This is only a preview of the January 1996 issue of Silicon Chip. You can view 22 of the 96 pages in the full issue, including the advertisments. For full access, purchase the issue for $10.00 or subscribe for access to the latest issues. Articles in this series:
Articles in this series:
Articles in this series:
|
SERVICEMAN'S LOG
The complaint seemed simple enough
Yes, it did sound simple. And, relatively
speaking, it was. The trouble was, it didn’t stop
there – it had brought all its gremlin mates along
with it. By the time I’d knocked them all over, it
was a major exercise.
This story concerns a Sanyo colour
TV set, model 6627 (79P chassis),
which lead me a merry dance with a
succession of faults – these in addition
to the original complaint.
The set belongs to a pensioner, one
of several among my regular customers, and whom I regard as being in
something of a special category. In
general, their equipment tends to be
older than average, for the very simple
reason that, for many, the cost of new
equipment is almost prohibitive. So
they keep their old units and call on
me to keep them going for as long as
possible.
Of course, I do my best to help
them, even though at times it taxes
one’s ingenuity and patience. (After
all, I’ll be old myself someday – and
no editorial comment, please). Anyway, this case was a classic example
of this sort of job and, as is typical,
involved a set that was over 10 years
old.
But the owner’s complaint seemed
simple enough – distorted sound. And
a quick check while he was there confirmed the complaint; the distortion
was quite bad. Even so, I reckoned
it should be a snack; that I would be
able to knock the job over in no time.
And that, as the reader has doubtless
guessed, was where I came a gutser.
Sound circuitry
The sound section in this set is quite
straightforward – see Fig.1. It consists
of a sound IF amplifier and demodulator IC (IC151), the latter feeding two
output transistors, Q151 and Q152.
These are both specified as 2SC2568
or 2SC2456.
80 Silicon Chip
My first step was to check the 220V
main HT rail, which came up spot
on, as did several secondary rails derived from it. OK, so where to in the
sound section? My first inclination
was to suspect one of the two output
transistors and, with more haste than
wisdom, I whipped them out and
tested them.
They both tested OK, which
served me right for rushing in. I
then did what I should have done
first – checked the voltages around
these transistors. And, yes there was
something wrong. The base voltage
of Q151 is shown on the circuit as
“80V-106V”, which seemed an unusually large spread. But that was
largely academic anyhow, because
the actual voltage was way down on
even the lower figure.
And this was where I encountered the first of several discrepancies between the set on the
bench and the circuit. And I
don’t meant bodgie repairs;
I’m referring to original components. The bias resistor
for Q151 (R151) is shown
as 39kΩ but the one in the
set was 27kΩ. Or, more correctly, it was coded 27kΩ.
In fact, it measured over
100kΩ.
Well that seemed like the
answer and I promptly fitted a
new 27kΩ resistor. That brought
the voltages back to within tolerance of those on the circuit and
wiped out most of the distortion.
And I say “most” because there
remained a niggling level. It was
nothing like the original but it
was enough to indicate that there was
still something wrong.
And that’s just about the nastiest
kind of fault I can imagine. It was at
such a level that, at times, on certain
pro
g ram material, one could kid
oneself that it wasn’t there. Then the
program would change and it was all
too obvious. There was nothing for it;
it had to be found.
So, with all the stage’s operating
voltages restored to normal, where
should I go from here? The IC seemed
the next most likely culprit. I had one
on hand and changing it was not a
particularly difficult job. But, alas, I
drew another blank.
Down to basics
It was time to really get down to
basics. I went right over the output
stage and, by one means or another,
checked each component in turn. And
in the process, I encountered another
circuit discrepancy; a diode, D153,
which was in the set but not on the
circuit. It has been drawn in on the
circuit shown here.
I paid particular attention to electrolytic capacitors C151 (1µF) and
C157 (2.2µF on the circuit, 4.7µF in
the set). Low value electrolytics are
always suspect. But these and all the
other components, except one, were
cleared.
That one component was C153, a
5600pF capacitor connecting to pin
13 of the IC. And I had left it until last
because, initially, I couldn’t identify it.
I had been looking for a small ceramic
capacitor or something similar but
without success. In the end, I had to
trace the copper pattern and, when I
found it, it was quite a surprise.
It wasn’t a ceramic capacitor and it
wasn’t 5600pF. It was an electrolytic
and it was 0.47µF; the biggest change
from the original circuit I had found
so far. More to the point, being an
electrolytic – and of very low value
to boot – it was a prime suspect and I
lost no time in reefing it out and fitting
a new one.
And that was the answer, with the
set now producing clean sound. And
to confirm it, the suspect electrolytic
showed sub
stantial leakage when
tested. So that looked like the end of
the exercise. I gave the set the usual
once over for general performance and
minor adjustments, then set it up on
the end of the bench and let it run.
The set carks it
Initially, it ran for several hours and
then, suddenly, I was aware that it
was completely dead, with no picture
and no sound. Well, I took another
punt: the horizontal output transistor
(2SD838 on the circuit, 2SD621L in
the set). And I picked it in one; it was
short circuit.
This failure, in itself, did not present any real problem, except that the
2SD838 was cheaper – at around $30
– than the 2SD621L ($42) but was no
longer available. It was something of
a slug for a pensioner but that’s life.
More importantly, I was concerned
as to why the transistor had failed. It
has a pretty hard life in this set. The
waveform shown on the collector is
1900V p-p which is high by any stan
dards. That means that the stage is
vulnerable to any spikes or rubbish
on the driving waveform. And from
experience, the most likely cause is
a failure in C483, a 1µF electrolytic,
which decouples the 220V rail to the
horizontal drive transistor, Q481.
Again, experience has shown that
this capacitor dries out, allowing all
kinds of rubbish to reach the driver.
So I pulled it and replaced it. And,
as an attempt at insurance, I upped
the value to 10µF. I can’t guarantee
how much it will help but it won’t
do any harm.
When I switched the set on again,
there was sound but no picture. So
what on earth could be wrong now?
My first reaction was to suspect the
operating voltages on the picture tube.
I fished out the probe and checked
the EHT. There were plenty of volts
there, something over 25kV, and so
I checked the screen voltage, focus
voltage and the RGB drive transistors.
All seemed OK.
I have experienced trouble in the
past around transistor Q191. This
forms part of the ACL (Automatic Contrast Limiter) circuit and the problem
concerns resistor R197 (220kΩ) which
goes high. And while the trouble
had never been anything like this, I
checked it, found it somewhat high
January 1996 81
Fig.1: the audio output stage in the Sanyo 6627. Note the additional diode
(D153) which has been drawn in between the base and emitter of Q151.
As well, R151 is now 27kΩ, C157 is now 4.7µF and C153 (top left) is now
a 0.47µF electrolytic.
and replaced it. But I wasn’t surprised
when it had no effect.
Next I did a waveform check, right
through the video chain, but could find
nothing wrong. I stopped and had a
think and a caffeine fix and went over
the checks I had made.
And suddenly I became suspicious.
I realised that all the voltages I had
measured – EHT, screen, focus, RGB,
etc – had all been marginally high. I
hadn’t taken as much notice of this as
I should have, the complete picture
failure suggesting a total loss of voltage
somewhere.
Now I went back to taws – the main
HT rail. And there was the answer, or
part of it. Instead of the previous spoton 220V, it was now 275V. It was only
a symptom but it was a start. I went
straight to the power supply and, after
a few preliminary checks, attacked
Q901, the power regulator. And that
was it; it was short circuit.
I fitted a new one and switched on.
And everything came up roses; 220V
on the HT rail and a picture on the
screen.
And that was the end of the drama. But why did the excessive HT
rail voltage create the effect it did?
Frankly, I don’t know. I considered a
number of likely reasons – including
the possible action of an over-voltage
protection circuit somewhere in the
system – but I’m afraid I was too fed
82 Silicon Chip
up with the set to want to spend any
more time trying to find out. I let it run
for another day or so, then called the
customer to come and collect it. And
I was glad to see the back of it.
Granted, I was lucky in one way. At
least those secondary faults occurred
while the set was still on the bench. If I
had returned the set immediately after
fixing the first fault – as I might have
done had the customer been in hurry –
then I would have had it bounce. And
that can generate bad will on the part
of the customer.
So let’s be thankful for small mercies.
The crook Telefunken
My next story is about a Telefunken
colour set. It used an ICC4 chassis and
while it had its problems, it wasn’t
quite the headache of the previous
story.
The set came from a colleague. He
passed it over to me for a couple of
reasons. First, he is not particularly
keen on servicing European sets and,
second, he was rather snowed under at
the time and didn’t want to be caught
with something that might take up a
lot of time. And I gathered that it had
been through several other organisations before it came to him.
The complaint was quite straightforward; it was completely dead. Fortunately, my colleague had a circuit,
although it was a trifle grotty in places.
And so I let myself be saddled with
the monster.
It was quite an elaborate set, with
most of the modern features: a very
impressive remote control system,
Teletext, and so on. As with any dead
set, the first thing to check is the rail
voltages and, by implication, the
power supply. So I went straight to
the power supply. And, yes, it was
completely dead.
The supply itself is a fairly standard
switchmode arrange
ment, the main
difference being that, in order to accommodate the remote control on-off
function, the supply runs continuously while ever the power point is on.
The set itself is turned on or off via its
12V rail and this comes from IP61, an
LM317T adjustable 3-terminal regulator. This regulator is in turn controlled
by a signal from pin 7 of IR25.
In addition to the aforementioned
12V rail, there is also a 13V rail, a 22V
rail and a 90V rail, the latter being the
main supply rail. And, at first glance,
there also appears to be a 17V rail emanating from the chopper transformer
(UP40).
In fact, this is something of a furphy;
the 17V rail is actually generated at pin
10 of the horizontal output transformer
and this apparently takes over from the
13V rail (which feeds the regulator)
once the horizontal stage fires up.
(Note the arrow configuration on the
17V block).
No voltage
More to the point, there was no
voltage on any of these rails. I moved
over to the primary side of the chopper transform
er (UP40). There was
voltage out of the bridge rectifier
and, in fact, this was applying some
350V across the main filter capacitor
(CP11 – 100µF). I traced this through
the primary winding, pins 9 & 1, of
the transformer to the collector of the
chopper transistor, TP32.
I subsequently spent some time
checking likely components around
this stage but could find nothing
wrong. But I did make one useful observation. With the CRO connected to
the waveform points indicated on the
circuit, I found that, at the moment of
switching on, there was a very brief
indication of activity but the waveforms vanished almost immediately.
The stage was trying to oscillate but
couldn’t continue.
Fig.2: part of
the switchmode
power supply in
the Telefunken
ICC4. IP61 is an
LM317 adjustable
3-terminal
regulator which
produces a +12V
rail. This +12V
rail is switched
off (to turn the
set off) when the
main control IC
pulls pin 2 of the
regulator low (via
a transistor).
This started a different train of
thought. Perhaps there was a short
circuit or overload on one of the rails
which was placing an unacceptable
load on the power supply?
First, I checked each rail with the
ohmmeter but found nothing suspicious. This was not conclusive of
course – there could still be a breakdown or leakage at the operating voltage, which would not show up with
an ohmmeter check. I also checked
the diodes supplying each of the rails.
Again I drew a blank.
Next, I checked the horizontal output transistor, TL37 (BU508A), which
connects to pin 2 of the horizontal
output transformer (UL65) and thence
to the 90V rail via pin 6. This checked
out OK.
On the basis of all these tests, and
assuming that the overload theory
was still a valid suspicion, the next
obvious step was to disconnect each
of the rails in turn. I started with the
90V rail by disconnecting the 0.22Ω
safety resistor, RP51, at pin 2 of UP40.
As it turned out, this was the wrong
way to do the right thing. It was right
because the power supply now show
ed signs of life. Each of the other rails
now came up, partially and briefly,
and then died away. (On reflection, I
suspect that the aforementioned weird
17V rail configuration had something
to do with this strange behaviour).
Unfortunately, disconnecting the
rail at that point was the wrong way
to do it, because it was directly on the
transformer pin and did not allow me
to check the 90V rail itself.
I restored the 0.22Ω resistor and
went back to the horizontal stage. This
is a very complex arrangement and
difficult to follow, both in the set and
on the circuit. But the 90V rail goes to
a choke (LL54), through diode DL56,
and thence to pin 6 of the horizontal
output transformer via LL57. And
LL54 provided a convenient place to
break the 90V rail and check it.
In fact, it did come good, in a similar manner to the way the other rails
had responded. The trail was getting
Fig.3: part of the horizontal output stage in the Telefunken ICC4. The 90V rail connects (via LL54, DL57 & LL57) to pin
6 of the horizontal output transformer, while pin 2 connects to the horizontal output transistor (TL37 – not shown).
January 1996 83
SERVICEMAN’S LOG – CTD
warmer now. I restored the connection
at LL54, then disconnected the rail at
pin 6 of the transformer – same result.
So the fault was either somewhere
on the other side of this transformer
winding, in a circuit connected to one
of the other windings,
or in the winding
itself. I restored the
pin 6 connection and
lifted the pin 2 connection. And it was
a different story this
time. I was now back
to the original fault,
with no voltage on
any of the rails.
By now, I was becoming more and
more suspicious of
the transformer itself
– so much so that I
went for broke and
pulled it out. My idea
was to check it for
shorted turns, which
I felt was the most
likely explanation.
But first I made a
routine check of each
winding with an ohmmeter. Well, they were
all intact individually
but when I happened to check between
winding 2-6 and winding 1-5, I struck
oil; there was a dead short between
them.
Naturally, there was only one answer to a fault like that; I needed a new
transformer. But that had me worried
initially because I knew of no current
Australian agency for Tele
funken.
Fortunately, a few enquiries revealed
that Hitachi use the same transformer,
a type 243445. In fact, as I understand
it, they actually make it and Telefunken buys it from them. Anyway, they
are readily available, and one was
obtained and fitted.
End of story? Not quite. Oh, the
switchmode supply leapt into life
alright at switch-on but there was one
little snag – the set was still dead.
A quick check with the voltmeter
provided the first clue; all the rails
were up and spot on, at least out of
the power supply. But there was no
12V rail out of pin 3 of regulator IP61.
The reason wasn’t hard to track down.
As mentioned earlier, IP61 is controlled by pin 7 of the remote control
IC, IR25. This control signal is fed
to pin 2 of IP61 via transistor TR74
(BC547B). And TR74 was shot – it was
as simple as that.
A replacement BC547B was fitted
and I finally had everything running
at full bore. And a very nice result it
was too. I gave the set the usual routine
adjustment check, let it run for a day
or so, and then passed it back to my
colleague to return to his customer.
It wasn’t going to be cheap, of
course, taking into account the new
transformer. But that’s the way it goes
SC
and I hope he was happy.
20 Electronic Projects
For Cars
Yes! Please send me ___ copies of 20 Electronic Projects For Cars
Enclosed is my cheque/money order for $________ or please debit my
❏ Bankcard ❏ Visa Card ❏ Master Card
Card No.
Price: $8.95 plus $3 for postage. Order by
phoning (02) 9979 5644 & quoting your credit
card number; or fax the details to (02) 9979
6503; or mail the coupon to Silicon Chip
Publications, PO Box 139, Collaroy, NSW
2097.
84 Silicon Chip
Signature________________________ Card expiry date_____/______
Name _______________________Phone No (_____)____________
Street
PLEASE PRINT
_________________________________________________
Suburb/town _____________________________ Postcode_________
|