Can Low Voltage kill a power supply?

[n33ck0]

Hi, everybody. 2 weeks ago ther was a case of low voltage in the
neighborhood. After calling the utility co. it was determined there was
too much demand on the grid. not anything specific in the house. So
here is my question can a low voltage situation kill a computer power
supply. Taking in account it is not a cheap one: I paid over $75.00US
when new its rated @ 500w and it never gave me a problem before. Also
its strange because the PS on a compaq desktop I also have didn't even
flinch. Nothing else got burned or failed in my house. I know all
components in that particular PC are good 'cause I tested it with
another PS. Thanks Also is there any way I can check if is still good?

 

[Rod Speed]

Hi, everybody. 2 weeks ago ther was a case of low voltage in the
neighborhood. After calling the utility co. it was determined there
was too much demand on the grid. not anything specific in the house.
So here is my question can a low voltage situation kill a computer power supply.

A properly designed power supply should shut down in that situation.

A badly designed power supply can be killed in that situation.

Taking in account it is not a cheap one: I paid over $75.00US when
new its rated @ 500w and it never gave me a problem before.

The price doesnt mean a lot on that basic design question.

Also its strange because the PS on a compaq
desktop I also have didn't even flinch.

Yeah, a properly designed power supply wont turn a hair.

Nothing else got burned or failed in my house. I know all components
in that particular PC are good 'cause I tested it with another PS.

Thats all the evidence you need that its dead.

Thanks Also is there any way I can check if is still good?

You just did.

 

[n33ck0]

A properly designed power supply should shut down in that situation.

A badly designed power supply can be killed in that situation.


The price doesnt mean a lot on that basic design question.


Yeah, a properly designed power supply wont turn a hair.


Thats all the evidence you need that its dead.


You just did.

Thank You Sir!. Suspicions confirmed.

 

[Ray Cassick \(Home\)]

A properly designed power supply should shut down in that situation.

A badly designed power supply can be killed in that situation.

Just a bit curious as to how you think a power supply can be damaged by a
low voltage situation?

Since all the regulation in a PS is designed to convert AC to DC and then
regulate those output voltages down to useable levels within a specified
tolerance I don't see how a low voltage could result in a dead PS. I DO see
how it can cause flakiness on the low voltage side since all outputs are
based upon a properly regulated input voltage tolerance, but all that should
result from a low input voltage is proportionally low output voltages.

Personally I would think that most of the damage would have been caused by a
backlash of higher voltage that can often occur after low voltage
situations. This higher than normal voltage inrush could happen faster than
the regulators are prepared to handle and cause a very quick spike to get
through the filters and fry the lower voltage side of the regulation circuit
perhaps.

 

[Sjouke Burry]

Just a bit curious as to how you think a power supply can be damaged by a
low voltage situation?

Since all the regulation in a PS is designed to convert AC to DC and then
regulate those output voltages down to useable levels within a specified
tolerance I don't see how a low voltage could result in a dead PS. I DO see
how it can cause flakiness on the low voltage side since all outputs are
based upon a properly regulated input voltage tolerance, but all that should
result from a low input voltage is proportionally low output voltages.

Personally I would think that most of the damage would have been caused by a
backlash of higher voltage that can often occur after low voltage
situations. This higher than normal voltage inrush could happen faster than
the regulators are prepared to handle and cause a very quick spike to get
through the filters and fry the lower voltage side of the regulation circuit
perhaps.

A switching supply will draw a bigger current from the
mains to compensate for less voltage, the output voltage
wont drop ,power wont drop, so you need more current.
When input is low enough ,and current becomes big enough,
either the safety cuts in, or the supply blows.
And that depends on the quality of the supply.

 

[Rod Speed]

Just a bit curious as to how you think a power supply can be damaged by a low voltage
situation?

Basically its unlikely that it was just a clean sag in the
mains voltage, its likely that there was a considerable
over voltage as well as the excess load was removed
or disconnected itself in the case of high motor loads etc.

The failure may be something as basic as
the internal fuse thats on the mains active.

Since all the regulation in a PS is designed to convert AC to DC and then regulate those
output voltages down to useable levels within a specified tolerance I don't see how a
low voltage could result in a dead PS.

See above.

I DO see how it can cause flakiness on the low voltage side

That shouldnt happen, the ATX specs require that the supply
shuts down cleanly when the output rails end up out of spec.

since all outputs are based upon a properly regulated input voltage tolerance, but all
that should result from a low input voltage is proportionally low output voltages.

Nope, because the ATX spec requires that the supply must
shut down cleanly when the output rails go out of spec.

Personally I would think that most of the damage would have been caused by a backlash of
higher voltage that can often occur after low voltage situations.

Sure, thats what is likely what killed the badly designed
supply. It couldnt handle that over voltage without dying.

This higher than normal voltage inrush

Its unlikely to be that, much more likely to just be the higher
voltage that killed it if it didnt just blow the internal fuse.

could happen faster than the regulators are prepared to handle

There is no regulation to handle inrush current, it
just goes thru the usual filter and gets absorbed
by the main caps that are used to rectify the mains.

and cause a very quick spike to get through the filters and fry the lower voltage side
of the regulation circuit perhaps.

That would normally fry the chopper etc. The output diodes
are generally pretty brutal devices that arent easily killed by
that sort of thing that would have to get thru the transformer.

 

[Rod Speed]

A switching supply will draw a bigger current from the
mains to compensate for less voltage, the output voltage
wont drop ,power wont drop, so you need more current.
When input is low enough ,and current becomes big enough,
either the safety cuts in, or the supply blows.

That increased current isnt what kills a badly designed supply.

The only current that goes up is from the mains and the diodes
that rectify the mains arent that marginal current capacity wise.

What actually kills a poorly designed power supply in that
situation is the inevitable associated mains surges as loads
trip out due to the low mains voltage, particularly motor loads.

 

[John Doe]

Hi, everybody. 2 weeks ago ther was a case of low voltage in the
neighborhood. After calling the utility co. it was determined
there was too much demand on the grid. not anything specific in
the house. So here is my question can a low voltage situation kill
a computer power supply. Taking in account it is not a cheap one:
I paid over $75.00US

If you are concerned about poor house current, consider buying a
voltage regulator (line conditioner). I've got a cheap one and it
seems to do a good job. I don't need battery backup. I think a
voltage regulator keeps things as stable as possible until it
completely shuts down. That way your power supply doesn't see low
voltages, it's all or nothing.

Good luck.

 

[David Maynard]

That increased current isnt what kills a badly designed supply.

Very well can.

The only current that goes up is from the mains and the diodes
that rectify the mains arent that marginal current capacity wise.

The problem is the flyback regulator. Low voltage on the filter caps means
an increased PWM duty cycle to compensate, stressing everything: drive
transistors, flyback transformer, flyback diodes, etc.

What actually kills a poorly designed power supply in that
situation is the inevitable associated mains surges as loads
trip out due to the low mains voltage, particularly motor loads.

Of course, that shouldn't make it past the input suppressors and filter caps.

 

[Rod Speed]

Rod Speed wrote

Very well can.

Nope, not when it didnt drop enough to cause the other system to turn a hair.

The problem is the flyback regulator. Low voltage on the filter caps means an increased
PWM duty cycle to compensate,
Yes.

stressing everything: drive transistors, flyback transformer, flyback diodes, etc.

That shouldnt kill anything in a properly designed power supply.

Of course, that shouldn't make it past the input suppressors and filter caps.

Shouldnt and didnt are too entirely separate matters.

MUCH more likely than your scenario killing the power supply.

 

[manny]

Just a bit curious as to how you think a power supply can be
damaged by a low voltage situation?

Since all the regulation in a PS is designed to convert AC to
DC and then regulate those output voltages down to useable
levels within a specified tolerance I don't see how a low voltage
could result in a dead PS.

Low AC voltage results in a higher duty cycle, which can cause
inductor saturation. When an inductor saturates it stops
opposing changes in current, and this can cause the current to
increase beyond the capacities of the driving transistors.

 

[Ray Cassick \(Home\)]

Thanks to all for the time and additional explanations.

I never did get into all the intricacies of switching PS so it was all
appreciated.

 

[David Maynard]

Nope, not when it didnt drop enough to cause the other system to turn a hair.

According to that kind of logic it didn't fail at all, yet it did.

That shouldnt kill anything in a properly designed power supply.

In theory nothing should... but it failed.

Shouldnt and didnt are too entirely separate matters.

That's funny considering you've just been singing the "properly designed
power supply" mantra.

MUCH more likely than your scenario killing the power supply.

Frankly, no. Surge suppression of the type you're arguing is relatively
simple and fairly static from design to design. I.E. the power mains don't
'change' when one goes from, say, a 400W design to a 450W. The flyback
circuit, however, is a whole different situation. Get that off so the core
goes into saturation with an unpredicted scenario and you've got fried
flyback. And it's no very hard to do, especially when plagiarizing,
'revising', or designing to power line 'specs' (meaning normal tolerances).

 

[Rod Speed]

Rod Speed wrote

According to that kind of logic it didn't fail at all, yet it did.

Wrong again. One PSU is properly designed and the other isnt.

In theory nothing should... but it failed.

Wrong again. One PSU is properly designed and the other isnt.

That's funny considering you've just been singing the "properly designed power supply"
mantra.

Yes, clearly the one that did die isnt properly designed.

Frankly, no.

Fraid so.

Surge suppression of the type you're arguing is relatively simple and fairly static from
design to design. I.E. the power mains don't 'change' when one goes from, say, a 400W
design to a 450W.

Yes, but you dont know that the badly designed power supply
would have survived that surge even in the 400W version.

Likely it never did the surge suppression properly.

The flyback circuit, however, is a whole different situation.
Get that off so the core goes into saturation with an unpredicted scenario and you've
got fried flyback.

It is however trivial to design the supply so that that cant
happen by monitoring the rectified mains voltage and shutting
the supply down when its low enough to produce that result.

And it's no very hard to do, especially when plagiarizing, 'revising', or designing to
power line 'specs' (meaning normal tolerances).

It is however trivial to design the supply so that that cant
happen by monitoring the rectified mains voltage and shutting
the supply down when its low enough to produce that result.

Not a shred of rocket science required at all.

 

[David Maynard]

Wrong again.

Yes, your statement was 'wrong' because, as you note below, one is not
right so comparing to "the other system" says nothing about whether low
voltage killed it.

One PSU is properly designed and the other isnt.




Wrong again.

Do you ever *think* about what's said or do you just invent whatever seems
convenient?

The statement was perfectly true. In THEORY, nothing should fail in a
properly designed power supply.

One PSU is properly designed and the other isnt.

Which is what I've said all along.

Yes, clearly the one that did die isnt properly designed.

Glad we agree on something so now you can stop singing "in a properly
designed power supply."

Fraid so.

'Fraid not.

Yes, but you dont know that the badly designed power supply
would have survived that surge even in the 400W version.

True, I don't. But surge suppression is trivial and relatively constant
from design to design while flyback design isn't so the likelihood of
screwing up one or the other or both is not the same.

Likely it never did the surge suppression properly.

You mean "likely" because you're in the middle of an argument and saying
"likely" seems a good thing to say.

It is however trivial to design the supply so that that cant
happen by monitoring the rectified mains voltage and shutting
the supply down when its low enough to produce that result.

That costs money. But, no, it isn't 'trivial' because it takes knowing at
what line voltage core saturation would occur (as but one consideration).

It is however trivial to design the supply so that that cant
happen by monitoring the rectified mains voltage and shutting
the supply down when its low enough to produce that result.

That costs money. But, no, it isn't 'trivial' because it takes knowing at
what line voltage core saturation would occur (as but one consideration).

Not a shred of rocket science required at all.

An opinion that proves you've never designed one.

And with that I am finished with this thread. You can not post how it's all
trivial and pretend you 'won' something.

 

[Rod Speed]

Rod Speed wrote

Yes, your statement was 'wrong'

Pathetic, really.

because, as you note below, one is not right so comparing to "the other system" says
nothing about whether low voltage killed it.

I never said that the low voltage is what killed it.

Do you ever *think* about what's said or do you just invent whatever seems convenient?

Have you ever actually managed to bullshit your way out of a wet paper bag ?

The statement was perfectly true.
Nope.

In THEORY, nothing should fail in a properly designed power supply.

Irrelevant to what killed that particular power supply.

Which is what I've said all along.

Liar. You claimed a particular failure mechanism without any
evidence that that is what actually killed that particular supply.

Glad we agree on something so now you can stop singing "in a properly designed power
supply."

Or I could tell you to go and **** yourself instead.

'Fraid not.

Fraid so.

True, I don't. But surge suppression is trivial and relatively constant from design to
design

Irrelevant when you dont know that the supply
that failed has adequate surge suppression.

And it aint just what is trivial that matters, there are plenty of examples
where stupid cost cutting sees inadequate surge suppression.

while flyback design isn't so the likelihood of screwing up one or the other or both is
not the same.

You dont know that the design was screwed up and that it
wasnt just cost cutting that produced the inadequate design.

You mean "likely" because you're in the middle of an argument and saying "likely" seems
a good thing to say.

Wrong, as always. I said that because even with the lower powered
versions of PSUs, anyone with a clue has noticed that some do have
inadequate surge suppression and other gross design deficiencys
like killing what is power from the PSU when the PSU fails.

That costs money.

So does surge suppression.

But, no, it isn't 'trivial' because it takes knowing at what line voltage core
saturation would occur (as but one consideration).

Even the most basic testing of the design would show that problem.

So it is indeed trivial to have the supply shut down when the
input voltage has dropped below what the design can handle.

That costs money.

So does surge suppression.

But, no, it isn't 'trivial' because it takes knowing at what line voltage core
saturation would occur (as but one consideration).

Even the most basic testing of the design would show that problem.

So it is indeed trivial to have the supply shut down when the
input voltage has dropped below what the design can handle.

An opinion that proves you've never designed one.

Wrong, as always. Its a fact that there isnt a shred of rocket
science required at all to have the supply shut down when the
input voltage is below what the design can handle.

And with that I am finished with this thread.

You have always been, and always will be, completely and utterly irrelevant.

You can not post how it's all trivial and pretend you 'won' something.

Only silly little children give a flying red **** about 'winning', child.