My first story this month concerns a Philips Matchline 29PT6361/79R. This particular set was manufactured in China in 2000 and uses the A10A chassis (the Asia-Pacific successor of the A8 chassis). It is a very sophisticated, high-performance set but can suffer from intermittent faults attributable to its Small Signal Panel (SSP) or SSB (Small Signal Board) - part No. 3139 178 6780.

The SSP is not really a repairable item for the faint-hearted (such as yours truly), as it has three miniature microprocessors. However, it is possible to change the 8-pin EEPROM IC (7066 M24C32-WMN6).

Getting back to those miniature microprocessors, IC7064 (SAA5067) is called the (ARTISTIC) PAINTER and is a control microprocessor with 100 pins. IC7301 (TDA8885) is called BOCMA and is a video microprocessor with 64 pins. And IC7651 is called MSP (for Multi-standard Sound Processor) and also has 64 pins. These are mounted on double-sided PC board measuring 150 x 100mm, along with eight other surface-mount ICs.

You need a lot of special resources and talent to repair these babies!

Items Covered This Month Philips Matchline 29PT6361/79R TV set (A10A chassis) Sanyo CPP3002-00 TV set (A3-A4 Series) Hitachi C28-P500R TV set (G7P chassis) Philips 33FL1880/79R TV set Philips 21PT118A Anubis SF TV set Panasonic TC33AV1 TV set (M16M chassis)

To find out whether the set is faulty, you can enter the SAM (Service Alignment Mode) by entering 062596 on the remote control, and then pressing the "OSD" (On Screen Display) button marked "i+"). This mode allows you to perform alignments and change option settings.

To get into the SDM (Service Default Mode) you punch in 062596 on the remote control again, followed this time by the "Menu" button. And to get into the CSM (Customer Service Mode), you press the MUTE key and any of the top control buttons on the TV simultaneously for at least four seconds.

The opening menu will display the set's operating hours (ie, how many hours it has been on) in hexadecimal - eg, 18H = 00011011 (binary) = 27 hours (decimal) - and the Software Identification of the main microcontroller. The second line shows the Error code buffer and this contains all the errors since the last time it was cleared. To clear the buffer, you activate "CLEAR ERRORS" in the SAM menu and exit via the "STANDBY" command.

Using this tool will let you store all the errors and can help in identifying intermittent faults, even when you are not there. Each code is listed in the Service Manual (and you can ignore Code 17!).

Anyway, all this is to give some background to a nasty fault encountered by a colleague of mine. The set came in with no picture but the sound was OK. The client said that the picture had been "getting pinker" before it disappeared altogether. By turning up the screen control on the flyback transformer, the screen showed a fully scanned grey raster with retrace lines but no OSD (On Screen Display).

Because of this, it wasn't possible to check the error codes in the buffer. Not to be outdone, my colleague was fortunate enough to have another similar set in for repair which he had just completed. He swapped over the SSP and was able to read the error codes on the other TV. These were 17, 23, 7 & 6 and of these, only error code 7 was pertinent. The description for this code is "BC-loop (Black Current) not stabilised" and it lists the possible defective components as the RGB amplifier, RGB guns or the RGB driving signals of the BOCMA (BIMOS One-Chip Mid-end Architecture) high-end video input processor (IC7301).

The BC-loop is part of the CRT drive circuits. In operation, the drive to the picture tube is continuously adjusted to prevent visible aging of the CRT and give "perfect" pictures. This is called "Continuous Cathode Calibration" and is achieved by comparing and monitoring two levels (Hi and Lo) of point black level stabilisation for each gun and altering the drive accordingly. The maximum current allowed is 100μA and this is fed to pin 30 of the BOCMA IC.

Having already eliminated the SSP and BOCMA IC, my colleague was left with the CRT socket and drive assembly and possibly the beam current limiting circuitry from the flyback transformer.

By measuring the voltages on the CRT Panel Board B, he found the drive voltages from the BOCMA to the RGB amplifier (pins 1-3 of IC7830, TDA6108Q) were all low by about 1V. In addition, the outputs to the CRT cathodes were all too high (by 50V) at 190V or so. Furthermore, the critical CUTOFF control voltage from pin 5 to pin 30 (BLKIN) of the BOCMA (Black Current Loop) measured 7V instead of 5.6V.

All the other voltages were substantially correct (considering there was a fault). My colleague replaced the IC but it made no difference. He then did a blanket check of all the pertinent components on the board (obviously not the additional parts involved with the SCAVEM circuit) with his digital multimeter but nothing showed up. By now, he was beginning to suspect the picture tube.

Next, he checked the signals with an oscilloscope. He found that RGB waveforms were arriving at the IC but nothing was going out to the cathodes.

My colleague then removed the CRT socket from the tube and re-measured the waveforms. They were now all reaching the cathode pins (8, 6 and 11). Finally, he found that pins 11 (blue cathode) and 12 (GND) on the CRT socket were short circuit.

That was when he asked me for my opinion. He was 99% sure that there was a cathode short inside the picture tube which, because of the cost, meant that this set would have to be scrapped.

First, I measured the CRT aquadag voltage to be about 12V and this is about what I would expect from a Philips TV with beam limiting. Next, I did my non-recommended method of checking CRTs (much to my colleague's disdain and concern for his new video output IC) - that is, I momentarily shorted each cathode in turn to ground. This gave intense colours for each gun.

I told my colleague that I thought that the CRT was probably OK and that I would bring my CRT analyser in the next day to confirm this. His response was "Well, what about the undoubted short inside the CRT gut?" but I didn't have an answer for this.

The next day, I brought in my ancient SWE-Check CRT analyser (OK, so all three of us are old). My colleague laughed when he saw this prehistoric piece of apparatus and nearly wouldn't let me put it on this 2-year old TV. I assured him that all I was going to do was check the emission and for shorts and promised not to blow the CRT up.

With bad grace, he finally allowed me to use my Heath Robertson "divining-rod" to check his pristine telly. Anyway, the good old SWE-Check analyser with the modified adaptor I had made showed there were no shorts at all and that the emission was excellent on all three guns. The cut-offs were spot on too and this was the same at 6.3V, 7.3V and 9V true RMS on the heaters (I wasn't game to go higher)!

My colleague still wasn't convinced, however, and pointed to the undeniable short on the CRT. I examined this very carefully and noticed that the short was deliberately welded inside the gun! Finally, it dawned on me what we were doing was wrong.

We were measuring the two adjacent pins on the righthand end of the tube socket which, when compared to the PC pattern for the CRT socket, looked as though they were pins 11 & 12. However, they were in fact pins 12 & 13, although the latter is not marked on the PC board. These are indeed both grounded and if we had thought about, it would be impossible for the blue cathode to measure +190V if it was indeed grounded!

Finally, to prove it wasn't the picture tube, my colleague plugged the suspect CRT into another working TV's circuits by placing the two sets back to back. And that finally proved that the picture tube was indeed perfect when it produced a good picture.

So where did that leave my colleague? Well, he wasn't able to substitute the CRT socket board (B) as none of the other sets he had at his disposal was identical. He had eliminated the SSP and the CRT and measured everything else, so having "done my bit", I left him to solve the puzzle.

Three days later I returned to find that the set had been fixed and returned to its owner. My colleague proudly showed me the offending part, a tiny glass diode (BAV21, D6633) from the CRT board. This diode is a clamp between the +200V supply and the green cathode.

"Well", I said, "how come you missed this when you did the first cold check of the CRT panel?" He gave it to me to measure, confident that it would check out OK.

Well, of course, I upset him because I produced my old Excalibur - my ancient analog meter which uses a 9V battery. It clearly showed the diode to be leaky in both directions (75kΩ) on the 100kΩ range. He then showed me how he had measured it originally, using the diode checker range on his "U-Beaut" DVM - it quite clearly measured OK on this range.

My colleague only discovered that it was in fact faulty when he later rechecked the part out of circuit on the high ohms range of his DVM. I think the world is just getting too high-tech for its own good!

The ancient Sanyo

Mrs Eva Ruddock is an elderly widow and thanks to a lifetime of paying taxes, now spends her days living in a "matchbox". Her telly is an ancient 34cm Sanyo CPP3002-00 (A3-A4 Series) and doesn't even sport a remote control! However, Eva reckons that she doesn't need a remote, as her room is so small!

Anyway, her beloved telly died the other day and she wheeled it all the way up to our workshop and asked me ever-so-nicely if I wouldn't mind fixing it.

Now, how could I say no? The set wasn't worth fixing but it was obvious that she really couldn't afford a new one. I told her I would see what could be done - what the hell, how hard could it be to fix this? It wasn't exactly the latest in high technology and that should make it easy!

And so I stuck it on my workbench and checked it out. There was no picture but the sound and On Screen Display, both on TV and AV input, were fine - all I was missing were luminance and chrominance.

This set uses two large ICs - a microprocessor and a jungle IC (IC101, LA7680). A check with the oscilloscope showed that plenty of luminance was arriving at pin 38, chrominance at pin 40 and sync signal to pin 33. However, nothing was coming out of pins 24 (Y), 23 (B-Y), 22 (G-Y) and 21 (G-Y). I also checked that the +9V rail was on pins 11 & 13 (Vcc).

When I fiddled with a pot at the rear of the set marked "Video Amp", I noticed that the set would occasionally produce a poor negative-looking picture on the screen in TV mode, with no horizontal or vertical sync. Unfortunately, I didn't have an accurate circuit for this set. However, I did have a poor photocopy of a later version (CPP 3012) but this has a lot of extra circuits that the CPP 3002 doesn't have.

However, I had repaired dozens of these sets over the years so why would I need any help?

I started checking for the more common faults, starting with the main B1 rail which was spot on at +130V. I then checked favourites such as R232 (Beam Limiting), D801, D731, R452, C402, C401, C232 and C233. That done, I measured all the voltages around IC101 and got involved in all sorts of mental anguish when I discovered that the voltage on pin 30 was considerably less than 7V.

I finally got over this "furphy" when I discovered that I could raise this voltage by playing around with the video input (besides which, in this model, it is only connected to R422 and C424 and these were OK).

At this stage, I couldn't make my mind up as to which area the fault might be in. Was the no picture due to no sync or the no sync due to no picture? Unfortunately, the block diagram inside the jungle IC was almost illegible and not very accurate.

Using an oscilloscope, I established that the line pulses on pin 26 from the flyback transformer and the vertical timebase were spot on. And there were horizontal and vertical pulses arriving at the microprocessor (IC701). I even got out the frequency counter and checked the crystals.

This was now getting incredibly frustrating. This was meant to be an easy repair on a well-known set that was now worth nothing. But damn it all - it was now eating at my pride and I really wanted to know what it was that was beating me.

Having tried the proper "high-tech" approach with meters, oscilloscopes and frequency counters - not to mention capacitance meters - I decided to try the old wet finger trick which, I should emphasise, you should never try yourself unless you are experienced and know exactly what you are doing (get it wrong and you might need repairing as well!). Anyway, I ran a wet finger up and down between pins 40 & 25 of IC101 (the maximum voltage here is only 9V) to see if any difference could be observed. And would you believe it, one wet serviceman's pinkie on pin 33 (only) resulted in a perfect colour picture.

Well, how could such a caveman's approach do this? It amazed me too, so I tried to substituting an electronic component for my finger. The "equivalent" turned out to be a 100kW resistor from pin 33 to ground!

After reconnecting the CRO and monitoring the sync input to this pin, I realised that all I doing was attenuating the voltage input a fraction. The voltage should be about 7V but this unit gave 7.2V without the resistor and 6.7V with the resistor in place.

Armed with this information, I went back to the video input divider circuit to the IC and found that R200 (10kW) was open circuit. I felt sure that this was the culprit and replaced it but it only made a very marginal difference and the old wet finger still gave the best result - and I wasn't about to donate that to my client!

Working back along the circuit, I eventually came to a video emitter follower stage that should have had 6V on its emitter. However, it was reading 6.7V and this seemed to be the only source of bias into this circuit. The base measured high as well, while the collector read over 13.9V. This was far too high.

Tracing it back further, I found that this voltage comes from an IC regulator (IC551, 7812). I checked the input to be 15.9V, so obviously it wasn't doing its job. Replacing this IC fixed the voltage, along with the rest of the set. And that, as they say, was that.

If I had had a better service manual, I might have taken more notice of the small variations in voltages - but it should be remembered that when there is a faulty component, it can affect lots of different circuits and give erroneous voltages and waveforms. The trick is trying to decide which are the critical ones. I must admit that I didn't think a figure of 0.3V difference on pin 30 of the jungle IC would have such consequences - especially as the waveforms didn't change much.

I didn't have the heart to charge Eva and she was stoked.

A tricky Hitachi

Problems associated with no picture have to be quickly divided into those that have OSD and those that don't.

The OSD is normally sourced directly from the microprocessor character generator and requires a clock, data and vertical and horizontal pulses to generate blanking pulses. This is normally fed into the jungle IC but on some sets can go directly to the CRT panel.

Recently I had a 1988 Hitachi C28-P500R (G7P chassis) where both the jungle IC TDA3562A IC501 and the microprocessor 50442-552ST IC001 were faulty, the former giving no picture and the latter no OSD. Without OSD, it's difficult to navigate through menus unless you are very familiar with the set. In this case, the microprocessor was no longer available but because the client had had the set from new, he was sufficiently familiar with it to continue to use it.

A difficult Philips

I have had cases (Teac & Sanyo) where the vertical pulse has been distorted by a fault in the vertical timebase that has also resulted in no picture or OSD. Currently, I am tearing my hair out (again) with a perplexing Philips 33FL1880/79R which has OSD plus picture in the PIP (Picture in Picture), as well as Teletext on Channel 7, but no picture on the main screen. I have traced the signals into one IC, TDA8443/C4 IC7395, but nothing is coming out the other end.

There are no error codes in the Service Diagnosis Mode, all pulses and waveforms into the IC are correct and the IC has been replaced. The problem is the same for AV as well as TV. I am waiting now until another set comes in and I can swap some of the modules. I suspect that the 100Hz "high-end" module may be the culprit or even possibly the 28-pin EEPROM. I'll let you know when I get to the bottom of this!

Another Philips set

When faced with a "no-picture" fault in a modern set, it can be very difficult to make a diagnosis backed up by measurements as everything is buried in large scale integrated (LSI) circuits. In fact, I have just finished a Philips 21PT118A Anubis SF set with exactly this problem.

I started at the CRT board and found that there were no signals coming in. Not wishing to repeat my colleague's mistake with the Philips Matchline, I thoroughly checked all the components before deciding that it must be a faulty jungle IC (TDA8360E, IC7225). Fortunately for me, it turned out that this was indeed the culprit, especially as the replacement wasn't cheap.