This is only a preview of the May 1989 issue of Silicon Chip. You can view 38 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:
Articles in this series:
Articles in this series:
Articles in this series:
Articles in this series:
|
THE WAY I SEE IT
By NEVILLE WILLIAMS
Are surges & spikes on the mains
a hazard to personal computers?
Don't be surprised if the following article raises
more questions than it answers. Unfortunately, that
seems to be the way it is when one tries to
quantify the problem of surges and spikes which
occur on the supply mains and the hazards they
present to personal computers.
First let me emphasise that I
boast no special expertise in the
subject. I am aware, though, of certain commonsense precautions that
can help safeguard personal computers in a home or office situation.
They are set out in a separate panel
for your consideration and I suggest
that you read through them at this
point.
My day-to-day involvement with
personal computers has mainly to
do with word processing and allied
tasks, where such precautions can
be observed without undue hassle.
They may be less practical in offices, in educational situations or in
other computing applications.
I set up my first word processor
about five years ago and considering the modest cost, it worked very
well. I subsequently replaced it
with a more ambitious system using
an Apple IIc. I'm still using that
system, along with its normal
monochrome monitor, Appleworks
and other software, external disc
drive, printer, modem, mouse and a
conventional RF line filter.
When first installed, a small
serial/parallel converter was also
included, allowing the Apple to
drive the same BMC BX-80 printer
that had formed part of the earlier
system.
44
SILICON CHIP
In the new situation, the BMC
printer failed on three separate occasions, with no obvious explanation but always with the same
symptoms: when self-tested, it
would print out its full character
set in the normal way but would not
accept drive from the computer.
When it failed a fourth time, so did
my patience and I replaced both it
and the associated converter with a
new Epson LX-800 printer, retrofitted to be Apple-compatible.
Since then, as per the old saying,
it's been "working like a bought
one" on both text and graphics!
There's a point to this tale: I
never did find out the real reason
for the repeated breakdown of the
original BMC printer. Predictably,
the suppliers of the printer and the
serial/parallel converter maintained that such failures were uncharacteristic of their respective
products. But they did seem to
agree on a couple of points:
• Such things do happen on occasions (what a brilliantly perceptive
remark)!
• Surges and spikes on the power
mains could well have had
something to do with the problem.
(Logical explanation or a cop-out?
Take your pick!)
That's about as far as their in-
terest seemed to extend. I gained
the distinct impression that when
something faiied (in this case a
printer), somebody, somewhere
replaced suspect ICs or suspect
boards until the fault disappeared.
If it failed several times, you fixed it
several times. The how, when,
where or why of the breakdowns
received scant consideration.
A major failure
I would probably have written
the whole thing off as "one of those
things" had it not been for a subsequent system failure of a quite different kind.
During that (for some) delightfully relaxed period after the
Christmas festivities, I fired up the
Apple II with the intention of
transcribing my hand-written notes
about Peter Wright's "Spycatcher"
(for the March issue).
Everything was behaving normally when, pausing to rethink a particular observation in my notes, I
heard an uncharacteristic "grunt"
from the printer, much as happens
when it is first switched on. But it
was already on and the system was
just sitting there waiting for me to
make up my mind.
When I did try to use the
keyboard, I realised that the computer had locked up. Unfortunately,
switching off and re-booting did not
fix the problem. A couple of foreign
semi-colons on what should have
been blank areas of the screen indicated that something was amiss
and while the word processor still
worked in a fashion, any attempt to
feed the text to either disc or
Surges On the Phone Lines
Reference: "Characteristics
and Rate of Occurrence of
Lightning-Caused Induced
Voltage Surges in a Telephone
Line" by D. Mackerras. Journal of
Electrical and Electronics
Engineering, Australia, Volume 5,
No.3, September 1985.
The line which provided the
basis for Mackerras' observations over a period of about four
years comprised about 3km of
overhead cable and 20km of
underground cable, including an
exchange. A sketch map accompanying the article shows the
position of nearby 50Hz high
voltage power mains.
It is not possible to adequately
summarise the contents but the
paper indicates that the pulse
energy can be attributed mainly
to the induced effect of ground
printer simply triggered another
lock-up.
At that point it dawned on me
that the computer could hardly
have picked a worse time to break
down. That became all the more
evident when I discovered that just
about every Apple PC repairman
that I could think of was on holidays
until after the new year break!
Fortunately, I was able to borrow
another Apple lie from a relative
for a few days and that allowed me
to carry on. But it also gave me time
to reflect on what had happened.
While I had to concede that the
failure might have been purely
spontaneous ["one of those things")
the circumstances did seem to suggest a high voltage transient on the
supply line.
This was subsequently supported
by the observations of an Apple
dealer service manager who, after
an initial examination, said that the
computer appeared to have suffered more than a simple failure.
With a bit of luck, they might be ·
able to repair it by replacing ICs, in
which case I could expect a bill for
around $100. Failing that, it could
be a case for board replacement
which would run to more like $250.
Asked whether he'd seen similar
(not cloud) lightning flashes and
to superimposed 50Hz energy
caused apparently by fault current or unbalanced voltage initiated by the same lightning
occurrence.
A typical recording, relative to a
flash within about 1 km of the
aerial section of the phone line
and close to the 11 0kV cable,
shows a prominent 9 cycles of
mains frequency at about 200V
peak-to-peak present on the
phone line . Other such events involved from 4 to 48 cycles at
amplitudes ranging from 1 5V to
320V.
More than half of all observed
lightning events involved multiple
surges, the mean average
number being 2. 8 with a mean
average separation of 1 38
microseconds.
damage before, the service manager said it was uncommon but certainly not unique. He still had a couple of other boards on the shelf
from computers which appeared to
have suffered a similar trauma. His
intention was to work through them
in detail "one of these days" and
bring them back into service.
In fact, he managed to salvage
my original board and when I got
the computer back, it seemed to be
working perfectly. That is, until I
attempted to feed the text down the
line to one of the SILICON CHIP computers. Then, what appeared to
leave my screen as perfect ASCII
copy showed up on theirs as badly
corrupted.
So my computer had to go back
for further attention, with a suggestion that there was still a fault in
the modem output port that had
been overlooked. Apparently, that
proved to be the case, leaving me
with an all-up repair bill of about
$170.
It seemed likely that the modem
port had been damaged at the same
time as the remainder of the system
but there was another possibility.
Thinking back, I realised that
while I had been dutifully unplugging the system from the mains, I had
As measured in the Queensland University Laboratory, at the
end of the underground section
of the cable , a maximum voltage
exceeding 320V occurred at a
rate of 10 events per year.
Events exceeding 650V occurred once per year, with the
possibility (extrapolated) of
1 400V once per 1 0 years.
In · the . Brisbane area, about
84% of all lightning events can
be expected in the December/
January period, with 79% of
those concentrated in the late
afternoon. from 2pm to 6pm .
The author lists 9 other
references on the same general
subject but warns that allowance
has to be made for measurement
criteria, line configuration and the
country or area in which the tests
are conducted .
been less diligent in disconnecting
it from the phone line. So during
some of the violent thunderstorms
that marked December on the east
coast, the system may well have
been exposed to lightning induced
pulses on the phone line.
I might not have been aware of
such damage until I actually tried to
use the modem. Either that, or a
voltage transient may have affected
the computer in a more subtle way,
rendering it liable to spontaneous
failure or vulnerable to, say, a
surge or spike on the power mains.
Alarming picture
If you have a mind to do so, it
isn't difficult to put together a deeply disturbing picture of what can
happen to your PC from such
eventualities.
Lightning was discussed in relation to NASA activities on page 30
of the November 1988 issue and
again, in more detail, in the
February 1989 issue [see also the
accompanying panel).
But what about spurious pulses
on the mains attributable to other
causes?
With this in mind, I phoned an acquaintance who is usually a mine of
information on such matters. He
MAY1989
45
Surges On the Power Mains
In a 1 984 brochure covering
their TRANSIL transient voltage
suppressors, Thomson-CSF
summarise the problem as
follows, presumably based on
observations in France. No
reference is made to reticulation
mishaps.
"Externally caused voltage
surges are of industrial or atmospheric origin and are
transmitted through the supply
cables by inductive or capacitive
coupling.
"On 220V mains, industrially
caused voltage surges can have
an amplitude of 5 to 1 0kV, with
durations of between 0. 1 and
recalled that a university student,
gaining work experience in the
Philips laboratory in Artarmon
(NSW), had once built up a device
to record the occurrence of spikes
on the 240V m::iins.
Using mechanical counters (it
was back in the 1960s) the circuit
was set up to sense the amplitude of
peaks in roughly 100V steps above
the normal peak of 350 volts to 1000
volts. Apparently, the device worked well but it was amazing to see
how many times the 1000 volt
counter registered.
Next on my list was a schoolteacher friend who has his own Apple Ilc and is involved with several
more of them at the school. No, he
had never experienced any similar
breakdown problems but said I:ie, "I
must tell you what happened the
other day around this neck of the
woods" - out Parramatta way.
It seems that a couple of bright
lads, poking around a nearby
building site, had come upon some
loosely rolled bare galvanised fencing wire. Sensing the opportunity
for a bit of "fun", they had carriPd
it upstairs in a block of units and
had landed it neatly, quoit-like,
across a nearby llkV power line.
What followed was an almighty
series of flashes and explosions,
which somehow defied the protective cutouts that were supposed to
operate in such an emergency.
When my ever-inquisitive friend ar46
SILICON CHIP
1 00 microseconds . Their occurrence is variable. Statistics show
that, in an area where there is a
lot of noise on the mains, 1 kV
surges occur between 1 00 and
1 000 times a year and 5kV
voltage surges occur on an
average of once a year.
"In the case of voltage surges
of atmospheric origin, lightning
can produce amplitudes of 30kV
for an average duration of 70
microseconds. In the case of
EMP (electromagnetic pulse)
voltage surges, the electrical
field can reach 1 00kV/m for
durations of 1 0 to 50 nanoseconds".
rived on the scene a few minutes
later, the charred llkV cable ends
were draped across the 415/240V
consumer wiring beneath. Said he:
"I wouldn't like my computer to
have copped that lot!"
He would have relished, even
less, the lot of residents out Manly
way, where a truck crashed into a
power pole and dropped a 33kV
feeder across the 415/240V mains. I
understand that the accident took
out quite a few microwave ovens
and TV sets in the area but it would
also have been a poor lookout for
any computers that happened to be
powered up at the time.
Distribution problems
When I raised the general subject with a technical officer attached to a major electricity distribution centre, incidents like those
above held no surprise for him.
They were par for the course.
As he said, "Our circuit breakers
are normally very fast and very effective, able to isolate a high
voltage line within a couple of
cycles. Having isolated a line", he
added, "they can test the circuit
again after about 60 seconds, to see
whether the problem had disappeared - as it might well have
done if a momentary short has been
caused by, say, a falling branch".
There had been a much larger
headache, he said, when they'd lost
a major underground cable during
the Christmas break. It failed suddenly, without warning, possibly
due to unreported damage and
subsequent moisture penetration. It
really upset things.
"Are incidents in the high voltage
network reflected in the 240V
mains?''
"Most definitely", he said.
"Variations, surges and spikes are
simply stepped down by the pole
transformers''.
He went on to say that a blackout
involving a major shopping centre
posed a particular problem, with
thousands of independent fluorescent lights and other appliances sitting across the mains. No-one ever
bothers to turn them off when the
supply fails. An enormous peak current flows at the instant supply is
restored, followed by complex current and voltage variations until
everything settles down.
In many cases, consumers in
areas surrounding large shopping
and industrial complexes share the
same high-voltage feeders - as a
practical and economic necessity.
What happens in the shopping or
industrial complex is reflected in
nearby consumer mains.
By the very nature of things, he
said, some consumers face greater
risks than others of mains supply interruptions and traumas.
Too many loose ends
As I said, it isn't hard to assemble a disturbing picture by lumping
together lightning strikes, reticula-.
tion accidents, load problems and
switching transients but in the process, I became aware of certain
"loose ends". Or if you prefer:
questions lacking ready answers.
For example, in what way are
computers supposed to be especially vulnerable to spikes and surges?
By simple failure of the DC power
supply, by the application of excessive voltage to other components, by anomalous injection into the data circuits, or all three?
A detailed documented analysis
would be much more helpful - and
convincing - than the usual bald
statement that computers are
vulnerable and that's that!
Again, there 's endless talk and
numerous references to spikes and
surges on the power mains. But try
Personal Computers - Typical Precautions
• Don't expose PCs to an environment which can generate
high electrostatic charges; eg,
synthetic carpets, plastic furnishings, etc, especially in a
naturally dry or air conditioned
atmosphere.
• Don't couple or uncouple
units or fiddle with plugs and
cables while the system is
powered up.
• Feed the system through a
line filter. It will hopefully liinit
to obtain a quantitative assessment
of the potential rate, amplitude or
duration of spikes and surges and
one is met with mainly anecdotal
answers.
A detailed analysis of all such
possible events would shed further
light on the whole subject. (In the
meantime, see the panels accompanying this article).
The w ay I see it
Doubts notwithstanding, it seems
to me that the precautions as listed
make good basic sense and are well
worth observing:
• Computers should not be exposed needlessly to large electrostatic
potentials, or physical changes
made to the set-up while it is switched on.
• RF filters , preferably including
a va ristor limiter, a re affordable
and should attenuate most highamplitude spikes from the microsecond to the low millisecond range.
If you want to build your own, ref er
most of the spurious pulses on
the power mains.
• If possible, avoid leaving the
system switched on during a
thunderstorm. Better still,
operate the entire system from a
multi-outlet cord or filter, which
can be readily removed from the
power point, should the need
arise.
• If connected to the telephone
system, unplug when not required but especially during
thunderstorm activity.
to the Mains Muzzler project in the
January 1989 issue of SILICON CHIP.
• If the system is not going to be
used for some hours, switch it off
until it is needed again. Reduce the
number of hours it sits across the
mains and you reduce in direct proportion its exposure to a reticulation mishap that the filter might not
be able to cope with.
• Don't just disconnect the modem
when you hear thunderclaps; leave
it unplugged until it is actually
needed.
What if the equipment needs to
operate continuously throughout
the whole working day? And during
the evenings, or the night as well? I
guess that one has to accept the added risk.
Either that or you can install a
more comprehensive - and expensive - line filter or conditioner, or
even an uninterruptible power supply. Few personal computer applications would justify that sort of
expense, however.
~
Lightning Monitor System
According to a report in a recent edition of the TV series
"Beyond 2000" , a system has
been set up to automatically
monitor the progress of
thunderstorm activity across
France.
Groups of receptors , positioned mainly on airfields, pinpoint
cloud and ground flashes , plotting them in real time as dots and
crosses respectively on a video
display, over which is superimposed a map of France.
The intensity, track and present location of thunderstorms
can be seen at a glance and
precautionary measures taken.
No less to the point, the system
provides a means by which computers and other lightning-prone
systems across France can be
isolated automatically as a storm
cell moves across their area.
Problems?
. .. and you
don .,t have our
J20page
catalogue . ..
At last . ..
a TRADE
catalogue for
the consumer
ARISTA .. . your one-stop problem
solver. Video plugs and sockets . . .
Video extension speakers .. . Video
flyleads . . . Video RF interference
filters ... Video splitters . . . Indoor
antennas . . . Video switching units
. .. Down converters .. . Video
speaker controllers .. . Video
camera lights ... Video tape
rewinders .. . Video cine adaptors
. .. Video head cleaners . . . Video
splicing kits .. . Video storage
cases ... Video dust covers ...
Video leads . .. Scart plug leads
... Video dubbing kits .. . Video
headphones .. . Video shotgun and
wireless microphone systems ...
Pre-amplifiers with video inputs . . :
Video camera stands ...
Just about anything
you want.
. . . Try us ... NOW!
Get your catalogua FREE
from your local ARISTA
dealer or send $2.50 P&H
and your return address to:
ARISTI\.
ELECTRONICS PTY LTD
PO BOX 191 , LIDCOMBE, NSW 2141
MAY.1989
47
|