This is only a preview of the October 1999 issue of Silicon Chip. You can view 37 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 "Backup Battery For Cordless Phones":
Items relevant to "Build The Railpower; Pt.1":
Items relevant to "Semiconductor Curve Tracer":
Items relevant to "Autonomouse The Robot; Pt.2":
Items relevant to "An XYZ Table With Stepper Motor Control; Pt.6":
Purchase a printed copy of this issue for $10.00. |
SERVICEMAN'S LOG
Did lightning strike at all?
The old saying that “lightning doesn’t strike
twice in the same place” has a new twist
these days. Did it strike at all? Many
insurance claims for lightning strikes on TV
sets are best described as dubious; and they
can put the serviceman in the hot seat.
Mr Lancia (not his real name) had
already sought service elsewhere
for his Toshiba 2529SH, 21 System,
63cm stereo TV receiver. It had been
purchased overseas but I only found
this out later. When he brought it in,
he said that it had died as the result of
a lightning strike and could I prepare
an estimate for insurance?
Estimates are tricky things and are
best avoided in other than exceptional
circumstances. The only way one can
“estimate” is to actually fix the set. But
by doing that, the estimate concept
disappears; it now becomes a factual
situation. Material costs have been incurred and hours of labour have been
spent. The estimate becomes the final
account. And if the customer doesn’t
like it, what then?
The best that I could hope for in
this current situation would be that a
lightning strike might produce enough
visual evidence to support a rough
guess as to the minimum cost. But
there is no way of knowing whether
the obvious damage is covering a host
of additional “invisible” problems.
In any case, if an insurance claim is
involved, it’s necessary to first determine the validity of the lightning
strike theory before going any further.
This set was probably made for the
Middle East market and certainly not
for Australia. I tried to order a service
manual from the Toshiba agent, only
to be told that they had never even
heard of this model, far less have any
parts for it. They were pretty emphatic
that there would be no support available in Australia for this set. I was on
my own.
I removed the covers and made a
visual check. Significantly, I was unable to see any signs of lightning damage. Encouraged by this, I switched
on; there was no sound or picture but
it was far from dead. First, I could
switch the power indicator LED from
“Standby” to “ON” from the remote
Fig.1: this circuit section from a Toshiba 3418DA helped solve the problem
with a Toshiba 2529SH, which isn’t sold locally. It shows the TA8783N
jungle IC and, in particular, the Vcc pins (6, 40, 61 and 63).
20 Silicon Chip
control, as well as from the front panel controls. Also, there was 124V on
the collector of the horizontal output
transistor (Q404) and the horizontal
drive transistor (Q4022).
But there was no voltage anywhere
on the 64-pin jungle IC (IC501, TA
8783N). I wasn’t familiar with this
device but it appeared that pin 39 was
the horizontal drive output and that
the crystal on pin 37 was part of an
oscillator. I couldn’t work out much
more without a circuit and I especially needed to find out which was the
supply rail pin (Vcc).
Next, I went to the power supply
and measured the rails on the output
from the chopper, Most had voltages
but it was far too complex to work out
what was what.
I did notice that a relay on the primary was switched off by a link across
the base emitter of its driver transistor.
When I removed this link the relay
began to chatter loudly. Once again,
without a circuit diagram, it was too
hard to find out how it was meant to
work, so I refitted the link.
This was about as far as I could go
and the set was put aside. I told the
client the situation and he liaised with
his insurance company, presumably
with a view to writing the set off.
Lateral thinking
Months later, I was fixing another
Toshiba for a different problem and I
needed technical assistance from the
local Toshiba agent. They weren’t able
to help directly, suggesting instead
that I try one of their larger agents in
Melbourne. I telephoned them and
they were able to help with my immediate problem.
Buoyed by this, I thought I’d push
my luck and asked if they had heard
of a 2529SH model. They hadn’t but,
as an afterthought, I asked them if they
had any Toshiba circuitry using the
TA8783N IC. The technician offered
to a look and after a minute or two he
came back and said he had a circuit
for a 3418DA which used this IC and
that I could buy a copy if I wished.
I agreed to go ahead and when the
circuit arrived I found that two Vcc
rails are fed to this IC: 9V on pin 40
and 12V on pins 6, 63 & 61. I took a
punt on pin 40 and traced this back
to the power supply before the trail
went cold. The power supply was
completely different from the 3418DA
circuit and the only chance I had now
was to trace out the circuit. This was a
fairly ambitious approach; it was very
complicated with about 20 transistors
and it was going to take some time.
I had already established that the
12V rail was derived from the horizontal output transformer via a 3-pin IC
regulator. This regulator is designated
as a UPC2412HF but I am unfamiliar
with it. However, I was still following
the idea that the 9V supply to pin 40
was critical.
Two issues now had to be resolved
before I spent more time on this theory: (1) was the set worth fixing?; and
(2) was I going to get paid for it?
To answer the first point, I hooked
up an external 9V power supply and
switched it and the TV set on together.
This produced both sound and picture and everything looked fine. This
confirmed that the fault was confined
to the 9V source – apart from the previously mentioned relay.
I revisited this part of the circuit
briefly and found that when the set
was on, the relay was silent – it was
only in the standby mode that it
chattered. I also worked out that its
function was to short out two large,
low value resistors which limited the
240VAC supply to the main bridge
rectifier.
Finding out how all this worked
and fixing the problem was going to be
time consuming. I phoned Mr Lancia,
and he said he was quite prepared to
pay for all my efforts if I could do it
for less than $300. He already had a
new set from the insurance company
and he could use this set as a second
set if I could fix it.
So in the next month or so, whenever I had some spare time, I would
draw the circuit of the power supply
module. The result was four voltage
sources: 124V, 8V, 16V & 17V, all of
which were present and correct. The
17V rail fed an NPN transistor (Q870)
which generated 9V at its emitter,
the 9V supply then going to pin 40 of
IC501. However, there was no 9V on
the emitter of Q870.
The transistor driving Q870 was
another NPN transistor, designated
Q871. Its base was controlled by a network of transistors and other assorted
devices. When I shorted Q871’s base
and emitter leads, its the collector
voltage rose and turned on Q870, restoring both picture and sound. I was
on the right track at last.
Safety circuit
But it was still complicated; the
network of transistors controlling
the base of Q871 was also part of the
remote control on/off power circuit,
including the safety cutoff. I already
knew that the remote system was
working properly, which left the safety
circuit to investigate.
From what I could work out, it
involved an SCR (D859), which was
controlled by the horizontal output
stage via a series of small unidentified
glass diodes from the 8V rail to the
16V rail. I tried disconnecting various
parts of this circuit, one at a time, to
see whether the set would fire up.
Items Covered This Month
• Toshiba 2529SH TV Set
• JVC 775AU TV Set
• JVC CX-60ME TV Set
• Hitachi C33-P900 TV set
• Mac LC630 computer
And it wasn’t until I disconnected
these diodes, which I realised were
unidentifiable zeners (D890, D891,
D892 & D893) that the circuit finally
fired up.
Because I couldn’t identify the diodes, I decided to reconnect them and
measure the voltage across each one.
Believe it or not, when I did this, the
set continued working and nothing I
could do would recreate the fault. I
froze them and heated them but they
wouldn’t fail. I assume that, in the
process of unsoldering them, the fault
had somehow been fixed.
Regretfully, the circuit involving the
relay was in the too hard basket. I tried
to draw the circuit but it used over 10
transistors, plus optocouplers, in all
sorts of bizarre configurations. I can
only guess it was some sort of current
sensor safety circuit but I couldn’t
work it out, so I left it as I found it.
I left the set on soak test for couple
of weeks before calling Mr Lancia. I
quizzed him on this and other parts
of the fault which I really couldn’t attribute to lightning. Finally, it emerged
that he had the same faults before and
that someone else had done the same
sort of thing as I had, which kept it
going until the recent storm. So, was
the storm really involved?
I don’t know. Nor did he explain
why he hadn’t returned the set to the
previous repairers, nor did he say who
they were.
For my money he was really a rather
October 1999 21
dodgy sort of bloke – but at least I was
paid for all the work I did.
Two elderly JVCs
Two elderly JVC TV sets came into
the workshop this week – one a 51cm
model and the other a 15cm battery/
mains portable. Both were over 10
years old and belonged to different
customers, and both sets were dead.
The 51cm unit was a 7755AU, while
the baby was a CX-60ME with an
attachable mains power supply.
Somewhat surprisingly, their circuit
diagrams were similar, with both employing an AN5900 pulse width modulator IC for the low voltage supplies.
I started with the baby, which I
found would still work with a 12.5V
bench power supply. I opened up the
AC adaptor (AA-60ME) and what was
left of resistors R09 and R10 told the
story. Fortunately, having the circuit
for both sets was a bonus. The two
resistors were in the emitter circuit
of the power output transistor Q03
(2SD1453), which turned out to be
short circuit.
I replaced these and applied power.
There were no fireworks; in fact there
was nothing – the unit was dead. Transistor Q03 takes its collector voltage
from the bridge rectifier, D01, via the
22 Silicon Chip
primary of transformer T03. And the
collector was sitting at 340VDC; the
unloaded voltage from the bridge and
the mains.
Nothing was oscillating and it is a
little obscure as to exactly how the
circuit worked. However, I figured
out that the problem was in the
starter circuit Q01, Q02 & Q05. The
ohmmeter indicated that Q01 had a
base-emitter short but the others were
OK. This circuit supplies the start-up
voltage to start IC01 oscillating until
the secondary 12V takes over.
I fixed this and the whole power supply started working but was
producing a fierce whistling sound.
This implied that it wasn’t oscillating properly, probably due to a leaky
electrolytic capacitor.
There were about 10 of these and I
started to hang additional capacitors
across the most likely ones until I
reached C22. Replacing this, a 47µF
16V unit, finally stabilised the circuit
and it behaved properly, giving a
12.5V output. I had already soldered
any potential dry joints and I left it
on soak test.
Moving on to the larger set, there
were no low tension voltages at all
from the power supply and a relay was
not activating. I checked the voltage
at the emitter of Q03, which is fed
directly from the bridge rectifier and
there was no 11V as indicated on the
circuit. Nor was there any voltage on
Q03’s collector, test point TP98, out
of bridge rectifier D11, or even from
the power transformer T02. It wasn’t
until I removed T02 that I found that
its primary winding was open circuit.
This didn’t look good but I persevered.
I fitted an external bench power
supply across the bridge rectifier and
ran it up to 62.5V, as indicated on the
circuit. This produced low voltage
rails of -28V, +12V and + 5V, the latter derived via IC001 (LA7930). But
the 110V main HT rail was low and
causing trouble in another part of the
power supply. This turned out to be
due to horizontal output transistor
Q551 (2SD1453), which was short
circuit.
I thought that replacing this would
be the end of the story but worse was
to come. While I now had sound, there
was no picture and there were blue
sparks inside the picture tube. All
the indications were that the picture
tube was down to air and this was
subsequently confirmed.
It was a good job I hadn’t ordered
a new T02 transformer, because the
sick picture tube was the death knell
for this set. There was no way one
could justify the cost of a new tube
(assuming one could be found) and
the labour costs involved. It was a
write-off.
That’ll teach me to muck about with
sets over 10 years old!
The big Hitachi
Mrs Belrose asked me to service her
78cm Hitachi C33-P900 (G8P chassis)
TV set at home because the prospect
of bringing it back to the workshop
was horrendous.
She complained of obscure intermittent faults such as intermittent
loss of picture, intermittent flickering
and a monochrome picture for three
minutes or so when cold. These faults
hardly ever occurred when the set
was warm, although occasionally it
would flicker for few seconds when
switching from the AV (audio/visual)
sockets to the tuner.
I removed the back of the set and
looked around. I was looking for the
AGC control and the associated circuits when I happened to see one of
several small black electrolytic capaci
tors on the video board. It had literally
spat the dummy – or, more correctly,
the electrolyte – all over the PC board.
Then I saw another, and another, and
realised all 30 or so electrolytics were
leaking badly.
Most, if not all, would to have to
be replaced. It was a big job, and one
that couldn’t be done in the customer’s home.
Though the set looked a lot younger, I learned from the service manual
that it was 10 years old. But apart
from these problems, the picture on
the 78cm picture tube was still good
which meant that the set was worth
fixing. After all, a replacement would
cost $1500 or more.
I wasn’t looking forward to taking
the set back to the shop. However, I
managed to round up three helpers
and it wasn’t too bad for the four of
us, considering the set’s dimensions
and weight (80 x 72 x 57cm and 64kg).
I started by replacing all the electro
lytics, mostly 10µF, on the video signal
board, hoping that would fix most of
the problems. It didn’t. I then examined the small electrolytics on all the
other boards. Two of these, C918 (1µF
50V) and C950 (470µF 60V), were
particularly bad but I also changed
a few other low-value units which
might be causing hum on the power
supply rails.
Altogether, I replaced 40 electrolytic capacitors, cleaned up all the corrosive electrolyte from them and felt
confident that I had cracked it – but I
hadn’t! The thought of replacing every
last one of the remaining capacitors
was too much to contemplate.
It was time to get technical. The
main clue I had was that it was temperature sensitive so, using the hairdryer
and freezer, I followed the video rails
from the tuner as they branched out
all over the set. Being a multi-system
set, it was fairly complex to follow.
During this procedure, I followed
a few false trails, especially at pin 8
of IC501, which I noticed earlier had
hum on it when connected to the
oscilloscope.
I had also used a signal source
plugged into the AV input sockets and
noticed that the stronger the signal,
the less chance there was of the symptoms occurring. For example, the set
was much better in the AV mode and
SBS (the strongest signal in this area)
was better than all the other channels.
Anyway, after messing around for
a long time around IC501, I found I
wasn’t really progressing and continued until I reached IC701 (TDA
2579A), which is the jungle IC. There
I noticed that freezing it caused the
horizontal frequency to change (I
could hear it) and then the picture
vanished. I gradually isolated the area
down to the components adjacent to
pin 6, especially C619 (2.2µF 50V).
Replacing this capacitor removed all
the remaining symptoms.
The easy bit was over – all I had to
do now was manhandle the set back
to Mrs Belrose!
Secondhand Mac
Our local school was given some
secondhand Macintosh com
puters
but one, an LC630, had died, so it
was brought in to me to see what I
could do.
I hadn’t worked on these before
and the first problem was figuring out
how to get inside the unit. It was like
a Chinese puzzle. This meshed with
that and that slid into this, to unlock
something else, and so on. However,
after a lot of time, I eventually managed to locate and remove the power
supply.
Eventually, I managed to unscrew
the final metal screening can and remove it. But, in so doing, my hands
touched a part of the PC board and
even though the set had been switched
off for hours, I received an awful
shock. I instinctively let go if it and the
wretched device tumbled to the floor.
After I had recovered with a very
stiff cup of coffee, I carefully picked
up the module and examined it
closely. Fortu
nately, it hadn’t been
damaged by the fall and the incident
had given me a major clue.
It was obvious that I had received
the shock from the fully charged filter
capacitor following the mains bridge
rectifier which meant that no current
was being drawn from it. This in
turn suggested that the power supply
wasn’t oscillating. So the first step was
to find the start-up circuit.
That wasn’t too difficult. I soon
found two 220kΩ resistors in series
(R16 & R17), one of which had gone
high. Replacing them (after shorting
out the remaining charge on the main
electroly
tic) completely fixed the
power supply. The only remaining
problem I had was trying to remember
how everything went back together
again.
I charged for two hours labour and
the bill came to $114. I was worried
the school might think that this was
too expensive, especially as one can
buy complete PC power supplies for
less than $60. However, the school
was delighted as they had been quoted
SC
over $200 for a replacement!
October 1999 23
|