UPS Recommendations?

What good is a UPS if you assume it is OK when battery is dead? That
is worse than no UPS. Newer design UPSes will not permit AC power if
the UPS battery is defective so that you remove a non-working UPS. A
desirable feature.

AC electricity comes directly from AC mains - not even through surge
protectors (UPS does not even claim such surge protection in numerical
specifications - that protection is promoted only by myths). If a UPS
battery is dead, you want UPS to not power anything - to announce it is
non-functional.

Typical UPS battery life expectancy is about 3 years. Serious UPS
systems feature a battery life of less than 20 years. Even automobile
batteries exposed to temperature extremes last seven and nine years.
Battery life expectancy is a function of battery quality and of how
battery charge is provided and maintained. But a computer grade UPS is
designed to be 'as cheap as possible' which is why plug-in UPS
batteries die quickly.

Is your UPS battery failing? With a digital volt meter and during a
load, determine UPS battery integrity quickly. There is no reason to
be swapping parts until something works. A meter identifies failures
quickly and can even report a failure before that failure happens.
Human then has sufficient time to find a new battery and prepare for
that failure.

Estimate about three years for a UPS battery. Figure even less time
if UPS is constantly switching to battery backup mode. UPS is for
computer data protection.

Meanwhile, low AC voltage does not cause power supply to work harder.
Computer power supplies even 30 years ago either worked just fine at
all voltages, or shutdown. Low voltage causing power supply failure is
nothing more than speculation, or a power supply that is missing
essential functions. All computer power supplies are required to work
just fine even when voltage drops so low that incandescent bulbs are at
less than 40% intensity. Required even 30+ years ago.

Suspect UPS battery is ready to fail after three years. Others have
demonstrated even faster failure rates. Remember, the plug-in UPS is
designed to be as inexpensive as possible.

Kyle wrote:
It stands to reason that a low input voltage to a PC power supply
might force the switching regulator circuits to work much harder than
normal to maintain rated outputs, causing overheating in the caps and
thus failures, particularly cheap branded PS units with cheap
capacitors. I've also seen cheap caps in more expensive PS units, ya
never know what you're gonna get any more.

As to AC power stability from the power utility, I see brief or
momentary power outages all the time in my area, I live in a large
midwestern US city and see power drops mostly on clear sunny days
(iow, not weather related). Oddly, I usually hear what sounds like a
very loud gunshot in the distance when power dropouts occur (I live
about 1 mile from the power substation). Since I work out of my home,
I invested in UPS protection for all critical systems, and some not so
critical systems (hehe, my gaming rig).

"Newer design UPSes will not permit AC power if the UPS battery is
defective so that you remove a non-working UPS. A desirable feature."

First, I don't belive that this is true of the class of UPS' that
consumers buy (let's say products under 1KVA and under about $150).

Further, the only way that you would even know if the battery was "dead"
(meaning insufficiently alive to run the UPS for at least a long enough
period of time for the computer to shut down) would be to actually run
off of battery power. Because a no-load measurement of the batttery
voltage doesn't really tell you anything.

The way that the retail UPS' work is that they run the AC mains through
a surge protector to the "surge protected" outlets. [I categorically
challenge your assertion that "UPS does not even claim such surge
protection" -- in fact, all of the APC retail class of UPS' have two
distinct sets of outlets, labeled (on my unit under this desk) "surge
protected" and "battery backup plus surge protection"].

The surge protected AC mains also goes to a relay. The load is
connected to the armature of the relay, which can switch between the
surge protected AC mains and the output of the inverter (which runs from
the battery). Normally, when AC power is "ok", the load is connected
directly to the surge protected AC mains, and the inverter is either not
running at all, or is only running for phase synchronization, producing
no power. When a "power line event" occurs, the inverter starts and the
relay switches the load from the surge protected AC mains to the
inverter output. The load is presumed to be capable of handling an
interruption of up to one AC power line cycle (16 milliseconds) for the
inverter startup and switching to occur. On most products the relay is
a mechanical relay and you can hear it "click" as the switch occurs.

There are "full time" UPS' that always run the load from the inverter,
but they are relatively expensive and they are not in the class of
products that are commonly sold in a retail environment for use with PC
class products.


w_tom wrote:
What good is a UPS if you assume it is OK when battery is dead? That
is worse than no UPS. Newer design UPSes will not permit AC power if
the UPS battery is defective so that you remove a non-working UPS. A
desirable feature.

AC electricity comes directly from AC mains - not even through surge
protectors (UPS does not even claim such surge protection in numerical
specifications - that protection is promoted only by myths). If a UPS
battery is dead, you want UPS to not power anything - to announce it is
non-functional.

Typical UPS battery life expectancy is about 3 years. Serious UPS
systems feature a battery life of less than 20 years. Even automobile
batteries exposed to temperature extremes last seven and nine years.
Battery life expectancy is a function of battery quality and of how
battery charge is provided and maintained. But a computer grade UPS is
designed to be 'as cheap as possible' which is why plug-in UPS
batteries die quickly.

Is your UPS battery failing? With a digital volt meter and during a
load, determine UPS battery integrity quickly. There is no reason to
be swapping parts until something works. A meter identifies failures
quickly and can even report a failure before that failure happens.
Human then has sufficient time to find a new battery and prepare for
that failure.

Estimate about three years for a UPS battery. Figure even less time
if UPS is constantly switching to battery backup mode. UPS is for
computer data protection.

Meanwhile, low AC voltage does not cause power supply to work harder.
Computer power supplies even 30 years ago either worked just fine at
all voltages, or shutdown. Low voltage causing power supply failure is
nothing more than speculation, or a power supply that is missing
essential functions. All computer power supplies are required to work
just fine even when voltage drops so low that incandescent bulbs are at
less than 40% intensity. Required even 30+ years ago.

Suspect UPS battery is ready to fail after three years. Others have
demonstrated even faster failure rates. Remember, the plug-in UPS is
designed to be as inexpensive as possible.

Kyle wrote:

It stands to reason that a low input voltage to a PC power supply
might force the switching regulator circuits to work much harder than
normal to maintain rated outputs, causing overheating in the caps and
thus failures, particularly cheap branded PS units with cheap
capacitors. I've also seen cheap caps in more expensive PS units, ya
never know what you're gonna get any more.

As to AC power stability from the power utility, I see brief or
momentary power outages all the time in my area, I live in a large
midwestern US city and see power drops mostly on clear sunny days
(iow, not weather related). Oddly, I usually hear what sounds like a
very loud gunshot in the distance when power dropouts occur (I live
about 1 mile from the power substation). Since I work out of my home,
I invested in UPS protection for all critical systems, and some not so
critical systems (hehe, my gaming rig).

Perhaps you might explain to me why the output filter caps in a
switching supply are typically rated at 105 degrees C? It is fair to
assume this rating is necessary because the current flowing in the
filter caps of a typical buck converter circuit is substantial and
does induce significant heat in the output filters. Now, consider the
scenario created by an undervoltage input condition, either the power
switching devices will fail due to the need to conduct more current to
maintain output voltages, or the output caps will fail due to higher
currents delivered from the coil, assuming a constant power
requirement at the output of the switching circuit (which requirement
is typical). I suppose in a sense your comment is pertinent in that,
according to some experts, it is, in theory, more likely the
semiconductor devices will fail before the output caps fail.
Overcurrent protection circuits limit the maximum current that can
flow in the output filter inductor typically. The need for these
circuits is two-fold, to minimize output voltage overshoot when the
load changes value rapidly, and to prevent overcurrent conditions in
the switching devices and the output caps.

To put it more simply, I've changed more caps in mobo Vcore switching
circuits than I've changed power MOSFETs to repair the circuit to
operational status. What I've seen is caps that can't take the load,
overheat, and blow their seals. Maybe you can more fully explain your
point or offer an alternative analysis.

--
Best regards,
Kyle
"Barry Watzman" wrote in message
...
| No.
|
| Capacitors are used for filtering. They would not react as you
hypothesize.
|
|
| Kyle wrote:
|
| wrote in message
| oups.com...
| | Yes, I understand that the cheaper ones don't use the battery
unless
| of
| | an outage, or dip. However, the power in my area is very
unstable.
| | The first month I had the UPS, it switched on over 100 times per
the
| | software counter. It's a little better now.
| |
| | I had 2 PSU's blow and other hardware instability while the bad
| battery
| | was in place. Replace the battery, or bypass the UPS and things
| work
| | fine. The battery was kicking out about 80V when switched on
per
| VO
| | meter. It did handle a 100W light okay. What else could it
have
| been
| | if not the battery? I still have it so I could test it if I
new
| how.
| |
|
| It stands to reason that a low input voltage to a PC power supply
| might force the switching regulator circuits to work much harder
than
| normal to maintain rated outputs, causing overheating in the caps
and
| thus failures, particularly cheap branded PS units with cheap
| capacitors. I've also seen cheap caps in more expensive PS units,
ya
| never know what you're gonna get any more.
|
| As to AC power stability from the power utility, I see brief or
| momentary power outages all the time in my area, I live in a large
| midwestern US city and see power drops mostly on clear sunny days
| (iow, not weather related). Oddly, I usually hear what sounds
like a
| very loud gunshot in the distance when power dropouts occur (I
live
| about 1 mile from the power substation). Since I work out of my
home,
| I invested in UPS protection for all critical systems, and some
not so
| critical systems (hehe, my gaming rig).

Some have reported that their UPS would not provide AC power until a
dead battery was replaced. Nothing claimed that function was standard
in all UPSes. But it appears to be a function in some newer UPSes - a
desirable function - and not difficult to design even to these computer
grade (least expensive) UPSes - not what you called 'full time'.

UPSes do not connect "through surge protected" outlets. Disconnect
surge protector components, then plug them into another wall receptacle
outlet. 'Surge protection' circuit remains unchange. Protector
circuit connects as if it was another light bulb on same AC circuit.
There is nothing - no surge protector - between AC mains and
'protected' outlets. If you know otherwise, then describe the
schematic connection (series mode devices) by breaking open a UPS.

If your UPS actually claims surge protection in numerical specs, then
list numbers that define protection for each type of surge. And good
luck. APC stopped discussing types of surges long ago so that
embarrassing questions are not asked. Protection is from a surge that
typically does not cause damage. Manufacturers wants you to assume and
then to promote myths.

Once APC did list a type of surge. Still no numbers that define
proteciton:
SURGE PROTECTION AND FILTERING
...
Normal mode clamping response time 0 ns, instantaneous
Normal mode surge voltage let through 5% of test peak voltage
when subjected to IEEE 587 Cat. A 6kVA test
Normal mode noise suppression Full time EMI/RFI filtering

You believe a protector circuit sits between AC mains and electronics
- a series mode protector? Good. Then numerical specs also define
number of dBs. None listed. No numbers. Why? Because no series mode
protection exists - nothing between AC mains and UPS relay. Protector
components - MOVs - are shunt mode devices. APC no longer mentions
surge modes so that you will not ask embarrassing questons. Well I am
asking. Where are these nubmers for each type surge? No numbers
because assumed protection does not exist.

Barry Watzman wrote:
"Newer design UPSes will not permit AC power if the UPS battery is
defective so that you remove a non-working UPS. A desirable feature."

First, I don't belive that this is true of the class of UPS' that
consumers buy (let's say products under 1KVA and under about $150).

Further, the only way that you would even know if the battery was "dead"
(meaning insufficiently alive to run the UPS for at least a long enough
period of time for the computer to shut down) would be to actually run
off of battery power. Because a no-load measurement of the batttery
voltage doesn't really tell you anything.

The way that the retail UPS' work is that they run the AC mains through
a surge protector to the "surge protected" outlets. [I categorically
challenge your assertion that "UPS does not even claim such surge
protection" -- in fact, all of the APC retail class of UPS' have two
distinct sets of outlets, labeled (on my unit under this desk) "surge
protected" and "battery backup plus surge protection"].

The surge protected AC mains also goes to a relay. The load is
connected to the armature of the relay, which can switch between the
surge protected AC mains and the output of the inverter (which runs from
the battery). Normally, when AC power is "ok", the load is connected
directly to the surge protected AC mains, and the inverter is either not
running at all, or is only running for phase synchronization, producing
no power. When a "power line event" occurs, the inverter starts and the
relay switches the load from the surge protected AC mains to the
inverter output. The load is presumed to be capable of handling an
interruption of up to one AC power line cycle (16 milliseconds) for the
inverter startup and switching to occur. On most products the relay is
a mechanical relay and you can hear it "click" as the switch occurs.

There are "full time" UPS' that always run the load from the inverter,
but they are relatively expensive and they are not in the class of
products that are commonly sold in a retail environment for use with PC
class products.

On Wed, 27 Sep 2006 19:39:17 -0400, Barry Watzman
wrote:

"Newer design UPSes will not permit AC power if the UPS battery is
defective so that you remove a non-working UPS. A desirable feature."

First, I don't belive that this is true of the class of UPS' that
consumers buy (let's say products under 1KVA and under about $150).

It is certainly true of my APC Back-UPS Pro 650

Peter Finney
Liphook
Hampshire
England

105C is a temperature rating. It may have as much to do with the
ambient temperature inside the power supply as with anything related to
the capacitor itself. But the current flow into/out of the filter caps
can be substantial ... in an extreme case, the caps must momentarily
supply nearly the entire power supply load (20 amps or more), although
only for a very brief time (microseconds).

The capacitors are on the output (secondary) side of the pulse switching
transformer. Therefore, they don't see and are not subject to a low
(undervoltage) condition on the AC power line input, as long as the
input doesn't go so low that the switching circuit can no longer produce
the necessary output on the secondary side (at which point the supply
should simply shut down completely).

The problems that occured with electrolytic caps (that caused massive
numbers of failures in Vcore supplies on motherboards, but also
elsewhere) were due to the use of an improperly formulated electrolyte
in the capacitors. These caps were truly defective the day that they
were manufactured, and you can't attribute anything about the way that
they behaved (or failed) to normal, properly manufactured parts.


Kyle wrote:

Perhaps you might explain to me why the output filter caps in a
switching supply are typically rated at 105 degrees C? It is fair to
assume this rating is necessary because the current flowing in the
filter caps of a typical buck converter circuit is substantial and
does induce significant heat in the output filters. Now, consider the
scenario created by an undervoltage input condition, either the power
switching devices will fail due to the need to conduct more current to
maintain output voltages, or the output caps will fail due to higher
currents delivered from the coil, assuming a constant power
requirement at the output of the switching circuit (which requirement
is typical). I suppose in a sense your comment is pertinent in that,
according to some experts, it is, in theory, more likely the
semiconductor devices will fail before the output caps fail.
Overcurrent protection circuits limit the maximum current that can
flow in the output filter inductor typically. The need for these
circuits is two-fold, to minimize output voltage overshoot when the
load changes value rapidly, and to prevent overcurrent conditions in
the switching devices and the output caps.

To put it more simply, I've changed more caps in mobo Vcore switching
circuits than I've changed power MOSFETs to repair the circuit to
operational status. What I've seen is caps that can't take the load,
overheat, and blow their seals. Maybe you can more fully explain your
point or offer an alternative analysis.

Your statement that "There is nothing - no surge protector - between AC
mains and 'protected' outlets" is simply wrong for the UPS' that I have
taken apart (mostly APC). The incomming line goes to a surge protector
(MOVs and inductors), and everything (both the UPS and those outlets
that don't get battery power) is on the output side of this surge
protector. I won't argue that on some UPS' this is not the case and
that your statement may be correct. But a "quality" UPS should include
surge protection, and the ones that I've opened up do. This is
internal, inside the UPS, it's not something external that the UPS' plug
into.

APC also has (I think; had for sure) a "connected equipment guarantee",
a type of insurance, in which they would cover damage to equipment
connected to their UPS' caused by surges.


w_tom wrote:

Some have reported that their UPS would not provide AC power until a
dead battery was replaced. Nothing claimed that function was standard
in all UPSes. But it appears to be a function in some newer UPSes - a
desirable function - and not difficult to design even to these computer
grade (least expensive) UPSes - not what you called 'full time'.

UPSes do not connect "through surge protected" outlets. Disconnect
surge protector components, then plug them into another wall receptacle
outlet. 'Surge protection' circuit remains unchange. Protector
circuit connects as if it was another light bulb on same AC circuit.
There is nothing - no surge protector - between AC mains and
'protected' outlets. If you know otherwise, then describe the
schematic connection (series mode devices) by breaking open a UPS.

If your UPS actually claims surge protection in numerical specs, then
list numbers that define protection for each type of surge. And good
luck. APC stopped discussing types of surges long ago so that
embarrassing questions are not asked. Protection is from a surge that
typically does not cause damage. Manufacturers wants you to assume and
then to promote myths.

Once APC did list a type of surge. Still no numbers that define
proteciton:

SURGE PROTECTION AND FILTERING
...
Normal mode clamping response time 0 ns, instantaneous
Normal mode surge voltage let through 5% of test peak voltage
when subjected to IEEE 587 Cat. A 6kVA test
Normal mode noise suppression Full time EMI/RFI filtering


You believe a protector circuit sits between AC mains and electronics
- a series mode protector? Good. Then numerical specs also define
number of dBs. None listed. No numbers. Why? Because no series mode
protection exists - nothing between AC mains and UPS relay. Protector
components - MOVs - are shunt mode devices. APC no longer mentions
surge modes so that you will not ask embarrassing questons. Well I am
asking. Where are these nubmers for each type surge? No numbers
because assumed protection does not exist.

Barry Watzman wrote:

"Newer design UPSes will not permit AC power if the UPS battery is
defective so that you remove a non-working UPS. A desirable feature."

First, I don't belive that this is true of the class of UPS' that
consumers buy (let's say products under 1KVA and under about $150).

Further, the only way that you would even know if the battery was "dead"
(meaning insufficiently alive to run the UPS for at least a long enough
period of time for the computer to shut down) would be to actually run
off of battery power. Because a no-load measurement of the batttery
voltage doesn't really tell you anything.

The way that the retail UPS' work is that they run the AC mains through
a surge protector to the "surge protected" outlets. [I categorically
challenge your assertion that "UPS does not even claim such surge
protection" -- in fact, all of the APC retail class of UPS' have two
distinct sets of outlets, labeled (on my unit under this desk) "surge
protected" and "battery backup plus surge protection"].

The surge protected AC mains also goes to a relay. The load is
connected to the armature of the relay, which can switch between the
surge protected AC mains and the output of the inverter (which runs from
the battery). Normally, when AC power is "ok", the load is connected
directly to the surge protected AC mains, and the inverter is either not
running at all, or is only running for phase synchronization, producing
no power. When a "power line event" occurs, the inverter starts and the
relay switches the load from the surge protected AC mains to the
inverter output. The load is presumed to be capable of handling an
interruption of up to one AC power line cycle (16 milliseconds) for the
inverter startup and switching to occur. On most products the relay is
a mechanical relay and you can hear it "click" as the switch occurs.

There are "full time" UPS' that always run the load from the inverter,
but they are relatively expensive and they are not in the class of
products that are commonly sold in a retail environment for use with PC
class products.

1) If an incoming AC mains line goes through MOVs, then no
electricity is delivered to that 'surge protected' receptacle. Know
what an MOV does. It conducts no electricity - acts like an open
switch - until voltage exceeds its threshold or let-through voltage.
That let-through voltage is typically 330 volts. How does 120 volts
appear on the 'surge protected' outlet when it takes 330 volts to start
current through the MOV? You said the incoming line goes through MOVs.
How? MOVs are open circuit switches when no surge exists?

Apparently the UPS was not yet opened to learn what the 'surge
protected' receptacle connects to. Obviously it does not connect
through MOVs. So where is this surge protection? Well, defined was
how MOVs connect. Remove them from the UPS and connect them to the
other outlet in a duplex wall receptacle. Same protector circuit.

2) Meanwhile, numerical specs from that UPS were not provided. Simply
providing those numbers would have proved your claim. You don't
provide numbers because the UPS does not claim to provide such
protection. APC once claimed to protect from one type surge. But it
did not and does not claim to protect from surges that are typically
destructive.

3) Meanwhile look at that "connected equipment guarantee". So full
of fine print exemptions that a claim will never be honored. But
again, they make subjective claims hoping you don't look at those fine
print details nor look at the numerical specifications.

What are components between AC mains and a 'surge protected'
receptacle? A fuse? Fuse is surge protection? Not for one minute.
You said MOVs are in that circuit. Again, they are not. Did you look
- or just assume?

Those too few MOVs - and again, where are numerical specs - do not
connect as assumed. Obvious: when an MOV is understood, it acts as a
normally open circuit switch. But again, where is this protection when
manufacturer does not even claim same in numerical specifications? No
surge protector components between AC mains and relay. No numbers
claim that protection exists. A guarantee chock full of exemptions. So
where is this series mode protection? If my statement is wrong, then
simply list 'in series' electrical components and post numerical
specs. You cite neither because no such devices and series mode
protection exist.

Meanwhile, some UPSes will not provide AC power when battery is not
functional. Battery life expectancy is typically three years. A
multimeter is a powerful tool to identify a failing battery before that
battery actually fails.

Barry Watzman wrote:
Your statement that "There is nothing - no surge protector - between AC
mains and 'protected' outlets" is simply wrong for the UPS' that I have
taken apart (mostly APC). The incomming line goes to a surge protector
(MOVs and inductors), and everything (both the UPS and those outlets
that don't get battery power) is on the output side of this surge
protector. I won't argue that on some UPS' this is not the case and
that your statement may be correct. But a "quality" UPS should include
surge protection, and the ones that I've opened up do. This is
internal, inside the UPS, it's not something external that the UPS' plug
into.

APC also has (I think; had for sure) a "connected equipment guarantee",
a type of insurance, in which they would cover damage to equipment
connected to their UPS' caused by surges.

Electricity does not go "through" MOVs. MOVs (metal oxide varistors)
are connected across the power line (in parallel with the load), not in
series with the load.

The fact that you don't know this very basic information leads me to
totally discredit your knowledge of the entire subject matter of this
thread.


w_tom wrote:

1) If an incoming AC mains line goes through MOVs, then no
electricity is delivered to that 'surge protected' receptacle. Know
what an MOV does. It conducts no electricity - acts like an open
switch - until voltage exceeds its threshold or let-through voltage.
That let-through voltage is typically 330 volts. How does 120 volts
appear on the 'surge protected' outlet when it takes 330 volts to start
current through the MOV? You said the incoming line goes through MOVs.
How? MOVs are open circuit switches when no surge exists?

Apparently the UPS was not yet opened to learn what the 'surge
protected' receptacle connects to. Obviously it does not connect
through MOVs. So where is this surge protection? Well, defined was
how MOVs connect. Remove them from the UPS and connect them to the
other outlet in a duplex wall receptacle. Same protector circuit.

2) Meanwhile, numerical specs from that UPS were not provided. Simply
providing those numbers would have proved your claim. You don't
provide numbers because the UPS does not claim to provide such
protection. APC once claimed to protect from one type surge. But it
did not and does not claim to protect from surges that are typically
destructive.

3) Meanwhile look at that "connected equipment guarantee". So full
of fine print exemptions that a claim will never be honored. But
again, they make subjective claims hoping you don't look at those fine
print details nor look at the numerical specifications.

What are components between AC mains and a 'surge protected'
receptacle? A fuse? Fuse is surge protection? Not for one minute.
You said MOVs are in that circuit. Again, they are not. Did you look
- or just assume?

Those too few MOVs - and again, where are numerical specs - do not
connect as assumed. Obvious: when an MOV is understood, it acts as a
normally open circuit switch. But again, where is this protection when
manufacturer does not even claim same in numerical specifications? No
surge protector components between AC mains and relay. No numbers
claim that protection exists. A guarantee chock full of exemptions. So
where is this series mode protection? If my statement is wrong, then
simply list 'in series' electrical components and post numerical
specs. You cite neither because no such devices and series mode
protection exist.

Meanwhile, some UPSes will not provide AC power when battery is not
functional. Battery life expectancy is typically three years. A
multimeter is a powerful tool to identify a failing battery before that
battery actually fails.

Barry Watzman wrote:

Your statement that "There is nothing - no surge protector - between AC
mains and 'protected' outlets" is simply wrong for the UPS' that I have
taken apart (mostly APC). The incomming line goes to a surge protector
(MOVs and inductors), and everything (both the UPS and those outlets
that don't get battery power) is on the output side of this surge
protector. I won't argue that on some UPS' this is not the case and
that your statement may be correct. But a "quality" UPS should include
surge protection, and the ones that I've opened up do. This is
internal, inside the UPS, it's not something external that the UPS' plug
into.

APC also has (I think; had for sure) a "connected equipment guarantee",
a type of insurance, in which they would cover damage to equipment
connected to their UPS' caused by surges.

On Thu, 28 Sep 2006 06:28:37 -0400, Barry Watzman
wrote:

The problems that occured with electrolytic caps (that caused massive
numbers of failures in Vcore supplies on motherboards, but also
elsewhere) were due to the use of an improperly formulated electrolyte
in the capacitors. These caps were truly defective the day that they
were manufactured, and you can't attribute anything about the way that
they behaved (or failed) to normal, properly manufactured parts.


http://www.burtonsys.com/bad_BP6/story1.html
http://www.niccomp.com/taiwanlowesr.htm
http://www.edn.com/article/CA255062.html?partner=enews
http://www.pcstats.com/articleview.cfm?articleID=195
http://www.pcmag.com/article2/0,4149,933571,00.asp
http://www.theinquirer.net/default.aspx?article=24596
http://badcaps.net/