This is only a preview of the February 1996 issue of Silicon Chip. You can view 26 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 "Woofer Stopper Mk 2 - Now It's Even Better!":
Articles in this series:
Items relevant to "Computer Bits":
Articles in this series:
|
SERVICEMAN'S LOG
The dingiest corner of a dingy room
I have a story about another long-in-thetooth set this month – one with a nasty sting
in its tail. And the situation wasn’t helped by
having to work in an anything-but comfortable
environment.
The set in question was a Pye model
48SL1, a 48cm set, fitted with what
Pye called a T38 chassis. It was, in
fact, a Philips KT3A-1 chassis, one of
several Philips KT3A chassis, some
of which were live. The KT3A-1 had
an earthed chassis, however, and this
particular model would be about 13
years old.
It was owned by a lady and the complaint, as nearly as I could determine
from her description, was a virtually
54 Silicon Chip
complete failure which possibly involved a hiccuping condition. Well,
that was fair enough and I didn’t anticipate that it would be a particularly
difficult job.
But there was one snag – the lady
insisted that the job be done in her
home; she didn’t want the set to leave
the house. Don’t ask me why but she is
not the first person I have struck who
had a thing about not letting a set out
of their sight. And, as I was to discover,
the lady was rather eccentric in other
ways as well.
I try to avoid house calls if possible.
It is impossible to take everything
one is likely to need for the job and
it invariably transpires that the one
thing you do need is back at the shop.
But the lady was insistent and, since
she was willing to pay any additional
costs, I agreed.
The lady’s house turned out to be
what was once undoubtedly a Victorian-style luxury home but which had
seen better days. But what really struck
me, even before I pulled the bell knob,
was the modern security fittings. The
door was fitted with heavy security
bars, was obviously fitted with more
than one lock, and every window was
fitted with heavy security shutters.
All of which should not have worried me except that, when I moved
inside, I realised that the security
Fig.1: the vertical output
stage of the Pye 48SL1.
The blanking pulse is
derived from the junction
of resistors R531 and
R532 and is fed to line
A51. Note the waveform
at this point.
to tackle a fault like that anywhere
away from the shop, let alone in this
Victorian chamber of horrors.
Naturally, the lady protested at this
suggestion but I explained, as politely
as I could, that there was no alternative; I needed equipment and facilities
which I simply could not provide
in her lounge room. So, finally, she
agreed, albeit reluctantly.
Back at the ranch
shutters not only kept out the burglars
but kept out the light as well. It would
not have been so bad if the rooms were
reasonably well lit. However, I doubt
that any of the light fittings boasted a
globe larger than 40 watts.
Again, I cannot explain why. I can
only assume that it was an attempt
to recreate what she imagined would
have been the dingy atmosphere of
the house in its heyday. It was a weird
setup; the only thing that seemed to
be missing was a black cat named
Salem!
But speculation aside, the result was
that I found myself down behind the
set, in the dingiest corner of a dingy
room, hoping that I could manage to
see what I was doing.
In fact, when my eyes became dark
adjusted, and with the aid of a hand
lamp, I was able to find my way around
without too much difficulty. These
chassis are well laid out and this,
cou
pled with the fact that I am
reasonably familiar with them,
also helped.
The hiccups
And so to the problem itself. My
original assumption was correct;
the set was hiccuping madly, which
invariably means an overload on the
power supply due to a breakdown
of some kind. But the question was,
where?
My first checkpoint was the main
electrolytic capacitor, C298, off the
bridge rectifier. This can produce
symptoms like this if it dries out
and, with a set of this age, it
was a prime suspect. But no; it
checked OK and there was about
350V across it, which was normal.
I also made a routine check for dry
joints but, as far as I could see, there
was nothing obvious.
The next step was to isolate the
horizontal output stage and the quickest way to do that was to pull the
deflection yoke plug, which carries a
protective link. That cured the hiccups
and allowed the main HT rail to come
up to a steady 131V.
So, the fault was somewhere in the
output stage. I narrowed this a little,
after replacing the yoke plug, by short
ing the base and emitter of the output
stage transistor, Q562. This also cured
the hiccups.
I spent some time checking various
possibilities. I disconnected the tripler
and, in turn, the various auxiliary
voltage rails off the output transformer secondary. And I went over the
transistor stage itself, checking all the
components around it. I even checked
the output transformer for shorted
turns but to no avail. And, remember,
all this was done in the confined space
and poor lighting I have previously
described.
I sat for a few moments and had a
bit of a think. Somehow, my thoughts
came back to the transistor itself
(Q562). Perhaps it had a weird fault in
it. I decided to pull it out and check it
or, if necessary, replace it.
I didn’t get that far. As I removed the
transistor there was the fault staring
me in the face; a black spot on the
insulating washer, where the voltage
had punched through. Fancy being
caught with that old chestnut.
I fitted a new washer, the hiccups
vanished, and I had a picture on the
screen. But it was a hollow victory;
the top half of the picture was riddled
with horizontal retrace lines.
I baulked at that. No way was I going
When I got back to the ranch, I
hoisted the monster onto my workbench and set to work. Since it
was obviously a vertical blanking
problem, I went over the circuit to
familiarise myself with the blanking
circuitry. It is fairly straightforward
really. A deflection pulse is taken off
the vertical deflection output stage
(Q530 & Q532) and goes to a pulse
processing stage (Q535).
This stage is biased so that it conducts only during the vertical flyback
period and delivers a series of square
pulses of about 1V amplitude to the
blanking section (pin 9) of the chomi
nance/luminance IC (IC192).
At least, that is the theory of the
circuit. And as far as I could determine, this was what appeared to be
happening. There was an appropriate
waveform at the vertical output stage
and a replica of it, somewhat attenuated, at the base of the processing stage
(Q535). And there were pulses out of
Q535 being applied to pin 9.
So why wasn’t the system blanking?
The only clue I had – if it could be
called that – was the discovery that
the problem varied with the height
control setting; reducing the height
would eliminate the lines, as would
increasing it beyond a normal setting.
And that, if it suggested anything,
pointed to the vertical stage.
As a result, I made a whole swag of
checks around this stage, including
changing transistors, likely electro
lytics and any resistors which were
marginally high. It was all to no avail.
Next, I went back to the shaping
stage and, in spite of what the CRO had
told me, I changed transistor Q535. It
wouldn’t have been the first time that
such a trick had paid off, contrary to
all the tests. But not this time. Nor did
a detailed check of all the associated
components.
In a fit of desperation, I hooked
up the CRO again and made another
check of the waveforms around this
February 1996 55
not contain much detail. It simply indicated
a square pulse with
an amplitude of 0.9V
and this amplitude
appeared to be correct.
But the manual gave
no indication of the
pulse width and this
was what I was now
querying. It wasn’t an
easy point to check.
Apart from the lack
of detail in the manual, the CRO wasn’t too
happy trying to resolve
the pattern. Pin 9 takes
in both vertical and
horizontal pulses and,
while in theory one
can resolve either one,
Fig.2: the vertical blanking pulse from the
according to the selectvertical output stage in the Pye 48SL1 comes
ed timebase, this is not
in on line A51 (bottom, centre) and is fed to
always so in practice
the base of Q535 via R540. The base bias on
this transistor is set by R529 and R534. The
and there was some
processed blanking pulse at the collector is
difficulty locking the
fed via D467 to pin 9 of the chrominance +
image.
luminance IC (IC192).
Nevertheless, now
that my suspicion was
aroused, the CRO patstage. There didn’t seem to be any
tern seemed to confirm it. And from
doubt about the waveform into Q535 this observation came the thought that
but closer examination of the pulses
Q535 was not being turned fully on
coming out made me suspicious. I during this portion of the waveform.
couldn’t be sure they were exactly as
That, in turn, directed my attention to
they should be. The waveform given
resistors R529 and R534, both 6.8kΩ.
in the manual – waveform 32 – did These set the bias for this stage; 15V at
the base – from the 30V rail – against
13V at the emitter.
Suppose I reduced that 15V bias
on the base? Suiting the action to the
thought, I unsoldered one end of R534
and substituted the nearest appropriate value to hand, which happened
to be 10kΩ.
And, presto! – the lines vanished.
Problem solved? Well, fault cured,
which is not exactly the same thing.
Naturally, there was a temptation to
leave the circuit like that, since it obviously worked. But I’m never happy
with such situations. Was I simply
curing the fault by brute force without
actually finding it?
While trying to decide how to resolve this question, the answer was
almost literally served up to me on
a plate – well, on the shop counter
to be correct. It was another Philips
set, this time with a KT3A-2 chassis,
which is virtually identical. Its fault
was simple enough and I soon had
it up and running, which provided
an excellent opportunity to make
comparative voltage and waveform
measurements.
In fact, I didn’t need to go that far.
As soon as I moved to this part of the
set the answer was plain to see; R534
in this set was 8.2kΩ. And it was obviously the original component, which
I subsequently confirmed by reference
to the KT3A-2 circuit. I fitted an 8.2kΩ
in place of my 10kΩ and it worked just
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.
56 Silicon Chip
Signature________________________ Card expiry date_____/______
Name _______________________Phone No (_____)____________
Street
PLEASE PRINT
_________________________________________________
Suburb/town _____________________________ Postcode_________
Fig.3: the power
supply circuitry
for the NEC
N-3540. Note
the “HOT”
and “COLD”
designations
and the COLD
secondary of
T601. IC651
is at top left
and portion of
IC1001 at right.
as well. Problem solved.
Obviously, the fault I’d been chasing was not the first in this type of
set. There must have been previous
cases which had prompted this modification. Nor does it answer all the
questions. Why did this set suddenly
develop the fault when it had obviously performed satisfactorily for all
those years?
I can only assume that the original
design was a bit marginal, so that minor changes in components as the set
aged were enough to tip the balance.
Anyway, that was the end of the
story. All that remained was to return
the set to the dingy recesses of the
customer’s abode. I hope I don’t have
to go back, although it was a valuable
lesson learned.
Hot & cold NEC
My next story is from the much
more convenient and familiar atmosphere of my own workshop. It
concerns an NEC colour set, model
N-3450, the 34 indicating 34cm. It was
fitted with a typical infrared remote
control system.
According to the owner, the set
was completely dead but he didn’t
think there was much wrong with it,
because the stand-by LED was on. He
also indicated that he didn’t want to
spend a great deal on a repair.
Apart from the fault itself, the
interest in the set concerns a rather
unusual circuit arrangement. It is a
live chassis arrangement but with
considerably more of it being live
than in most cases. And the circuit is
clearly marked “HOT” and “COLD”, as
appropriate – see Fig.3. Well, at least
one is warned.
As is usual, the mains connects
straight to a bridge recti
f ier and
thence to a switchmode power supply, involving a trans
former T601
plus a switching transistor and error
amplifier in one package (IC601). The
primary winding of T601 and one of
two secondary windings is on the
HOT side, but the other secondary is
COLD. This provides a 20V rail via
diode D650.
Back on the HOT side, the output
from IC601 is the main HT rail at 115V.
This supplies the horizontal output
section, consisting of horizontal driver
transistor Q501, horizontal output
stage Q502, and the primary of the
output transformer, all still on the HOT
side. The input to Q501 is from the
main IC (IC701) via transformer T503,
which has a COLD primary and a HOT
secondary. All the output transformer
secondaries are COLD.
Having digested all that, I turned
my attention to the problem itself.
The fact that the stand-by LED was
on suggested that at least some part
of the power supply was working.
And, in fact, checks confirmed that the
previously mentioned 20V and 115V
rails were functioning.
On the other hand, there was no
horizontal waveform on any part of
the horizontal system. This made
me suspect that the fault could be in
the remote control system; either the
remote control receiver (PWC 3607C)
or the microprocessor (IC1001) which
it controls. In other words, the set was
simply not being switched on.
With the aid of the CRO, I established that the remote control receiver
was working and delivering a signal to
pin 14 of IC1001 (not the easiest path
to trace on the circuit). However, there
was no signal coming out on pin 33 of
February 1996 57
YOU CAN
AFFORD
AN INTERNATIONAL
SATELLITE TV
SYSTEM
SERVICEMAN’S LOG – CTD
SATELLITE ENTHUSIASTS
STARTER KIT
YOUR OWN INTERNATIONAL
SYSTEM FROM ONLY:
FREE RECEPTION FROM
Asiasat II, Gorizont, Palapa,
Panamsat, Intelsat
HERE'S WHAT YOU GET:
●
●
●
●
●
●
400 channel dual input receiver
preprogrammed for all viewable satellites
1.8m solid ground mount dish
20°K LNBF
25m coaxial cable
easy set up instructions
regular customer newsletters
BEWARE OF IMITATORS
Direct Importer: AV-COMM PTY. LTD.
PO BOX 225, Balgowlah NSW 2093
Tel: (02) 9949 7417 / 9948 2667
Fax: (02) 9949 7095
VISIT OUR INTERNET SITE http://www.avcomm.com.au
YES GARRY, please send me more
information on international band
satellite systems.
Name: __________________________________
Address: ________________________________
____________________P'code:
__________
Phone: (_______) ________________________
ACN 002 174 478
58 Silicon Chip
IC1001 to switch the set on.
The signal from pin 33 is applied to
transistor Q1072, then to Q1071 to turn
it on – see Fig.3. Q1071 functions as a
voltage regulator, generating a 12V rail
from the 20V rail. This 12V rail powers
IC701 which contains the horizontal
oscillator and this feeds horizontal
driver transistor Q501.
And that is how the set is turned on
and off – by switching this 12V rail.
With no 12V rail, there is no signal to
drive the output stage or, in fact, any
other function depending on IC701.
Voltage checks
OK, so why no signal on pin 33? It
could be a fault in IC1001 of course
but I wanted to check everything else
before I pulled that out. And the first
and obvious check was the voltage
supplying this IC. It is a 5V supply,
derived from a 3-terminal voltage
regulator (IC651) operating from the
20V rail.
Well, it was delivering voltage all
right – too much voltage; it was closer
to 8V than 5V. At the same time, I was
prompted to look more closely at the
20V rail. In fact, that 20V figure is a
nominal one. According to the circuit,
it can vary from 22.4V on stand-by to
18.7V when the set is running. This
is why I was deceived when I first
confirmed that this part of the set was
working.
With the set switched off, that rail
should have been at 22.4V, whereas
it was slightly less than 20V, a value which had appeared to be close
enough at first glance.
But it was the 8V at the regulator
output which was the real clue. I
pulled IC651 out and replaced it. And
that was it – there were now normal
input and output voltages and the
set was up and running. IC651 had
broken down and was acting more
like a resistor than a regulator, thereby
placing a heavier load on the 20V rail
and applying excessive voltage to the
microprocessor.
And it would appear that it was
that excessive voltage which upset
the microprocessor. The 5V rail feeds
several pins on this IC and it is not
surprising that the excessive voltage upset some of the internal logic
functions.
After all, it was not without good
reason that the rail was regulated in
SC
the first place.
|