Items Covered This Month Philips 33FL1880/75R FL1.1S TV set. Philips Series 900 KR684 KL9A3 TV set. Philips 29PT6231 (A8.0A chassis) TV set. Philips 34PT6361/79R (A10A chassis) TV set. Fender Hot Rod Deluxe guitar amplifier

I've written before at length on the Philips 33FL1880/75R (FL1.1S chassis), a set I am not impressed with when it comes to servicing.

Recently, I had one come in to the workshop that really baffled me. After a power surge, Mr Phelps' set would no longer show a picture on any channel or input. However, the sound was still present, along with the on-screen displays and the Teletext (Supertext on Channel 7). And, strictly speaking, there was a picture but only in the small PIP (Picture-In-Picture) mode - it wouldn't transfer to the main picture. The brightness and contrast controls all worked.

After removing the back, I could just reach service pins S23, S24 and S25 but there were no error codes displayed on the SDM mode. I had a couple of 29FL1880/75R sets in the workshop at the time and I swapped as many modules as I could but the few that were compatible made absolutely no difference.

Next, I hooked up a colour bar generator to the AUX AV SCART input and fired up the oscilloscope. I then had to balance the chassis very precariously on its front edge before starting to trace the video signal.

I followed the signal from pin 18 of IC7219 (TEA6414) to pin 15 of IC7365 (TDA4650), where it splits into Y, R-Y and B-Y signals before going into IC7366 (TDA4660). From there, the signals are fed into video control IC7395 (TDA8443A) but that's as far as they went. There was no output on pins 19, 20 or 21 to the "100Hz High End Box" (the 100Hz High End Box steps up the 50Hz field rate to 100Hz and includes noise reduction circuitry).

This seemed like a good start and so I proceeded to analyse this IC in depth. I checked all the voltages to all the pins and every one except the three outputs was spot on. Next, I checked the sandcastle input to pin 24, then checked the SCL and SDA lines to pins 13 and 14 for digital noise before moving on to check the Frame Blanking to pin 3. Thus far, I was drawing a total blank but I couldn't find any signal so I spent a great deal of time following this back via Q7183 to IC7258 (HEF4094), which is an 8-bit shift resistor and output latch.

Eventually, I worked out that the blanking pulse was used to switch this IC from RGB input to colour difference. In disgust, I then replaced IC7395 but it made no difference.

Well, if it wasn't the IC, perhaps it was the load. The output goes via a limiting resistor into the 100Hz High End Box module, where it feeds a series of surface-mounted transistors. These are biased on and off by IC7210, an HEF4066 quad bilateral switch. I replaced this IC but that didn't help and I couldn't fault the circuit. I then tried running the set without this module plugged in but it wouldn't work.

Next, I tried heating and freezing around this area and noticed that when it got very hot, a very faint picture appeared. However, all this achieved in the end was to damage the components on the module itself.

By this time, Mr Phelps was impatient for a report on the set's progress and what it meant for him. I wasn't in really much of a position to advise accurately, so I could only advise him that, in terms of labour and parts, it was going to be expensive. At his insistence, I gave him an approximate figure for the repair and this enabled him to go back to his insurers and make a claim (this was not the only appliance to be damaged by the power surge).

The figure I supplied, along with a note about parts availability, was enough for Mr Phelps to get a new set. And in the process, I scored the old one.

A break at last

All I was really interested in was finding the cause of the fault. I put the set aside and started asking around. I interrogated the web, Philips and other service centres but I was getting nowhere until a competitor offered to lend me a similar set which was in his workshop for repair. It was giving a very dull picture and he thought that the picture tube might be going "flat".

Perhaps this was the break I needed! I soon found that the cause of the problem in his set was due to the wrong flyback transformer having been fitted, the incorrect part giving low EHT and low heater filament voltages (a 29xxx version had been used). I also replaced C2523 (8.2nF) and C2504 (470pF) just in case. The fact is, the confusion surrounding these sets is due to the poor service manuals that are made available. Nothing in them is clear, especially the differences between models, a lot of information is missing and there are outright errors in some places.

Once his set was going, I could start swapping large chunks of it without worry. (Swapping many boards with the 29xxxx version just didn't work as there are just too many differences). I had already swapped some of the smaller modules, along with the microprocessor and several ICs (after mounting IC sockets), especially the suspect video control ICs. Nothing made any difference until I swapped the EEPROM (IC7137, X24CO4P1) and High End Box module together, when suddenly there was a main picture. At last!

Next, I went into the Service Mode (by shorting S23 and S24) to check the option codes. Option Code 1 was 010 (= FQ816/MS Multi-system type of tuner and PIP module fitted) and Option code 2 was 005 (= NTSC(1) and 100Hz high-end box(4) fitted). The option codes for these sets are not published as far as I am aware and the lists of options are difficult to comprehend.

I then refitted the old EEPROM which I had already tested in a 29FL1880 set (FL1.0S chassis Option 1 = 010, Option 2 = 017 - I think). That done, I punched in the new option code numbers before storing them into memory with the "Personal Preference Store" key on the front panel of the set. This too restored the picture.

My problem in relating this story is that the repair was done in dribs and drabs over many months and I can't remember exactly what the original option codes were when the set came in. However, I think they were 026 and 001, which in the absence of definitive data might have been correct. The additional "4" for the 100Hz High End Box would account for why the video control IC was switched off and gave no output.

As mentioned before, the original 100Hz High End Box had been damaged due to the excessive heat I had subjected it to. This double-sided board was hardly a delight to fix - the metal work plus the location of the module make access extremely difficult.

Anyway, I eventually discovered that R3210 (22Ω) was burnt up because of a short in IC7400 (TDA2579B) and surface-mount transistor 7104, which feed off the 13V rail. The latter fault gives a no-colour symptom.

Still in the woods

Unfortunately, I am still not completely out of the woods, as the set is now showing (a few months later) intermittent error codes at switch-on from cold only - and sometimes not even showing a picture. The new error numbers include 15 and 10, neither of which are listed in the service manual, and also error 09, which is IC7430 (TDA4680) which has already been replaced. In addition, there is error 05 which is IC740 (the SDA9088 PIP processor) and error 08 which is IC7324 (TDA4670), both of which appear to work properly.

I checked the +5V V START and reset rails and changed C2071 from 33μF to 330μF but unfortunately the problem still remains unresolved. It all goes to show that we now have to cope more and more with both hardware and software faults.

Another Philips set

In total contrast, I also had a 1986 Philips Series 900 KR684 KL9A3 TV come into the workshop. It too had no picture - the raster and sound were OK but the set actually came in for a completely different reason. In fact, the complaint was that the focus was intermittent and I found that the arcing tripler needed replacing.

The set had always been used with the remote and otherwise left in the standby position. However, switching the mains power off highlighted another problem. The backup battery (1675) had dropped from 2.4V to just 0.4V. Replacing that fixed the memory backup, after which and the picture tuning and sound all had to be reset.

Fortunately, I still have a few spares for this set and so I started off by swapping the modules. When I fitted the chrominance-luminance module, it fixed the "no-picture" problem and so I naturally figured that the problem lay there. I checked the luminance delay line and the voltages but was finally forced to dredge up the oscilloscope.

It took a long time for the penny to drop that the module that worked was somewhat different from the one that didn't. The latter had an extra two sockets (N4 and N5) and the former had a link from 1N4 to 3N4.

When I fitted this extra link, the set once again performed correctly and so I followed the N4 lead to the switches - both physical and an HEF4066 (IC706P) - on the SCART interface panel. However, this wasn't the problem area and it wasn't until I followed the video switch line (1N5) into the dusty bowels of the set under the CRT that I found the problem - the RGB buffer wasn't switching and so no RGB signal was coming from the SCART socket either.

By measuring the voltages on this module, I soon discovered that none of the transistors were being biased on, even though there was +13V and -24V supply rails to the board. Finally, I discovered that a single 10kΩ resistor (R3616) fed all the eight transistors and this was open circuit. A new one soon fixed the problem.

The third Philips

A 1999 Philips 29PT6231 (A8.0A chassis) came in with no sound and no monitor output. This turned out to be a classic textbook repair which demonstrated just how diagnostic software should work.

The error code was 014 which points to IC7430 (MSP3410D). A few quick voltage checks around this IC showed that the +8V on pin 39 had gone "missing", although the +5V on pin 7 was OK. Tracing back from pin 39 to Q7431 showed that there was no voltage on the collector of this transistor (BC337-25). However, there was 13V on the other side of R3464 and this 15Ω fusible resistor proved to be open circuit.

Replacing it fixed both symptoms, as the +8V rail also fed the IO switching circuits.

The seaside Philips

I had to attend to a 2000 Philips 34PT6361/79R (A10A chassis) which was on the top floor of a block of units that overlooked the beach. The view was fantastic but the 3-year old TV was already rusty due to the salt-laden air.

Mrs Allenby was complaining that the remote wouldn't work "after a while" and that the set would go dead after a few hours. If she then let the set cool down for about three hours, it would work normally again.

When I arrived, the set was working but not the remote control and when I switched it off, it wouldn't restart. This was a blow as it meant that I couldn't access the error codes.

The Small Signal Panel (SSP) board can cause problems in these sets and, as service centre, we are now expected to repair it ourselves - including changing the 100-pin sub-miniature surface-mounted microprocessor ICs! I can cope with the 8-pin EEPROM but change the other ICs on the off chance that this would fix the fault without compounding it? - not me! Even the exchange boards sometimes have problems and I need the confidence of the warranty.

In this case, I took the SSP with me to send off to Philips and left a loan set. Ten days later, another board arrived and I went back and fitted it. Because I never saw the on-screen displays in the SDM mode, I just had to assume that the correct option codes were in the replacement. I realigned the geometry and adjusted the tuning of the set and left, but not before giving a lecture on the disastrous location of the set and the dangers of on-shore winds corroding the set to pieces. In fact, I strongly advised that they cover the set over when not watching it. I noted the error codes as 16, 13, 17.

A week later, a distraught Mrs Allenby called me back, saying the set was doing exactly the same thing as before. Oops!

I returned as soon as possible to witness the same story and this time I removed the entire chassis and shipped it off to Philips. It looked as though it would take longer this time to get it back as they had run out of SSPs! But initially I was informed no fault could be found, although after soak testing it the symptoms began to show again. Finally, they found the cause of the problem - one of the 5-pin 5V IC regulators (IC7967, SI-3050C) was failing when it got hot, damaging the SSP.

I returned the set, realigned it again and noted that the error numbers that were still there after clearing the buffer were 16, 13 and 17. I am assured these are normal.

When I left this time, I emphasised that they were not to cover the back of the set when they were watching it - only when they weren't! After all, the back cover doesn't have ventilation slots purely for decoration.

I haven't heard from them for three months now so - cross fingers - it's still going OK.

And now for a change of scene, here is a contributed story from A. P. of Kuranda, Qld. I'll let him tell it in his own words.

Fender guitar amplifier

Dave brought his circa 1996 Fender Hot Rod Deluxe 30W valve guitar amplifier to me saying that it hadn't sounded quite right for some time. He demonstrated the problem by playing his electric guitar through it. The high and middle range notes sounded fine but the bass was quite distorted.

I also noticed that, even without any input signal, there was a strange rustling noise coming from the loudspeaker. This noise remained constant, regardless of the settings of the volume and tone controls. Dave said that the amplifier sometimes produced this rustling sound when he used it at home but never when he was playing at a gig. He blamed it on "dirty power" at home but I thought that the amplifier might be oscillating supersonically.

Not having worked much with valve equipment, the first thing I did when I was alone was to turn it on and cup my hand over one of the valves to feel its warmth. However, I was disappointed - the valve I had picked was cold! It was V4, one of two 6L6GC output valves and the other valves were all glowing nicely. Well, that explained the distortion.

Could V4 have a blown heater? I tried putting the other output valve, V5, in the V4 position. It came up with the heater on.

On the strength of this test I ordered two 6L6GC equivalents. My plan was to replace both valves to ensure that they remained a matched pair.

When the new valves arrived, I installed one in the V4 position and turned the amplifier on. Its heater stayed dark but then suddenly came on when I touched the valve!

I took the back off and this revealed a large vertical PC board with most of the components on it. In addition, there was a long, narrow PC board mounted horizontally, copper side up, and this carried the power amplifier and one of the preamplifier valves.

The solder side of the power amplifier board was readily accessible and I quickly found that two pins of the V4 socket moved in their holes when I wiggled the valve. Closer inspection showed that the solder on these pins had cracked.

It was easily fixed and for good measure, I remelted the solder on the pins of all the valve sockets. A few of the pins moved with a slight click when I did this, relieving tensions that had been built in during manufacture. This solved the problem with the distorted bass response but the rustling sound was still there.

Because the rustling didn't change with the setting of the volume control, I suspected that the problem was in the power amplifier. I tested this by plugging my signal tracer into the PREAMP OUT socket - there was no rustling. Conversely, when I put a 6.5mm plug into the POWER AMP IN socket, the main speaker continued to rustle. That cleared the preamplifier and also exonerated the switch in the POWER AMP IN socket that disconnects the preamplifier from the power amplifier.

Next, I tried replacing V5 with the other new valve but this made no difference.

Up to this point, I had been working without the circuit diagram but now I'd tried all the easy things and it was essential to have it. Fortunately, when I phoned Dave to bring him up to date on progress, he said he had the user manual and it had a circuit diagram.

Now non-technical people sometimes confuse block diagrams with circuit diagrams, so I wasn't hoping too hard. However, a couple of days later, Dave dropped by with the manual and I found that it did indeed have a clear circuit diagram, which included test voltages at many points. There was also a comprehensive layout diagram.

The power amplifier has just one other "bottle" apart from the output valves: V3, a 12AX7A dual triode, connected as a phase splitter for the push-pull output stage. Because it was easy, I swapped V3 with another 12AX7A (V1 from the preamplifier section). The rustling noise remained, so I installed a plug in the POWER AMP IN socket to isolate the preamplifier from the power amplifier, in case the original V3 was still causing problems. However, the rustling noise persisted.

The power amplifier has feedback from the centre tap of the secondary of the output transformer to the grid of V3B via a resistor/capacitor network. I needed to disable this feedback if I was to pinpoint the source of the rustling and it occurred to me that I could do this safely without having to remove any PC boards from the chassis by simply removing both output valves.

I did this, and found that although the signal at the grid of V3A was clean, there was rustling at the anode. There was also rustling at the anode of V3B.

V3A's anode connects to the +392V Y supply via R5 (82kΩ 0.5W), while V3B's anode connects to Y via R58 (100kΩ 0.5W). I now suspected that the fault was in one of these resistors or in the Y supply itself.

Unfortunately, I couldn't connect my signal tracer directly to the Y supply because the voltage exceeds its maximum DC rating, so I looked for an indirect method.

Reference to the circuit showed that this amplifier has a standby switch, S5, which disconnects the high voltage supplies at the secondary of the transformer. This allows you to turn on the sound cleanly after the heaters have brought the cathodes up to operating temperature. However, it also allowed me to do a rough test of the Y supply. With my signal tracer connected to the anode of V3A and the rustling sound in full evidence, I threw S5 to standby. The rustling continued for about a second or two, then faded as the reservoir capacitor discharged.

The fact that the rustling didn't stop dead when I switched to standby exonerated the power transformer. I now felt - admittedly without very much justification - that the rest of the Y supply was probably also OK.

Now for those two resistors. I began by discharging the four high-voltage filter capacitors before measuring R57 and R58 with an ohmmeter. Frankly, I wasn't really expecting to see anything. Instead, I suspected that the resistors would look fine on the meter but that one or the other was breaking down under the strain of about 200V across it.

I was already trying to think of ways to test for this but I needn't have bothered: R58 was spot on at 100kΩ but R57, which is marked on the circuit as 82kΩ, came in at 220kΩ!

At this stage, I still hadn't actually seen R57. Instead, I was measuring it from the copper side of the board, courtesy of the detailed layout diagram, and the resistor itself was obscured by V3's socket. Of course, it was possible that R57 was supposed to be a 220kΩ resistor but, due to a design change or printing error, was shown differently on the circuit diagram.

Anyway, I removed the narrow board and examined R57. It was indeed marked as 82kΩ and looked to be perfectly OK, with no charring or cracks. I replaced it and the rustling sound ceased!

I don't know why Dave thought that the rustling noise only occurred when he used the amplifier at home, since it would have been present all the time. Perhaps the crowd noise at a live gig was masking out the problem?

I also don't know whether the rustling noise was generated in the resistor itself, perhaps due to high voltage stress, or whether the noise was a result of V3A being incorrectly biased. Of course, I could have tried replacing R57 with a genuine 220kΩ resistor to see whether this was the case but by this time I just wanted to declare the job done.