This is only a preview of the June 1992 issue of Silicon Chip. You can view 46 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:
Items relevant to "Multi-Station Headset Intercom; Pt.1":
Articles in this series:
Articles in this series:
|
SERVICEMAN'S LOG
From a snack to a 3-course meal!
The above gastronomical analogy may appear a
trifle strange but it's quite appropriate in the
circumstances. What presented itself as a simple
snack, turned into a full meal and a severe case
of technical indigestion.
The device involved was a Samsung
50cm colour TV set, model CB5025J,
and the fault was plain for all to see.
The antenna socket was broken out of
the tuner, which is quite a common
one with this particular Korean-made
unit. And this same tuner is used in a
whole range of other sets: Goldstar,
Orion, AWA, NEC, Sharp and Tandy,
to name just those I can recall offhand.
So, as I implied above, it looked
like a snack; particularly as the set
was under warranty and it was a simple job to replace the tuner. In these
circumstances, I didn't pay a great
deal of attention to the owner's comments when he showed me what had
happened. Had I done so, I would at
least have been better prepared for
subsequent events.
Our loss is
your gain
Fracarro's
ESVU2. The
really low
loss VHF/
UHF mixer
with areasonable
price tag.
Peter C. Lacey Services Pty. Ltd.
P.O. Box 678 (74Fulton Ad.) Mount Eliza 3930
Tel:03 787 2077 Fax:(03) 787 3460 ACN006893438
36
SILICON CHIP
In fact, he had stuffed the socket
back in position and secured it with
sticky tape and a prayer. And although
the broken connection was not restored, there was apparently enough
capacitive coupling to provide a picture of some kind, even if it was snowy.
And, he added, there wasn't any colour.
Well, I didn't take much notice of
that last remark; what else could one
expect with a signal that weak? So I
simply went ahead and fitted a replacement tuner. And, of course, this
resulted in a nice bright, snow-free
picture; one which anyone would be
happy to watch - except for one problem. There was still no colour!
Suddenly the owner's, comment
took on a lot more meaning. The lack
of colour was obviously a separate
fault and, had I taken the trouble to
question him, I would have learnt
that this had been obvious before the
tuner socket collapsed.
A fair dinkum fault
Anyway, that was history. I now
had to track down a fair dinkum colour fault. Fishing out the circuit, I
concentrated on IC501 (TDA3561A)
which performs the decoding functions. I established that the video signal goes in on pin 3 and that, according to the CRO, there was a normal
signal at this point. The gating pulse,
to admit the 10Hz reference burst,
goes in on pin 8 and this was also
normal.
The next logical check was around
the reference oscillator crystal, X501,
between pins 25 and 26. And here it
seemed that I had struck oil; this time
the CRO report was negative - there
was no reference oscillator signal.
So, was it a faulty crystal? This
seemed the most likely cause and, in
any case, was simple enough to check.
I fished a 4.43MHz crystal out of stock
and fitted it. Result - still no colour,
even though I now had a reference
oscillator signal.
First whinge
However, this signal was very low
in amplitude. And that brings me to
the first of several whinges I have
about the circuit and the data supplied by Samsung. The circuit and
manual are seriously devoid of data
of this kind, which meant that I was
·on my own. I took a punt on a signal
of about 2V p-p and, in the event, this
estimate was pretty well spot on. But
right now, I had only a fraction of this
value.
I woke up when I took a closer look,
through the jeweller's loupe, at the
minute markings on the crystal I had
taken out. The crystal was not marked
4.43MHzbut8.86MHz. Whythisvalue
is used I don't know but it appears to
be the latest trend.
"
I didn't have such a crystal, which
meant that I had to order one from
Samsung, and this put the job back a
couple of days. But it arrived fairly
promptly and I lost no time in fitting
it. And this confirmed my previous
estimate of the reference oscillator
amplitude; it was almost exactly 2V
p-p. But that's all it did- I still had no
colour.
After some thought, I reasoned that
the new crystal, while almost certainly
within tolerance, was sufficiently different from the original to prevent it
from locking onto the burst signal with
the setting used forCV0l, a 22pFtrimmer in series with the crystal.
In any case it is good practice, when
a new crystal is fitted, to adjust this
trimmer so that the reference oscillator is running as close as possible to
the burst frequency. This then ensures
"test points", I would have been an
hour
ahead.
HATIIIX-ADJ
CSl4
Anyway, I went ahead and bridged
TSOI
A524
SIO
these "test points". This should have
RS:ZO
produced colour in one form or an12K
~~ I
other. If the reference oscillator's free~
HATIIIX-AllJ
running frequency is sufficiently accurate, it can be an otherwise normal
C521 221
0$01
pattern, but with the colour slowly
1N4141
drifting from one hue to another as
R521
m tlO
the phase of the oscillator drifts, relative to the burst.
More usually the error is a good
deal greater, producing horizontal colour bars, superimposed on the vertical bars, which also drift in colour as
the phase changes. And the greater
the degree of error, the more horizontal colour bars.
But while that may be all very fine
(201
:ol.4111'
in theory, it was oflittle practical value
in this case. There was still no sign of
colour in any form after I bridged pins
23 and 24.
So where to from here? As I have
already noted, I had checked the sigFig.1: the colour decoding ciFcuitry in the Samsung CB5025J. IC501
nal into the chip at pin 3 and the burst
is the colour decoding chip & the crystal is at top left, between pins
gating
signal at pin 8, and found both
25 & 26. The delay line (DL501) is located directly above the chip.
Note the poor qualify of the diagram which makes troubleshooting
to be correct. But what about the sigjust that much more difficult.
nal out of the chip? This appears at
pin 28 and goes to the delay line,
DL501.
Or that's what should have happened. But there was nothing at pin
that the oscillator locks in reliably, and no waveforms. Nor is it always
28. And, in view of all the other tests
even on a weak signal.
strictly correct and , to make it even I had made, I was forced to conclude
The standard procedure for this harder, the printing quality is often so that I had a faulty chip. Fortunately, I
adjustment is to first disable the lock- poor that it is literally impossible to had a chip in stock, so I simply fitted
ing function between the burst signal read the component values. The cir- it.
and the reference oscillator. This lets cuit used to illustrate this article is a
And that fixed it? No way - there
the reference oscillator run free and typical example.
was still no colour and I was looking
Try to read the IC component at a brick wall with no ideas left. All I
the set's behaviour is then monitored
using a standard vertical colour bar number and type number directly could think of was to talk to the maesabove the righthand end of the chip tro at Samsung and hope that he had
pattern from a colour bar generator.
diagram. The type number inside the some ideas.
The reference oscillator is then adjusted to produce the closest approxi- diagram - top left corner - is much
The result was a lengthy discusmation to a stable and correct colour clearer but is wrong. The correct des- sion during which we covered the
ignation, as noted earlier, is TDA- history, the symptoms and all the
display.
3561A, not TDA3562A as shown here.
things I had tried. In fact, we thrashed
Second whinge
This mistake caused me to order the the whole problem about pretty thorWell that's the broad theory. But wrong component on a previous oc- oughly. But it was all to no avail; they
how does one unlock this particular casion, resulting in an inconvenient were just as much against the brick
delay. (There is a TDA3562A. It is wall as I was. So I was on my own.
system? And this brings me to the
second of the whinges I have about used in circuits designed to decode
I went over the circuit again, both
the data supplied. The manual states both SECAM and PAL signals).
on paper and on the board, checking
But back to the unlocking proce- voltages, resistors and capacitors,
that the system may be unlocked by
bridging test points 23 and 24 but dure, I eventually worked it out by many for the umteenth time in case I
gives no indication as to where, on reverting to basics and doing a lot of had missed something. I drew a com, the board, these test points are lo- circuit tracing. But I wasted over an · plete blank.
cated. Nor are they indicated on the hour in the process and that's not
funny. And where did I finish up? At Transcendental meditation
circuit or anywhere on the board.
The system falls down in other ways pins 23 and 24 on the chip. So, if the
When I run into a brick wall like
too. The circuit carries no voltages manual had said "pins" instead of this, I find that it is often beneficial to
ll!1..J
1m
;
-- -
JUNE 1992
37
SERVICEMAN'S LOG - CTD
put the offending device to one side
and, as far as it is concerned, go into
transcendental meditation mode. And
by that I don't mean squatting on the
ground and contemplating one's navel. (Physically impossible at my age
anyway)!
No , I mean tackling a completely
different job and banishing the problem completely from one's mind - at
least as far as this is possible. And it is
surprising how often it works.
More to the point it worked this
time. Suddenly, when I least expected
it, I realised that there was one component I hadn't checked - the chrominance delay line. And I hadn't checked it because it was obvious that it
couldn't possibly have any bearing
on the problem. For one thing I
couldn't recall ever having had any
problems with chroma delay lines;
luminance delay lines, yes, but not
chroma types.
In any case, what kind of a fault
would create these symptoms? An
open circuit could create problems
further down the video chain but how
could it kill the signal at the IC output
pin?
Nevertheless, the thought kept nagging away at me and I was grasping at
straws. Fortunately, I had a fair stock
of these devices on hand, mostly salvaged from scrapped sets, and it was
only a few moments work to prove
that my theory was wrong.
Full colour
Only it wasn't wrong. The set came
to life in full colour and, of course,
there was now a nice clean signal of
around 0.5V at pin 28. So, in some
mysterious way, the original delay line
was loading pin 28 to the point where
it was producing virtually zero output.
And I use the word mysterious deliberately. No one, including those at
Samsung, has yet been able to offer
any explanation. The temptation is to
assume that the input transducer had
developed a short circuit. But, as I
understand it, these transducers are
piezoelectric devices, presenting a
capacitive load to the input signal.
How could they develop a short circuit?
So the mystery remains. But at least
the fault had been found and the job
completed. Well, almost completed. I
had previously fiddled with the crystal trimmer (CV0l) and so I unlocked
the reference oscillator/burst system
again and set the crystal up properly.
In fact, it was a fair way out, though
still locking on a strong signal.
And that's pretty well the end of
the story. But I can only classify this
job as a real "stinker" and I think
readers will appreciate why I described it as a snack that turned into
three course meal; three faults - two
of them nasty ones - where I had
expected only one simple one.
However, I must add one more
thought. A lot of the trauma I experienced was due to the shortcomings of
the manual and circuits, as I explained
in the story itself.
But, in fairness to Samsung, I'm
happy to report that they have taken
these criticisms - and mine were not
the only ones - very much to heart.
The latest manuals are a far cry from
the ones I have criticised. They are
now well printed and contain all the
voltages, waveforms and other data
that we have come to expect from
other manufacturers. Full marks mate!
The Panasonic portable
PR.OM t,.. 'SN~C.\< -ro ~
38
SILICON CHIP
~~5 COU'R.~e:. N\E.A\.-~oe
My next story is about a Panasonic
35cm colour set, model TC-1401AR.
It, too, was still under warranty, being
virtually brand new, and having been
in the field for only three months.
The owner's complaint was, basically,
that there was insufficient brightness
but he went on to expand on this (and
I did listen carefully this time) saying
that, in his opinion, the brightness
control was not working at all.
And he added that the set was fitted with a remote control and that
other functions which it controlled,
such as contrast, colour and volume,
all seemed to work normally. Naturally, I couldn't hazard a guess as to
the nature of the problem and simply
suggested that he bring it in.
When he did, I turned it on while
he was there and, yes, it was exactly
as he said; no brightness control but
all other control functions normal.
More to the point, the effect was the
same whether the command was via
the remote control or the front panel
control. I also noticed that the bright-
R1142
181<
IC1102
CDLOUR31 ·
BRIGHT30
01110
P-+-+tH-+-¥N_..._-i
R1144
12k
IC601
VIDEO/CHROMA
J188
JUNGLE
LINK
P--o--o~-018 BRIGHT
J45
LIN
L__
---------~-:>--f.
5V
1
Fig.2: this simplified diagram shows the main components in the
brightness control circuit of the Panasonic TC-1401AR. Unfortunately,
Link J45 is not shown on the circuit and link J188 was hidden on the
board.
ness was virtually normal, although
perhaps just a trifle on the dark side.
Whep I finally dug out the manual
and opened the set, I went first to
IC1102 (MN15142TEAI),amicroprocessor which handles the functional
commands from both the front panel
and the remote control. In greater detail, the colour saturation control voltage comes out on pin 31 and the brightness control voltage on'pin 30.
On checking these voltages, pin 31
showed a varying voltage, from about
1.5V to 6V, depending on the setting
of the colour control. But the brightness control voltage sat at a steady 5V,
regardless of the brightness control
setting. Well, at least these measurements fitted in with the observed
symptoms.
But why was it happening? After
some preliminary checking around the
IC, without finding anything obvious,
I felt it was most likely an internal IC
fault. After all, it was only a single
function that had failed.
AUSTRALIAN MADE TV TEST EQUIPMENT
10% discount on all test equipment.
15% discount on two or more items.
SHORTED TURNS TESTER
,
Built-in meter to check EHT transformers including
split diode type, yokes and drive transformers.
$78.00 + $4.00 p&p
So I decided it was worth replacing
the IC. Granted, I had to order one and
there are 42 pins involved, but this
job has become pretty much routine
these days - one like this taking about
10 minutes. And, more often than not,
the IC is the culprit.
Unfortunately, it wasn't the culprit
this time. Oh well, it was worth a try
and a spare IC wouldn't go amiss. I'll
use it eventually.
My next approach was to track
down the source of that fixed 5V on
pin 30. Tracing the circuit (see Fig.2)
from pin 30, I came first to diode
Dl 110, which is shunted with an 18kQ
resistor, R1142. And from the pin 30
side of the diode is a 12kQ resistor,
Rl 144, to the 5V rail. So the 5V on pin
30 was legitimate enough, except that
it was from a relatively high impedance source.
Following diode Dl 110 was a 27kQ
resistor, Rl 143, after which the circuit ran through a maze of twists and
turns to pin 18, marked "Bright", of
IC601, which is described as a "video/
chroma jungle" IC. And, fairly obviously, it was the voltage on pin 18
which controlled the brightness. But
ANTRIM
TOROIDAL TRANSFORMERS
HIGH-VOLTAGE PROBE
Built-in meter reads positive or nega- .. .
tive voltages from 0-50kV. For checking EHT and other HT voltages .
. $98.00 + $5.00 p&p
DEGAUSSING WAND
Strong magnetic field. Double insulated for your
safety with momentary switch operation. 240V
AC 2.2 amps. As important as a soldering iron!
$75.00 + $10.00 p&p
REMOTE CONTROL TESTER
Designed to test infrared or ultrasonic ,
control units . Supplied with extension ·
infrared detector lead. Output is via a LED
and piezo speaker. $84.00 + $4.00 p&p
~~
~ _TUNER REPAIF!_~
--.:?"
~
'
From $19 exchange
plus post & pack
Cheque, Money Order, Visa, Bankcard or Mastercard
216 Canterbury Rd,
Revesby, NSW 2212,
Australia.
Phone (02) 774 1154
Fax (02) 774 1154
•
•
•
•
Now made in Australia
Standard range _15VA to 2kVA
Large ex-stock quantities available
Competitive pricing available for OEMs
& resellers
• Simply the best transformers all round
HARBUCH ELECTRONICS PTY LTD
90 George St, Hornsby, NSW 2077.
Phone (02) 476 5854 Fax (02) 476 3231
JUN E 1992
39
TETIA TV TIP
AWA video recorder AV11
Symptom: very poor vertical hold
on own recordings or EE picture but
OK on pre-recorded tapes. Any light
to dark transition in picture content
shows several white lines paralleling the outline.
Cure: C116, a 0-47µF 50V electrolytic capacitor, defective. This capacitor forms a bypass on part of
the tuner/IF AGC circuit and the fault
the voltage on pin 18 was, not surprisingly, also locked solid at 5V.
causes the IF stage to go into oscillation which destroys the vertical
sync pulse. The capacitor tests OK,
with no leakage and correct capacity at low frequency, but fails at the
much higher intermediate frequency.
TETIA TV Tip is supplied by the
Tasmanian branch of the Electronic
Technician's Institute of Australia.
Contact Jim Lawler, 16 Adina St,
Geilston Bay, 7015.
to secure another component to the
board.
The other way to open a circuit
The missing link
without cutting a track is to find a
At this point, it was clear that I had series component which can be lifted.
to separate these two pins before I In this case, there was a small induchad any hope of determining the tor, L1113, shown near pin 30 of
source of this voltage. Now, in theory, ICl 102 on the circuit and also close to
it should have been easy to break this this point on the board. It took only a
circuit. This board carries a number few seconds to lift this.
oflinks, designated with a "J" number,
This removed the 5V from pin 30 of
which make it easy to open a circuit ICl 102 but still left it on pin 18 of
for testing. In this case the link was
IC601. So the search area had been
narrowed to a handful of components:
designated J188.
diode Dl 110 and resistors R1142,
This link would have made things
Rl 143 and Rl 144. As already menvery simple, had I been able to find it.
But search as I might, I simply could tioned, Rl 144 goes directly to the 5V
not locate J188. I did find it eventu- rail and, in effect, feeds pin 30. I asally but only after I had solved the sume that, when functioning norproblem by other means. It was on the mally, pin 30 can pull this value down,
edge of the board and hidden under a as required, via R1144.
So, with Ll 113 disconnected, I shut
layer of Silastic which had been used
the set down and proceeded to check these
few components individually. And to help
isolate them, I lifted
R1142. I then checked
diode D1110 , which
proved to be perfectly
OK. The same applied
to Rl 142 and Rl 143.
This leftR1144. And
thi s was the breakthrough . Instead of
measuring 12kQ on the
meter, it registered a
dead short. And I do
mean a dead short; it
wasn't a few hundred
ohms, or even a few
ohms - it was zero.
~ i f ~. . . .
Now resistors don't
-nu~ OWNE:.~'s C.OM?LP..\NT wA.s,
go short circuit and so
l!,ft...StCAL-L-Y, iHi\T i\-\1;::."RE: WA..-S
I began a detailed ex\I\\SUFFIClENT :B'R.\GH,NE'SS
amination of the board
40
SILICON CHIP
and copper tracks involved. And this
emphasised something I hadn't taken
much notice of before. In series with
the 5V rail feeding Rl 144 was another
of the links that I had mentioned earlier, J45 . Strangely enough, this is not
shown on the circuit but is shown on
the component side of the layout diagram.
The copper track in which this is
fitted runs parallel with the track in
which Rl 144 is fitted. More to the
point, R1144 and J45 sit next to each
other, with their solder pads side by
side. In fact , one pair of pads is joined
to provide the necessary connection
between the 5V rail and R1144.
The two pads at the other end of
these components are separate - or at
least they should have been. But they
weren't. The solder pad at the pin 30
end ofR1144 was -yes you've guessed
it - bridged across to the J45 pad
alongside it. (The tracks are only 3mm
apart).
It wasn't simply a sliver of solder
but a genuine solid bridge which had
flowed between the two pads during
the wave soldering process. Little
wonder there was no brightness control; pin 30 of ICl 102 was connected
directly to the 5V rail.
Of course it was easily fixed; a quick
touch with the iron and the set was
better than new. And I make that point
quite deliberately; this fault must have
been present when the set was made
and had passed through all tests without being deh:icted. It would appear
that the sub-brightness control had
been set to produce a typical brightness level but no-one had checked to
see whether the brightness control was
working.
As it was, I now had to reset the
sub-brightness control (R312 on pin
18 of IC601) to establish an adequate
brightness range.
So how did it happen? It probably
has to do with the way manufacturers
check their sets as they come off the
assembly lines. As I understand it,
most factories now employ a sampling-type final check, with most sets
given only a cursory examination to
confirm that they are actually working. Only a small a percentage (or
sample) are actually given a thorough
check out. It's a far cry from the old
days in a radio factory but that's the
way it now is.
Anyway, I had another satisfied
customer.
SC
|