Items Covered This Month Sony KV-XF29M35 TV set (BG-3S chassis SCC-U24A-A). Panasonic TC-48P10 TV set (M15L chassis).

Recently, I was doing some work for our local Sony agent and a couple of "newish" flat-screen TVs came in. Now when I say flat screen in this instance, I don’t mean plasma displays. I mean flat-front glass CRTs in silver cabinets as used in the Sony KV-XF29M35 (BG-3S chassis SCC-U24A-A). This set is actually circa 1999 but looks brand new and has great performance. It’s also considered to be a really reliable model, so I was surprised when two came in together with different faults.

The first had no sound, which seemed easy enough, so I decided to fix this one first. This actually turned out to be no TV sound – all the AV inputs and outputs worked OK, as did the volume control and all the on-screen sound menus.

Armed with an audio probe, I quickly confirmed that there was no output from the tuner/IF/decoder block (TU101) at pins 30 and 31. However, touching this point resulted in a healthy hum from the speakers.

I then set the TV into its service mode with the remote sequence: "Display" - "5" - "-Vol +" - "Power" and checked the data displays with that of the manual. In particular, I looked at anything to do with the sound; eg, OPB (Optional Bits). I also made sure that CCIR B/G PALD was selected as the TV system.

Everything was correct, so I felt it just had to be the tuner/IF/decoder (TU101) that was faulty. I ordered a new one in and fitted it but it made no difference – there was still no sound from the tuner IF decoder. So much for my snap diagnosis.

There really wasn’t much left to check apart from the supply rails to the various pins.

First, I had to get the chassis into a safe service position. This isn’t easy but with the rear end on the bench and the rest of the chassis up in the air, I could finally access the PC pins on the tuner pack. There are supposed to be 31 pins but nine are not connected, which makes it very difficult to determine which pin is which.

I began by confirming that the 5V, 9V and 30V were present. I also found the SDA and SCL digital buses to be 4.75V each, with activity indicated on the CRO.

The picture was perfect, so the RF AGC (automatic gain control), AFT (automatic fine tuning) and VIF (video intermediate frequency) connectors had to be OK. So what was left?

Well, "NC" normally means "no connection" to me but the circuit indicates that these pins are connected! I felt that they needed investigating but there was also PLL-SW and DET OUT (detector out) signals to check – and what if the output load was affecting something?

The circuit also shows IC203# but the "#" means that the device isn’t fitted in this model – and it wasn’t. Clasping at straws, I swapped over the B3 Audio Processor/AV switch from the other set but it made no difference.

The DET out (pin 22) goes via R105 to transistor Q101, which provides the video output signal. This was OK but what about the PLL-SW on pin 23? Well, this went via R29 1kΩ to pin 27 of microcontroller IC001 and this pin is marked on the circuit as "0V" and "ST" (for stereo). However, I measured the voltage on this pin as 4.9V, so at last I had a clue.

Similarly, there were three other pins on the microcontroller that were of interest: pin 29 (mute) at 0V; pin 28 (SAP = Sound Audio Processor?) at 0.3V; and pin 26 (audio switch - forced mono) at 4.9V. The latter is connected via R030, R108 and R103 to pin 26 of the tuner but it is all marked "#" which again means "not fitted".

The "VOL CONT" line from pin 49 of IC001 to tuner pins 17 & 25 checked out OK too, in that R104# and R107# are not fitted. This all took time to investigate and I also changed IC001 and its companion EPROM (IC003) to no avail.

By now, I felt that I had already gone the extra mile for what was, after all, a simple problem. I was getting nowhere except for the incorrect voltage on pin 27 of the microcontroller. I tried shorting this out but it made no difference – there was still no sound.

All I had left were the nine "no connection" pins on the tuner to check out; ie, pins 16-20, 24-26 and 29. In fact, I had already checked most of these but there was one left that I had overlooked. This was pin 24, the Reset line, which measured 1.7V.

This was an interesting value as it was neither really a low (0V) nor a high (at 5V). I tried shorting it alternatively to Ground and the +5V rail but the sound still stubbornly refused to come on.

In operation, the Reset pin (pin 24) is supposed to go low for a fraction of a second and then rise to 5V. This pin is connected to IC100 (580743AL), which is a surface-mounted device, via D100. The time constant is determined by C104.

I tried heating and freezing the device and then switching the set off and on. Suddenly there was sound! Replacing IC100 completely fixed this awkward fault.

The second Sony

The second Sony came in with the initial fault of being intermittently dead and, surprisingly, an intermittent pink menu! This latter symptom really had us amazed, as there is no menu to change the colours from the normally blue, white, black and yellow. The picture and sound were otherwise perfect.

In the end, this set was fixed by a team effort, with no less than four technicians having a go. Perhaps it might have been better if one person had stayed with it all the time but it changed hands several times due to various circumstances within the service company; ie, the Christmas rush, illness and even a retirement.

The first technician noticed that tapping the set would make it go on and off and change the menu colour. By careful trial and error, the sensitive area was reduced to the Teletext module (V1), so he removed the module and reworked it completely. Afterwards, when it was refitted, the set no longer seemed to intermittently switch on and off but the menu colour still required attention.

After a lot of consultation, he decided to replace the Y/C Jungle IC (IC301, CXA2130S). By the time the IC arrived, technician number two was now on the job and he replaced the high-density 48-pin device. This wasn’t all that easy because there were lots of surface mounted components around it. However, although I didn’t observe him do it, the completed job looked fine – the only trouble was that the set now didn’t work at all! It was completely dead.

It was now that I was given the job. Thanks, mate!

The first thing I did was examine it very carefully everywhere there was fresh solder. The guy had done an excellent job and had also reworked microcontroller IC001 as well. The exact symptom we had was that the power LED (D3906) on the H3 function board came on yellow but nothing else, which only implies that there is voltage on the +15V standby line.

I fished out the meter and started at the power supply to find that the standby voltage was low (at 7V). And although it was arriving at pin 4 of IC002, nothing was coming out on pin 5, so I ordered a new MM1319AFBE 8-pin surface mounted IC. When I refitted the new one, the 5V was there initially but dropped over a period of 15 minutes.

Despite unsoldering many of the pins of the microcontroller, EPROM and jungle ICs, it made no difference. I was informed that it was essential for the SDA and SCL data lines to be at 5V for the set to work. I ordered and fitted a new EPROM and microcontroller IC but it made no difference. I then tried disconnecting as much as I could from the 5V rail but was getting nowhere.

By now, we were beginning to discuss whether or not the set should be written off. We had already wasted too much time on it for the job to be economical. Fortunately, about this time, a new technician was hired to replace the bloke who had retired. The new boy on the block had to be tested, so he was given this set to see what he was made of – not that we were really expecting anything much in view of the problems we were having.

To our astonishment, he had the set working in half an hour!

So what was the problem? He had found that the reset line from IC002 to the microcontroller was the source of the problem. R043, a surface-mounted resistor, had a tiny solder dag that was just shorting to pin 12 of IC001. This pin is normally grounded, which explains everything – no wonder the voltage stayed low.

Still, it’s a wonder to me that no-one (including yours truly) picked it up. It made me and the other two technicians look a bit stupid and the new guy is well in with the service centre. Still, such is life and he deserves it.

Jousting with Phil

It was a gorgeous day; too good to be working in fact – but hey, isn’t every day?

I was on my way to fix a set for Phil, an old client of mine. Phil’s set is a Panasonic with a TC-48P10 M15L chassis. It’s an old set but so are we - and I like the guy.

I pulled up outside his place and he was digging in the garden. "Finally buried your missus, Phil?", I quipped. "Nope – I’m preparing it for you", he replied, "in case you charge too much". "Now, now", I countered, "you know I only charge you double what it’s worth!" "Yup, that’s why I’ve got to get you listed down on my tax form as a dependant", he shot back.

Anyway, down to business. Phil’s set had no picture but the raster and sound was there and also the On Screen Display.

All these series of Panasonics look very similar. They are housed in black plastic cases with the same controls on the front and their features can be confusing. I asked for his remote control.

"You’ve still got it from last time", he replied. I let that comment slip by and when I removed the back I could see that it was my mistake – this set was a "non-remote" model and no IR receiver was fitted.

Unfortunately, there are no AV inputs to try an external source of video, so really there wasn’t much I could do without an oscilloscope. I knew I was going to get some verbiage when I told Phil the good news about it going to the workshop but it had to be. He made a few comments about "licensed thievery" and I gave him back as good as I got.

Back at the shop, I fired up the CRO and followed the video from the video detector (IC101, pin 19) to Q601, Q302 and finally pins 15 & 16 of IC601 (AN5601K), the Y/C jungle IC. But there were no RGB outputs from pins 24, 25 and 21.

By now, I was feeling pretty confident about an early and easy outcome to this problem. I was sure that it was all a matter of beam limiting and/or contrast control that was the cause.

The most common suspect is R525 (12Ω) which goes high but it proved to be OK in this case. I checked the subcontrast control (R302) and found that the voltage on pin 6 of IC601 varied over the expected range (ie, 2-4V). Similarly, the voltage range for the brightness control checked out OK on pin 18.

By now, all the obvious causes had been eliminated so I ordered in a new AN5601K and laboriously replaced this 42-pin high density IC. To my dismay, this only made things worse because it took away the on-screen display. Suspecting a crook new IC, I obtained another one but it made no difference.

Despite the access difficulties, I decided to make a series of voltage measurements on the underside of the IC. To make things easier, Mrs Boss wrote down the results as I proceeded and compared them with what they should be.

In a fault condition, you can’t expect the two sets of figures to exactly match but you can make a note of the ones that are significantly different and then examine the circuit for clues.

Most of the readings were very close to the chart in the manual (which, I might add, was for a different model but was the only one with the voltages drawn). For some reason, The TC-48P10 manual doesn’t show voltages.

The significant voltage was on pin 14 ("DATA BLK"), which should have been 0.1V but was in fact 1.1V. It was only 1V out but this was my only real clue – that and the fact it was connected to some sort of blanking control circuit which might give the symptoms I was getting.

Pin 14 is connected via D604 to R641 and the 12V rail. It also goes to Service Switch S601 via D604. I could measure 1.7V on the anode of D604 but hang on – in the Normal position of the Service Switch this pin should be connected to ground? So how was this possible? The ohmmeter didn’t measure low resistance to ground either, so I reasoned that the problem had to be in the service switch S601 itself.

This was a blow, as I doubted I could get this as a spare part from Panasonic considering its age. Anyway, I removed the switch assembly and dismantled it but could find nothing wrong with it apart from the fact that it needed a bit of a clean. I reinstalled the switch but nothing had changed, so I measured the connection at the switch and at the junction of R641 and D604.

This revealed that there was no continuity between these two points so there had to be a break in the copper track on the PC board. I couldn’t see any break but by following the very fine track and scraping away the Shellac every few centimetres, I finally narrowed the break to a very slightly discoloured 2.5mm of track. Fitting a fine wire link across this section restored the circuit and the picture. For some reason or other, this section of track had corroded right through.

Now the only problem was letting Phil know the cost of reality. Yes, I know I am in for a real hard time!