Are PC surge protectors needed in the UK?

"J.J." wrote in :



Does anyone have a reference to HDDs getting corrupted by power
events on the mains power supply?



References, no, practical, hands on experience, yes. I've seen this
several times, not only the HDD, but I've had transient power problems
take out motherboards too.
Examples:
Case 1: Home computer (this one) tree fell across a main line (11kva I
think), caused a surge prior to the stepdown transformer kicking out,
corrupted an almost new 40 gig hdd. Fortunately, an LLF fixed it.
Case 2: Engraving lasers at work, fed from the bus, kept killing HDD's
and motherboards. Ultimately traced to transient voltage spikes,
installed an AVR UPS. Failures were occuring once to twice a week, after
the AVR UPS was installed on each machine, we have had Zero failures, in
over a year.



--
Anthony

You can't 'idiot proof' anything....every time you try, they just make
better idiots.

Remove sp to reply via email

Hard drives can be corrupted for various reasons based also
upon what the filesystem is. For example, if using FATxx
filesystems, then a loss of electrical power at the right time
can even erase files from that drive. Just another reason why
the technically informed want NTFS filesystems on drives; not
FAT.

Transients should never be a problem to disk drives or
memory. Based upon how these devices are connected, then a
differential type transient would be required to cause
damage. But all minimally acceptable power supplies must have
the essential function called overvoltage protection - that
makes a differential transient not possible.

That is the theory as well proven by power supplies even 30
years ago. Reality is the gross profits obtained by dumping
inferior supplies in North America where so many computer
assemblers don't even have basic electrical knowledge. Many
clones are not built and sold missing the essential
overvoltage protection because the assembler only understands
one specification - dollars. It's called a bean counter
mentality. If the power supply is sold on the cheap, (ie full
retail price is less than $60), then this and other critical
functions are simply *forgotten*. Does not matter. Consumer
is only to be fleeced.

If the destructive differential transient does occur, there
is no overvoltage protection circuit to protect that hardware
- do to power supply purchased by a bean counter. No
problem. Myth purveyors then quickly blame speculated surges,
and recommend overpriced, typically undersized, and
ineffective plug-in protectors.

Up front - does the power supply specifically state that
overvoltage protection is provided? If not, then it probably
is a man-made disaster just waiting to destroy disk drive,
data, and other computer components.

This overvoltage protection is something completely
different from another disk drive threat to FAT filesystems -
blackouts and brownouts.

"J.J." wrote:
Does anyone have a reference to HDDs getting corrupted by
power events on the mains power supply?

"w_tom" wrote in message
...
Hard drives can be corrupted for various reasons based also
upon what the filesystem is. For example, if using FATxx
filesystems, then a loss of electrical power at the right time
can even erase files from that drive. Just another reason why
the technically informed want NTFS filesystems on drives; not
FAT.

Transients should never be a problem to disk drives or
memory. Based upon how these devices are connected, then a
differential type transient would be required to cause
damage. But all minimally acceptable power supplies must have
the essential function called overvoltage protection - that
makes a differential transient not possible.

That is the theory as well proven by power supplies even 30
years ago. Reality is the gross profits obtained by dumping
inferior supplies in North America where so many computer
assemblers don't even have basic electrical knowledge. Many
clones are not built and sold missing the essential
overvoltage protection because the assembler only understands
one specification - dollars. It's called a bean counter
mentality. If the power supply is sold on the cheap, (ie full
retail price is less than $60), then this and other critical
functions are simply *forgotten*. Does not matter. Consumer
is only to be fleeced.

If the destructive differential transient does occur, there
is no overvoltage protection circuit to protect that hardware
- do to power supply purchased by a bean counter. No
problem. Myth purveyors then quickly blame speculated surges,
and recommend overpriced, typically undersized, and
ineffective plug-in protectors.

Up front - does the power supply specifically state that
overvoltage protection is provided? If not, then it probably
is a man-made disaster just waiting to destroy disk drive,
data, and other computer components.

This overvoltage protection is something completely
different from another disk drive threat to FAT filesystems -
blackouts and brownouts.

A brownout is generally recognized as planned voltage cuts - or
undervoltage.
My utility in the US routinely implements 3% and 5% voltage cuts to shave or
reduce peaks (and thereby save $' when buying power).

There is also the issue of spikes and harmonics.

Given a current laptop/desktop will be ~ $800USD to ~$4000 for a high
performance system and a UPS sells for $40 - why subject a computer
to unexpected power events? What is the value of the data on the hdd(s)?


"J.J." wrote:
Does anyone have a reference to HDDs getting corrupted by
power events on the mains power supply?

In article ZsAHc.14750$oh.13125@lakeread05,
"someone" writes:

A brownout is generally recognized as planned voltage cuts - or
undervoltage.
My utility in the US routinely implements 3% and 5% voltage cuts to shave or
reduce peaks (and thereby save $' when buying power).

There is also the issue of spikes and harmonics.

Given a current laptop/desktop will be ~ $800USD to ~$4000 for a high
performance system and a UPS sells for $40 - why subject a computer
to unexpected power events? What is the value of the data on the hdd(s)?

This just isn't an issue in the UK, about which the original question
was asked. Maybe it's more of a problem in the US for some reason?

--
Andrew Gabriel

No, there's no hardware nor data worth saving in the UKG.

A 3 or 5% reduction in voltage, also known as a brownout to
the utility, is totally irrelevant to electronics and
especially irrelevant to computers. A computer works just
fine even when incandescent bulbs dim to less than 40%
intensity. Even demanded in Intel specifications. IOW what
the utility calls a voltage reduction is full power to the
computer. Utility would have to decrease voltage more than
20% for a computer to see a brownout. But if utility voltage
drops that low, then electric motors may be damaged. IOW
voltage too low to damage electric motors is even full power
to a computer - which demonstrates how resilient a computer
really is.

BTW, utility does not institute a voltage reduction to save
money. Voltage reductions are a last ditch effort to avoid
rolling blackouts.

Spikes and harmonics are (or should be) irrelevant to a
computer. Again, because the computer is so resilient.
However that internal computer protection assumes the building
has a 'whole house' protector so that spikes cannot overwhelm
computer internal protection.

All of which is irrelevant to HD protection. Either the
power supply will output correct power or it will shutdown.
This, of course, assumes the computer assembler had basic
electrical knowledge and did not install those 'defective by
design' $25 or $40 power supplies. But again, this was
explained earlier.

There is nothing cost effective adjacent to the computer. No
UPS nor power strip protector that will protect computer
hardware. Computer internal protection assumes the building
has implemented a 'whole house' protector on AC mains
connected less than 10 feet to central earth ground.
Protection as it was even done and well proven before WWII.

someone wrote:
A brownout is generally recognized as planned voltage cuts -
or undervoltage. My utility in the US routinely implements
3% and 5% voltage cuts to shave or reduce peaks (and
thereby save $' when buying power).

There is also the issue of spikes and harmonics.

Given a current laptop/desktop will be ~ $800USD to ~$4000
for a high performance system and a UPS sells for $40 - why
subject a computer to unexpected power events? What is the
value of the data on the hdd(s)?

w_tom wrote:
BTW, utility does not institute a voltage reduction to save
money. Voltage reductions are a last ditch effort to avoid
rolling blackouts.

Or to avoid the expense of bringing additional generating capacity
online, thereby saving money...

All of which is irrelevant to HD protection. Either the
power supply will output correct power or it will shutdown.

A PSU shutting down is not irrelevant to HD protection. Shutdown at the
wrong moment (especially with the wrong operating system) and you end up
with a badly trashed filesystem.

I have operated computers much further from 120V with absolutely no problems.

Once at the end of 100ft extension cord along with 1500W heater, lights
dimming.

The other time at 130V, light bulbs popping every month.

Even on a generator nearly out of gas I had a minute to shutdown my system, no
corruption occured.

"someone" wrote in message
news:ZsAHc.14750$oh.13125@lakeread05...

A brownout is generally recognized as planned voltage cuts - or
undervoltage.
My utility in the US routinely implements 3% and 5% voltage cuts to shave or
reduce peaks (and thereby save $' when buying power).

There is also the issue of spikes and harmonics.

Given a current laptop/desktop will be ~ $800USD to ~$4000 for a high
performance system and a UPS sells for $40 - why subject a computer
to unexpected power events? What is the value of the data on the hdd(s)?

"w_tom" wrote in message
...
A 3 or 5% reduction in voltage, also known as a brownout to
the utility, is totally irrelevant to electronics and
especially irrelevant to computers. A computer works just
fine even when incandescent bulbs dim to less than 40%
intensity. Even demanded in Intel specifications. IOW what
the utility calls a voltage reduction is full power to the
computer. Utility would have to decrease voltage more than
20% for a computer to see a brownout. But if utility voltage
drops that low, then electric motors may be damaged. IOW
voltage too low to damage electric motors is even full power
to a computer - which demonstrates how resilient a computer
really is.

BTW, utility does not institute a voltage reduction to save
money. Voltage reductions are a last ditch effort to avoid
rolling blackouts.

True - my opinion as well - perhaps you could convince my local utility re
this procedure being used on a daily basis.

Spikes and harmonics are (or should be) irrelevant to a
computer. Again, because the computer is so resilient.
However that internal computer protection assumes the building
has a 'whole house' protector so that spikes cannot overwhelm
computer internal protection.

All of which is irrelevant to HD protection. Either the
power supply will output correct power or it will shutdown.
This, of course, assumes the computer assembler had basic
electrical knowledge and did not install those 'defective by
design' $25 or $40 power supplies. But again, this was
explained earlier.

There is nothing cost effective adjacent to the computer. No
UPS nor power strip protector that will protect computer
hardware. Computer internal protection assumes the building
has implemented a 'whole house' protector on AC mains
connected less than 10 feet to central earth ground.
Protection as it was even done and well proven before WWII.

Perhaps you could explain the half dozen UPS I have seen that operated
correctly and interrupted close in electrical faults.
Naturally the UPS were scrap after the electrical event - but the protected
electronics were ok.
$40 UPS vs $800 desktop or in one situation $200 - $300 UPS vs. $5000 of
servers.

someone wrote:
A brownout is generally recognized as planned voltage cuts -
or undervoltage. My utility in the US routinely implements
3% and 5% voltage cuts to shave or reduce peaks (and
thereby save $' when buying power).

There is also the issue of spikes and harmonics.

Given a current laptop/desktop will be ~ $800USD to ~$4000
for a high performance system and a UPS sells for $40 - why
subject a computer to unexpected power events? What is the
value of the data on the hdd(s)?

Was UPS between AC mains and computer? No. UPS and
computer both connect to AC mains just like light bulbs. In
fact it would be same protection if both computer and UPS
shared same wall receptacle. Any transient from the
receptacle confronts UPS and computer equally. However
protection inside a UPS is often so grossly undersized that a
surge too small to damage a computer might still damage the
UPS. Furthermore, some computers can even act as surge
protectors - shunt a destructive surge so that it does not
seek earth ground via other computers.

Until you define specific circuits - including how every
wall receptacle is wired, then I cannot provide more
information.

I cannot say exactly why that particular event happened.
But above is one reason why a UPS may be damaged and computer
is not. Computer power supplies have internal protection.
Protection so sufficient that there is little adjacent to a
power supply that can enhance protection. But computer
internal protection can be overwhelmed if destructive
transients are not earthed before entering the building.

Bottom line is this. You had UPS failure. Therefore you
have no effective surge protection. Even surge protectors
must not be damaged due to a surge.

To provide a better answer, do as I do - autopsy the dead
body. Replace the defective part to learn what has actually
been damaged. Autopsy only complete when the failed unit is
fully functional.

If a server farm has no 'whole house' protection and a single
point earth ground, then no UPS or plug-in protector is going
to do anything better. In fact, it is just not a reliable
operation if 1) every incoming utility line does not enter at
the common service entrance all connected to the single point
earth ground and 2) building does not have necessary 'whole
house' protector on incoming AC mains. From Sun Microsystems
planning guide:
http://www.sun.com/servers/white-pap...ning-guide.pdf
Lightning surges cannot be stopped, but they can be diverted.
The plans for the data center should be thoroughly reviewed
to identify any paths for surge entry into the data center.
Surge arrestors can be designed into the system to help
mitigate the potential for lightning damage within the data
center. These should divert the power of the surge by
providing a path to ground for the surge energy. Protection
should be placed on both the primary and secondary side of
the service transformer. It is also necessary to protect
against surges through the communications lines. The
specific design of the lightning protection system for the
data center will be dependent on the design of the building
and utilities and existing protection measures.

If you are suffering transient damage, then the human is
reason for failure. What Sun writes is so well proven and
understood that it was standard even before WWII. Protection
is only as effective as its earth ground.

As for your brownouts - if any voltage is too low for a
computer, then the utility has grossly violated national
standards. A PUC call would create a massive response - if
your AC voltage drops so low as to be problematic to a
computer.

someone wrote:
"w_tom" wrote in message
...
A 3 or 5% reduction in voltage, also known as a brownout to
the utility, is totally irrelevant to electronics and
especially irrelevant to computers. A computer works just
fine even when incandescent bulbs dim to less than 40%
intensity. Even demanded in Intel specifications. IOW what
the utility calls a voltage reduction is full power to the
computer. Utility would have to decrease voltage more than
20% for a computer to see a brownout. But if utility voltage
drops that low, then electric motors may be damaged. IOW
voltage too low to damage electric motors is even full power
to a computer - which demonstrates how resilient a computer
really is.

BTW, utility does not institute a voltage reduction to save
money. Voltage reductions are a last ditch effort to avoid
rolling blackouts.

True - my opinion as well - perhaps you could convince my
local utility re this procedure being used on a daily basis.

Spikes and harmonics are (or should be) irrelevant to a
computer. Again, because the computer is so resilient.
However that internal computer protection assumes the building
has a 'whole house' protector so that spikes cannot overwhelm
computer internal protection.

All of which is irrelevant to HD protection. Either the
power supply will output correct power or it will shutdown.
This, of course, assumes the computer assembler had basic
electrical knowledge and did not install those 'defective by
design' $25 or $40 power supplies. But again, this was
explained earlier.

There is nothing cost effective adjacent to the computer. No
UPS nor power strip protector that will protect computer
hardware. Computer internal protection assumes the building
has implemented a 'whole house' protector on AC mains
connected less than 10 feet to central earth ground.
Protection as it was even done and well proven before WWII.

Perhaps you could explain the half dozen UPS I have seen
that operated correctly and interrupted close in electrical
faults. Naturally the UPS were scrap after the electrical
event - but the protected electronics were ok. $40 UPS vs
$800 desktop or in one situation $200 - $300 UPS vs. $5000
of servers.