Power supply hot to touch


K

kony

Hi All,

I just took the side of my case to improve CPU temps and noticed two things:

1) The PSU is very hot. I mean to the point of burning your hand almost (as
is the exhaust air from the back).

Either:

A) It's junk and will fail soon
B) The fan and/or case intake isn't adequate for the load.

2) I think the fan on the bottom is blowing hot air into the case, not out.

?? If it has adequate exhaust area, that shouldn't be an
issue.

Removing the side dropped CPU and both core temps by about 10C. I only have
one 120mm exhaunst fan and the same at the front for intake (I don't think
either of these work very well - I can hardly feel any air move past them).

This suggests the case itself impedes airflow, but it
doesn't necessarily mean the PSU is of acceptible quality or
capacity.

Now as for the PSU, this worries me because I'm used to it being a little
warm, but not hot like this. It's a Dynex 500W (specs below), my old blew
and this is the only brand Bestbuy carries (nearest place to get a PSU same
day). For the longest time I thought both fans on the PSU were blowing
outwards (bottom one sucking hot air from the case and the rear one blowing
it out).

I would not carry any confidence at all on only what bestbuy
carried. There are ample PSU reviews online, if you can
take that PSU back for refund that seems the best option.

If the PSU is blowing more hot air into the case, it's likely why my CPU
temps are high - the fan is right over the hsf for the cpu.

PSU specs:
--------------------------------------------------------------------------
Voltage: +3.3V +5V +12V1 +12V2 -5V -12V +5Vsb
Max Load: 35A 40A 16A 18A 0.5A 0.8A 2.5A
Min load: 0.5A 2.0A 1.0A 1.0A 0A 0A 0A
Regulations: ±5% ±5% ±5% ±10% ±10% ±5%
Ripple: 50mV 50mV 120mV 120mV 100mV 120mV 50mV
Noise & Ripple 100mV 100mV 200mV 200mV 200mV 200mv 100mv
Fan Speed Control Inside
--------------------------------------------------------------------------

My setup:
---------------------------------------------------------------
M/B: MSI K9N4SLI
CPU: Athlon 64 X2 5200+ (2600Mhz, 13 x 200, stock values)
RAM: 2x 512MB Kingston HyperX DDR2 800Mhz
Gfx: PCI-E RADEON X1600
HDD1: Western Digital WD800JB PATA 7200RPM
HDD2: Western Digital WD2000JB PATA 7200RPM
DVD: Sony/Optiarc AD5170A DVDRW

Vcore: 1.3v (stock voltage - it won't run below this)
RAM: 1.95v (again, stock voltage)

The heat is a function of fan speed, when considering
external PSU case temp, but it is also likely it's cheap
junk and should be replaced.
 
Ad

Advertisements

S

Skeleton Man

Hi All,

I just took the side of my case to improve CPU temps and noticed two things:

1) The PSU is very hot. I mean to the point of burning your hand almost (as
is the exhaust air from the back).
2) I think the fan on the bottom is blowing hot air into the case, not out.

Removing the side dropped CPU and both core temps by about 10C. I only have
one 120mm exhaunst fan and the same at the front for intake (I don't think
either of these work very well - I can hardly feel any air move past them).

Now as for the PSU, this worries me because I'm used to it being a little
warm, but not hot like this. It's a Dynex 500W (specs below), my old blew
and this is the only brand Bestbuy carries (nearest place to get a PSU same
day). For the longest time I thought both fans on the PSU were blowing
outwards (bottom one sucking hot air from the case and the rear one blowing
it out).

If the PSU is blowing more hot air into the case, it's likely why my CPU
temps are high - the fan is right over the hsf for the cpu.

PSU specs:
--------------------------------------------------------------------------
Voltage: +3.3V +5V +12V1 +12V2 -5V -12V +5Vsb
Max Load: 35A 40A 16A 18A 0.5A 0.8A 2.5A
Min load: 0.5A 2.0A 1.0A 1.0A 0A 0A 0A
Regulations: ±5% ±5% ±5% ±10% ±10% ±5%
Ripple: 50mV 50mV 120mV 120mV 100mV 120mV 50mV
Noise & Ripple 100mV 100mV 200mV 200mV 200mV 200mv 100mv
Fan Speed Control Inside
--------------------------------------------------------------------------

My setup:
---------------------------------------------------------------
M/B: MSI K9N4SLI
CPU: Athlon 64 X2 5200+ (2600Mhz, 13 x 200, stock values)
RAM: 2x 512MB Kingston HyperX DDR2 800Mhz
Gfx: PCI-E RADEON X1600
HDD1: Western Digital WD800JB PATA 7200RPM
HDD2: Western Digital WD2000JB PATA 7200RPM
DVD: Sony/Optiarc AD5170A DVDRW

Vcore: 1.3v (stock voltage - it won't run below this)
RAM: 1.95v (again, stock voltage)
--------------------------------------------------------------

Given the specs above, is it likely the PSU is cheaply built and being run
to capacity ? (hence the heat)

Chris
 
S

Skeleton Man

One thing I just noticed doing some math - how can you get 724W of power
from a PSU rated for 500W ?

3.3V x 35A = 116W
5.0V x 40A = 200W
12.0V x 16A = 192W
12.0V x 18A = 216W
-----------------------
Total: 724W

Am I calculating something wrong ?

Chris
 
I

Ian D

Skeleton Man said:
One thing I just noticed doing some math - how can you get 724W of power
from a PSU rated for 500W ?

3.3V x 35A = 116W
5.0V x 40A = 200W
12.0V x 16A = 192W
12.0V x 18A = 216W
-----------------------
Total: 724W

Am I calculating something wrong ?

Chris
Those are the maximums each rail can supply. In the specs it usually
states what specific combinations can supply. It's very unlikely the
3.3V or 5.0V demands will be anywhere near the PSU capabilities.
The important spec these days is for the 12V rails as they power the CPU
and video card(s).
 
M

Mike Walsh

The specs are for each voltage when it is running alone. E.g. if you are using only 3.3 volts it will put out 116 watts and if you are using only 5 volts it will put out 200 watts, but the combined output of 5 volts and 3.3 volts together might be only 220 watts. Also the ratings will vary depending on if the rating is for peak power when cold or continuous power when hot.
 
P

Paul

Skeleton said:
Hi All,

I just took the side of my case to improve CPU temps and noticed two things:

1) The PSU is very hot. I mean to the point of burning your hand almost (as
is the exhaust air from the back).
2) I think the fan on the bottom is blowing hot air into the case, not out.

Removing the side dropped CPU and both core temps by about 10C. I only have
one 120mm exhaunst fan and the same at the front for intake (I don't think
either of these work very well - I can hardly feel any air move past them).

Now as for the PSU, this worries me because I'm used to it being a little
warm, but not hot like this. It's a Dynex 500W (specs below), my old blew
and this is the only brand Bestbuy carries (nearest place to get a PSU same
day). For the longest time I thought both fans on the PSU were blowing
outwards (bottom one sucking hot air from the case and the rear one blowing
it out).

If the PSU is blowing more hot air into the case, it's likely why my CPU
temps are high - the fan is right over the hsf for the cpu.

PSU specs:
--------------------------------------------------------------------------
Voltage: +3.3V +5V +12V1 +12V2 -5V -12V +5Vsb
Max Load: 35A 40A 16A 18A 0.5A 0.8A 2.5A
Min load: 0.5A 2.0A 1.0A 1.0A 0A 0A 0A
Regulations: ±5% ±5% ±5% ±10% ±10% ±5%
Ripple: 50mV 50mV 120mV 120mV 100mV 120mV 50mV
Noise & Ripple 100mV 100mV 200mV 200mV 200mV 200mv 100mv
Fan Speed Control Inside
--------------------------------------------------------------------------

My setup:
---------------------------------------------------------------
M/B: MSI K9N4SLI
CPU: Athlon 64 X2 5200+ (2600Mhz, 13 x 200, stock values)
RAM: 2x 512MB Kingston HyperX DDR2 800Mhz
Gfx: PCI-E RADEON X1600
HDD1: Western Digital WD800JB PATA 7200RPM
HDD2: Western Digital WD2000JB PATA 7200RPM
DVD: Sony/Optiarc AD5170A DVDRW

Vcore: 1.3v (stock voltage - it won't run below this)
RAM: 1.95v (again, stock voltage)
--------------------------------------------------------------

Given the specs above, is it likely the PSU is cheaply built and being run
to capacity ? (hence the heat)

Chris

Some interpretations:

1) Your previous power supply blew. Perhaps the motherboard was overloading it.
The motherboard could be overloading the current supply.

2) The supply is getting hot, because of a fault within the supply
itself. There could be a short in there somewhere.

3) The power supply fan has stopped turning.

For (1), I would use the clamp-on DC ammeter I own, to measure
all rails, and see if excessive motherboard current draw is
present. Because the clamp-on meter is a non-contact device,
it can measure and characterize a computer rapidly, without
the need to snip wires.

For (3) you can check visually, as well as audibly. If the
fan slowed down, perhaps you remember it running faster at one
time.

For (2), that would be harder to check. I might disconnect the supply
from the motherboard, plug it into a Kill-a-Watt meter, and see
if the unloaded dissipation is in the 10W to 20W region or not,
with PS_ON# connected to COM.

In other words, if you really want to know what is going on,
it'll take a bit of test equipment. Slapping new power supplies
in there, over and over again, may not fix it.

A Kill-a-Watt meter is cheap, and is worthwhile to have around the
house. The clamp-on DC ammeter is much more expensive.

http://ecx.images-amazon.com/images/I/210M82BKVAL._AA183_.jpg
http://www.amazon.com/P3-International-Kill-Electricity-Monitor/dp/B00009MDBU

Paul
 
Ad

Advertisements

W

w_tom

One thing I just noticed doing some math - how can you get 724W of power
from a PSU rated for 500W ?

You are already doing more than most manufacturers intend. Most
computer assemblers who buy power supplies look only two numbers:
watts and dollars. So many power supplies are marketed to people who
would not do numbers.

ATX specs note that one voltage may output a maximum amperage only
when the other voltage load is much smaller. Charts even exist for
these relationships.

Did excessive loading cause an old supply to fail? Maybe if the
supply was defective when purchased. For any properly designed supply
- no load can damage the supply. In fact, every supply can have
shorted together all outputs without any supply damage. Some specs
even define how large that wire must be to short together all outputs
- without damage. Power supply must provide any load or shutdown.
Damage is not listed among the possibilities - if a supply is properly
designed.

How much was that new supply? One benchmark; if less than $60 full
retail, then is probably missing essential functions. Obviously the
reverse is not true: that a $60 power supply is sufficient.

Did that supply come with a long list of numeric specs that list its
many required functions? Better supplies will claim to provide those
required functions - in writing. If not, then assume some required
functions are missing. Normal is a computer booting when a supply is
defective. You can identify a bad supply but cannot 'see' a good one.
Latter requires a simple tool.

One chassis fan should be sufficient cooling for any computer even
operating in a normal 100 degree F environment. No, that is not
normal for you but must be normal for any computer. Does not matter
which direction fans blow. Relevant factor is air flow - the number
of CFMs that move through a chassis. Obviously if both 120 mm fans
blow in and no other exhaust is available, then number of CFMs can be
reduced.

That is chassis temperature - not to be confused with your other
objective - CPU temperatures. CPU temperature analysis starts with a
spec for that heatsink fan combo - degree C per watt. What does that
manufacturers state in writing?

Power supply should typically work without a fan in a 70 degree F
room. Then we push air through with a fan to make supply even more
robust, so that it will work at room temperatures above 100 degree F,
AND to move sufficient air through the chassis. Same fan solved two
problems. Some supplies are not so robust. But appreciate an old
rule for all semiconductors. If you touch it and don't leave skin,
then it is not too hot.

Still, prefer your power supply case to be less than 100 degrees F
when in a 70 degree room because doing that is so easy. How hot is
that power supply box - in numbers? How much air is really moving
through the chassis (do fans operate in series or are they 'fighting'
each other)?
 
K

Ken Maltby

kony said:
Either:

A) It's junk and will fail soon
B) The fan and/or case intake isn't adequate for the load.



?? If it has adequate exhaust area, that shouldn't be an
issue.

So you see no problem with a PSU that has fans blowing air
out, but no way for much air to go into the PSU? Personally,
I would think competing fans might actually make the PSU
hotter.

This suggests the case itself impedes airflow, but it
doesn't necessarily mean the PSU is of acceptible quality or
capacity.

True, at least in the since that the hot air from the PSU
had no place to go , but onto the CPU heat sink, with the
side on. With the side off, there would be a path that allows
it to flow, and even hot air can carry away heat, if it is
moving air.
I would not carry any confidence at all on only what bestbuy
carried. There are ample PSU reviews online, if you can
take that PSU back for refund that seems the best option.



The heat is a function of fan speed, when considering
external PSU case temp, but it is also likely it's cheap
junk and should be replaced.

How about: it's a function of the air flow through the
PSU case? If you have a box with competing fans;
where both are trying to blow air in or trying to blow
the air out, you aren't going to have much air flow. The
best practice, with multiple fans, is to have intake fans in
one corner of the box and the exhaust fans in the opposite
corner, so the air flows through the whole box. Since
hot air rises, it is natural to have the intake at or near the
bottom and the exhaust at or near the top.

But you knew that, didn't you, Kony?

Luck;
Ken
 
J

John McGaw

Skeleton said:
Hi All,

I just took the side of my case to improve CPU temps and noticed two things:

1) The PSU is very hot. I mean to the point of burning your hand almost (as
is the exhaust air from the back).
2) I think the fan on the bottom is blowing hot air into the case, not out.

Removing the side dropped CPU and both core temps by about 10C. I only have
one 120mm exhaunst fan and the same at the front for intake (I don't think
either of these work very well - I can hardly feel any air move past them).

Now as for the PSU, this worries me because I'm used to it being a little
warm, but not hot like this. It's a Dynex 500W (specs below), my old blew
and this is the only brand Bestbuy carries (nearest place to get a PSU same
day). For the longest time I thought both fans on the PSU were blowing
outwards (bottom one sucking hot air from the case and the rear one blowing
it out).

If the PSU is blowing more hot air into the case, it's likely why my CPU
temps are high - the fan is right over the hsf for the cpu.

PSU specs:
--------------------------------------------------------------------------
Voltage: +3.3V +5V +12V1 +12V2 -5V -12V +5Vsb
Max Load: 35A 40A 16A 18A 0.5A 0.8A 2.5A
Min load: 0.5A 2.0A 1.0A 1.0A 0A 0A 0A
Regulations: ±5% ±5% ±5% ±10% ±10% ±5%
Ripple: 50mV 50mV 120mV 120mV 100mV 120mV 50mV
Noise & Ripple 100mV 100mV 200mV 200mV 200mV 200mv 100mv
Fan Speed Control Inside
--------------------------------------------------------------------------

My setup:
---------------------------------------------------------------
M/B: MSI K9N4SLI
CPU: Athlon 64 X2 5200+ (2600Mhz, 13 x 200, stock values)
RAM: 2x 512MB Kingston HyperX DDR2 800Mhz
Gfx: PCI-E RADEON X1600
HDD1: Western Digital WD800JB PATA 7200RPM
HDD2: Western Digital WD2000JB PATA 7200RPM
DVD: Sony/Optiarc AD5170A DVDRW

Vcore: 1.3v (stock voltage - it won't run below this)
RAM: 1.95v (again, stock voltage)
--------------------------------------------------------------

Given the specs above, is it likely the PSU is cheaply built and being run
to capacity ? (hence the heat)

Chris

I have no doubt that the PSU is cheaply built. Had you ever heard of
Dynex before you wandered into BestBuy?

But even a cheap PSU wouldn't be expected to blow hot air from its
bottom fan into the computer case -- it definitely should be sucking air
from above the CPU and ejecting it out the rear of the case.. Could be
that one of the assembly line workers was having an especially bad day
and inserted it backwards...
 
W

w_tom

But even a cheap PSU wouldn't be expected to blow hot air from its
bottom fan into the computer case -- it definitely should be sucking air
from above the CPU and ejecting it out the rear of the case.

First, the original ATX specs called for blowing air into the case
from power supply. Later, that standard changed; reversed airflow.

Second, relevant number is CFM through the chassis. Basic science
(read application notes from fan manufacturers) makes it obvious.
Relevant to chassis internal temperatures is the parameter called CFM
and temperature where air enters the chassis.

One standard 80 mm or 120 mm fan provides more than enough airflow
even assuming no heat gets transferred through a metal chassis.
Second fan (if in parallel) provides only a marginal temperature
improvement. But it two fans are blowing in so as to reduce the
total CFMs, then a problem has been created. Simplest answer for
chassis cooling starts with these numbers. How many CFM are being
moved through that chassis? Airflow direction is irrelevant. CFMs are
relevant.


Meanwhile, as others noted, that power supply should not be hot to
touch. Hot? How hot? How many have assumed 130 degrees F or
higher? Without numbers, everyone must speculate.
 
S

Skeleton Man

Meanwhile, as others noted, that power supply should not be hot to
touch. Hot? How hot? How many have assumed 130 degrees F or
higher? Without numbers, everyone must speculate.

Around 50C/122F

Chris
 
Ad

Advertisements

L

larry moe 'n curly

Skeleton Man wrote:
w_tom said:
1) The PSU is very hot. I mean to the point of burning your
hand almost (as is the exhaust air from the back).

Around 50C/122F

Did it cool down, or were you exaggerating in your original post?
Because 50C isn't that hot.

Still you should check the fans by comparing the air flow from that
PSU against the flow from a known good PSU. Also unplug the AC power
cord from the PSU (important!) and try rotating each fan with
something nonmetallic, like a chop stick or plastic straw, to see if
they spin as easily as the fans in a good PSU.

Is it possible that a PSU fan is right next to the CPU fan, causing
each one to get very little intake air?

When PSUs develop bad electrolytic capacitors, the transistors and
diodes connected to them may turn on and off much more slowly and
cause overheating. See www.badcaps.net.
 
K

kony

So you see no problem with a PSU that has fans blowing air
out, but no way for much air to go into the PSU? Personally,
I would think competing fans might actually make the PSU
hotter.

What I meant was that there aren't any PSU made that blow
into the case instead of out, AFAIK, but even if there were
it would not make the PSU hotter it would tend to make it a
little cooler running at the expense of the rest of the
system getting hotter.

Instead what seemed more likely was that this is one of many
PSU that have the bottom fan but also have some slits on the
rear such that some of the intake air is blown back into
case... this is usually not a big problem as it is a small
percentage compared to what is exhausted out the back, but
if the case intake is very poor or the PSU exhaust area is
mostly obstructed then the % would start climbing. Overall
I still suspect a combination of poor psu and poor case
intake.


True, at least in the since that the hot air from the PSU
had no place to go , but onto the CPU heat sink, with the
side on. With the side off, there would be a path that allows
it to flow, and even hot air can carry away heat, if it is
moving air.

Yes, taking the case side panel off should improve the
situation but it should already be fairly obvious whether
the case has ample front intake (and perhaps a side fan or
grate) or not. I suppose we should determine for certain if
the PSU fan were really reversed and blowing the wrong
direction, but "think the fan is blowing into the case"
seems fairly ambiguous to me, it should be pretty obvious by
looking at the fan and the blade slope/curvature, which way
it's blowing when spinning even if not powered so the fan
isn't spinning at the moment.

How about: it's a function of the air flow through the
PSU case?

That too, but given a fast enough fan the PSU still
shouldn't be that hot... just too noisey. Then again "hot"
is a bit of a relative perception without a temp reading
number. Even so, it takes a fairly poor setup for the PSU
casing to feel more than barely warm except when it's
completely off in S5 state the 5VSB current generation
without a fan running can tend to make some a little warmer.

If you have a box with competing fans;
where both are trying to blow air in or trying to blow
the air out, you aren't going to have much air flow.

You'd certainly have less, but unless the rest of the case
were very poor and no passive intake (or exhaust in this
case) it should still be providing enough airflow that the
PSU doesn't get terribly hot, unless it's quite poor,
inefficient, or the fans aren't running fast enough.

Note that most OEMs do use this "competing fan" arrangement
on the majority of their PC systems, with both the PSU fan
and the rear case/passive-CPU fan both blowing outward and
they manage to do it even without much noise.

The
best practice, with multiple fans, is to have intake fans in
one corner of the box and the exhaust fans in the opposite
corner, so the air flows through the whole box.

Absolutely, with one exception that the gain by doing this
can be offset by the increased noise emission having a fan
in the front where the sound can escape and be more easily
perceived by the user. Even then, when that intake fan is
forced through the HDD rack it does a much better job of
keeping multiple drives cool.

Since
hot air rises, it is natural to have the intake at or near the
bottom and the exhaust at or near the top.

Hot air rising isn't much of a concern when it's being
actively moved by a couple fans. More significant is what
you already wrote, that the intake (whether it be a fan or
just passive) and exhaust are at opposite *corners* of the
system as much as reasonably possible.
 
K

kony

One thing I just noticed doing some math - how can you get 724W of power
from a PSU rated for 500W ?

3.3V x 35A = 116W
5.0V x 40A = 200W
12.0V x 16A = 192W
12.0V x 18A = 216W
-----------------------
Total: 724W

Am I calculating something wrong ?

Chris

A more reasonable rating for this PSU might be something
like:

3.3V x 25A { combined, comes out of 25A @ 5V current }
5.0V x 25A
12.0V x 20A
==========
365W (plus other rails = ~ 385W @ 25C)
That's just a ballpark guess.

Dynex being generic is likely rating for peak momentary load
at 25C, so we can fairly well ignore the specs except
possibly to see what they suggest the ratio is of 5V:12V
current to see whether the PSU is more suitable for old 5V
centric CPU power or newer (now several years so not /that/
new anymore) 12V centric.

Often with consumer grade PSU having more than one 12V rail,
the manufacturer is suggesting it can output up to the rated
current on one rail or the other but not both - often some
brands will also list a total 12V current rating.

Unfortunately this PSU has unrealistically high 5V AND 12V
current ratings for a generic 500W PSU. They probably used
an old design that is not very suitable for powering a
semi-modern system, split the 12V rail to make it look like
almost double it's true capacity.

Even so, it is not uncommon for good PSU to have a
cumulative true wattage below the summed values on the
sticker. It can be a matter of heat generation, the max
thermal load it can take, or transformer saturation, or
switching transistor current ratings or heatsinking among
other things. IOW, a fairer test for such a PSU is to load
it based on the ratio of 5V:12V implied by the label to the
point where it is outputting 500W (including some load on
the other rails), and unfortunately probably that at 25C
room temp - then see if it can survive that for a few
power/thermal cycles or if it pops.

If it passes that test it has merely survived long enough
not to disqualify itself (if ripple were within spec)
_yet_, but for long term use a more careful scrutinization
of the parts would be prudent, or at least buying from a
manufacturer you trust to have conservatively rated the unit
and used higher quality parts. Often it's a very bad sign
when the manufacturer didn't even put their own name on the
label instead selling relabeled units to some gypsy outfit.
If the gypsy outfit wanted to create high value PSU they
would want the (good) manufacturer reputation on the PSU and
be an importer instead of relabeler.
 
K

kony

In other words, if you really want to know what is going on,
it'll take a bit of test equipment. Slapping new power supplies
in there, over and over again, may not fix it.

While this may be true, some PSU just aren't worth the time
to fix or isolate failure point. Slapping in a better new
PSU might be the answer if the case ventilation isn't
horrible.
 
P

Paul

kony said:
While this may be true, some PSU just aren't worth the time
to fix or isolate failure point. Slapping in a better new
PSU might be the answer if the case ventilation isn't
horrible.

I was referring to an internal short inside a motherboard.
That could be drawing more current than normal from the
supply. Checking consumption, even if only with a Kill-a-Watt,
might give you some idea whether there is more to the problem
than a cheap power supply. I find a report of one blown supply,
and a second one running hot, to be suspicious. It might
not be the fault of the supply.

Paul
 
Ad

Advertisements

S

Skeleton Man

Did it cool down, or were you exaggerating in your original post?
Because 50C isn't that hot.

It's still burning hot the touch (painful to leave your hand on the front of
the PSU casing for more than a few seconds) so maybe my temp probe is fauly
(I tried the same probe on two different meters and it gives the same
reading).
It's only reading 45C at the base of my heatsink too, which again is burning
hot to the touch and the software (lavalys, core temp and every other app I
have tried) reports CPU temp as around 60C.
Still you should check the fans by comparing the air flow from that
PSU against the flow from a known good PSU. Also unplug the AC power
cord from the PSU (important!) and try rotating each fan with
something nonmetallic, like a chop stick or plastic straw, to see if
they spin as easily as the fans in a good PSU.

I checked the fan (turns out there is only one 120MM at the bottom, the back
is just like a mesh with lots of holes) and it's fine. The power supply is
less than 6 months old (but more than 30 days I doubt I could return it
now). It appears I was wrong about it blowing into the case - it just felt
like it (probably from HSF).
Is it possible that a PSU fan is right next to the CPU fan, causing
each one to get very little intake air?

The case fan, PSU fan and CPU fan are all in the same area. 120mm case fan
is mounted top rear of the case, with the power supply fan directly above it
at a 90 degree angle (ie. pointing down into the case). The CPU fan is
mounted at the top rear of the board about 6 inches of clearance between it
and the case/psu fans.

See photo: http://www.guestwho.com/fans.jpg

You can't see the PSU fan in the photo, but it's mounted inside the PSU at
the bottom (not screwed to the outside of the case like some).
You can see the HDD bay at the bottom right of the photo - I have a 120mm
intake fan infront of this.

The case in question is the Silvertone Temjin TJ04 - it didn't come with a
power supply, I was using the 350W supply from my old setup (pentium 4,
single 2.4Ghz) until I blew - that's when I bought the Dynex and apparently
got screwed over large because I paid like $130 before tax and the same
model sells on ebay for $25.

I knew it wasn't a big brand name, but I needed a power supply that day and
all they had was Dynex - no brand name parts (I asked). I live in a small
town so even bestbuy/fs is like a 90 minute drive - the stores here only
carry no-name components.
When PSUs develop bad electrolytic capacitors, the transistors and
diodes connected to them may turn on and off much more slowly and
cause overheating. See www.badcaps.net.

I'm well aware of this. I had a previous power supply fail with bad caps but
it didn't overheat and slowly die - one day it was running fine, the next it
stopped (no sparks, no bang, just silently failed and took out the
transistors and god knows what else - I stopped troubleshooting when I found
the transistors were gone).

Chris
 
K

kony

I was referring to an internal short inside a motherboard.
That could be drawing more current than normal from the
supply. Checking consumption, even if only with a Kill-a-Watt,
might give you some idea whether there is more to the problem
than a cheap power supply. I find a report of one blown supply,
and a second one running hot, to be suspicious. It might
not be the fault of the supply.

Paul

We don't know anything about the old one, but I would tend
to think a short in the motherboard would cause some
instability if bad enough to overtax a reasonably mated PSU.
Probably some melting of the connector block and
discoloration of the effected area, you can't burn off an
extra few dozen watts without some sign of it on the parts.

I wouldn't be surprised of a 3rd PSU blows if it is also
some generic 500W. While the aforementioned system isn't
high end by today's standards it is hungry enough to wipe
out a junk PSU in anywhere from a few hours to a few
months... kinda the reason why they should be avoided. I
suppose my main point is that until a known good quality PSU
is tried, and apparently the poor case ventilation problem
corrected, it would be hard to isolate the variables
further. With the present PSU running so hot it certainly
won't have a good lifespan either way and ought to be
replaced even if there were some other component problem.
 
K

kony

The case fan, PSU fan and CPU fan are all in the same area. 120mm case fan
is mounted top rear of the case, with the power supply fan directly above it
at a 90 degree angle (ie. pointing down into the case). The CPU fan is
mounted at the top rear of the board about 6 inches of clearance between it
and the case/psu fans.

See photo: http://www.guestwho.com/fans.jpg

From the picture it looks a lot like that rear fan is mostly
blocked by the case wall having only some tiny holes
stamped out of it. While it's a lot of work I feel at least
one thing that needs done is the parts pulled out and that
rear fan hole cut out.

If the front is equally obstructed, it too ought to be cut
out. As for the bezel in front, I can't see it in that or
another picture I saw but it wouldn't be surprising if it is
also in need of some modification considering they didn't
put too much thought into the rear exhaust area.
You can't see the PSU fan in the photo, but it's mounted inside the PSU at
the bottom (not screwed to the outside of the case like some).
You can see the HDD bay at the bottom right of the photo - I have a 120mm
intake fan infront of this.

The case in question is the Silvertone Temjin TJ04 - it didn't come with a
power supply, I was using the 350W supply from my old setup (pentium 4,
single 2.4Ghz) until I blew - that's when I bought the Dynex and apparently
got screwed over large because I paid like $130 before tax and the same
model sells on ebay for $25.

Try taking it back, at worst they say no. Tell them it
might be fraud that a system using less than 500W can't run
from this PSU.
I knew it wasn't a big brand name, but I needed a power supply that day and
all they had was Dynex - no brand name parts (I asked). I live in a small
town so even bestbuy/fs is like a 90 minute drive - the stores here only
carry no-name components.

The caps also start becoming lossy from dielectric leakage,
rapidly heating up themselves, the upstream parts too, then
pop.


I'm well aware of this. I had a previous power supply fail with bad caps but
it didn't overheat and slowly die - one day it was running fine, the next it
stopped (no sparks, no bang, just silently failed and took out the
transistors and god knows what else - I stopped troubleshooting when I found
the transistors were gone).

I suspect it's just a junk PSU plus poor case cooling, but
we are ignoring one other possiblity, that Dynex has some
warranty on the PSU... just doesn't do much good to prevent
downtime while it's being RMA'd if it does.
 
Ad

Advertisements

W

w_tom

First, ATX standard originally blew air into chassis so that CPU
heatsink required no fan. When CPU heatsinks got their own fan, then
ATX spec reversed that power supply airflow direction.

Second, does not matter whether power supply is a big name or no
name. What matters if it contains essential and required functions.
If it did not come with a long list of numeric specs in writing (major
name or no name), then suspect those functions are missing. When a
supply is misssing essential functions, the manufacturer forgets to
claim any functions exist - in writing.

Third, a computer can boot and the power supply is defective. Many
make this major mistake; assume the power supply is OK because a
computer booted. Good chance that the other supply was defective when
computer booted. Then failure got worse so that computer eventually
stopped working. Failure always existed.

Failures with a bang - rare. Failures with any visual indication -
extremely rare. Most all failures require simple tools such a
multimeter to be observed.

Fourth, if supply is hot, then you have burned skin. What the hot
feels as hot is perectly ideal temperatures to ICs. Do you have
second or third degree burns? If not, then supply was not hot; only
warm. However, to say more (to provide a useful reply), significant
other design details are necessary.

Fifth, see that 120mm fan on picture left side? Remove it. You
should discover only trivial chassis temperature difference assuming
the power supply fan is working properly. To work, that left side fan
must blow into chassis; same as power supply. If blowing out, then
airflow is not across the chassis - bad. And if blowing in, then
holes on other side of chassis must be large enough for airflow from
both fans.

Smarter location for that fan is someplace farthest from power
supply so that air incoming from power supply is outgoing via that
other fan; air travels across the entire chassis.
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Similar Threads

Case Fan Questions 6
Athlon 64 X2 5200+ Temps 12
(weird) Power Supply failure, reason? 4
PSU 1
CPU and MB temprature too high?? 2
Removing PSU fan? 11
sucks or blows revisited 11
low power from 12v rail 3

Top