A brand-new 2003 Panasonic TX-68PS72A (MX12 chassis) requiring
urgent warranty attention appeared on my bench. The set wouldn’t start but the
+141V, +15V, +12V, +9V, +5V and +3.5V rails were all being delivered by the
primary switchmode power supply and were apparently all reaching their correct
There was also +141V on the collector of the line output
transistor (Q551, 2SC5517000LK) and on the horizontal driver transistor Q541,
which told me that the latter wasn’t being switched on. This was due to the fact
that no pulses were coming from pin 19 of the jungle IC (IC601, C1AB00001715). I
then checked the voltage rails going to IC601. In this case, there should be +5V
to pins 11, 42 and 74; and +9V to pins 25, 67 and 45. The latter (ie, to pin 45,
HVCC) is the all-important +9V rail for the horizontal oscillator. This is
derived from the 12V rail but measured just 4V on the actual pin.
A close inspection of the circuit revealed that the +12V rail
is fed through two very special diodes – D401 & D402 – and then a small
choke (L401). These diodes, type MA2C029WAF, are called "Diode Varistor Double
1.24V" and the only trouble was that D401 had 6.76V across it instead of only
1.24V. A new diode restored the voltage and the set then functioned
Items Covered This Month
- Panasonic TX-68PS72A TV (MX12 chassis)
- Panasonic TX-68P100Z TV (MD2 chassis)
- Sanyo C29PK81 TV (AA1-A29 chassis)
- Philips VR788/75 Hifi VCR
- Panasonic TC-68V80A TV (MX-4M chassis)
- Philips 29PT9418 TV (MG3.1 chassis)
The streaky TX- 68P100Z
Mrs Staples’ 3-year old Panasonic TX-68P100Z flat screen TV
(MD2 chassis) had an horrendous picture, with incredible 8cm streaking after
every object on the screen. In fact, at first glance it looked as though it was
some deliberate digital effect. The streaking wasn’t confined to any particular
colour but could be varied slightly with the contrast control.
Well, I checked the digital board by substitution and I could
see, on the oscilloscope, that the RGB waveforms going into the three output ICs
on the CRT "L Board" were clean. However, they were decidedly crook going to the
CRT cathodes. I checked for ripple on the +140V and 210V rails to the three ICs
and they were fine but the 12V was a different story.
At this point, my attention was grabbed by a 3-pin device
that’s normally used in low-voltage switchmode power supplies. This device is a
TL431CLP zener linear IC controller and, in this unit, is designated D354. It is
fed with 2.5V from pin 1 of each of the three output ICs and controls the 12V
rail via R387, a 470Ω
This circuit is designed to protect the rest of the set from
CRT flashover by shorting the 12V rail and causing the set to shut down.
However, in this instance, a partial failure in D354 had caused these symptoms.
A new one fixed the problem.
Faults that haunt
It’s amazing how some common faults come back to haunt you in
Mr Cabezas was complaining about his Sanyo C29PK81 (AA1-A29
chassis) and was waffling on about the colour being "wrong" and "going funny".
Now, having been in the trade for far too long, I "kinda" switch off when people
waffle on about symptoms in their TV sets, as most are incredibly poor in
describing what they see.
For example, back in the early 1950s when RCA was developing
colour TV in the USA, the public was invited to examine a whole row of colour TV
prototypes in a large corridor and grade the quality of the pictures according
to various subjective criteria. One of the sets was not a TV set at all – it was
just a window looking outside onto a garden. 60% of the people who filled in the
questionnaire ticked that the picture was poor and the colour unnatural!
Many people use the word "unwatchable" when describing picture
faults. Does this mean that they are blind? In fact, some 10% of the population
is colour blind and about half of them don’t even know it. So I guess it’s just
you and me who really know what good colour is all about . . . and I am
beginning to wonder a bit about you!
Anyway, I told Mr Cabezas to bring his TV in, whereupon he
promptly wedged the heavy set into the back seat of his Falcon and, judging by
the time he took to arrive, drove it around on two wheels! If the set hadn’t
been damaged before, both it and the car certainly were now.
Finally, we got it onto my workbench and I hooked it up to an
antenna and looked at the picture. He was dead right – the picture was . . .
well, funny. It was intermittently dark, sometimes going negative and with the
colours appearing to split as though the static convergence was being adjusted.
That said, it really is hard to accurately convey the effect on the picture.
OK, so where should I start? While I was thinking and dithering
about this, the picture became so dark that it eventually "went out" altogether.
Great, out of the frying pan into the fire. Had the picture tube gone? No; it
was still there!
After I had recovered from my panic attack, I decided to work
from first principles. What did I have? Well, for starters, stereo sound. In
addition, the filaments of the tube were alight and there was no smell of
burning. "Wonderful", I thought. "It can’t be that bad".
Using a multimeter, I found that there was bags of screen
voltage on the G2 grid of the tube. However, the collectors of the output
transistors (and thus the cathodes) were all high at nearly 200V, which meant
that these transistors were well and truly switched off.
To check the CRT, I chose one hapless cathode and momentarily
shorted it to ground. This gave a fully scanned raster of bright colour, so I
now knew that the set’s vertical and horizontal scanning and EHT were all OK.
What more could Mr Cabezas want? Well, a good picture would be nice, I
Fortunately, by now, a few bells were beginning to ring. And
then I remembered – of course, it was just had to be resistor R1792
(120kΩ), which biases all
three output transistors from the 200V rail.
Well, that fixed the problem and I mentally kicked myself for
not remembering this rather well-known fault. I guess there is no fool like an
old fool! I also replaced R485 (180kΩ) for good measure, as it can also give weird dark
Mr Cabezas arrived to collect his set almost immediately,
forcing it "none too gently" back into his long-suffering Falcon and
disappearing in a cloud of blue tyre smoke.
Philips hifi VCR
The main worry with blown switchmode power supplies is that you
have diagnosed and replaced every faulty component, and determined (if only
roughly) why it blew up in the first place, before switching it back on.
Recently, I had a 1999 Philips VR788/75 Hifi VCR brought in
after a block of units was struck by lightning. Unfortunately, Mr Ford was an
old age pensioner and did not have household contents insurance, so he was faced
with either buying a new one or getting the old one fixed. And while he could
buy a cheap mono VCR, watching movies was his only pastime and indulgence and so
he chose to get his hifi VCR fixed.
When I removed the cover, I could see that the power supply on
the right of the 2-piece board was severely damaged – even the optocoupler’s
plastic case had exploded! At this stage, my main concern was whether this was
mainly a mains-borne lightning surge or a surge that had also come through the
aerial socket. If it was the latter, then the tuner could also be U/S which,
combined with the damaged power supply, would make it too expensive too fix.
My next step was to dismantle the entire unit to get the
motherboard out. As I did so, I examined it very carefully, looking for telltale
black marks, but none could be found. So far so good!
Getting back to the power supply, fuse FS001 was now just a few
encrusted black shards of glass protruding from the metal end-caps. In addition,
0.39Ω resistor R5106 had a
black hole in its side, while the 2SK2632 FET had lost its face, as had
optocoupler PC5101, transistor Q5102 (2SD2144S) and transistor Q5302
This lightning hit had been violent, although ironically had
not affected the TV set which was still merrily working.
Having removed the five obviously-damaged components, I used a
multimeter to check some of the other parts in the power supply. This quickly
showed bridge rectifier D5001 to be short circuit and so I continued to check
all the remaining active components, including the diodes. Because of the severe
damage to the optocoupler, I was concerned that the surge had also taken out
every single active component on the secondary (cold) side of this device.
Despite this, I couldn’t find anything else that was faulty.
Before ordering new parts, I checked a pile of similar wrecked
Philips VCRs I had stored – mostly mono VR299/75 models. However, pickings were
minimal – most had found their way to this, their last resting place, because of
similar component failures. Acting on a whim, I decided that I would choose one
of these and rebuild its power supply too – after all, the parts were very
similar and it was just as easy to order for two as for one.
When the order arrived, I changed all seven parts for the VR788
and only the four I felt were required for the VR299 – ie, the fuse, the FET,
the bridge rectifier and resistor.
Well, predictably, I had guessed correctly for the former and
messed up for the latter. In retrospect, I should have replaced the same
components, even though they measured OK in circuit. Mr Ford’s video was now
performing perfectly while the mono VR299 was chucked back onto the junk pile
(after blowing away the smoke)!
Ms Hardy’s service call
Ms Hardy requested a service call for her Panasonic TC-68V80A
TV set (MX-4M chassis), which was dead. When I arrived, it was easy to see the
reason for the set’s demise. She had an absolute waterfront to the Pacific Ocean
and electricity and sea air just do not mix well.
The set was about seven years old and was pretty rusty.
Unfortunately, I had the temerity to suggest that it was past its use-by-date,
especially in this environment. Miss Hardy soon put me right on that score and I
was ordered to take it away and "jolly well fix it, my man"!
After nearly killing myself by carrying the set down the stairs
to the visitors’ carpark area about 15km away, I wasn’t sure which was in worse
condition – the TV or myself.
Anyway, back at the workshop, you could see the green gangrene
of salt-water corrosion extending into the bowels of the set from the rear,
though fortunately it had only seriously penetrated about 50mm into the
components on the PC boards.
My first step was to thoroughly clean the boards, which left
them looking pretty good and with only superficial damage. The brunt of the
attack was on the AV H Board, which protrudes vertically (probably saving the E Board behind), and on the Power Supply P Board.
After replacing the boards, I switched the set on and was
amazed to see it actually fire up and give a picture with sound. The result was
somewhat intermittent, though – especially the picture width – and the power
supply was making a few noises. I substituted a good power supply from another
TV and everything looked pretty good, so I left it on test whilst I tackled the
original power supply.
The components around IC802 HA17555 looked a bit sorry, so I
set about removing and testing them. They all measured OK but I replaced zener
diodes D821 (8.2V), D819 (11V), D830 (9.1V) and D831 (4.7V) just to make sure.
However, I couldn’t replace D836 because I didn’t have one in stock.
Back in the set, the power supply was now a lot more stable but
other things had started to go wrong. There was noise in the left-channel
speaker, the colour was fading and there was no picture on the AV2 input. I
swapped the AV (audio/video) module over and these symptoms all cleared up, so
the fault was in this module.
There isn’t much on the component side of the AV module except
IC3001 (M51321P), transistor Q3005 and a few assorted capacitors. I socketed the
IC and swapped both it the transistor but this had no effect. Surely it couldn’t
be the surface mounted components on the rear of the board, because these
weren’t facing the sea and would have been protected by the board itself?
There are three ICs and nine transistors (all surface mount
devices) on the rear of the board. A quick check showed that all were being fed
from the +12V supply rail, so I decided to look carefully at the symptoms and
solve each problem in turn.
By now, the noisy left-channel sound had become intermittent
"no-sound". All three ICs on the rear of the board are involved with AV
switching, while part of IC3001 is involved with the stereo sound. Using an
audio probe, I could hear the TV sound arrive at pin 18 but it was low on pin 1
compared with that from the right channel on pin 9.
A glance at the circuit showed that pin 1 fed Q3001 and this
transistor is muted by Q3002 on its base. I shorted the base-emitter junction of
Q3002 (to turn it off) and the sound came up. I then found that freezing Q3002
made the sound come and go and replacing this surface-mounted transistor fixed
the sound problem.
Next on the list was the intermittent colour problem. This was
now almost a "no-colour" problem, which was good for me as I had a definite
The set employs what looks like a really weird circuit to split
the colour and luminance, and you have to wonder why they did it that way. It
involves two TC4066BFN analog switches and five transistors. It’s only when you
look at the "upmarket" version of this set, which uses the same board but is
fully loaded with all the components that are missing from the model I was
working on, that you see that the extra SVHS inputs require the switching.
I followed the signal from pin 14 of IC3001 to emitter follower
Q3005 and thence to pin 3 of IC3004. However, at this point, it wasn’t switching
properly to pin 4, so the colour signal stopped there.
Before replacing this surface-mounted IC, I checked the
switching control voltage to pin 5 from Q3013. That proved to be a wise move,
because it wasn’t there.
R3068 and R3069 form a voltage divider to the base of Q3013 and
the only voltage I could detect was on Q3013’s base. But where was this phantom
voltage coming from? It turned out to be Q3013 itself – it was leaky and so the
transistor was switching itself on! Replacing it fixed the "no-colour"
So far so good – that left the "no-picture" on AV2 problem to
The AV2 switching is carried out by IC3003 and a check with an
oscilloscope soon showed that the video was arriving at pin 9. However, there
was no output from pin 8. Due to an error in the circuit diagram, pin 6 (the
control pin) is shown disconnected whereas, in actuality it is connected to the
12V rail via R3040 (10kΩ).
R3040 measured OK and there was voltage on pin 6, so I replaced IC3003 which
fixed the problem (much to my relief).
By the way, I couldn’t obtain the correct IC (TC4066BFN) from
my local supplier, as the ones they supplied were (surprisingly) too wide!
Instead, I had to order originals from Panasonic.
Despite my earlier work, the power supply was still playing up.
I replaced IC802 (HA17555 – the CMOS version of an NE555 timer) to no effect.
Next, I tried heating and freezing around this area and this caused some
dramatic effects, often switching the power supply completely off.
Transistors Q808 and Q809 were proving to be rather sensitive,
so I decided to replace them one at a time. Both proved to be intermittently
internally leaky and replacing them finally fixed the power supply problem.
The only problem left was Ms Hardy. I phoned her about the
considerable cost of the repair and added extra for every time she
condescendingly used the term "my good man".
I am now in traction after re-delivering the set up all those
flights of stairs. I certainly won’t be having anything further to do with Ms
Hardy’s set if I can possibly help it!
The arcing Philips TV
A Philips 29PT9418 (MG3.1 chassis) came into the workshop with
the customer complaining that it was dead! Well, it wasn’t quite dead – it just
wouldn’t start up.
The slow start-up procedure with this set is indicated by a LED
display, starting with red (standby), then green and finally yellow before the
picture and sound come on. In this instance, it would go through this process
but instead of giving picture and sound, the red LED would flash, indicating
that the set was in "Protection Mode". Unfortunately, because I don’t have the
"Dealer Service Tool", I was unable to read the error code.
After removing the covers, I found that I could measure the
+141V rail as it rose until the red LED started flashing, at which point it
would cut off. I checked the line output transistor (Q7421, BU2520PX) and this
proved to be OK. However, an oscilloscope connected to the collector of this
transistor momentarily showed some ringing before the set closed down.
By now, I was satisfied that the flyback (or line output)
transformer was the cause of the problem and so I started to remove it. There is
a support screw from a rear panel to the transformer which has a plastic collar.
When I removed this, I noticed it was carbonised on the inside.
As this set came from a beachside address, I suspected it had
been arcing around the collar. Anyway, I carefully cleaned away the carbon and
switched it back on. This time, success – the picture and sound came on but I
could still detect the telltale hiss of arcing around the flyback transformer. A
new one fixed the fault properly.