Checking power supply fuse

[Terry]

I know electricity but very little about electronics. The schematics
of computer supplies I have looked at all have the fuse in series with
everything in the primary circuit.

I was telling a friend that instead of opening up the power supply to
check the fuse, you should only have to check the back of the computer
with the cord unplugged. If you get continuity between the hot and
neutral that tells you 100% that the fuse is good. It doesn't
necessarily tell you the power supply is good, but it will tell you if
the fuse is good.

My friend says that is not true. Who is right?

 

[larry moe 'n curly]

I was telling a friend that instead of opening up the power supply to
check the fuse, you should only have to check the back of the computer
with the cord unplugged. If you get continuity between the hot and
neutral that tells you 100% that the fuse is good. It doesn't
necessarily tell you the power supply is good, but it will tell you if
the fuse is good.

My friend says that is not true. Who is right?

Your friend, because the AC inputs go into a diode bridge that blocks
the current from the ohm meter. I verified this with ATX PSUs from two
different manufacturers and measured roughly 500K, or about the value
of a typical bleeder resistor..

If the fuse blows, it's often because a transistor or chip on the high
voltage side has gone bad.

A primer on fixing PC PSUs (old but still very useful):

www.nutsvolts.com/PDF_Files/PSRepair.pdf

PSUs can retain a high voltage charge on their big capacitors long
after the AC cord has been unplugged, especially if the PSU is
defective. Also some of the heatsink may be connected directly to
high voltage (I've seen this in Antec SmartPowers and Truepowers and
some Enermaxes).

 

[Terry]

Your friend, because the AC inputs go into a diode bridge that blocks
the current from the ohm meter. I verified this with ATX PSUs from two
different manufacturers and measured roughly 500K, or about the value
of a typical bleeder resistor..

500k is still a complete circuit. A good fuse will always give you a
continuity reading. A bad fuse will always read open? Am I missing
something?

If the fuse blows, it's often because a transistor or chip on the high
voltage side has gone bad.

A primer on fixing PC PSUs (old but still very useful):

www.nutsvolts.com/PDF_Files/PSRepair.pdf

PSUs can retain a high voltage charge on their big capacitors long
after the AC cord has been unplugged, especially if the PSU is
defective. Also some of the heatsink may be connected directly to
high voltage (I've seen this in Antec SmartPowers and Truepowers and
some Enermaxes).

This may be true, but you would not be able to read it (from outside
the case) if the fuse is bad.

 

[Paul]

500k is still a complete circuit. A good fuse will always give you a
continuity reading. A bad fuse will always read open? Am I missing
something?

http://www.pavouk.org/hw/en_atxps.html

If you get a reading of infinity, either one or both of the bleeder
resistors is open, or the fuse is open. Infinity is not conclusive
for fuse state. If the supply was damaged by lightning, there may
be no "logic" to what blows, so you cannot really say that bleeders
never fail.

And, in any case, to read 500K, you need "high power ohms". To get
a reading, two diodes in the bridge rectifier have to be forward
biased. If your meter uses low power ohms, you might get a reading
of infinity, even when the fuse is good. So, you need to know
something about your meter. By using two meters, setting one to volts,
and the other to ohms, you can investigate what kind of open circuit
test voltage is being used.

Paul

 

[Terry]

http://www.pavouk.org/hw/en_atxps.html

If you get a reading of infinity, either one or both of the bleeder
resistors is open, or the fuse is open. Infinity is not conclusive
for fuse state. If the supply was damaged by lightning, there may
be no "logic" to what blows, so you cannot really say that bleeders
never fail.

And, in any case, to read 500K, you need "high power ohms". To get
a reading, two diodes in the bridge rectifier have to be forward
biased. If your meter uses low power ohms, you might get a reading
of infinity, even when the fuse is good. So, you need to know
something about your meter. By using two meters, setting one to volts,
and the other to ohms, you can investigate what kind of open circuit
test voltage is being used.

Paul

Looking at the schematic you provided, the first thing in the circuit
is a fuse. If you get continuity, then 100% the fuse is good.

Like you say, you can get infinity for more than one reason. Infinity
does not necessarily mean the fuse is bad, but continuity does always
mean the fuse is good.

My statement still seems true to me.

 

[Paul]

Looking at the schematic you provided, the first thing in the circuit
is a fuse. If you get continuity, then 100% the fuse is good.

Like you say, you can get infinity for more than one reason. Infinity
does not necessarily mean the fuse is bad, but continuity does always
mean the fuse is good.

My statement still seems true to me.

I don't understand why the state of the fuse is important to you.

If you want unambiguous results, the best way is to get as close as
possible to the source. The results of a "distant" measurement, are
only as good as your understanding of the physical construction of the
thing. For example, if there was carbon scoring on the unit somewhere,
maybe you could get your 500K reading that way. The only unambiguous
reading to me, is to take fuse in hand and measure it. That eliminates
a lot of other (obscure) possibilities. If I was to rely on such a
remote measurement, I'd at least want to visually examine the area
in question, to see if there are other ways to interpret the results.

In your case, I would treat a non-infinite reading as an "indication"
that the fuse could be good, but not a guarantee. For example, if
I smelled a burning smell coming from the unit, I'd have to open
it up and see what kind of a mess was in there. Say that something
exploded, and blew conductive material around...

Generally, when I'm using a multimeter, I'd debugging something. Like
a detective, I'm gathering evidence. I never trust one single piece
of evidence by itself. If several pieces of evidence point in a
particular direction, then there may be a theory to explain all the
observations.

To give you another silly example, take an ordinary filament lightbulb.
Occasionally, one end of the filament comes loose, and flops around
inside the bulb. The filament is still welded to one leg of the bulb,
but the other end is loose. Now, I could measure the light bulb, and
get a zero ohm reading, I could tap the bulb lightly on the table,
and get an infinite reading. Should I have concluded, from the first
measurement, that the bulb was good ? In your view, the bulb would be
good. Yet, if we take the two readings together, we have a body of
evidence. A possible explanation, is one end of the filament is loose.
A third measurement would be visual inspection of the bulb, for signs
of a loose filament.

Paul

 

[Conor]

500k is still a complete circuit. A good fuse will always give you a
continuity reading. A bad fuse will always read open? Am I missing
something?

Yes. The fact the fuse is on the LIVE side of a circuit, not the
neutral.

 

[Terry]

I don't understand why the state of the fuse is important to you.

It would keep you from having to open the power supply to check the
fuse.

Checking the fuse would be about the only repair many

If you want unambiguous results, the best way is to get as close as
possible to the source. The results of a "distant" measurement, are
only as good as your understanding of the physical construction of the
thing. For example, if there was carbon scoring on the unit somewhere,
maybe you could get your 500K reading that way. The only unambiguous
reading to me, is to take fuse in hand and measure it. That eliminates
a lot of other (obscure) possibilities. If I was to rely on such a
remote measurement, I'd at least want to visually examine the area
in question, to see if there are other ways to interpret the results.

In your case, I would treat a non-infinite reading as an "indication"
that the fuse could be good, but not a guarantee. For example, if
I smelled a burning smell coming from the unit, I'd have to open
it up and see what kind of a mess was in there. Say that something
exploded, and blew conductive material around...

Generally, when I'm using a multimeter, I'd debugging something. Like
a detective, I'm gathering evidence. I never trust one single piece
of evidence by itself. If several pieces of evidence point in a
particular direction, then there may be a theory to explain all the
observations.

To give you another silly example, take an ordinary filament lightbulb.
Occasionally, one end of the filament comes loose, and flops around
inside the bulb. The filament is still welded to one leg of the bulb,
but the other end is loose. Now, I could measure the light bulb, and
get a zero ohm reading, I could tap the bulb lightly on the table,
and get an infinite reading. Should I have concluded, from the first
measurement, that the bulb was good ? In your view, the bulb would be
good. Yet, if we take the two readings together, we have a body of
evidence. A possible explanation, is one end of the filament is loose.
A third measurement would be visual inspection of the bulb, for signs
of a loose filament.

Paul

I am sure you can come up with a few examples where that is true. I
don't think you could be able to do that with a fuse, which is what
were are talking about.

 

[larry moe 'n curly]

I was telling a friend that instead of opening up the power supply to
check the fuse, you should only have to check the back of the computer
with the cord unplugged. If you get continuity between the hot and
neutral that tells you 100% that the fuse is good. It doesn't
necessarily tell you the power supply is good, but it will tell you if
the fuse is good.

My friend says that is not true. Who is right?


500k is still a complete circuit. A good fuse will always give you a
continuity reading. A bad fuse will always read open? Am I missing
something?

It depends on where the 500K bleeder resistor is located. If it's on
the main circuit board it's probably after the fuse, but if it's on a
small separate board or built into the AC plug, it may come before the
fuse, as in this 400W Ultra Value:

http://farm1.static.flickr.com/74/196076358_abe5adea14_b.jpg

and this 300W cheapo (see area marked "EMI filter"):

http://farm1.static.flickr.com/41/75125024_161da61ad7_b.jpg

 

[Paul]

It depends on where the 500K bleeder resistor is located. If it's on
the main circuit board it's probably after the fuse, but if it's on a
small separate board or built into the AC plug, it may come before the
fuse, as in this 400W Ultra Value:

http://farm1.static.flickr.com/74/196076358_abe5adea14_b.jpg

and this 300W cheapo (see area marked "EMI filter"):

http://farm1.static.flickr.com/41/75125024_161da61ad7_b.jpg

The resistor on the back of the second one, could be similar to
the resistor used across the rails, in this EMI filter design.
I suspect it could be a bleeder for the caps in the EMI filter
or something. At least, I don't see another explanation for a
resistor right across the rails.

http://www.cor.com/PDF/N.pdf

A resistor placed on the back of the EMI filter wouldn't help be a
bleeder for the main caps, because the bridge rectifier would
prevent reverse current flow. And now that I think about it,
the big caps would appear as a slowly charging waveform, if you
connected a high power ohmmeter. So maybe what is happening, in fact,
is the 500K reading is coming from the EMI filter stage, and
because the ohmmeter is low power ohms, it cannot even see the
bleeders after the bridge rectifier.

So where exactly is the fuse in those things ? Is it soldered
to a PCB or did they use a fuse holder ?

Paul