This is only a preview of the July 1995 issue of Silicon Chip. You can view 31 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 "A Low-Power Electric Fence Controller":
Items relevant to "Run Two Trains On A Single Track":
Items relevant to "Satellite TV Receiver; Pt.3: Setting Up A Ground Station":
Articles in this series:
Items relevant to "Build A Reliable Door Minder":
Articles in this series:
Articles in this series:
|
SERVICEMAN'S LOG
Well, it looked like that at first
It is probably just as well that neither I nor
my colleagues are medical practitioners.
At least if we make a wrong diagnosis, we
usually get a second chance & the first error
can be corrected.
My first story this month is about a
video recorder – a Panasonic model
NV-SD10A which belongs to the local
primary school. This is a fairly recent
model, released only a couple of years
ago. And while it has nothing to do
with the story, it is worth mentioning
that it is fitted with the very much
upgraded “K” deck.
This deck is a considerable improvement on earlier decks which, in
most brands, employed some five or
six belts, and possibly three motors,
to perform all the necessary functions.
And as can be imagined, the belts were
a common cause of problems.
The “K” deck has eliminated most
of the belts – it uses only one if my
memory is correct – and has proven
to be a very reliable unit.
But that is by the way. This particular recorder had other problems.
According to the teacher who brought
it to the shop, it could not be turned
on. Well that seemed to be a straightforward enough description, even if
a little quaint. I would have simply
described it as completely dead.
And it was too. I plugged it in while
the teacher was there and there was
no sign of life of any kind; no clock
display or indicator lights and no response to any of the function buttons.
All of which was deceptively simple.
I imagined a fairly obvious power
supply fault, or even just a faulty fuse.
So the teacher left it with me adding,
as he left, that there was no particular
hurry for the machine; they had others
that could fill in.
Power supply checks
It was a day or so before I could
get at it, then I went straight to the
power supply, which is a switchmode
type. A preliminary check revealed
nothing obviously wrong. The fuse
was intact and mains voltage was
reaching the right terminals. Next, I
checked the various voltage rails out
of the supply (45V, 14V, 12V, 12.3V, 5V,
-29V, etc). And they were all present
and correct.
That put a completely different slant
on things. If the power supply was
delivering all the necessary voltages,
then the failure was in some other
section. Fair enough but how would
a failure in one section shut the whole
thing down, causing it to appear completely dead?
The most logical answer was that
the fault was somewhere in the
management section, possibly in the
microprocessor (IC7501) – see Fig.1.
This determines and initiates the correct sequence of events for any user
command. It also drives the clock and
other front-panel displays.
Of course, this was simply a broad
assumption. Exactly where or how this
failure was occurring was what I had
to track down.
What followed was a long and laborious voltage checking procedure.
Fig.1: a section of the microprocessor control circuitry in the National MV-SD10A VCR. IC7502 is at extreme right &
its output at pin 3 connects to pin 21 of the microprocessor (IC7501) at top left. The 4.7V shown at pin 3 of IC7502
somehow becomes 4.8V at pin 21 of IC7501.
68 Silicon Chip
As far as possible, I tried to confirm
that the various voltages at the power
supply were being applied to all the
points where they were supposed to
be. But, in particular, I concentrated
on the voltages applied to the microprocessor.
This eventually provided a clue.
Pin 21, which is labelled “Reset”,
is marked as 4.7V but, in fact, was
showing only about 2.4V. Back tracking from there brought me to IC7502,
a PST7026. This 3-legged device is
similar in appearance to a small signal
transistor but is rather more complicated, the circuit diagram showing that
it contains a couple of op amps and
various other circuit blocks.
More to the point is its function. I
wasn’t sure of this at the time. All that
the circuit indicated was that it is fed
from one of the 5V rails – shown as
4.8V at one point and 4.7 at another
- and delivers 4.7V to pin 21 of the
microprocessor. But obviously, there
had to be more to it than that. Its real
job is a delay function. At mains (repeat mains) switch-on, it pulls pin 21
low momentarily, before applying the
indicated 4.7V. This, apparently, is the
“reset” function.
Before I go any further, a word about
some of the voltages quoted. It is not
unusual to find quite silly voltage
figures on many circuit diagrams. Usually, the differences are only nominal
but I have known them to be quite
significant and misleading. It was this
kind of mistrust that confused matters
later on.
Nevertheless, I felt sure that 2.4V
in place of 4.7V was too great a difference. And, with very little else in
the circuit, the IC seemed the likely
culprit. It’s a very inexpensive device
costing less than $2, so the easiest
thing to do was change it.
Unfortunately, I had none in stock,
so I placed an order. And that was
the next hurdle. They were on back
order and it would probably be several weeks before delivery. Well, the
teacher had said there was no hurry,
but I contacted him and explained the
situation. He was quite understanding
and confirmed that there was no great
hurry at that time.
It was just as well because it was
over two months before new stocks
arrived. Unfortunately, this is not a
rare occurrence; it happens all too
often these days.
When my order arrived, I lost no
time in fitting it. When I switched the
machine on, it immediately burst into
life. The clock and all other indicators
came up correctly and a quick check
of all the various functions indicated
that they were working perfectly.
Problem solved?
So that solved that problem. Or did
it? Out of curiosity, I went back to the
output of the new IC, expecting 4.7V,
or something close to it. In fact, it was
only about 3.6V. And that was about
as awkward a figure as one could imagine. While ob
viously adequate to
allow the set to function, it was less
than shown on the circuit.
So was the circuit wrong and what
was the allowable tolerance on this
voltage, or the spread of the IC tolerance? I had no way of knowing, so
I simply let the machine run for the
next couple of days. During this time, I
made and replayed several test recordings and switched the machine off and
on at the mains at regular intervals. It
never missed a beat.
But I still wasn’t happy and was
trying to decide what best to do when
the teacher rand to say that their situation had changed somewhat. Some
programs that they wanted to record
clashed in time and they needed all
their machines. Could they have this
one back some time in the next week?
I explained the situation: that the
July 1995 69
SERVICEMAN’S LOG – CTD
but, worse than that, I couldn’t find
it; there was no sign of it anywhere
near its electrolytic mate, adjacent
to IC7502.
Nor could I see it anywhere else on
the board. In the end, I had to resort
to tracing the copper pattern on the
PC board. And that’s where I found
it; a surface mounted type behind the
microprocessor (IC7501).
I put the meter across it in-situ and
measured something over 1kΩ. That
wasn’t conclusive, of course, so I
pulled it out and switched the machine
on. It leapt into life and, more to the
point, I now had close to 4.7V on pin
21. Problem solved.
I’m not sure of the exact role of this
capacitor and the recorder seemed
quite happy without it. But, of course,
it had to be replaced. I didn’t have
a surface mounting 0.1µF capacitor,
so I substituted a small disc ceramic
type which I’m confident will do just
as good a job. Or perhaps even better.
No, this wasn’t one of my better
efforts but I’m telling the story for the
benefit of any reader who encounters
a similar problem in this model.
Life’s little mysteries
machine was working, apparently
quite reliably, but that I had some reservations about the job. If he wanted to
take it and try it in those circumstances
he was welcome. After all, it might just
as well have a trial run at the school
as on the bench.
The machine returns
And so the machine was duly collected. And, as I later learned, it performed faultlessly during their special
recording sessions and for a couple of
weeks thereafter. Then the teacher was
back on the phone again with the news
that it was exhibiting exactly the same
fault as before.
So I told him to bring it back in. He
was rather worried about the cost but
I assured him that I would stand by
the job, as the real fault had not been
found.
Naturally, I went straight to the
output of IC7502 and confirmed what
70 Silicon Chip
I feared. It was down to 2.5V again. So
what now?
I took another look at the circuit.
Assuming that the re
placement IC
was not at fault, there appeared to be
only two things that could be loading
the voltage on pin 21 and pulling it
down: (1) a fault in the microprocessor (IC7501); or (2) one of two small
capacitors shown connected to pin 3
of IC7502.
I tended to discount IC7501, if only
because it performed normally when
fed with the correct voltage. And that
left the two capacitors – C7511 (a
0.22µF 50V electrolytic) and C7510,
shown simply as 0.1µF.
Well, if it was going to be one of
these, it would be the electrolytic.
Wrong again; it took only a few moments to pull it out and fit a replacement. There was no difference.
So that left only the 0.1µF as the
last hope. I wasn’t very confident
My next story comes from a colleague and is another of life’s little
mysteries. The problem was eventually solved but with no really satisfactory explanation. This, more or less, is
how he told it to me.
The set was a National model TC2697 and the customer’s complaint
was that the picture rolled occasionally. What a horrible word that “occasionally” is – how does one tackle
an “occasional” fault? With difficulty
might be the best answer.
Anyway, all I could do was set it
up and let it run, hoping that when
it misbehaved I would see it and gain
some insight into the cause. So that’s
what I did but trying to keep one eye
on it while working on other jobs is a
near impossible task.
It was a couple of days before there
was any hint of trouble and then it
was only a glimpse out of the corner
of the eye. Did that picture roll? I
wasn’t sure. But patience paid off;
eventually it rolled when I was looking directly at it. It flicked one frame
then, a few seconds later, it flicked
two more frames.
After that, it seemed to settle. I
watched it for some time but it was
rock steady. This didn’t help very
much, except to confirm that the fault
was in the set, rather than due to local
interference. But I was no closer to
even guessing what was causing it.
In general terms, of course, I suspected a fault in the vertical deflection
system, or a sync fault. Following this
latter thought, I checked the hold control behaviour. Weak sync pules will
normally make the hold control setting
quite dodgy but not in this case; it was
locking up solidly. I even contrived
to run it on a very weak signal and
it still locked up positively. OK, rule
that one out.
The first real clue came by chance.
The workshop happened to be very
quiet – I wasn’t running the sound
–and I was studying a circuit on the
bench when I thought I heard a faint
splat. At the same time, I thought I
glimpsed the picture flick from the
corner of my eye.
I moved in closer and watched the
screen directly. Sure enough, the picture flicked again and there was the
splat at the same time. There wasn’t
much doubt about it now; I had a
problem somewhere in the horizontal
output stage and this was triggering
the vertical stage.
Until now, I hadn’t even taken
the back off the set, preferring not to
disturb anything until I had seen the
fault. Now that I had seen it and had
a clue, it was time to look inside. I set
up a mirror so that I could watch the
screen while looking in the back of the
set. It was a good setup but didn’t help
much. The picture flicked a couple of
times and I could hear the splat but I
couldn’t see anything.
Naturally, I concentrated around the
horizontal output transformer and the
ultor cap, but to no avail. And again,
I hesitated to disturb anything until I
had pinpointed the source. This approach continued for several episodes,
without any success.
The next step was born of desperation. I closed all the blinds on the
windows, turned out the lights and
checked the inside of the set again. I
didn’t have to wait long; the picture
rolled and I heard the splat.
I saw the light from the flash but not
the flash itself. I had been watching
the ultor cap but it wasn’t there. Anyway, after a few more rolls and splats
I finally spotted it and it was quite
tiny. It involved the horizontal output
transformer but in a very strange way.
The transformer is a fairly conven-
tional type, consisting of a rectangular
ferrite core made in two halves. The
windings are on one half and the two
halves are glued together and held
with a strong steel clip, which sits in
a groove in each piece of ferrite. And
this spring is floating; it is not at earth
or at any other potential.
But the layout is such that it is
within about a millimetre of an
aluminium bracket-cum-heatsink,
which is mounted on the chassis. This
bracket carries the horizontal output
transistor. And, by some mechanism,
a charge was building up on the clip,
eventually becoming strong enough to
jump the small gap to the bracket. The
set would then behave normally until
the charge built up again.
By what mechanism this was happening I don’t know. The only explanation that seemed reasonable at the time
was that any piece of metal in a strong
electric field can acquire a charge from
it. I understand that this effect can
be quite a problem for power supply
linesmen working in the proximity of
very high voltage power lines.
If this was the explanation, then it
seemed reasonable to cure the fault by
simply connecting the clip to chassis.
And no sooner said than done. I prised
up one end off the clip, slipped a short
length of copper braid under it, and
connected it to chassis.
And that fixed it – no more sparks,
no more splats and no more rolling. I
ran the set for a couple of days, positioned so that I could hear as well as
see the fault, and felt confident that it
would not occur again. The set was
then returned to the customer.
Complete failure
Which was all very fine – except
that, about three months later, the set
failed completely. The cause – failure
of the horizontal output transformer.
Yes, I know, you told me so. Or did
you? And, if so, I’d still like an explanation.
The obvious one would be that the
winding had broken down to the core.
OK, but ferrite is supposed to be a
non-conductor and, if the suggestion is
that it could provide sufficient leakage
at the high voltage involved, then this
is a new one on me.
Otherwise, I might have decided
to replace the transformer in the first
place. But could I have justified the
cost on the basis of the symptoms
described. In hindsight, yes, although
a new transformer isn’t cheap. The
new one cost over $100 but that was
all it cost the customer. I waived any
labour charges in the circumstances.
After all, fair’s fair.
Well, that’s my colleague’s story. I
must confess that it’s a new one on
me. Doubtless he’ll know better next
SC
time – and so will I.
July 1995 71
|