Power surge safety

  • Thread starter Thread starter Happy
  • Start date Start date
Computers are like TVs. They never turn off (mine is rare as it has a REAL on/off switch as well). This is mainly so you can turn it on from external devices.

You unplug computers if you have third would power. A switch only activates the active wire. There are two more wires un switched.

And we are talking about power surges. A surge is a high energy / short duration pulse. It's the consequence of

Lightening hitting a long way away
A factory turning big electrical machines off or on. You need to live near the factory. There are special light bulbs for people in this situation.

But in cities the council protects you (because they have to protect their own equipment).

A surge is too short to damage most things. But ICs have a low voltage tolerance as they are very small circuits and high voltage just burns through.
 
Again, if AC mains power cycling is destructive, then it is
destructive whether computer is powered on or powered off.
Again, that cycling power is still powering circuits inside
the computer. Power cycling is not destructive. Anything
that power cycling might do to be destructive is even
eliminated by functions in the power supply controller. My
perspective is different. I ask what component can and cannot
be damaged. I know how the subsystems work. AC power
cycling is only destructive in myths.

Something sometimes preceding a power cycle or brownout can
be destructive. This destructive transient might damage the
computer - powered on or powered off. BIOS setting would do
nothing to avoid this damage.

Effective solutions provided in the previous post dated 22
Feb 2005.

The BIOS change does not protect hardware. The plug-in UPS
(once we read through the facts and add a few industry
standards they forget to mention) only protects data from
blackouts and brownouts. We still don't build new homes as if
the transistor exists. Protection that already exists inside
appliances (including that computer) assumes you have earthed
an incoming transient at the service entrance. IOW install
the highly recommended 'whole house' protector - that costs
about $1 per protected appliance. The plug-in UPS does not
even claim to provide that hardware protection - no matter
what subjective, color glossy, sales brochures imply without
numbers.
 
I think they are refering too an improper shutdown. Comp shuts down on power loss, power comes on comp tries to boot but power goes off and another improper shutdown, et al.

An improper shutdown runs the risk of corrupting the disk, not often but it does happen. It always happens on a Fat32 disk (though the damage is very minor on the "every" part), it will usually happen (again very minor) on NTFS (but not usually on servers as they don't cache writes) but the transaction roll back will usually cope. But every now and then there is massive corruption. I do lots of improper shutdowns on purpose and I expect to lose data on a disk every two years. But I do sweat a bit waiting for the reboot each time.

As windows loads only a read only disk driver at boot time it would require a few seconds of power on to make a difference, one has to wait till disk writing starts.
 
Power loss should never corrupt disk hardware. A disk drive
sees no difference between an unexpected power loss or a
normal shutdown. When voltage drops normally or unexpectedly,
then the disk drive 'computer' shuts down - without damage.
This was even how 1960s disk drives (with those large voice
coils that could take out a finger) worked. The disk drive
does not receive a special message when power is about to be
removed. The disk drive only knows to shutdown when DC
voltage drops. Loss of power is a normal shutdown to disk
drives today as it was 35 years ago.

Some obsolete disk file systems (ie FAT) could be corrupted
by a power loss. For example, if writing an update to a file
when power is lost, then an FAT file system could erase a new
file (obviously) and also delete the older existing file. So
yes, if not using a stable disk file system such as NTFS, then
a UPS is necessary - to protect data. Again, the UPS is not
providing hardware protection.

Yes, I once encountered a computer so poorly designed that
a power loss almost caused loss of human life. The designer
did not even have a degree - but he was inexpensive. I was
appalled that this late 1970s design was even permitted.
Why? Power loss and power cycling even 30 years ago must not
harm hardware. Defacto standards are that old. It applies to
PC today. Power loss and power cycling, even 30+ years ago,
must not harm hardware.

"Power cycling does not harm hardware" runs contrary to
myths. And so I then ask these myth purveyors to ID the
component damaged by power cycling - and why. They don't
know. They don't know what those internal components do.
They just *know* power cycling must be destructive. That is
the difference between an engineer and a myth purveyor. An
engineer says so from both experience AND from knowing how the
internals work. The myth purveyor just knows power cycling
must be destructive. The myth purveyor assumes using the same
logic that said WMDs also existed.

In the meantime, do you know about the design in a power
supply controller that also locks out excessive power
cycling? Just another fact often unknown to those myth
purveyors. Just another reason why power cycling is not
hardware destructive.

Again, a plug-in UPS is for data protection from two types
of power problems - blackouts and extreme brownouts. Those
other power problems are solved (made irrelevant) elsewhere.
 
On Fri, 11 Mar 2005 00:10:08 -0500, w_tom wrote:
[please excuse snippage]
The BIOS change does not protect hardware. The plug-in UPS
(once we read through the facts and add a few industry
standards they forget to mention) only protects data from
blackouts and brownouts
Click to expand...

Agreed including proper grounding being essential.
Again, if AC mains power cycling is destructive, then it is
destructive whether computer is powered on or powered off.
Again, that cycling power is still powering circuits inside
the computer. Power cycling is not destructive. Anything
that power cycling might do to be destructive is even
eliminated by functions in the power supply controller. My
perspective is different. I ask what component can and cannot
be damaged.
Click to expand...

In the incident I related (power on/off repeatedly), I lost my external
modem and my self-powered hub. Taking into consideration what you've said,
both of those items had separate plugs. I will readjust my thoughts on this
as with reflection, being attached to the computer probably had nothing to
do with their loss. Instead it was a failure of their own power supplies to
withstand the fluctuations.

I've had horses for many years. Have seen many strange incidents from
lightning at the boarding barns. I've tried to learn a healthy respect and
caution for lightning events.

Again, thanks for your knowledge and information on this topic. I've
enjoyed reading it and learning from it.
 
Power loss should never corrupt disk hardware. A disk drive
sees no difference between an unexpected power loss or a
normal shutdown.
Click to expand...

With the exception of drive heads possibly landing/crashing under these
conditions, I would expect this to be correct from a hardware perspective.

Move the scenario from power surge to every day usage and improper
shutdowns. Corruption of the system can happen. David summed up my
concerns over this.
 
On disk drives even 30 years ago, heads had to retract after
power was removed - the normal shutdown procedure. Heads were
controlled by a voice coil - almost the size of a football.
IOW a disk drive maintains sufficient power internally so that
when a power switch was cut off, only do head start
retracting. No difference between a normal power off and an
unexpected power off.

Hard drives today require so little power to retract those
heads. IOW when power is removed, the disk drive maintains
more than sufficient power to retract those head long before
the disk platter eventually stops spinning. Unexpected power
loss is also called normal power off - same thing - from the
perspective of disk drive heads.

Yes, even 30+ years ago when a retracting head could
seriously harm fingers, the normal shutdown was same as an
unexpected power loss. We simply clicked the power switch
off. Only then did the disk drive start retracting heads.
Today's disk drives do same so much easier because the heads
require so much less electrical power.

Can corruption happen? Yes. Unexpected power loss can
cause loss of unsaved data. With poor filesystems such as
FATxx, saved data on disk can even be deleted. But unexpected
power loss does not damage hardware. The UPS is only for data
protection; not for hardware protection. Even disk drives
will properly relocate heads after power has been removed - as
was standard even 30+ years ago.
 
Generally power cycling can occur coincidentally with other
events such as a destructive transient. Same transient that
may cause modem damage might also cause repeated power cycling
as the utility tries to restore power to a 'now damaged'
electric network. IOW power cycling would not do the damage.
That coincidental and previously unknown transient may have
caused the hardware damage. Transient that would not be
destructive if a 'whole house' protection system was installed
- especially the single point earth ground.

Why a single point ground? Same reason why horses can be
electrocuted when a nearby tree is struck. Lightning seeks
earth some 4 kilometers west. So it strikes a tree east of
the horse. Shortest path westward is up that horses hind legs
and down horse's fore legs. Horse is killed because it became
a good path for lightning to flow west.

Same applies to a house protection system. If using multiple
grounds, then a transient can rise up on the left side ground,
find destructive paths through appliances, then reenter earth
via the right side earth ground.

How to make a single point ground beneath the horse?
Surround the barn with a buried halo ground wire. Then earth
beneath the horse becomes one big single point ground. Horse
can be killed by the same concept that also kills a computer.
Therefore better buildings also install halo grounds or Ufer
grounds. Earth ground - not the UPS - being so important for
computer hardware (and horse) protection.

Sharon said:
On Fri, 11 Mar 2005 00:10:08 -0500, w_tom wrote:
[please excuse snippage]
The BIOS change does not protect hardware. The plug-in UPS
(once we read through the facts and add a few industry
standards they forget to mention) only protects data from
blackouts and brownouts
Click to expand...

Agreed including proper grounding being essential.
Again, if AC mains power cycling is destructive, then it is
destructive whether computer is powered on or powered off.
Again, that cycling power is still powering circuits inside
the computer. Power cycling is not destructive. Anything
that power cycling might do to be destructive is even
eliminated by functions in the power supply controller. My
perspective is different. I ask what component can and cannot
be damaged.
Click to expand...

In the incident I related (power on/off repeatedly), I lost my external
modem and my self-powered hub. Taking into consideration what you've said,
both of those items had separate plugs. I will readjust my thoughts on this
as with reflection, being attached to the computer probably had nothing to
do with their loss. Instead it was a failure of their own power supplies to
withstand the fluctuations.

I've had horses for many years. Have seen many strange incidents from
lightning at the boarding barns. I've tried to learn a healthy respect and
caution for lightning events.

Again, thanks for your knowledge and information on this topic. I've
enjoyed reading it and learning from it.
Click to expand...
 
Back
Top