Cheap/generic PSUs - any good? (was: Low power socket A)

[Mark: csiphs]

For what it's worth, everyone I know except one has generic
PSUs in their systems (myself included). This coveres probably
somewhere in the range of 30-40 systems, ranging from P4 1.6's
to A64 3200's and dual-MP2800 systems. The only one I've known
to fail under normal circumstances is when a friend of mine
plugged an (overclocked) XP2000 Palomino into an ancient (and
known to blow under high loads) 230W PSU I'd given him to power
an old Pentium-1 class machine. Pop and smoke, but nothing
damaged.

(1) If there is no load applied to a PC's power supply then after a few
seconds it burns up unless it has protective circuitry. I had thought that
cheaper/generic PC PSUs tended to lack this protective circuitry.

(2) I had also thought that the cheaper/generic PC PSU's were more likely
to permit a surge of current through the motherboard if and when the PSU
failed. If I understand this correctly then the motherboard could also get
destroyed.

Are these two dangers no longer a problem with cheap/generic PSUs?






[crossposted to electronics & PC builders groups]

 

[Excalibur]

One other problem is cheap PSU's, is that when you load down one of the
rails (ie the 12v rail) that it will affect the other voltage rails. In the
goog PSU's, one rail does not affect the other. In my opinion, you get what
you pay for. If you buy $2,000.00 plus of computer, and then throw in a
$30.00 PSU, you are asking for trouble. If you bought a Ferrari F50, would
you put regular unleaded in it?? or maybe an engine from chevette??

For what it's worth, everyone I know except one has generic
PSUs in their systems (myself included). This coveres probably
somewhere in the range of 30-40 systems, ranging from P4 1.6's
to A64 3200's and dual-MP2800 systems. The only one I've known
to fail under normal circumstances is when a friend of mine
plugged an (overclocked) XP2000 Palomino into an ancient (and
known to blow under high loads) 230W PSU I'd given him to power
an old Pentium-1 class machine. Pop and smoke, but nothing
damaged.

(1) If there is no load applied to a PC's power supply then after a few
seconds it burns up unless it has protective circuitry. I had thought
that
cheaper/generic PC PSUs tended to lack this protective circuitry.

(2) I had also thought that the cheaper/generic PC PSU's were more likely
to permit a surge of current through the motherboard if and when the PSU
failed. If I understand this correctly then the motherboard could also
get
destroyed.

Are these two dangers no longer a problem with cheap/generic PSUs?






[crossposted to electronics & PC builders groups]

 

[Ed Light]

Fun when the fan stops. Had it happen.


--
Ed Light

Smiley :-/
MS Smiley :-\

Send spam to the FTC at
[email protected]
Thanks, robots.

 

[Jon Danniken]

Fun when the fan stops. Had it happen.

Yep, also fun when capacitors blow and take out the motherboard - had that happen too.

Jon

 

[David Maynard]

]

As I mentioned previously, if you have a specific PSU you
can, with confidence, recommend based on it running a very
similar system (to the extent that power distribution among
the different rails is also similar), for over a year, that
might be relevent... at least it would suggest same
make/model might suffice, for a year. Success with same or
different generics running lower-powered old systems is not
relevant.

Well, this would cover about every PSU I've ever bought.

... and yet generic PSU cause problems quite often, it just
seems that you have a golden touch with them.

For what it's worth, everyone I know except one has generic
PSUs in their systems (myself included). This coveres probably
somewhere in the range of 30-40 systems, ranging from P4 1.6's
to A64 3200's and dual-MP2800 systems. The only one I've known
to fail under normal circumstances is when a friend of mine
plugged an (overclocked) XP2000 Palomino into an ancient (and
known to blow under high loads) 230W PSU I'd given him to power
an old Pentium-1 class machine. Pop and smoke, but nothing
damaged.

(1) If there is no load applied to a PC's power supply then after a few
seconds it burns up unless it has protective circuitry. I had thought that
cheaper/generic PC PSUs tended to lack this protective circuitry.

Sounds like an urban myth.

Without a load the PSU can't regulate (an inherent characteristic of
switching power supplies) but, while I can't say I've seen 'every' power
supply out there, I've never seen one 'burn up' from being unloaded.

(2) I had also thought that the cheaper/generic PC PSU's were more likely
to permit a surge of current through the motherboard if and when the PSU
failed. If I understand this correctly then the motherboard could also get
destroyed.

There are two basic types of output 'protection'. One is over-current
protection (OCP). That is to protect the power supply from a fault external
to the PSU pulling more current than the PSU can provide and wouldn't
protect the motherboard since, if it's pulling fault current then, there's
already something wrong.

I's be surprised if there are many, if any, PSUs that don't have OCP
because that's built into the switching regulator ICs that everyone pretty
much has to use and, btw, this is what causes the symptom of the front
panel power switch won't turn it back on after the 'no boot' with 'the fans
immediately stopped spinning' power on fault. The OCP needs to be reset by
turning off the PSU with the rear PSU power switch, or unplugging it, for
some brief period of time, say 10 seconds, so the internal caps can bleed off.

The other is over-voltage protection (OVP). That circuit should clamp the
output to a 'safe' level if something in the PSU fails causing it to put
out excessive voltage. That would only matter if the fault specifically
caused an over-voltage rather than, say, the outputs going dead to begin
with. It's effectiveness is complicated by how fast it responds and how it
controls the voltage.

Are these two dangers no longer a problem with cheap/generic PSUs?






[crossposted to electronics & PC builders groups]

 

[Dave Platt]

(2) I had also thought that the cheaper/generic PC PSU's were more likely
to permit a surge of current through the motherboard if and when the PSU
failed. If I understand this correctly then the motherboard could also get
destroyed.

There are two basic types of output 'protection'. One is over-current
protection (OCP). That is to protect the power supply from a fault external
to the PSU pulling more current than the PSU can provide and wouldn't
protect the motherboard since, if it's pulling fault current then, there's
already something wrong.[/QUOTE]

I've read one report which indicated that some cheaper/generic PC PSUs
do seem to be skimping, on the input (AC mains) side of things. These
cheap supplies have only a limited amount of RF filtering and
protective circuitry on the mains side - just enough to keep them from
feeding unacceptable (illegal) amounts of RF hash back into the mains.
They have less (often little or no) protection against high-voltage
spikes, inductive kickback, and mains-born RF noise. As a result,
they're somewhat more likely to suffer damage if there's a significant
high-voltage spike/surge on your mains (as may happen at the start of
end of a power brownout or outage).

 

[w_tom]

40 cars run stop signs and have never killed anyone. That
proves stopping is not necessary? Such reasoning without
consulting underlying facts and concepts is called junk
science reasoning. Same junk science reasoning here proves
all that essential power supply circuitry is not required.

All power supplies even 30 years ago required over voltage
and over current protection. As noted by another, some
functions are even provided in power supply controller chips.
Is the function enabled? Only way a consumer can say yes is
if the power supply claims, in writing, that such necessary
functions are provided. If not in writing, then the less than
1% who do such analysis cannot and will not tell others of
missing functions. If power supply does not claim to contain
such functions, then assume essential functions do not exist.

Consumers sometimes suffer failures due to inferior power
supplies. Then they blame failures on 'things that must exist
because they cannot be seen'. Ie power surges. No wonder
some can claim they don't need no stink'in internal
protection. The motherboard caused all his problems. He just
knows.

Switching power supplies can or cannot be operated with no
load. A function defined in power supply manufacturer data
sheets. However a no load condition never damages any
minimally acceptable power supply.

AC mains surges do not pass through switching power
supplies. AC mains first gets filtered. Then gets converted
to 300+ volt DC - more filtering. Then get converted to high
frequency AC. Then passes through an isolation transformer.
Then get converted to DC again. Then gets filtered again.
Power surges on AC mains will pass through all this? Of
course not. Instead destructive surges use other paths that
completely bypass the power supply.

It is very profitable to dump inferior power supplies. Some
even use junk science reasoning to *prove* they know better
than engineers and 30+ years of experience.

 

[David Maynard]

There are two basic types of output 'protection'. One is over-current
protection (OCP). That is to protect the power supply from a fault external
to the PSU pulling more current than the PSU can provide and wouldn't
protect the motherboard since, if it's pulling fault current then, there's
already something wrong.

I've read one report which indicated that some cheaper/generic PC PSUs
do seem to be skimping, on the input (AC mains) side of things. These
cheap supplies have only a limited amount of RF filtering and
protective circuitry on the mains side - just enough to keep them from
feeding unacceptable (illegal) amounts of RF hash back into the mains.
They have less (often little or no) protection against high-voltage
spikes, inductive kickback, and mains-born RF noise. As a result,
they're somewhat more likely to suffer damage if there's a significant
high-voltage spike/surge on your mains (as may happen at the start of
end of a power brownout or outage).
[/QUOTE]

Yes, that's the more likely place for 'cost savings' in super el-cheapo
PSUs because it takes 'extra', or better, parts to include it.

 

[Watson A.Name - \Watt Sun, the Dark Remover\]

[snip]

(1) If there is no load applied to a PC's power supply then after a few
seconds it burns up unless it has protective circuitry. I had thought that
cheaper/generic PC PSUs tended to lack this protective circuitry.

Of the many cheapo generic AT SMPSes that I've worked on, I cannot
remember ever seeing one without an OVP circuit. This usually quickly
shuts down the PS as soon as the voltage rises over the protection
point, around 5.5VDC.

(2) I had also thought that the cheaper/generic PC PSU's were more likely
to permit a surge of current through the motherboard if and when the PSU
failed. If I understand this correctly then the motherboard could also get
destroyed.

Are these two dangers no longer a problem with cheap/generic PSUs?

No longer?? They have never been a problem, AFAIK.

[crossposted to electronics & PC builders groups]

 

[James Sweet]

(1) If there is no load applied to a PC's power supply then after a few
seconds it burns up unless it has protective circuitry. I had thought that
cheaper/generic PC PSUs tended to lack this protective circuitry.

I've never seen one that did that, and I've worked with a lot of power
supplies.

(2) I had also thought that the cheaper/generic PC PSU's were more likely
to permit a surge of current through the motherboard if and when the PSU
failed. If I understand this correctly then the motherboard could also get
destroyed.

That's entirely possible, I've seen it happen twice, don't know how cheap
the PSU's were.

Are these two dangers no longer a problem with cheap/generic PSUs?

Whether they are or not, there's plenty of reasons not to get the cheapest
ones you can find. As with audio equipment, there's plenty of stuff out
there with ridiculous wattage ratings on it, but you'll pay accordingly for
a *real* 300W output.

 

[Mark]

40 cars run stop signs and have never killed anyone. That
proves stopping is not necessary? Such reasoning without
consulting underlying facts and concepts is called junk
science reasoning. Same junk science reasoning here proves
all that essential power supply circuitry is not required.

All power supplies even 30 years ago required over voltage
and over current protection. As noted by another, some
functions are even provided in power supply controller chips.
Is the function enabled? Only way a consumer can say yes is
if the power supply claims, in writing, that such necessary
functions are provided. If not in writing, then the less than
1% who do such analysis cannot and will not tell others of
missing functions. If power supply does not claim to contain
such functions, then assume essential functions do not exist.

Consumers sometimes suffer failures due to inferior power
supplies. Then they blame failures on 'things that must exist
because they cannot be seen'. Ie power surges. No wonder
some can claim they don't need no stink'in internal
protection. The motherboard caused all his problems. He just
knows.

Switching power supplies can or cannot be operated with no
load. A function defined in power supply manufacturer data
sheets. However a no load condition never damages any
minimally acceptable power supply.

AC mains surges do not pass through switching power
supplies. AC mains first gets filtered. Then gets converted
to 300+ volt DC - more filtering. Then get converted to high
frequency AC. Then passes through an isolation transformer.
Then get converted to DC again. Then gets filtered again.
Power surges on AC mains will pass through all this? Of
course not. Instead destructive surges use other paths that
completely bypass the power supply.

It is very profitable to dump inferior power supplies. Some
even use junk science reasoning to *prove* they know better
than engineers and 30+ years of experience.

Best response I've seen so far... obviously this guy knows his stuff.
And knows his basics!

Nice one,

Mark

 

[w_tom]

I have been shocked by the B+ that powers vacuum tubes.
Provides an idea where that experience comes from?
Electro-therapy can make you smarter? Maybe - or not.

 

[Terran Melconian]

I's be surprised if there are many, if any, PSUs that don't have OCP

because that's built into the switching regulator ICs that everyone pretty

Some at least have defective protection if they have it at all. I've
seen reviews of PSUs which included pictures of the charred remains of
some models which failed the overload test. Let me see if I can find
it....

Ah, here we go: http://www6.tomshardware.com/howto/20021021/index.html

They claim that 3 out of the 21 power supplies they tested caught on
fire and/or had components explode when loaded WITHIN the specifications
given by the manufacturer, much less when overloaded.

 

[JAD]

Its a CRAP SHOOT, as I have said many times.... I have as many 70$ antecs as
I do Generic cheapos in the recycle bin. Sure you would have better 'luck'
paying more for a better made unit, but it is far from worry free situation.
The 'warrenty' thing is a joke also.
Has anyone had the machine replaced from a defective PSU(other than 30 days
into a barebones tested system and the like). The only ones I have seen get
anywhere, ended up costing more in court costs than the machine was worth.

I hate these extra worries....AMD can give big headaches, because of the
power requirments....as time passes the standard will need to change, people
are routinely running 4 hd's -2 gigs ram- burner -DVD- huge vidcards -OC'n
(which I have no Idea why, 8mhz to 16mhz was something to accomplish, now
it's 'set a bios switch' and boot, whoopie!) and lights and whistles and
bells with a turbine for a cooler. PSU's will be under the gun soon.

For what it's worth, everyone I know except one has generic
PSUs in their systems (myself included). This coveres probably
somewhere in the range of 30-40 systems, ranging from P4 1.6's
to A64 3200's and dual-MP2800 systems. The only one I've known
to fail under normal circumstances is when a friend of mine
plugged an (overclocked) XP2000 Palomino into an ancient (and
known to blow under high loads) 230W PSU I'd given him to power
an old Pentium-1 class machine. Pop and smoke, but nothing
damaged.

(1) If there is no load applied to a PC's power supply then after a few
seconds it burns up unless it has protective circuitry. I had thought that
cheaper/generic PC PSUs tended to lack this protective circuitry.

(2) I had also thought that the cheaper/generic PC PSU's were more likely
to permit a surge of current through the motherboard if and when the PSU
failed. If I understand this correctly then the motherboard could also get
destroyed.

Are these two dangers no longer a problem with cheap/generic PSUs?






[crossposted to electronics & PC builders groups]

 

[Dave Platt]

Some at least have defective protection if they have it at all. I've
seen reviews of PSUs which included pictures of the charred remains of
some models which failed the overload test. Let me see if I can find
it....

Ah, here we go: http://www6.tomshardware.com/howto/20021021/index.html

A decade or so ago, I worked for the company which designed the
chipset for the 3DO videogame system. The original prototype was a
big stack of boards (about 18" by 24") full of PLDs and discrete logic
chips, wire-wrapped together. It took a _lot_ of power to run it.
There were only two prototypes, both difficult and expensive to build,
and we couldn't afford to

The power was provided by a bank of PC-type switching supplies, one
for every board or two.

The hardware guys put these supplies through a long, extensive
qualification process before approving them for use with the
prototype. The test included a prolonged run at full rated power,
into a set of oil-filled water-cooled dummy loads), transient tests,
and tests to make sure that the overvoltage- and overcurrent-
protection circuitry worked rapidly and safely.

It was amazing how many PC-type PSUs failed their qualification
tests. They either could not deliver rated current without the
voltage sagging or bouncing, shut down abruptly for no reason, didn't
keep the voltages within spec when the protection circuitry operated,
or just rolled over and died.

At least a few of the supplies could very well have resulted in the
destruction of the prototypes, if they'd been put into use. Under
certain conditions (even when used within their ratings) they'd allow
the rail voltages to pop up to above the "absolute maximum" Vcc spec
for the PLDs on the board. We could have ended up with a very
expensive Chinese New Year celebration in the lab.

Fortunately, thanks to the qual, we never lost a board.

 

[Ron Reaugh]

I have been shocked by the B+ that powers vacuum tubes.

How did your cylindrical head get to that voltage?

 

[Jon Danniken]

I have been shocked by the B+ that powers vacuum tubes.
Provides an idea where that experience comes from?
Electro-therapy can make you smarter? Maybe - or not.

I stuck my pinky in the anode lead of a tube tester when I was in my teens; then I
pushed the big red button.

Got thrown a good three feet from that one; good thing I was still young.

Jon

 

[David Maynard]

Some at least have defective protection if they have it at all. I've
seen reviews of PSUs which included pictures of the charred remains of
some models which failed the overload test. Let me see if I can find
it....

Ah, here we go: http://www6.tomshardware.com/howto/20021021/index.html

They claim that 3 out of the 21 power supplies they tested caught on
fire and/or had components explode when loaded WITHIN the specifications
given by the manufacturer, much less when overloaded.

Yes, well, loaded "WITHIN the specifications" wouldn't cause an
over-current protector to kick in, now would it?

I didn't say the things were properly designed and those are good examples
of how just having 'protection' (even when it's listed in the
specifications) doesn't necessarily mean squat if it isn't done right.

It would seem from the examples that those PSUs were built with way
under-rated components.

 

[James Sweet]

Yes, well, loaded "WITHIN the specifications" wouldn't cause an
over-current protector to kick in, now would it?

I didn't say the things were properly designed and those are good examples
of how just having 'protection' (even when it's listed in the
specifications) doesn't necessarily mean squat if it isn't done right.

It would seem from the examples that those PSUs were built with way
under-rated components.

More likely the wattage ratings on them were way inflated, ever seen those
"600 WATT" pc speakers with the 0.75W amp chip in them?

 

[James Sweet]

Its a CRAP SHOOT, as I have said many times.... I have as many 70$ antecs as
I do Generic cheapos in the recycle bin. Sure you would have better 'luck'
paying more for a better made unit, but it is far from worry free situation.
The 'warrenty' thing is a joke also.
Has anyone had the machine replaced from a defective PSU(other than 30 days
into a barebones tested system and the like). The only ones I have seen get
anywhere, ended up costing more in court costs than the machine was worth.

I hate these extra worries....AMD can give big headaches, because of the
power requirments....as time passes the standard will need to change, people
are routinely running 4 hd's -2 gigs ram- burner -DVD- huge vidcards -OC'n
(which I have no Idea why, 8mhz to 16mhz was something to accomplish, now
it's 'set a bios switch' and boot, whoopie!) and lights and whistles and
bells with a turbine for a cooler. PSU's will be under the gun soon.

It's not just AMD anymore, the latest Intel chips are power hogs as well.