On 13 Nov 2004 13:00:51 GMT, Arno Wagner <me (AT) privacy (DOT) net>
wrote:


Well there's labeled capacity then there's actual capacity.
We can pretty much ignore the label as percentages can't be
directly applied to that at all except with an known entity,
a specific power supply. Instead looking at actual
sustainable capacity, the PSU will run indefinitely. When
looking at peak capacity, there will be either an overheat
limit, an under (or over) voltage limit, or on a really
nasty power supply, extreme ripple (worse situation,
eventually killing parts). ON a decent PSU you probably can
expect surge 120% of capacity, but on a cheap PSU you might
not be able to expect even 80% sustained capacity.


Enermax is a bad example, their sustained rating for MTBF
is only 70% of the labeled wattage. That makes their $100
460W PSU inferior to even a $4 Sparkle 350W. Power supply
reviews from the likes of Tom's Hardware have also revealed
their actual capacity to be far lower than the Enermax
labeled capacity in testing.

In article <ON4ld.78715$HA.24737@attbi_s01>, Clyde
<lughclyde (AT) attbi (DOT) comedy> writes
--
Roger Hunt

With the case side off & power off
o Carefully locate the CPU fan
o Avoiding SATA/other cables, carefully press the centre of the fan rotor
o Verify you can do this during the initial noise phase you report

Now repeat this with the power on & the PC making the noise.

Slowing the fan down with your finger may negate the noise, in which case
you have identified the cause. Don't touch the blades, they can break.

You might be able to perform the same test without touching the fan,
but the power on & the side off the case - allowing better noise location.


Inductor whine tends to be a whine:
o PSUs can suffer it - coils/inductors vary in quality
---- switched-mode PSUs "chop-up" voltage & harmonics can be audible
o Motherboards can suffer it - coils/inductors on the board itself
---- some Intel boards, P-Pro & P2, I recall used to suffer this
o The whine tends to be high-pitched - like a TV flyback transformer.

Regular noise can be caused by a cable touching a fan:
o Axial fans do not produce much suction, but a cable can get brushed

Induction sucking noise can be caused by obstruction to a fan inlet:
o Axial fans are susceptible to any intake obstruction within about 1"
---- one reason why round-wire grills are preferable re noise on intake fans
o Outlet is suspceptible to a lesser extent
---- so in both cases ensure airflow paths are relatively clear

Buzzing suggests resonance:
o It could be an out of balance fan on the heatsink
---- verify fan fitment on the CPU heatsink
---- perhaps perform the test I suggest to help identify the cause
o The fan need not necessarily be out of balance
---- the heatsink/motherboard/case have a particular natural frequency
---- if a harmonic of the fan rpm hits that natural frequency it causes excitation
---- excitation is typified by *buzzing resonance*
o Verify the heatsink is correctly seated, the fan isn't damaged
o Verify the motherboard is correctly secured, and items in the case are secure

Resonance is particularly possible with "clip-in case fan brackets":
o These typically clip into a case but without particularly sound retention
o Fans will excite this clip-in fitting, often into resonating against the case
o Again, touching the clip with the side off would help identify noise

Resonance can be caused if the case has low-airflow punched grills:
o Typically a punched case grill produces 85% of the airflow resistance in a PC
o Airflow through such grills can be reduced to 45-55% of no-grill
o Round wire-grills achieve 89-91% of the airflow of a no-grill condition
o Round wire-grills also minimise noise re no sharp edges & reduce resonance

Overtightening securing screws can be as much a problem as undertightening.
Some motherboards spin their fans to maximum speed during start-up, and then
the temperature control profile (or BIOS settings) take over & regulate fan speed.

Points the finger towards a resonance issue during the high speed fan period.
Watch fingers, but should be relatively easy to pin-down and fix :-)
--
Dorothy Bradbury
www.dorothybradbury.co.uk for quiet Panaflo & NMB fans

In comp.sys.ibm.pc.hardware.misc kony <spam (AT) spam (DOT) com> wrote:
[...]

Looks a bit different in the c't tests. And MTBF at 70% is not
uncommon in quality power electronics. They are just being
honest, as opposed to other PSU manufacturers.

The $4 sparkle is more likely to have 250W actual capacity and
to blow up at 300W (happened to c't with several cheap PSUs
before the maximum load was reached). Enermax will just
switch off, as any quality PSU will do.

Do you have a reference to a test of Enermax PSUs that shows
this effect? So far I only saw them comming out among the best.

Arno
--
For email address: lastname AT tik DOT ee DOT ethz DOT ch
GnuPG: ID:1E25338F FP:0C30 5782 9D93 F785 E79C 0296 797F 6B50 1E25 338F
"The more corrupt the state, the more numerous the laws" - Tacitus

In comp.sys.ibm.pc.hardware.misc David Maynard <dNOTmayn (AT) ev1 (DOT) net> wrote:
Yes, this is a perfectly correct labeled 300W (400W peak) PSU.
What do you not like about this? Since HDD spin-up generates much
higher load, doing this type of design is appropriate in
PC PSU design. Or did they claim 400W _sustained_ on the
box?

Arno
--
For email address: lastname AT tik DOT ee DOT ethz DOT ch
GnuPG: ID:1E25338F FP:0C30 5782 9D93 F785 E79C 0296 797F 6B50 1E25 338F
"The more corrupt the state, the more numerous the laws" - Tacitus

"Roger Hunt" wrote:
And even if they won't allow it be lent, they may often allow it to be bent.

Jon

Arno Wagner wrote:

Where did you get the ideal it's 'labeled' a 300 Watt supply? It's labeled
and sold as a "Q-TEC 400W Dual Fan Gold."

Putting on the sticker that +5V is 30 Amps when it's capability is 25 Amps,
or that 3.3V is 20A instead of 14A, or that 12V is 16A instead of 12A is
not "perfectly correct labeled."

I'm going by what a purchaser said he got from Q-Tec when he emailed them
because I can't, for the life of me, find a single amp rating on their
entire web site.

http://www.qtec.info/products/produc...=13023&specs=1

'Peak' without a time interval means nothing. Tell me it's 400 watts for 30
seconds and maybe I'll buy that's for spin up time but just "400 watts
peak," if they had bothered to even mention 'peak', could be for 5
nanoseconds for all anyone knows. Btw, what is it that's supposedly being
'spun up' on the 3.3V rail so that it's labeled 20 Amp when it's really 14?

It's simply ka-ka to label and sell a PSU by a meaningless 'peak' rating.

On 14 Nov 2004 04:50:19 GMT, Arno Wagner <me (AT) privacy (DOT) net>
wrote:

Not "just being honest" compared to all though.


A Typo, should've read "$45".

Do you have reference that shows them outputting their rated
wattage, sustained? If they can't sustain their rated
wattage isn't it a bit of a fraud?

On 14 Nov 2004 04:52:27 GMT, Arno Wagner <me (AT) privacy (DOT) net>
wrote:


No, it's NOT even correctly labeled for a 300W PSU. 300W
PSU typically has 180-200W combined 3V+5V capacity, not
150W. Decent 300W PSU can sustain more than 150W but that
"400W" PSU, apparently can't.

kony wrote:

I think you're misinterpreting the MTBF number being at 70% load. That does
not mean it can't sustain 100%; they've just the load point where they did
the MTBF rating, is all.