then just as I had quoted for the repair and made an appointment to return the set, the fault (error 2) reoccurred. This fault then reoccurred three more times, each over four hours apart.

My next line of attack was to monitor the +135V rail with a digital multimeter that has peak-hold to see if it increased over the next few hours. However, this didn't reveal anything as the meter kept timing off but I was fairly happy that this wasn't the problem area.

In the meantime, the client, on my suggestion, came up with an unofficial modification they had got from somewhere in the old country. It involved cutting the ground track to one of the filaments of the CRT, both on the CRT board C and the deflection board (A) around the flyback transformer, thus completely isolating the picture tube heaters and apparently "discarding its overcurrent protection" - though I don't quite see why in this case.

The LED flashing twice indicates a fault in the secondary excess-current protection in the line output stage. If the current increases across R572, the extra voltage (1.2V) switches on Q571 and Q574. The resulting signal is then applied to pin 18 of the jungle chip (IC301) and pin 52 of microprocessor IC001, which switches the set off.

Apparently, if there is a flashover or irregular short inside the CRT (eg, a heater to cathode short), the voltage across the filaments increases with respect to ground. This causes excess current to flow through the flyback transformer and switches the set off as described above.

Anyway, I did the modification while not expecting too much but was pleasantly surprised when it was still operating over 72 hours (continuously) later. The set was then returned to its owner with the rider that the picture tube could be on the way out.

Subsequently, I have found similar modifications in some Sony TVs to desensitise the current protection circuit by changing the 1Ω feed resistor (R572) to the line output stage to 0.56Ω.

I think it would be a great help if I could just get a copy of this circuit from Sony.

Wide-screen Philips TV

A large wide-screen Philips TV greeted me one Monday morning, as I stared at my once vacant bench sipping my coffee. How it got there over the weekend I don't know but I suspect Mrs-she-who-has-to-be-obeyed had a hand in it.

Anyway, this beautiful state-of-the-art (1995) 76cm Philips 32PW-9631/05, fully imported from Belgium and expensive, was dead. Could I do a Lazarus on her?

I removed the large back to be confronted with a pile of electronic circuitry that was being eaten alive by that dreadful cancer called salt corrosion. Still, on the bright side, it wasn't completely dead - the front stand-by LED was flickering.

The chassis number printed on the back said S5GFL2.30E which I translated to be GFL2.30E.

My next step was to try to determine how far the power was getting into the set but the chassis was unfamiliar and the access wasn't the best. I ordered a service manual but unfortunately it arrived with two pages missing which, as it turned out, contained critical information.

After I had discovered the concealed screws and clips (and breaking only a few), I managed to lower the large turret on the right (Audio Amplifier A Board, IO Dolby U Board, AFU Z Board, and ECO Low Power Standby AU Board). I also discovered how to lower the Service Selection Board I.

That done, I tackled the "Large Signal Panel" where I discovered that the +141V was cut off from the line output transistor via the line switch FET (7480). This, in turn, is controlled by the "Slow-down" and "Fast-down" protection circuits, the former coming from the "Small Signal Panel" (SSP) via plug S/L11 (pin 21). However, all was revealed when I examined the small signal panel. This board is double-sided with very fine tracks and it was very corroded. Similarly, the Teletext board (AQ) and the Source Selector (I) were also corroded.

I removed the SSP and saw that it was probably beyond economic repair. I then tried to obtain a new or exchange board from Philips but it is no longer available. And so, after a "heart-to-heart" with the customer, and a lot of crying on their part over the expense, they agreed that they would pay me to repair this panel.

First, I removed it and washed it in hot water to remove the salt. I then use Nifti to remove the dirt and grease before thoroughly drying it. That done, I washed it in "metho", after which the board was thoroughly cleaned using a professional PC-board cleaner solvent. Finally, I left a very fine spray of CRC 2-26 all over the board surfaces which I then blew off with an air-compressor.

Next, I got out my continuity tester and checked what felt like thousands of interconnections across and through this double-sided board, fitting links where necessary. The two worst were the corroded connections from S31 pin 1 (+5V standby) to the I/O board and the Power-On Reset line between Q7214 and R3205. Once I had repaired these two, I was able to power up the set and get a picture!

However, the picture was very poor with no sound. There was east-west distortion and the picture-in-picture (PIP) was barely working.

To diagnose this problem, I first connected service pins 2 and 1 on socket S98 on the SSP to get into the "Service Alignment Mode". This gave all sorts of information, including the Operating Hours, ID, Option Codes and Error Codes, plus alignment details. It also reported Error 013, IC7814 TMP47P443VN, which is on the I/O module (I).

Unfortunately, this was the start of a huge amount of time wasting from misinterpreted data. I followed up each of the remaining four faults, looking for possible shared causes like missing power rails and corroded tracks, etc but got nowhere.

Sound problem

The "no-sound" problem should have been the easiest fault to locate. Indeed, it started well when I traced the sound from the tuner (F) to the stereo decoder and from there to the sound processing circuitry. I was then able to trace the sound via the source selection switching and all the way to the AFU module (Z).

Personally, I thought I was particularly brilliant to do this, considering very few of the circuits in the service manual I had exactly corresponded to the circuits in the set. But now I was really stumped because I had nothing on the AFU (Audio Feature Unit) at all, as the pages were missing. That really made things difficult, as this is a large module containing eight ICs for Dolby Surround Sound before continuing onto the "Dynamic Enhancement" circuitry and audio amplifiers (plus headphone amplifiers).

As for the other problems, the east-west circuitry was only just working and the digital software controls had little effect. And the picture-in-picture was not centred and only half there.

I was about to give up in disgust after all the effort I had put into this repair when I had a bit of luck. Stuck on the picture tube are the option codes for each variation of this set and I noticed that the numbers didn't correlate with the model I had on the bench. Initially, I didn't think that this meant much but, lacking any other ideas, I wrote down the ones programmed into my set and then programmed the numbers from the label into the set.

And would you believe it? - after storing them and restarting the set, all these faults were fixed simultaneously, apart from a little fine tuning. Perhaps this set is just a little too smart for its own good!

Dealer service tool

Finally, the service manual makes frequent references to a "Dealer Service Tool". This is a very special 2-way remote control (RC7150) that can interrogate and diagnose dead sets! And it can automatically program them as well! However, this tool is just too expensive for me and is not readily available.