This is only a preview of the December 1987 issue of Silicon Chip. You can view 39 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. Items relevant to "100W Amplifier Module to Build":
Articles in this series:
Articles in this series:
Articles in this series:
Articles in this series:
Articles in this series:
|
All this
•
ID
a brand new set
Most service faults fall into a particular pattern: the fault
in a brand new set; the intermittent fault; the multiple
fault; and even the contradictory fault where replacing a
faulty component makes the situation worse instead of
better. Most of us have experienced these from time to time,
but how many have found them all in one set?
Well, that's the gist of my main
story this month and I think it sets
some kind of record. See what you
think.
It all started with a call from a
dealer colleague for whom I do service work, including warranty work
on new sets. This concerned a
brand new TV set - a National
TC2258 - which he had just taken
out of its carton, for display on the
showroom floor, only to find that it
appeared to be completely dead.
And, since National is one of
the brands I am authorised to
service, it was clearly my
baby.
My colleague duly delivered the
set and I dug out the appropriate
manual. As well as the model
number already quoted, some
readers may recognise the set as
using the National chassis type
M14H. The relevant portion of the
circuit is reproduced herewith to
assist readers in following the
story.
My initial reaction to the job was
somewhat blase - I assumed that
it would be a relatively straightfor-
t)J
ALL 11-\lS IN
62
SILICON CI-IIP
~
~N't>
ward fault without any serious
hassles. After all, it was completely
dead. A preliminary check revealed
that there was a normal HT rail at
around 113V, but no sign of
horizontal deflection, EHT, or any
other functions derived from this
part of the circuit.
I stoked up the CRO and began
checking through the horizontal circuit. The horizontal oscillator is
part of of a 'jungle' type chip,
IC601, AN5600k-R, a 42-pin
monster which delivers the horizontal oscillator signal from pin 41.
This part of the circuit is not shown
here but the output from this pin
goes to the base of transistor Q500,
the horizontal driver stage. This
drives the output stage, Q501 , via
transformer T500 (bottom left hand
corner of circuit).
The CRO established that the
signal was coming out of the chip
OK and was being applied to the
base of Q500. But that was all;
there was nothing at the Q500 collector. Nor was the reason hard to
find; a quick check with the meter
showed that there was no voltage at
this point. This voltage is normally
applied via R551, an 820Q safety
resistor, although the purpose of
the safety resistor in this line is far
from clear.
Anyway, this was the problem,
the resistor being open circuit.
Judging by appearances, this
resistor might have been nothing
more than a conventional 2W
metallised type but, in any case, I
had nothing like this value, quite
apart from any special qualities it
might have. Anxious to get the set
working, I made up a string of three
1W resistors: two 330Q and a 2200,
making 880Q, which I reckoned was
near enough for a test.
Protesting squawk
~
SE"t"....
I fitted the string in place and
switched on. The result was a loud
,j.
D30,
"'""
.
.,,.
IO<)
-
.
"",
· 0, 47
ffiID
''"
560P500V
""
TlHl5768
IC401
AN5521
l502
=-tl.1!5lA02
r
~p
SOO'
Fig.1: relevant portion of the National TC2258 circuit. Transistors Q500 and Q501 are at bottom left, Q503
and Q504 at bottom right, and D510 and TPE21 roughly top centre.
protesting squawk from the line
output transformer, then silence;
the set was as dead as ever before.
Further checking revealed that the
horizontal output transistor, Q501,
had failed. Fortunately, I had a
replacement type on hand, and this
was duly fitted. But, before trying
again, I did what I should have done
the first time and checked the value
of the substitute resistor string.
I need hardly add that Murphy
had been at work. What I had taken
for 2200 was, in fact, a 22000
resistor with a somewhat dubious
colour band. My immediate reaction was to wonder whether this
mistake on my part had caused the
output transistor to fail, or whether
it had been faulty all along, but only
became apparent when I restored
the voltage to the driver transistor.
This was the first of many such
questions I was to ask myself before
I finshed with the set.
I fitted a correct value resistor,
crossed my fingers, and switched
on. And this time all was well - no
signs of protest, normal sound, and
a first class picture when the tube
warmed up. The only snag now was
the 8200 safety resistor. I had
ordered a replacement from National, along with other parts, but
when the order arrived there was
no resistor - it had been put on
back order.
Naturally, I didn't want a brand
new set to go out into the field with
a bodgie string of resistors in it but,
on the other hand, the dealer
wanted to put it on display. The upshot was that I explained the situation to him and we agreed that I
return the set to him for display but
on the understanding that the
resistor would have to be replaced
before the set was delivered to a
customer.
And that was more or less that or so I thought. I'm not quite sure
what happened over the next few
days except that it appears that my
dealer friend must have left the set
switched off most of the time, turning it on only when a customer
showed interest in it. Anyway, it
was some time before he had occasion to run it for any length of time.
When he did, he was on the
phone in short order. It appeared
that the set had run for only three
quarters of an hour, then stopped.
He had switched it off for about 10
minutes, then switched it on again,
whereupon it had played normally
for another three quarters of an
hour. He had found that this pattern could be repeated indefinitely
and that a pause of as little as five
minutes could restore performance,
but that this seemed to shorten the
playing time.
And so the set landed back on my
bench. My first move was to simply
run it to confirm the dealer's
description, and to watch for any
symptoms which might provide a
clue as to the nature of the fault.
This approach paid off. After about
40 minutes I noticed that the picture was starting to jitter slightly
from side to side, suggesting possible instability in the horizontal circuitry. A few minutes later there
was another protesting squawk and
the set shut down.
Then, just as the dealer reported,
switching the set off for a few
minutes was all that was necessary
to restore performance. So what
now? There was little doubt in my
mind that it was the horizontal
system failing, the real question being why. Once again I decided that
a CRO would be the best form of attack, at least initially. And this
time, I selected a triple trace instrument, a BWD Model 525.
I connected one probe to the base
of transistor Q500, one to this transistor's collector, and one to the collector of the output transistor,
Q501. With the set running cold,
the waveforms were pretty much as
indicated in the service manual, so
it was simply a matter of waiting
until the picture started to jitter.
My first attempt was not very
successful. I was concentrating on
the output stage waveform when
DECEMBER1987
63
the picture started to jitter, but this
seemed to be perfectly stable. It
was only a moment before the set
shut down that I looked closely at
the waveform at the base of Q500,
and realised that it appeared to
have changed shape.
The change was not very marked
and I put the set through a couple
more cycles before I was sure that
this was so. When I was, I removed
the probe from the collector of
Q501 and connected it to the
horizontal output pin, pin 41, of
IC601. Then I put the set through
another cycle.
This time, results were more encouraging. Initially, the waveform
at pin 41 was very close to that
shown in the manual, in both
amplitude and shape, although it
differs significantly from that at the
base of Q500, due mainly to the RC
network in the base circuit.
But as the set approached its
shut-down condition, the waveform
at pin 41 changed significantly,
both in amplitude (which was
decreasing) and shape. However,
as I noted previously, this had only
a marginal effect on the waveform
at the base of Q500. Nevertheless,
as the signal from pin 41 continued
to change, it eventually reached a
point where the set shut down.
Which was all very interesting
up to a point, but what did it mean?
My knee-jerk reaction was to blame
the chip but, while I didn't entirely
rule out the possibility, I quickly put
that idea on hold. I am coming to the
conclusion that most chips are pretty reliable these days and I am less
inclined to replace them than I once
was. In any case, the idea of unsoldering 42 pins doesn't particularly appeal unless all other
possibilities have been exhausted.
At this point I took some time off
to study the circuit and form a
clearer picture of how this part of it
worked. This didn't seem to help
much at the time, although it did
prove useful later. But it did inspire
me to make a voltage check of the
chip, both when the set was :.unning normally, and when it was
about to fail.
Initially, all voltages were within
normal tolerance, so I let the set
run for about half an hour, then
checked them again. Most of them
64
SILICON CHIP
-- .
A-r Tl-lAT TIME 'I l)E:Ol)e:D
lT ~s TIME TO A'SAN't>OI\) -ms
SC\ ENT\l=\C A??~OAC H P\ \\\'t>
Re~oR"t TO il'\E ?1<\N\\1" \V~ ••. .
showed only minor differences but
one had dropped significantly. Pin
42 is marked 8.5V on the circuit
and had actually read 8.35V when
the set was cold. But now it was
down to 6V and still falling. When it
reached 5.4V, the set shut down.
Now I felt I was getting
somewhere, and my previous study
of the circuit was proving useful.
As nearly as I can make out the
8.5V on pin 42 is the supply for the
horizontal oscillator. It is derived
from the 113V rail via R511, a 6.Bkn
3W resistor (approx. top centre of
the accompanying circuit). From
this point, a line runs (left) to pin 42,
and right to R536 (1000} and thence
to the emitter of Q503 (bottom
right).
Protection circuit
Now Q503, together with Q504,
is part of an over-voltage protection
circuit, designed to operate if the
main HT rail should rise significantly above 113V. It works like this.
Near the top centre of the circuit is
a voltage divider from the 113V rail
to chassis, consisting of a 18 7kQ
resistor (R527), and a 20kQ resistor
(R528}, both 1 %. The voltage at the
junction (no load) should be approximately 10.9V.
This junction in connected to a
zener diode, D510, the exact value
of which is not stated but is obviously somewhere around 10.9V. The
other side of the zener goes to the
base of Q504, which normally has
no forward bias and is turned off. If
the voltage rises at the zener junction, the zener will conduct, turn on
Q504 and Q503, and pull down the
8.5V rail feeding the horizontal
oscillator. As a result, the horizontal circuit and those circuits
operating from it are shut down.
At least, that was what I deduced
from my study of the circuit. Some
of the details were obviously still
missing, but I reckoned I had
enough to go on for the time being.
There was either a genuine overvoltage condition on the HT rail,
causing the shut-down circuit to
function as intended, or there was a
temperature sensitive fault in the
protection circuit which was
creating the false alarm.
I had already checked the HT rail
on several occasions and it had
always been spot on, but I went
through the motions again with
special attention to the shut-down
condition. I also checked the
voltage at test-point TPEZ 1, and
even the EHT. They all remained
rock steady right up until the set
shut down. I did note, however, that
the voltage at zener D510 was
somewhat lower than I calculated
for a simple divider arrangement.
So it looked like a false alarm. All
I had to do was find out why. My
first step was to apply some freezer
to what I felt were likely to be
temperature sensitive components
in this part of the circuit: zener
diode D510, zener diode D502,
diode D513, and transistors Q503
and Q504. None of the diodes
responded to this, but both transistors did. If either one was
sprayed as the 8.5V rail was dropping towards shut-down, the voltage
would rise to normal.
This proved to be a rather surprising finding, in view of subsequent events.
Next I monitored the various
voltages applied to Q503 and Q504
as shut-down approached. I checked them first in the cold condition
and found them to be virtually spot
on. Then I let the set run and watched for any changes. An obvious
change was at the emitter of Q503
which was virtually the same point
as pin 42 of the chip, so this fell
gradually as shut-down approached. The same applied to the collector of Q504, and the base of Q503,
which are connected to the same
line.
But the real puzzle was that
there was no change to the voltage
on the base of Q504 (and the collector of Q503), where one would expect to find the 'alarm' voltage real or false - needed to turn on
these two transistors and pull down
the 8.5V rail. At this point I decided
to abandon the scientific approach
and resort to the primitive; ie,
check each of the dozen or so components in this part of the circuit
individually.
The resistors were easy enough
to check in situ, and all came up
well within tolerance. Transistors,
on the other hand, are best removed from the circuit for testing. I
pulled Q504 out first and found that
it was not only faulty, but faulty in a
rather unusual way. It had a base
to emitter short and an open circuit
collector.
The discovery was gratifying, of
course, but was also puzzling. I
couldn't relate the fault to the
symptoms and, in particular, I
couldn't reconcile the discovery
with the fact that spraying the transistor with freezer appeared to
cure the fault.
But there was no point in dwelling on this. I fished out a suitable
replacement transistor and fitted it ,
then tried the set again. Result: the
set wouldn't even start. So it was
back to the component by compo-
nent check. I lifted zener diode
D502 and it checked OK. Then I
tried D513, and this proved to have
a high resistance leak. So that was
replaced.
Still the set refused to work. I
removed Q503 and checked it, but
could find nothing wrong with it. I
was feeling deperate now and
decided to lift zener diode D510 and
thus render the protection circuit
inoperative. This took only a moment and the set came good immediately, with all voltages and
waveforms normal. What was
more, it continued to run for the
next several hours, with no significant change to any of these
parameters.
wrong value replacement for R551?
Apart from stating that I don't
think this last failure happened in
this way, I really can't answer
these questions.
In a sense, I suppose, all this is
rather academic. The faults have
been found and the set repaired,
and that is all that really matters.
At least that is the practical approach and, as I have commented
on previous occasions, there is a
limit to how much time one can
spend mulling over the whys and
wherefores of circuit behaviour
under fault conditions.
No more squawks
Which didn't leave much to
suspect except the zener, D510. As
I said earlier, there was no indication as to the exact value of this, but
I reasoned that it was probably a
10.BV type. So, without bothering to
check the old one, I fitted a 10.BV
type out of stock and tried again.
And away went the set like a
bought one!
Fortunately, the manual
describes a check for the protection
circuit. It calls for the application
of a progressively higher voltage,
from an outside source, to test point
TPE21, until the set shuts down,
which it should at about 11.BV. I did
this and it came out spot on.
And that was more or less the
end of the job. I ran the set all day
for the next couple of days, then
returned it to the dealer. It has
given no trouble since .
Naturally; I was glad to have
solved the problem and to have the
set off my hands, but I was less than
happy about the reasons for what I
found. To start with, why were
there so many faults in one set, and
a brand new o:i:J.e at that? A faulty
safety resistor (R551), a faulty
horizontal output transistor (Q501),
a faulty protection transistor
(Q504), a faulty diode (D513), and a
faulty zener diode (D510).
Did all these faults occur independently, or did one fault trigger all the rest? Or perhaps only
some of them? And if one fault triggered the rest, which one was it?
And did I destroy Q501 by fitting a
'RE..\I\Vll'JG ~ FA\n\FU\Ol-!) 'Pt\\\..\~S
K9 ....
And now for a change of scene
from a very new set to a very old
one: the old faithful Philips K9. The
owner had been a customer since
he bought the set, over 11 years ago
now, and while it has developed the
usual faults common to this model,
it has given good service and has
plenty of life yet.
This particular problem is interesting because it demonstrates
that there is always something new
to be encountered, even in a model
about which one imagines one has
seen all the tricks, as I did about
this one. It also emphasises that it
pays to be suspicious, perhaps
more so than I was on this occasion.
The fault itself was fairly
straightforward; the set was completely dead due, in turn, to the fact
that the power supply was also
DECE/IIHE/l 1987
65
VL 170
:~
C175
....--->,;N,--_, I
I
I
I
I
!~
;~
I
I
I
Fig.2: skeleton circuit of the Philips K9 power supply switching transistor
network. Why was it tough on the transistor?
completely dead. Since this is not
an uncommon situation with this
model, I keep a spare power supply
board on hand for a quick test. By
plugging in a known good board I
can quickly establish whether the
power supply is itself faulty, or
whether it is being shut down by a
fault elsewhere in the set.
In this case the test clearly indicated a faulty power board and,
from there, it was a short step to
pinpointing the real culprit. It proved to be the switching transistor,
TS170 (2N472), which had shorted
collector to emitter and blown the
2A fuse , VL170, in the process.
Simply replacing these two components would have been enough to
put the set back into operation, but
experience has taught me that
there are several other faults
which can initiate this failure (note:
TS 170 can be a BUl 26 in some
models, or even a BU326 if the
original has been replaced in the
field}.
The main off enders are dry
joints, particularly in those parts of
the circuit where they can cause
sparking and spikes. Joints to
transformer T182 are common offenders , as are those to the main
smoothing capacitors, Cl 78a and
Cl 78b. All these possibilities were
thoroughly checked and ruled out,
the faulty components replaced, the
set given a test run, and returned to
the owner.
All went well for about three
months and I had virtually forgotten about the job until the owner
turned up at the shop again with
the K9 in the back of his utility. It
was completely dead again and it
turned out to be exactly the same
fault. Naturally, I was suspicious,
and I went over the board again,
determined to make sure there was
nothing I had missed the first time
around.
In fact I found nothing, and I was
forced to the conclusion that it was
mere coincidence, unlikely though
TETIA CORNER
Rank Arena C2239 (B2 Chassis)
Symp'tom: No picture but sound
OK. Rather dark. raster with
retrace lines on channel, but raster
becomes lighter off channel.
Brightness and contrast controls
have no effect. Y signal disappears
at pin 5, IC701. Voltage at pin 2
IC701 much higher than it should
be.
66
SILICON CHIP
Cure: TR402 short circuit. This
transistor is one of two forming a
multivibrator that generates the
vertical blanking signal. When it
fails the blanking is turned hard on,
and IC702 is permanently
blocked.
This information supplied by The
Electronics Technicians' Institute
of Australia (Tasmanian branch).
this seemed. So I fitted another new
transistor and fuse, gave the set a
test run, and returned it to the
owner. And, again, all went well for
another three months. Then the set
was back in the shop with the same
fault.
No coincidence
Well, that settled it. There was
no way I could accept a third
failure as coincidence; there just
had to be a more subtle fault that
was responsible. Since I felt fairly
confident that I had excluded the
more common causes of this problem, such as dry joints, the alternative approach seemed to be to get
the set running and make as many
dynamic tests as possible.
So I replaced the transistor and
fuse, turned the set on, and started
with some voltage checks. Unfortunately, the circuit diagram gives
very little information in this
regard, so the best I could do was
check the voltages which were
marked - which all turned out to
be within tolerance - then make as
many other checks as I could think
of and try to relate them to what I
imagined they should be. Once
again I found little to make me
suspicious.
Having drawn a blank with the
voltages, the next thing I wanted to
do was check the various
waveforms around this section.
But, once again, the circuit is
noticeably lacking in such information. Such waveforms as are shown
seemed hardly appropriate to the
problem. In particular, I would
have liked to have known the appropriate waveform for the collector of TS 170.
Fortunately, there seemed to be a
way out. I plugged in my stock
power board and connected the
CRO to it. In so doing, I had to take
into account that the negative
reference point is not the chassis
but a point on the circuit marked as
the reference point for certain
voltage measurements (marked in
red}, and which is close to the
negative end of the main bridge
rectifier.
The resultant waveform was
basically a square wave, but with
some ringing on the leading edge.
More exactly, the basic square
wave had an amplitude of about
550V, while the ringing had a peakto-peak amplitude of about 100V, or
50V above the 550V, making a total
peak value of 600V. I made a note of
this and was about to plug in the
set's own board when I had a stroke
of luck.
Another K9 came in for a
relatively simple fault, providing a
golden opportunity for another
reference. In fact, it produced a
waveform almost identical with
that from my own board, so I felt
reasonably confident that this was
how it should be. All that remained
now was to check the suspect
board.
And this gave me my first real
clue. In general terms, the
waveform was the same, at least in
shape, but the values were
significantly different. The square
wave portion was now up to 600V
while the ringing amplitude had
also increased, now running at a
good 200V p-p, or 100V above the
square portion, giving a peak
amplitude of 700V.
Considering the set's history over
the last few months I felt sure that
this was the most likely cause. But I
still had to find out why it was happening and, in the process, prove
that it wasn't just a normal spread
of component values. And now that
I had something definite to go on I
began to recall some suggestions I
had heard about causes of excessive ringing and possible
destructive spikes.
Among the suspect components is
R175 (120} and R174 (560), both in
the base circuit of TS 170, which, if
they go high, can cause excessive
ringing. Also in the base circuit is
C177 (4.7µF} which apparently can
cause trouble if it drops its value.
These were checked and found to
be spot on. Less likely suspects are
R182 (0.330 or, in some cases, 100),
and R176 and R177 (18kn}. These
were also checked and cleared.
So what now? With the base and
emitter circuits seemingly cleared,
what about the collector circuit?
Among other things this contains
three filter networks consisting of
resistors, capacitors and diodes.
The failure of any one of these
would surely have some effect on
the waveform.
I started with D176, R173 and
Cl 93, all of which checked out OK.
So did D179 in the adjacent network, but R190 (150k0) was a different matter - it was open circuit! I replaced this and then,
before returning the board to the
set, checked out the third network
consisting of D178, R189 and Cl 75.
All were OK.
I put the board back into the set,
hooked up the CRO, and switched
on. Up came a perfect waveform; ie,
identical to the other two boards.
And that, as far as I was concerned, was all the proof I needed. I
have since handled several more
K9s and have made a point of
checking the waveform as a matter
of routine. All have been essentially
the same as my own stock board.
So that was it. I put everything
back together, ran the set for a couple of days, then returned it to the
customer. Only a few weeks have
elapsed since then so it is much too
early yet to boast. All I can do is
keep my fingers crossed and hope.
-1e1gscHlADIResistor kits for
Laboratory * Service * Workshop
These comprehensive resistor kits are ideal for the engineer, technician
and the dedicated enthusiast.
No more wasting time trying to find odd value resistors.
No more high procurement costs for small order quantities.
Have them at your fingertips when you need them , neatly stored for
quick selection in transparent rectangular tubes or compartments, clearly marked with the IEC Series resistance values.
Kits are available with 50 to 586 resistance values and containing from
1,210 to 26,000 resistors, in surface mount (MELF) or leaded types.
This is our Mini-MELF (LMM
0204-24) lab assortment
containing 6500 resistors.
For further information contact:
1
f~
~~\
CRUSADER ELECTRONIC COMPONENTS PTY LTD
~I//J\\\"":
81 Princes Hwy, St Peters , NSW 2044 . Phone (02) 516 3855 , 519 6685, 517 2775. Telex 123993 . Fax (02) 517 1189 .
APPOINTED DISTRIBUTORS
Sydney: George Brown & Co Pty Ltd , 519 5855. Geoff Wood Electronics Pty Ltd , 810 6845. Wollongong: Macelec Pty Ltd,
29 1455 . Canberra: George Brown & Co Pty Ltd , 80 4355. Newcastle: Novocastrian Electronic Supplies, 62 1358.
· Melbourne: Nalmos Pty Ltd , 439 5500. Jesec Components Pty Ltd , 598 2333. George Brown & Co Pty Ltd, 878 8111.
Brisbane: L. E. Broughen & Co, 369 1277 . Colourview Wholesale Pty Ltd , 27 5 3188. St Lucia Electronics, 52 7 466.
Adelaide: Protronics Pty Ltd, 21 2 3111. Perth: Simon Holman & Co, 381 41 55. Protronics Pty Ltd , 362 1044.
DECEf\lBER 1987
67
|