Email Address:

Lost your password?

This is the legacy website; please use the new website.

Serviceman's Log

Variety is the spice of life.

by the TV Serviceman

  • Akai CT-2167A TV set
  • JVC AV32X25EVS TV set (MF II chassis)
  • Panasonic NN-S453WF microwave oven
  • Tektronix 465 oscilloscope
  • Goldstar Models OS-7040A & OS-9040D oscilloscopes
  • Panasonic TX-60P82A TV set (MX10 chassis)
  • Panasonic TX-80VO3A TV set (MX12 chassis)

I sold an Akai CT-2167A TV receiver in 1996 to Michael Selley. This was a 51cm stereo unit with teletext, made by Kong Wah in China and it had performed flawlessly for eight years. But now, it was paying me a return visit with what seemed like a trivial fault – ie, lack of height (letterbox).

Click for larger image

Michael had phoned me in advance and I initially told him that this might be due to a transmission in 16:9 format from the TV station. However, he quickly informed me that it was now like this all the time.

We have an enormous number of people complain about under-scanned screens these days. Most feel that they have paid for a certain size screen and aspect ratio and therefore it should be fully scanned, regardless of the TV stations. The TV stations, on the other hand, couldn’t give a monkey’s, as long as you are watching their ads, so we get a lot of nuisance calls for this.

Michael’s other problem was his remote control which was in near fatal condition, having survived (just) three young children (who like chocolate and Coca-Cola) and one dog. I eventually got it to go but only after stripping it down and cleaning out eight years of goo and corrosion. I also had to fix several dry joints and the battery springs before reassembling the case (which was missing large chunks of plastic).

Now for the TV itself. After removing the back, I began by checking all the main voltage rails: +106V, +25V, +15V, +7V, +5V, +180V, +22V and +12V. These were all as expected so I then set about checking the voltages around IC401, the AN5521 vertical deflection amplifier. Again, I could find nothing amiss, although access to this stage is quite difficult.

Next, I checked the voltage on the deflection yoke and this was 11.4V (half the Vcc to the output stage), as you would expect. I then measured the voltages on pins 29-35 of the AN5601k jungle IC (IC301) and again they were not far removed from the figures listed in the service manual.

However, you cannot put much store on the validity of the data in service manuals these days, as they are often full of errors and contradictions. This particular manual shows voltages on the block diagrams and the adjustments chart that contradict the circuit diagram in at least half a dozen cases. For example, it says on page 6 to adjust B+ to 106V ± 0.5V on TP405, which the circuit diagram clearly shows as +105V.

Out of ideas

By now, I was running out of ideas, so I adjusted the VR401 height control to its end stop to see what effect this had. The picture was now only 100mm from the top and bottom but I really didn’t learn much from this other than that the height control was still working. I then got the oscilloscope out and checked waveforms 15 to 21 and these were all OK too.

This job was becoming more difficult by the minute. Despite all my measurements, I still had very little to go on. And from what little I did have, it was impossible to tell what was significant.

At this point, I decided to change all the electrolytic capacitors in the vertical deflection amplifier, as they can give trouble. These include C344, C331, C338, C326 and C401 but changing them made no difference. I then moved onto the resistors and replaced R421 as it was critical in supplying 12V to the jungle IC. This too made no difference, so I also tried heating and freezing the components but I continued to draw blanks.

Feeling increasingly desperate, I then went back through the notes and measurements I had made, looking for anything that might give a clue. I had already noticed that Vcc on pin 29 of the Jungle IC measured 10.54V whereas the circuit showed it as 12V. I tried shorting out the R421 I had already replaced but the rail only rose to 11.3V and it still made no difference to the picture.

Next, I connected an external power supply and wound it up to the full +12V but this also turned out to be a furphy. Pin 35 (feedback) measured 4.46V instead of 5.02 but changing C321, the only component on this pin, still made no difference.

What’s more, pin 33 (vertical oscillator) measured 0.88V instead of 1.2V but I was now totally cynical about the voltages marked on the circuit diagram. For that reason, I initially didn’t regard the difference as important, especially as waveform 18 was correct in shape, although a little low at 1.5V peak-to-peak instead of 2.2V. Still, it was my only real clue so far. I checked D305 as OK and then took a look at R331. This is marked 180kΩ on the circuit but is only 51kΩ in the set. I checked the parts list which also had it listed at 51kΩ.

Anyway, I removed it from the set and measured it. It was high, the meter showing a reading of about 70kΩ! I replaced it with a 56kΩ resistor (I didn’t have a 51kΩ unit in stock) and the set immediately began vertically overscanning.

Resetting the height control finally restored everything to normal and put an end to my misery.

The pretentious JVC

A JVC AV32X25EVS TV with the pretentious name of "InteriArt" (employing an MF II chassis) came in under warranty, its owner complaining of a loud "popping" noise from the loudspeakers when the set was switched off.

It turned out that the set had been purchased back in 2002 and the fault had been present right from day one. It was only now that the customer had decided to bring it in. God knows why he had left it so long before complaining – perhaps it was because the warranty period was coming to an end?

Being blessed with an original service manual for a change, I could see that the audio amplifier was pretty complex, employing about 14 ICs and its own microprocessor. The subwoofer, which was the main "popping" source, was fed by power amplifier IC601 and controlled by a complex network of seven muting transistors and 11 diodes by five command rails: A_mute, audio_mute, Amp_mute, D_mute and centre (these are repeated for the left and right channel amplifiers and the audio output).

What ever happened to the KISS principle (Keep It Simple, Stupid)? This circuit seemed to be over engineered and unnecessarily complex.

I began by checking the ±9V, 8V, 10V, 24V and ±27V rails and these were all OK but I did find some major errors in the circuit diagram. This shows +27V coming out D954 and going all the way to pins 13, 14, 15 & 16 of connector CN004. However, on the next page, pins 13 & 14 are shown at -27V!

Click for larger image

I then found a transistor (Q601) in the muting circuit whose function was described as "power on/off det". This seemed a good place to start making and comparing measurements, until I noticed that its emitter voltage was marked as +4.9V on the circuit while in practice it is fed via a forward-biased diode (D629) straight from a 9V IC regulator (IC605). This and many other mistakes in the circuit made it very difficult to troubleshoot this complex circuit.

What I did find were two electrolytic capacitors (C606 and C607) in series with Q601’s emitter. Their job is to maintain a charge on the mute line until the audio amplifier has powered down, after the set is switched off. On the circuit diagram, they are marked as two 220μF 16V electrolytics but in the set itself, I found one 220μF cap-
acitor and one 1μF capacitor. So which was correct – the set or the circuit diagram?

Because the service manual had other mistakes, I tried replacing these two capacitors with the same values as those already fitted. However, this made no difference, so I moved on and started testing other parts of the same circuit.

In the end, nothing I did was making any difference, so I went back to capacitors C606 and C607. What if it was the circuit that was correct and they both should be 220μF?

And that was it! The wrong value had been put in during manufacture and the service manual was correct (the spare parts list has these down as C1606-07, both being 220μF 16V as well).

Dead microwave oven

Michael, our microwave specialist (and all round good bloke he insists I inform you), was telling me about some of the problems he had been having with a couple of late-model Panasonic microwave ovens with inverter power supplies.

In particular, the circuit diagram errors are a problem. He showed me two circuit diagrams (ie, for NN-S453WF and NN-S553WF models) where the oven lamp and turntable motor would never work if connected as shown, since both sides of these components are connected to the mains Active via Power Relay A (RY2). And in another case, the QPQ schematic of the NN-S560WF series of ovens shows the mains actually being shorted out via the same Power Relay A (RY2)!

He also told me an interesting story about a Panasonic NN-S453WF which came in dead. The PC track from pins 1 & 3 of CN1 (mains input) to Power Relay A (RY2) had evaporated on the Digital Programmer Circuit (DPC). In addition, the PC fuse or "fuse pattern" (PF2) had blown, along with the track from Q223’s collector to the relay coil.

There is a "Troubleshooting Guide" that comes with the service manual which says: "(1). Remove the jumper wire PF1; (2). Insert the removed jumper wire PF1 to PF2 pattern and solder it. If both PF1 and PF2 fuse patterns are open, replace the DPC".

Apart from the confusion when you look at the board as to which fuse pattern is which (he could only see PF2 and PF3 marked on this board), the answer invariably is to change the DPC. The circuit only shows three fuses and these fuses can also blow if the interlock safety switches are off-centre.

Michael tried repairing the DPC first by fitting the links where the track had blown. When he refitted the board into the oven, he found that the oven worked but the lamp didn’t and the turntable motor kept on turning when the door was closed.

After replacing the lamp (probably the culprit for all this), he noticed that the new one stayed on all the time and so he decided to order and change the DPC (Part No: F603L5Q40QP). Both the oven lamp and turntable motor are controlled by Power Relay A (RY2) from Q223’s collector (a surface-mounted transistor, DTC123JA), which in turn is driven from pin 9 [(E)P53] of microprocessor IC1 (MN101C589EL).

The new DPC looked perfect and there are no published installation instructions for initialising it or setting it up, so he just fitted it to find it made no difference! The lamp and motor were still on which was annoying.

After making absolutely certain that no mistakes had been made, he ordered yet another DPC. This time the board worked flawlessly and all the faults were cleared.

In the end, he found that Q223 was faulty on both the original board and the first replacement board.

Second-hand scopes

My luck turned for the better recently, as I had the good fortune to purchase some secondhand Tektronix oscilloscopes at give-away prices.

For example, I recently acquired a 465 (I won’t tell you the price – you will weep), which is a 100MHz delay CRO, circa 1974. Unfortunately, this unit wasn’t showing many signs of life when I first got it but with a bit of fiddling, I managed to get two stationary dots on the screen with the Beam Finder (and it wasn’t in the X-Y mode)!

From this, it was obvious that the A and B horizontal timebase sweeps weren’t working. When I removed the covers, the only clue I had was what looked like a "slightly-cooked" 33Ω resistor in the centre of the board underneath the CRO.

I had a copy of the manual for the Tektronix Model 466 but was dismayed to find that is quite different to the Model 465. So I got onto the Internet and found at least two sources for service manuals. One was, an extremely helpful site in the United States where for only US$9 I could FTP download 198Mb of scanned manuals. The expensive part for me was the dial-up 56k connection and the time it took. The other site was Denis Cobley at, who can supply a scanned service manual on CD for AUD$25 from Tektronix right here in Sydney.

Denis was very helpful too and based on my description, suggested that the burnt resistor was in fact 22Ω (not 33Ω). He also said that the 1000μF electrolytic capacitor next to it may be the culprit and that I should start my investigations with the sweep logic IC.

Different ball game

I have to say that repairing oscilloscopes is a different ball game to fixing TVs, with a complete set of new buzz words to learn!

Armed with the 300-odd page service manual, I established that the resistor was R1220 and mine measured 27Ω instead of 22.1Ω. The capacitor was C1220 and is an axial type rated at 10V. It measured perfectly OK but there was no -8V being fed to it. In the end, I replaced it anyway – when I finally managed to track down a 1000μF axial capacitor.

I eventually found that the cause of the missing -8V rail (and +5V rail as well) was due to a faulty CR1561 bridge rectifier. Initially, I replaced this with four 1N5408 diodes which are rated at 3A each but they ran warm. The original 15-0488-00 bridge is rated at 200V 1.4A. I eventually managed to get and fit a 6A bridge (KBU602) for just $1.55 from WES Components.

This didn’t completely solve the problems, however. Both the X and Y amplifiers were now working and I could centre the traces but there was still no horizontal sweep.

I had been told that no specialised parts were available for this now 30-year old CRO, so all I could do was hope that nothing critical was at fault. Anyway, the U870 Sweep Control IC seemed like a good place to start my investigations. A good tip Denis gave me was to start by reading the concise circuit description in the manual before doing anything. This I did and I prayed that it wasn’t the IC itself, as I would have Buckley’s chance of getting a new one.

Having read the manual, I started by measuring the DC voltages around the IC before checking the waveforms with another CRO. The 465 manual had very few waveforms in it but the 466 – which has its similarities – did have waveforms printed on the circuits.

I soon discovered that the only waveform was on pin 1 of the IC (TD in Auto Sense Input) and around transistors Q862 and Q864. All waveforms fizzled out after diode CR863. I had incorrect voltages on pin 6 (Auto gate timing), pin 8 (holdoff output) and pin 14 (sweep reset).

The sawtooth sweep generator based on Q1030 & Q1036 (Miller integrator) and Q1012 & Q1014 (multivibrator) wasn’t oscillating. However, there wasn’t much point in swapping transistors and FETs from the A sweep circuit to the B sweep, as that wasn’t working either.

A few quick voltage checks showed that Q1024’s collector was incorrect at -0.42V instead of -1.7V. This made me suspect the time/div switch (S1150), especially as you could momentarily see a fully-scanned waveform in detail by rotating it quickly. I also suspected Q1030 (an N-channel FET) but swapping it with Q1090 made no difference.

Click for larger image

The only other inconsistencies I found were incorrect collector voltages on Q854 and Q804. However, according to an article on servicing the 465 portable oscilloscope in TEKSCOPE ". . . the sweep circuit contains several feedback circuits and is difficult to troubleshoot unless you break the feedback loop. A convenient means of doing this is to pull the Disconnect Amplifier (Q1024) out of its socket. This causes one sweep to be generated and often provides a rapid clue as to what portion of the circuit is in trouble. The horizontal amplifier is push pull and can be checked by the usual method of shorting the two sides by means of a jumper".

I tried removing Q1024 as suggested but it made no difference. However, I couldn’t comprehend the latter piece of advice and was too chicken to try it – there just wasn’t enough detail.

So now I was stuck. With old Tektronix gear, you don’t just rush out and buy parts and fit them on a hunch as they can cost and an arm and a leg. However, my money was on U870 as being the most likely culprit. The part no: is 155-0049-01 and I found out it costs $A160 + GST + freight from a company in Victoria.

Fortunately, I got a lucky break. I managed to get access to a Model 475 oscilloscope, which uses the identical IC as U600. And swapping the two ICs over transposed the faults from one oscilloscope to another.

Then, unbelievably, I had a second break. I managed to obtain a brand new IC for only $US25 from Sphere Research Corporation in Canada, via the Internet. Now all I need is to get the Option 05 TV Sync Separator retrofit kit (as fitted in the 465B), which is necessary for TV and video work. Does anyone know where to get one?

Another problem

I had an additional small problem with the Ch2(Y) vertical preamp. The calibration knob was set in the "UNCAL" position and wouldn’t switch to "CAL". The reason was simple enough. The coupling shaft was not engaging the shaft of the control and was slipping.

The difficult part was access to tighten the allen key grub screws. I eventually used a very long 0.05-inch allen key to reach and tighten the grub screws, only to have the aluminium collar break apart and fall into the equipment. Quelle horreur!

Well, what to do? There was Buckley’s chance of getting a new one. However, this control is not that important as long as I could get the switch back to the "off" position.

Well, I tried all sorts of special tools to coax this rotary switch back into the "off" position. I got it so close but extra force was needed to move the switch which I just couldn’t supply.

In the end, I gave up and decided to just glue the plastic coupler left onto the shaft so that the knob wouldn’t pull out of the mechanism. To do this, I applied a large drop of superglue onto a very thin long screwdriver and very gingerly inserted it between the PC boards until I got some of the glue onto the shaft. I then pushed the knob and its long shaft into position, so that the plastic coupler went over the control shaft and glue. Finally, I wiggled the knob back and forth to spread the glue around.

As I was doing this, the superglue suddenly set hard (as it does) and the knob, shaft and control all engaged and I managed to switch it off easily. Since then, this control works well but I am reluctant to use it in case it breaks again.

I also had a problem on a Tektronix 475 with the Ch2(Y) position centring control. The beam was at the extreme top of the picture with the control turned completely anticlockwise.

The circuit shows this control as varying the control voltage from +8V to -8V. This is applied to the junction of Q272, Q278 and Q282, Q288 in the third cascade amplifier.

Well, I could get it to go positive but I couldn’t get it to go very negative at the junction of Q272 and Q278. Remembering that everything is socketed, I decided to switch the transistors one at a time with those of the Ch1(X) vertical preamplifier. And when I swapped Q278 with Q178 I transposed the problem to the other beam.

Q278 turned out to be open circuit and fitting a BC558B fixed the problem in the short term. The only thing is that Q278 is a 2N4261 PNP Si transistor (Part No. 151-0434-00 or 151-0202-00) which has a frequency response in excess of 2GHz, whereas a BC558 only goes to 150MHz (and this is a 200MHz CRO). However, I managed to also get this from Sphere Research Corporation in Canada – the freight cost more than the parts!

PS: point your web browser to for a very useful Tektronix parts cross-reference data list.

Goldstar scopes

I also had to repair some ex-university Goldstar Oscilloscopes, models OS-7040A and OS-9040D. These are both 40MHz delayed sweep CROs.

The OS-7040A had intermittent "no display" and "no scan" symptoms and was fairly easy to fix, with all the trouble being dry joints on the power supply and EHT board. Reworking all the soldering on this board, including under the EHT screening cans, fixed all the problems.

Click for larger image

The OS-9040D was a different story. There was no display on this unit but without a service manual, I was pretty well stuck. And so I went back to the Internet, where I tracked down a scanned copy of the "Operation Manual" which fortunately has a basic circuit in the back.

I started by checking the power supplies which provide +195V, +55V, +12V, ±5V, -12V and +32V. Next, I tried measuring the voltages to the tube. Unfortunately, I didn’t know what to expect because there are very few voltages on the circuit.

To eliminate the tube, I swapped it with the unit in the OS-7040A and confirmed that there was EHT and high negative voltages on the cathode grid and focus pins. These were all OK but there was still no beam. I also established that both the timebases were oscillating correctly and giving waveforms to the deflection plates.

There are about 38 internal adjustment controls inside this CRO, some of which are not even marked on the circuit. At this stage, I was only interested in the ones affecting the display tube, which are the "CRT Bias" (VR617) and "HV Adj" (VR618). I marked their wiper positions and measured the voltage on them before adjusting them.

I was in luck – realigning these two controls restored the beam and after adjusting the focus (VR113) and astigmatism (VR616) controls, I got a beautifully sharp trace. Obviously, someone had been fiddling!

Finally, if anyone has the service manuals for any of these models, please contact me via SILICON CHIP.

Two more Panasonic TVs

I recently had two late model Panasonic TV sets come in with very similar fault themes.

The first was a 2001 Panasonic TX-60P82A (60cm picture tube) employing an MX10 chassis. The set was dead and it didn’t take a mountain of intelligence to see that the flyback transformer (T501, ZTFN34001A) was cactus, just from looking at its condition.

However, having quoted for and "completed" the job, a secondary symptom which had been masked by the first appeared – the set would now cut out after five minutes.

When in doubt, I start by measuring the voltage supply rails. Just about all were spot on except for the 5V on TPAS which is critical because it feeds the microprocessors. Instead, the voltage output from IC885 (AN78L05-TA) read 5.6V, which is 0.6V or 12% too high. Replacing this 3-terminal IC fixed the problem, which was a relief.

The second set was a later 2003 Panasonic TX-80VO3A (80cm picture tube), employing an MX12 chassis. Both the sound and picture on this set were fine but it too kept cutting out after 3-4 minutes, with the picture and LED pulsating.

All eight voltage rails checked out OK but the power control IC (IC801) was getting hot. And, as I quickly discovered, the symptoms became worse during testing. The pulsating frequency began to change and then the set died completely after momentarily displaying a white line across the screen.

I replaced the vertical output IC (IC451, TDA8177) which improved the symptoms a little but it still kept turning off after 3-4 minutes and IC801 was still getting very hot. In the end, more by luck than good judgement, I spotted a small crack on T801, the chopper transformer ferrite former. This was undoubtedly causing the power supply to work harder and therefore get hotter.

A new transformer fixed everything up properly.

Share this Article: 

Privacy Policy  |  Advertise  |  Contact Us

Copyright © 1996-2019 Silicon Chip Publications Pty Ltd All Rights Reserved