Life expectancy

What's probably the life expectancy of my A7V333 motherboard if I take good
care of it? It has 2 years on it right now. I run the computer for the
most part constantly except when I leave town or do something with the
inside of the computer. Current MB temperature is at 30 C.

In article URjEd.32449$3m6.5163@attbi_s51, "Travis King"
wrote:

What's probably the life expectancy of my A7V333 motherboard if I take good
care of it? It has 2 years on it right now. I run the computer for the
most part constantly except when I leave town or do something with the
inside of the computer. Current MB temperature is at 30 C.

If the case temp is not excessive, the electrolytic caps should
be good for 10 years. Solder joints under stress, could be
anyone's guess (more likely to happen with a P4 retail heatsink
and its high clamping forces).

In ten years, you can expect several PSU failures, and any one of
those PSU failures could damage the motherboard.

If you have a lot of lightning storms, or bad quality power,
that could influence how long the mobo lasts. Look carefully
at any modem, cable modem, ADSL wires etc, to see if there are
any protection devices to take a (nearby) lightning hit, before
it gets to the motherboard. For example, on a phone line, there
may be a carbon block at the entry point, and you could enhance
that by using a second protection device nearer the computer.
For the really paranoid, a wireless network would reduce the
wiring exposure to just the power lines. A real ($1K purchase
price) UPS would reduce the risk of an AC power event from
getting you, and would help protect the PSU from getting
damaged. Cheap UPSes offer no protection at all, as they are
actually SPS (standby power supplies) - they are a "straight wire"
to power spikes, and the unit only cuts over to batteries if
the AC power dies for enough milliseconds.

On the motherboard itself, the Vcore circuit is the circuit under
the most stress. If the MOSFETs are cool to the touch, that is
a good sign. I've never read any MTBF estimates for switching
regulators on motherboards, so don't know whether they are
good for a 1 million hour MTBF or not.

Large BGA packages also have a rating, for solder joint
reliability. For example, a BGA with 750 pins, will last for
about 10 years, with a certain daily temperature variation.
From the Via web page:

* 552-pin BGA VT8366A North Bridge
* 376-pin BGA VT8233 South Bridge

so you have little risk of a failure there (caps will fail
first).

Handling the processor a lot (removal, regrease, reposition
heatsink) will cut into the life expectancy, if say the
processor gets cracked, and it happens to overload the Vcore
circuit. If the processor has the rubber bumpers on the top
of the chip, that will cut that risk a bit.

I would say your biggest exposure, is to external factors.

Paul

Paul wrote:

In article URjEd.32449$3m6.5163@attbi_s51, "Travis King"
wrote:


What's probably the life expectancy of my A7V333 motherboard if I take good
care of it? It has 2 years on it right now. I run the computer for the
most part constantly except when I leave town or do something with the
inside of the computer. Current MB temperature is at 30 C.


If the case temp is not excessive, the electrolytic caps should
be good for 10 years. Solder joints under stress, could be
anyone's guess (more likely to happen with a P4 retail heatsink
and its high clamping forces).

In ten years, you can expect several PSU failures, and any one of
those PSU failures could damage the motherboard.

If you have a lot of lightning storms, or bad quality power,
that could influence how long the mobo lasts. Look carefully
at any modem, cable modem, ADSL wires etc, to see if there are
any protection devices to take a (nearby) lightning hit, before
it gets to the motherboard. For example, on a phone line, there
may be a carbon block at the entry point, and you could enhance
that by using a second protection device nearer the computer.
For the really paranoid, a wireless network would reduce the
wiring exposure to just the power lines. A real ($1K purchase
price) UPS would reduce the risk of an AC power event from
getting you, and would help protect the PSU from getting
damaged. Cheap UPSes offer no protection at all, as they are
actually SPS (standby power supplies) - they are a "straight wire"
to power spikes, and the unit only cuts over to batteries if
the AC power dies for enough milliseconds.

On the motherboard itself, the Vcore circuit is the circuit under
the most stress. If the MOSFETs are cool to the touch, that is
a good sign. I've never read any MTBF estimates for switching
regulators on motherboards, so don't know whether they are
good for a 1 million hour MTBF or not.

Large BGA packages also have a rating, for solder joint
reliability. For example, a BGA with 750 pins, will last for
about 10 years, with a certain daily temperature variation.
From the Via web page:

* 552-pin BGA VT8366A North Bridge
* 376-pin BGA VT8233 South Bridge

so you have little risk of a failure there (caps will fail
first).

Handling the processor a lot (removal, regrease, reposition
heatsink) will cut into the life expectancy, if say the
processor gets cracked, and it happens to overload the Vcore
circuit. If the processor has the rubber bumpers on the top
of the chip, that will cut that risk a bit.

I would say your biggest exposure, is to external factors.

Paul

I would agree, and add that IME motherboards are far more tolerant of
external factors than one has any right to expect ;-)

Our Cottage PC runs an Asus P2B-S, manufactured in 1998 and in service
for almost 6 years. We use it on weekends in winter, but the cottage is
not heated while we are away, so the system experiences repeated thermal
stress cycles - it's common for the inside temperature to be -20C or
lower when we arrive. I discourage the kids from powering up the PC
until the place has warmed up, with limited success :-)

Power is unreliable at the Cottage, and we experience frequent
thunderstorms, however despite a lack of protective measures the only PC
failure which has occurred there to date was a sound card that stopped
working after lightning struck a tree behind the building - no doubt due
to a spike induced in the 40' cable running from the sound card to a
stereo system on the other side of the room. Hardly surprising, but
damage was limited to the sound card.

I expect the Cottage P2B-S to be the first of my numerous P2B series
boards to eventually fail, but perhaps not since I use several in my lab
and subject them to frequent CPU swaps and other hardware changes. My
primary system runs a P2B-DS and has been in service 7x24 since October
1997 except for occasional shutdowns for hardware upgrades or fan
service. The power supply refused to restart after a shutdown in 2002
and was replaced, but no other failures have occurred. I have a total of
11 P2B series boards in regular use, with zero motherboard failures to date.

I'd better start saving my pennies since replacing all my systems when
the electrolytic caps fail in 2008 will be expensive ;-)

P2B

YOu want to worry about something? Worry about your data. Your MB is
obsolete, Your OS is obselescent and going obsolete and the only thing
that's of any value is your data. All the stuff about UPS's is theory. Are
you running NASA and a lunar lander mission? If you are you should be shot
for having only a single system. Given enough time EVERY piece of equipment
will fail including $1000 UPS's.
I power my system on and off all the time. I don't shut down when lightening
storms are overhead and I never worry about the telephone lines or cable.
But I live where the power wiring is fairly new ~15 years and its above
ground. If lightening is going to get you its going to get you. I have
heard claims that buried power and phone cables are susceptable to problems
but I've never seen proof.

"Paul" wrote in message
...
In article URjEd.32449$3m6.5163@attbi_s51, "Travis King"
wrote:

What's probably the life expectancy of my A7V333 motherboard if I take
good
care of it? It has 2 years on it right now. I run the computer for the
most part constantly except when I leave town or do something with the
inside of the computer. Current MB temperature is at 30 C.

If the case temp is not excessive, the electrolytic caps should
be good for 10 years. Solder joints under stress, could be
anyone's guess (more likely to happen with a P4 retail heatsink
and its high clamping forces).

In ten years, you can expect several PSU failures, and any one of
those PSU failures could damage the motherboard.

If you have a lot of lightning storms, or bad quality power,
that could influence how long the mobo lasts. Look carefully
at any modem, cable modem, ADSL wires etc, to see if there are
any protection devices to take a (nearby) lightning hit, before
it gets to the motherboard. For example, on a phone line, there
may be a carbon block at the entry point, and you could enhance
that by using a second protection device nearer the computer.
For the really paranoid, a wireless network would reduce the
wiring exposure to just the power lines. A real ($1K purchase
price) UPS would reduce the risk of an AC power event from
getting you, and would help protect the PSU from getting
damaged. Cheap UPSes offer no protection at all, as they are
actually SPS (standby power supplies) - they are a "straight wire"
to power spikes, and the unit only cuts over to batteries if
the AC power dies for enough milliseconds.

On the motherboard itself, the Vcore circuit is the circuit under
the most stress. If the MOSFETs are cool to the touch, that is
a good sign. I've never read any MTBF estimates for switching
regulators on motherboards, so don't know whether they are
good for a 1 million hour MTBF or not.

Large BGA packages also have a rating, for solder joint
reliability. For example, a BGA with 750 pins, will last for
about 10 years, with a certain daily temperature variation.
From the Via web page:

* 552-pin BGA VT8366A North Bridge
* 376-pin BGA VT8233 South Bridge

so you have little risk of a failure there (caps will fail
first).

Handling the processor a lot (removal, regrease, reposition
heatsink) will cut into the life expectancy, if say the
processor gets cracked, and it happens to overload the Vcore
circuit. If the processor has the rubber bumpers on the top
of the chip, that will cut that risk a bit.

I would say your biggest exposure, is to external factors.

Paul

Everything man makes turns to dust right?

I don't believe in tempting fate. There are plenty of low quality UPS and
surge protectors out there. I will generalise and say most low cost consumer
stuff is complete rubbish until I am proven otherwise.You can have a *good*
surge protector for as little as $US30. Many surge protectors are only good
for 1 surge!A good true online UPS for an average power computer can cost
$US250 or less.

RAID 1 costs 1 extra disc drive and a controller capable. It will protect
you from drive failures.

There is no replacement for quality proven backups kept off-site. This is
most important.

My sisters house got blown by lightning - oven, fax, x computers, TV's etc.
The only thing that worked afterwards was the traction engine.

Prevention is better than cure.



"notritenoteri" wrote in message
...
YOu want to worry about something? Worry about your data. Your MB is
obsolete, Your OS is obselescent and going obsolete and the only thing
that's of any value is your data. All the stuff about UPS's is theory.
Are
you running NASA and a lunar lander mission? If you are you should be shot
for having only a single system. Given enough time EVERY piece of
equipment
will fail including $1000 UPS's.
I power my system on and off all the time. I don't shut down when
lightening
storms are overhead and I never worry about the telephone lines or cable.
But I live where the power wiring is fairly new ~15 years and its above
ground. If lightening is going to get you its going to get you. I have
heard claims that buried power and phone cables are susceptable to
problems
but I've never seen proof.

"Paul" wrote in message
...
In article URjEd.32449$3m6.5163@attbi_s51, "Travis King"
wrote:

What's probably the life expectancy of my A7V333 motherboard if I take
good
care of it? It has 2 years on it right now. I run the computer for
the
most part constantly except when I leave town or do something with the
inside of the computer. Current MB temperature is at 30 C.

If the case temp is not excessive, the electrolytic caps should
be good for 10 years. Solder joints under stress, could be
anyone's guess (more likely to happen with a P4 retail heatsink
and its high clamping forces).

In ten years, you can expect several PSU failures, and any one of
those PSU failures could damage the motherboard.

If you have a lot of lightning storms, or bad quality power,
that could influence how long the mobo lasts. Look carefully
at any modem, cable modem, ADSL wires etc, to see if there are
any protection devices to take a (nearby) lightning hit, before
it gets to the motherboard. For example, on a phone line, there
may be a carbon block at the entry point, and you could enhance
that by using a second protection device nearer the computer.
For the really paranoid, a wireless network would reduce the
wiring exposure to just the power lines. A real ($1K purchase
price) UPS would reduce the risk of an AC power event from
getting you, and would help protect the PSU from getting
damaged. Cheap UPSes offer no protection at all, as they are
actually SPS (standby power supplies) - they are a "straight wire"
to power spikes, and the unit only cuts over to batteries if
the AC power dies for enough milliseconds.

On the motherboard itself, the Vcore circuit is the circuit under
the most stress. If the MOSFETs are cool to the touch, that is
a good sign. I've never read any MTBF estimates for switching
regulators on motherboards, so don't know whether they are
good for a 1 million hour MTBF or not.

Large BGA packages also have a rating, for solder joint
reliability. For example, a BGA with 750 pins, will last for
about 10 years, with a certain daily temperature variation.
From the Via web page:

* 552-pin BGA VT8366A North Bridge
* 376-pin BGA VT8233 South Bridge

so you have little risk of a failure there (caps will fail
first).

Handling the processor a lot (removal, regrease, reposition
heatsink) will cut into the life expectancy, if say the
processor gets cracked, and it happens to overload the Vcore
circuit. If the processor has the rubber bumpers on the top
of the chip, that will cut that risk a bit.

I would say your biggest exposure, is to external factors.

Paul

MOst surge protectors aren't much good they just don't operate fast enough.
If U insist on protection a good UPS which convers AC to DC charges a
battery pack then provides AC out the other side with an inverter is the
best bet for my money. You can buy these in Canada where I am for about
$200. They have to be maintained and the batteries do go after about 5
years. As to RAID 1 I think you really need 4 drives, controller that is
smart and the software but maybe your definition is different than mine.

"Mercury" wrote in message
...
Everything man makes turns to dust right?

I don't believe in tempting fate. There are plenty of low quality UPS and
surge protectors out there. I will generalise and say most low cost
consumer
stuff is complete rubbish until I am proven otherwise.You can have a
*good*
surge protector for as little as $US30. Many surge protectors are only
good
for 1 surge!A good true online UPS for an average power computer can cost
$US250 or less.

RAID 1 costs 1 extra disc drive and a controller capable. It will protect
you from drive failures.

There is no replacement for quality proven backups kept off-site. This is
most important.

My sisters house got blown by lightning - oven, fax, x computers, TV's
etc.
The only thing that worked afterwards was the traction engine.

Prevention is better than cure.



"notritenoteri" wrote in message
...
YOu want to worry about something? Worry about your data. Your MB is
obsolete, Your OS is obselescent and going obsolete and the only thing
that's of any value is your data. All the stuff about UPS's is theory.
Are
you running NASA and a lunar lander mission? If you are you should be
shot
for having only a single system. Given enough time EVERY piece of
equipment
will fail including $1000 UPS's.
I power my system on and off all the time. I don't shut down when
lightening
storms are overhead and I never worry about the telephone lines or
cable.
But I live where the power wiring is fairly new ~15 years and its above
ground. If lightening is going to get you its going to get you. I have
heard claims that buried power and phone cables are susceptable to
problems
but I've never seen proof.

"Paul" wrote in message
...
In article URjEd.32449$3m6.5163@attbi_s51, "Travis King"
wrote:

What's probably the life expectancy of my A7V333 motherboard if I
take
good
care of it? It has 2 years on it right now. I run the computer for
the
most part constantly except when I leave town or do something with
the
inside of the computer. Current MB temperature is at 30 C.

If the case temp is not excessive, the electrolytic caps should
be good for 10 years. Solder joints under stress, could be
anyone's guess (more likely to happen with a P4 retail heatsink
and its high clamping forces).

In ten years, you can expect several PSU failures, and any one of
those PSU failures could damage the motherboard.

If you have a lot of lightning storms, or bad quality power,
that could influence how long the mobo lasts. Look carefully
at any modem, cable modem, ADSL wires etc, to see if there are
any protection devices to take a (nearby) lightning hit, before
it gets to the motherboard. For example, on a phone line, there
may be a carbon block at the entry point, and you could enhance
that by using a second protection device nearer the computer.
For the really paranoid, a wireless network would reduce the
wiring exposure to just the power lines. A real ($1K purchase
price) UPS would reduce the risk of an AC power event from
getting you, and would help protect the PSU from getting
damaged. Cheap UPSes offer no protection at all, as they are
actually SPS (standby power supplies) - they are a "straight wire"
to power spikes, and the unit only cuts over to batteries if
the AC power dies for enough milliseconds.

On the motherboard itself, the Vcore circuit is the circuit under
the most stress. If the MOSFETs are cool to the touch, that is
a good sign. I've never read any MTBF estimates for switching
regulators on motherboards, so don't know whether they are
good for a 1 million hour MTBF or not.

Large BGA packages also have a rating, for solder joint
reliability. For example, a BGA with 750 pins, will last for
about 10 years, with a certain daily temperature variation.
From the Via web page:

* 552-pin BGA VT8366A North Bridge
* 376-pin BGA VT8233 South Bridge

so you have little risk of a failure there (caps will fail
first).

Handling the processor a lot (removal, regrease, reposition
heatsink) will cut into the life expectancy, if say the
processor gets cracked, and it happens to overload the Vcore
circuit. If the processor has the rubber bumpers on the top
of the chip, that will cut that risk a bit.

I would say your biggest exposure, is to external factors.

Paul

In article , "notritenoteri"
wrote:

MOst surge protectors aren't much good they just don't operate fast enough.
If U insist on protection a good UPS which convers AC to DC charges a
battery pack then provides AC out the other side with an inverter is the
best bet for my money. You can buy these in Canada where I am for about
$200. They have to be maintained and the batteries do go after about 5
years. As to RAID 1 I think you really need 4 drives, controller that is
smart and the software but maybe your definition is different than mine.

At the risk of starting one of those whiny surge protector threads,
the devices used can actually react fast enough. Some of them conduct
in nanoseconds. The problem is with the physics - it is hard to get
the dumped energy into the ground fast enough. Ground wiring is
a secondary consideration in building construction, and even adding
an extra extension cord to a surge protector can render it useless,
due to the inductance of the cord preventing the dumped energy from
getting to ground in time. (The Tripplite ISObar outlet has an
insurance policy, that in invalidated if another extension cord is
used with the product.)

As for the UPS, the UPS acronym is used rather carelessly by the
marketing people. Your typical cheap home UPS is actually an SPS.
An SPS has no filtering capability to speak of, and uses a relay
to switch operating modes. It looks like this:

SPS - Normal operation SPS - Battery operation

AC ---X---X----AC AC ---X X----AC
/
DC DC

You can tell an SPS, because it remains stone cold to the touch
while operating. After all, in normal operation, AC is just being
passed through the device, via a relay.

A real UPS looks like this, and the path stays the same
all the time:

AC -----DC------AC

Such a UPS gets warm/hot, because the inverter on the output
making the AC, dissipates energy just like the switcher inside
your ATX power supply. A real UPS typical costs $1K (for no
good reason that I can see).

Paul


"Mercury" wrote in message
...
Everything man makes turns to dust right?

I don't believe in tempting fate. There are plenty of low quality UPS and
surge protectors out there. I will generalise and say most low cost
consumer
stuff is complete rubbish until I am proven otherwise.You can have a
*good*
surge protector for as little as $US30. Many surge protectors are only
good
for 1 surge!A good true online UPS for an average power computer can cost
$US250 or less.

RAID 1 costs 1 extra disc drive and a controller capable. It will protect
you from drive failures.

There is no replacement for quality proven backups kept off-site. This is
most important.

My sisters house got blown by lightning - oven, fax, x computers, TV's
etc.
The only thing that worked afterwards was the traction engine.

Prevention is better than cure.



"notritenoteri" wrote in message
...
YOu want to worry about something? Worry about your data. Your MB is
obsolete, Your OS is obselescent and going obsolete and the only thing
that's of any value is your data. All the stuff about UPS's is theory.
Are
you running NASA and a lunar lander mission? If you are you should be
shot
for having only a single system. Given enough time EVERY piece of
equipment
will fail including $1000 UPS's.
I power my system on and off all the time. I don't shut down when
lightening
storms are overhead and I never worry about the telephone lines or
cable.
But I live where the power wiring is fairly new ~15 years and its above
ground. If lightening is going to get you its going to get you. I have
heard claims that buried power and phone cables are susceptable to
problems
but I've never seen proof.

"Paul" wrote in message
...
In article URjEd.32449$3m6.5163@attbi_s51, "Travis King"
wrote:

What's probably the life expectancy of my A7V333 motherboard if I
take
good
care of it? It has 2 years on it right now. I run the computer for
the
most part constantly except when I leave town or do something with
the
inside of the computer. Current MB temperature is at 30 C.

If the case temp is not excessive, the electrolytic caps should
be good for 10 years. Solder joints under stress, could be
anyone's guess (more likely to happen with a P4 retail heatsink
and its high clamping forces).

In ten years, you can expect several PSU failures, and any one of
those PSU failures could damage the motherboard.

If you have a lot of lightning storms, or bad quality power,
that could influence how long the mobo lasts. Look carefully
at any modem, cable modem, ADSL wires etc, to see if there are
any protection devices to take a (nearby) lightning hit, before
it gets to the motherboard. For example, on a phone line, there
may be a carbon block at the entry point, and you could enhance
that by using a second protection device nearer the computer.
For the really paranoid, a wireless network would reduce the
wiring exposure to just the power lines. A real ($1K purchase
price) UPS would reduce the risk of an AC power event from
getting you, and would help protect the PSU from getting
damaged. Cheap UPSes offer no protection at all, as they are
actually SPS (standby power supplies) - they are a "straight wire"
to power spikes, and the unit only cuts over to batteries if
the AC power dies for enough milliseconds.

On the motherboard itself, the Vcore circuit is the circuit under
the most stress. If the MOSFETs are cool to the touch, that is
a good sign. I've never read any MTBF estimates for switching
regulators on motherboards, so don't know whether they are
good for a 1 million hour MTBF or not.

Large BGA packages also have a rating, for solder joint
reliability. For example, a BGA with 750 pins, will last for
about 10 years, with a certain daily temperature variation.
From the Via web page:

* 552-pin BGA VT8366A North Bridge
* 376-pin BGA VT8233 South Bridge

so you have little risk of a failure there (caps will fail
first).

Handling the processor a lot (removal, regrease, reposition
heatsink) will cut into the life expectancy, if say the
processor gets cracked, and it happens to overload the Vcore
circuit. If the processor has the rubber bumpers on the top
of the chip, that will cut that risk a bit.

I would say your biggest exposure, is to external factors.

Paul

On Sun, 09 Jan 2005 19:50:06 -0500, (Paul) wrote:

......
wiring exposure to just the power lines. A real ($1K purchase
price) UPS would reduce the risk of an AC power event from
getting you, and would help protect the PSU from getting
damaged. Cheap UPSes offer no protection at all, as they are
actually SPS (standby power supplies) - they are a "straight wire"
to power spikes, and the unit only cuts over to batteries if
the AC power dies for enough milliseconds.


I think that you are being unduly negative about cheap "UPSes". You
are absolutely correct that they are not true UPSes, since:
a) they normally connect the mains power (effectively) straight
through to the controlled devices, thus offering no _intrinsic_
protection from spikes; and
b) they have to switch to inverter mode when the power fails, which
takes finite time.
However:
a) almost all SPSes include at least as much separate surge protection
on the mains supply line as a decent standalone surge protector would
provide; and
b) All except the cheapest, no-name, ones switch fast enough so that a
normal computer system power supply does not "notice" the transient
power loss (though network switches, hubs, and the like may "glitch").
For most home and SOHO users, an SPS will provide cost-effective
protection against most of the data loss problems which might
otherwise be caused by brownouts and/or power outages, while their
built in (but unrelated) surge protection circuitry is a _lot_ better
than nothing as insurance against damage from power line spikes.

Please respond to the Newsgroup, so that others may benefit from the exchange.
Peter R. Fletcher

What you describe as describe as "all SPSes include at least
as much separate surge protection on the mains supply line as
a decent standalone surge protector would provide;" is really
near zero protection for numerous reasons. Both the plug-in
UPS and power strip protector have the same protector
circuit. And both are typically so grossly undersized to be
ineffective. Always start with the numbers. In this case
joules.

In another post and in those previously cited posts are
examples of how joules define protection. Notice that so many
plug-in protectors AND their plug-in UPS counterparts may be
rated at 345 joules. As joules increase, the life expectancy
of that protector increases exponentially. IOW if the plug-in
protector is good for two same size surges, then the 1000
joule 'whole house' protector is good for something on the
order of 300 of those same size surges.

Then it continues farther. The plug-in protector has no
earth ground. Therefore the manufacturer avoids the entire
topic altogether. This is how one identified ineffective (and
grossly overpriced - yes grossly overpriced) plug-in
protectors. 1) No dedicated wire connection to earth ground
AND 2) manufacturer avoids all discussion about earthing.

Further details will be provided in response to Milleron.
But the plug-in protectors are on the order of 10 and 50 times
more expensive per protected appliance. So yes, what you are
calling cheap protectors are really overpriced and expensive
protectors that also are not effective.

BTW, UPSes switch in milliseconds. (One must be careful to
buy power supplies with numerical specs that read: Hold up
time, full load: 16ms. typical). Surges do their damage and
are done in microseconds. 300 consecutive surges could pass
through a UPS before the UPS even considered switching to
battery power. Plug-in UPSes have one function - data
protection. They do not provide the hardware protection so
often implied.

You want a UPS that also provides hardware protection? That
is typically the building wide UPS that also makes this all so
important 'less than 10 foot' connection to earth ground.
Plug-in UPSes are for data protection; not for hardware
protection.

"Peter R. Fletcher" wrote:
On Sun, 09 Jan 2005 19:50:06 -0500, (Paul) wrote:
......
wiring exposure to just the power lines. A real ($1K purchase
price) UPS would reduce the risk of an AC power event from
getting you, and would help protect the PSU from getting
damaged. Cheap UPSes offer no protection at all, as they are
actually SPS (standby power supplies) - they are a "straight wire"
to power spikes, and the unit only cuts over to batteries if
the AC power dies for enough milliseconds.

I think that you are being unduly negative about cheap "UPSes". You
are absolutely correct that they are not true UPSes, since:
a) they normally connect the mains power (effectively) straight
through to the controlled devices, thus offering no _intrinsic_
protection from spikes; and
b) they have to switch to inverter mode when the power fails, which
takes finite time.
However:
a) almost all SPSes include at least as much separate surge protection
on the mains supply line as a decent standalone surge protector would
provide; and
b) All except the cheapest, no-name, ones switch fast enough so that a
normal computer system power supply does not "notice" the transient
power loss (though network switches, hubs, and the like may "glitch").
For most home and SOHO users, an SPS will provide cost-effective
protection against most of the data loss problems which might
otherwise be caused by brownouts and/or power outages, while their
built in (but unrelated) surge protection circuitry is a _lot_ better
than nothing as insurance against damage from power line spikes.

Please respond to the Newsgroup, so that others may benefit from the exchange.
Peter R. Fletcher

On Fri, 14 Jan 2005 14:23:43 -0500, w_tom wrote:

What you describe as describe as "all SPSes include at least
as much separate surge protection on the mains supply line as
a decent standalone surge protector would provide;" is really
near zero protection for numerous reasons. Both the plug-in
UPS and power strip protector have the same protector
circuit. And both are typically so grossly undersized to be
ineffective. Always start with the numbers. In this case
joules.

Unless you live in an area with frequent major thunderstorms, your
surge protector will mainly be blocking the much smaller surges and
spikes on the power line caused by local appliances starting and
stopping. If you have a lighting strike near enough to damage
unprotected equipment in your house, you may anyway want to replace
your surge protectors (or surge-protector containing SPSes), for some
of the reasons you imply - you can do that quite frequently for the
cost of one good enough not to need replacing!


In another post and in those previously cited posts are
examples of how joules define protection. Notice that so many
plug-in protectors AND their plug-in UPS counterparts may be
rated at 345 joules. As joules increase, the life expectancy
of that protector increases exponentially. IOW if the plug-in
protector is good for two same size surges, then the 1000
joule 'whole house' protector is good for something on the
order of 300 of those same size surges.

See above.

Then it continues farther. The plug-in protector has no
earth ground. Therefore the manufacturer avoids the entire
topic altogether. This is how one identified ineffective (and
grossly overpriced - yes grossly overpriced) plug-in
protectors. 1) No dedicated wire connection to earth ground
AND 2) manufacturer avoids all discussion about earthing.

I live in England - all power circuits installed in the last 40 or 50
years here are wired with a separate earth ground, and all half-way
decent UK surge protectors will shunt common-mode surges to this
ground. I must admit that I had forgotten that this (wired grounds to
every power point) "aint necessarily so" in the rest of the world.


Further details will be provided in response to Milleron.
But the plug-in protectors are on the order of 10 and 50 times
more expensive per protected appliance. So yes, what you are
calling cheap protectors are really overpriced and expensive
protectors that also are not effective.

BTW, UPSes switch in milliseconds. (One must be careful to
buy power supplies with numerical specs that read: Hold up
time, full load: 16ms. typical). Surges do their damage and
are done in microseconds. 300 consecutive surges could pass
through a UPS before the UPS even considered switching to
battery power. Plug-in UPSes have one function - data
protection. They do not provide the hardware protection so
often implied.

I think that it was perfectly clear from my earlier response that I
did not believe that _the_ _SPS_ _component_ of a cheap "UPS" provided
any surge or spike protection.


You want a UPS that also provides hardware protection? That
is typically the building wide UPS that also makes this all so
important 'less than 10 foot' connection to earth ground.
Plug-in UPSes are for data protection; not for hardware
protection.

I stand by my original statement, though with the proviso that plug-in
surge protectors, to be reasonably effective, must have a separate
wired ground connection (which can be, and in the UK usually will be,
a connection to a properly wired ground circuit in the house wiring),
and must shunt common mode spikes to it. If your
system/application/business is so critical that you want a guarantee
of 24/7 operation through any reasonably conceivable electrical storm,
then the sort of approach to power protection that you advocate makes
perfect sense, but most people are not in that position. As I think
you or someone else says elsewhere, hardware can be replaced at finite
(and often quite low) cost - lost data may be irreplaceable, and
can't always be backed up to the minute. Relatively cheap UPSes
provide a very high level of data protection gainst brownouts and
power outages and (by virtue of their separate surge protection
circuitry) a significant, though incomplete, level of hardware
protection from the effects of "normally" dirty power. I would be very
concerned if someone read your posts and concluded that, if (s)he
can't afford to spend $1,000+ on whole-house power protection, (s)he
should not bother with anything less.


"Peter R. Fletcher" wrote:
On Sun, 09 Jan 2005 19:50:06 -0500, (Paul) wrote:
......
wiring exposure to just the power lines. A real ($1K purchase
price) UPS would reduce the risk of an AC power event from
getting you, and would help protect the PSU from getting
damaged. Cheap UPSes offer no protection at all, as they are
actually SPS (standby power supplies) - they are a "straight wire"
to power spikes, and the unit only cuts over to batteries if
the AC power dies for enough milliseconds.

I think that you are being unduly negative about cheap "UPSes". You
are absolutely correct that they are not true UPSes, since:
a) they normally connect the mains power (effectively) straight
through to the controlled devices, thus offering no _intrinsic_
protection from spikes; and
b) they have to switch to inverter mode when the power fails, which
takes finite time.
However:
a) almost all SPSes include at least as much separate surge protection
on the mains supply line as a decent standalone surge protector would
provide; and
b) All except the cheapest, no-name, ones switch fast enough so that a
normal computer system power supply does not "notice" the transient
power loss (though network switches, hubs, and the like may "glitch").
For most home and SOHO users, an SPS will provide cost-effective
protection against most of the data loss problems which might
otherwise be caused by brownouts and/or power outages, while their
built in (but unrelated) surge protection circuitry is a _lot_ better
than nothing as insurance against damage from power line spikes.

Please respond to the Newsgroup, so that others may benefit from the exchange.
Peter R. Fletcher


Please respond to the Newsgroup, so that others may benefit from the exchange.
Peter R. Fletcher