Twittering sound!

  • Thread starter Thread starter Graeme
  • Start date Start date
G

Graeme

Yes, I know, what a silly subject. But let me explain.
My company's IT dept has just built me two PCs. One for work, the other for
home (where I sometimes work also). They are both pretty much the same spec;

Case + 400W PSU (unbranded as far as I can tell)
Gigabyte 8IPE1000 mobo.
3GHz HT P4
2 x 512M RAM
10/100 Ethernet
120GB 7200 HDD
Floppy
Old 16x DVD ROM
Win XP

Then; work's PC has an Geforce MX440, and the home PC has an FX5700Ultra and
a CDRW drive.

Since these machines have been built, I have mainly used the home PC. And
this is what I find:

During normal (word, excel, etc) use, the CPU core temp just reaches about
30C. During heavy use, this reaches 46C. The CPU fan speed is about 3300rpm.
The FX5700Ultra runs at about 41C when idle, and upto 68C during heavier use
(QuakeIII, Dawn demo, etc). These temp's I'd assume are OK. There is also a
case fan (blowing out of the case). The standard perforated metalwork of the
case (where the fan mounts) seemed very restrictive to me, so I've removed
that which increases the air-flow substantially.

Now to the twittering. This machine makes a high-pitched (10-15kHz???)
whistle from somewhere. This type of noise is proving impossible to trace,
but is also annoying. I've tried touching various 'magnetic' components on
the mobo, and video cards, but none of these would appear to be the cause.
But the weirdest thing is;
1) Just scrolling through a pdf I noticed that the PC would emit a little
chirp (not through the speakers) when changing page. Not always, just
sometimes. But always on the same page boundary.
2) Running excel and dragging a SUM box made a very interesting twittering
noise (like above, but repeating) when the mouse button was held down and
the SUM box stretched. Releasing the button stopped the noise.

So, what's the casue? Is it normal that the various switching power supplies
in this PC should make this noise during load changes? Is the 400W PSU not
adequate? Do you think I'm stupid ;-) ?

Oh, and MEMTEST86 runs ok for 24 hours. The system is otherwise stable,
although sometimes on start-up I get a Microsoft message appologising for
Wndows Explorer encountering a problem and having to close, and then the
nVidia notification area icon does not appear. The nVidia driver was the
latest as of one week ago.

Well, thanks for reading (if you've stayed the course and got this far), and
thanks for any comments. And of course, happy holidays (if that's part of
your religion).
 
Graeme said:
Yes, I know, what a silly subject. But let me explain.
My company's IT dept has just built me two PCs. One for work, the other for
home (where I sometimes work also). They are both pretty much the same spec;

Case + 400W PSU (unbranded as far as I can tell)
Gigabyte 8IPE1000 mobo.
3GHz HT P4
2 x 512M RAM
10/100 Ethernet
120GB 7200 HDD
Floppy
Old 16x DVD ROM
Win XP

Then; work's PC has an Geforce MX440, and the home PC has an FX5700Ultra and
a CDRW drive.

Since these machines have been built, I have mainly used the home PC. And
this is what I find:

During normal (word, excel, etc) use, the CPU core temp just reaches about
30C. During heavy use, this reaches 46C. The CPU fan speed is about 3300rpm.
The FX5700Ultra runs at about 41C when idle, and upto 68C during heavier use
(QuakeIII, Dawn demo, etc). These temp's I'd assume are OK. There is also a
case fan (blowing out of the case). The standard perforated metalwork of the
case (where the fan mounts) seemed very restrictive to me, so I've removed
that which increases the air-flow substantially.

Now to the twittering. This machine makes a high-pitched (10-15kHz???)
whistle from somewhere. This type of noise is proving impossible to trace,
but is also annoying. I've tried touching various 'magnetic' components on
the mobo, and video cards, but none of these would appear to be the cause.
But the weirdest thing is;
1) Just scrolling through a pdf I noticed that the PC would emit a little
chirp (not through the speakers) when changing page. Not always, just
sometimes. But always on the same page boundary.
2) Running excel and dragging a SUM box made a very interesting twittering
noise (like above, but repeating) when the mouse button was held down and
the SUM box stretched. Releasing the button stopped the noise.

So, what's the casue? Is it normal that the various switching power supplies
in this PC should make this noise during load changes? Is the 400W PSU not
adequate? Do you think I'm stupid ;-) ?
<snipped>

You better have them re-check the machine...the cpu temp seems awfully
high!
Also, to trace down noises...as silly as it sounds all you need to do is
put a rubber tube in your ear and use the other end to snoop...it really
works~
 
loads more snipped
You better have them re-check the machine...the cpu temp seems awfully
high!
Also, to trace down noises...as silly as it sounds all you need to do is
put a rubber tube in your ear and use the other end to snoop...it really
works~

I've heard the trick with the rubber tube, I just need to go and get one.

What sort of CPU temperature would be 'normal'? 30C didn't seem high to me.
 
kony said:
It IS normal. CPU should be fine at least up to 60C.


Dave

Thanks for the feedback. I can't seem to get the CPU above 47C running
everything I have.

I'm more concerned about the FX5700Ultra core temp being 71C (max reported).
But the nVidia utils have a slowdown core temp set at 129C, so it's still
got a lot more to go.

I have noted the PSU model is marked as 'Winpower 400W'. I can see these
units for sale at about £12, but can also see 400W supplies for £30+. Do you
know if Winpower is good/bad? Would this unit be capable of supplying my
system? (3GHz P4, 1G RAM, 1 HDD, 1 DVD, 1 CDRW, FX5700Ultra, 10/100 and good
ol' floppy)
 
I did mention 68C, but that was in relation to the FX5700Ultra core
temperature.

Higher-end video cards do generally run hotter than the CPU's they're
paired with in a good system, is simply a matter of less airflow and
room, weight limitations on the video card 'sink. It might help to
leave off the case PCI slot cover under the card, and as many PCI
slots empty below it as possible.
 
Thanks for the feedback. I can't seem to get the CPU above 47C running
everything I have.

I'm more concerned about the FX5700Ultra core temp being 71C (max reported).
But the nVidia utils have a slowdown core temp set at 129C, so it's still
got a lot more to go.

That upper-end temp would usually be to prevent immediate damage, but
running long-term below that would likely cause problems... If the
heatsink fan is the typical cheap sleeve-bearing type you might want
to put a drop of thick oil (like gear oil) or very thin grease on it
BEFORE it fails, as a preventative measure... once the fan starts
making noise, wearing out a dry bearing, it's never quite as good as
if it'd stayed lubed all along.

In the other post I mentioned removing the case slot cover, it can
help a lot when the case has inbalanced airflow, and most do, with
more exhaust than intake. In a case with balanced flow it may help to
put a fan on the case side-panel or just fab a simple bracket to hang
the fan from the slot-cover screws.

I have noted the PSU model is marked as 'Winpower 400W'. I can see these
units for sale at about £12, but can also see 400W supplies for £30+. Do you
know if Winpower is good/bad? Would this unit be capable of supplying my
system? (3GHz P4, 1G RAM, 1 HDD, 1 DVD, 1 CDRW, FX5700Ultra, 10/100 and good
ol' floppy)

I vaguely recall those being Leadman, not too good. I may be wrong
though, but the generic label and low price would also suggest it's a
poor power supply, and considering the rest of the parts in your
system it's out of place in the system... I'd replace it, keep it as a
spare for emergencies only. If it has sleeve-bearing fans they, like
the video card fan, ought to be lubed before they start failing.

Dave
 
kony said:
That upper-end temp would usually be to prevent immediate damage, but
running long-term below that would likely cause problems... If the
heatsink fan is the typical cheap sleeve-bearing type you might want
to put a drop of thick oil (like gear oil) or very thin grease on it
BEFORE it fails, as a preventative measure... once the fan starts
making noise, wearing out a dry bearing, it's never quite as good as
if it'd stayed lubed all along.

In the other post I mentioned removing the case slot cover, it can
help a lot when the case has inbalanced airflow, and most do, with
more exhaust than intake. In a case with balanced flow it may help to
put a fan on the case side-panel or just fab a simple bracket to hang
the fan from the slot-cover screws.



I vaguely recall those being Leadman, not too good. I may be wrong
though, but the generic label and low price would also suggest it's a
poor power supply, and considering the rest of the parts in your
system it's out of place in the system... I'd replace it, keep it as a
spare for emergencies only. If it has sleeve-bearing fans they, like
the video card fan, ought to be lubed before they start failing.

Dave

Thanks for the comments.
The FX5700 is brand new (10 days), but I'll see if it's possible to lube it
up now.
The rear case fan seems to work well- fixed at 2000rpm (not too loud).

The original point of my posting was to find out about the twittering noise
during some simple operations. Well, using the rubber tube trick, it is
almost certainly coming from either the P4 or the on-board local power
supply for the CPU (you know, that series of toroidal inductors and caps).
It is not from the system PSU, or the FX5700.

Back to the PSU, what would be a good make/model for this system?
Thanks again.
 
loads more snipped


I've heard the trick with the rubber tube, I just need to go and get one.
I use 2' vacuum hose when I look for engine noises. Any auto supply
will cuy a piece for you. Should be <$1 IIRC.
 
Thanks for the comments.
The FX5700 is brand new (10 days), but I'll see if it's possible to lube it
up now.

Just remember that thick lube is better than thin...it'll more likely
stay where it's supposed to, and it will get thinner sitting on a hot
heatsink.
The rear case fan seems to work well- fixed at 2000rpm (not too loud).

The original point of my posting was to find out about the twittering noise
during some simple operations. Well, using the rubber tube trick, it is
almost certainly coming from either the P4 or the on-board local power
supply for the CPU (you know, that series of toroidal inductors and caps).
It is not from the system PSU, or the FX5700.

Most often I've heard high-pitched sounds from 2 areas, loose
transformers and damaged ball-bearings... Non damaged bearing would
make a more consitent sound but when one or more balls has a flat
spot, very minor temp and RPM change seems to get it moving, or stuck,
and intermittent noise results. The power supply inductors are likely
far too thick and sparsely wound to have a similar resonant effect as
a loose transformer, though there isn't a whole lot left that I'd
expect to make such a noise in that area... it's "possible" that a
capacitor was installed backwards and makes such a high-pitched noise,
I've seen it happen, but it's a rare thing and pretty easy to spot.
I'd always supposed that one backwards would explode but apparently
if installed that way new, it'll reform itself if the voltage is low
enough, work a bit instead of exploding, but never be quite properly
functioning.

It's all low-voltage though, if you can't pinpoint the noise any other
way then it's time for the good 'ole touch-testing.

If you're using onboard sound it could possibly be the power supply...
not usually but it can happen, voltage fluctuations can even make
chips produce noise, or so it seemed. Underclocking the CPU as much
as possible "might" help, but less likley since the CPU is using 12V
and not the other non-related motherboard circuitry.

Back to the PSU, what would be a good make/model for this system?
Thanks again.

Beyond the 3GHz CPU the system isn't "yet" particularly demanding...
will be using a fair bit of power from both 12V and 5V->3V rails, a
fairly balanced load to a certain extent, so at the present time any
quality 300W should work, but these days I'd recommend buying a bit
more margin for the future, and having excess capacity to allow
running under peak capacity generally allows longer lifespan. Several
makes are good, it might depend on what you find a deal on. A few
would be PC Power & Cooling, Delta, Sparkle/Fortron, Antec, and
others. If it were my system I'd get a minmum of 350W, probably 400W
just to be sure it'd handle any kind of upgrade in the future.
Sparkle is usually the best value but in the 400W & up range they have
only rear air intake so for most cases the Antecs in 380W+ might be
ideal, but again you really need a decent 300W, the rest is breathing
room, but with higher wattages also you often get a fan on the bottom
which also helps remove CPU heat instead of having more of it
recirculate. Generally I'd avoid anything with less than 200W
combined 3V+5V rating and be sure the rear exhaust fan uses dual
ball-bearings, bu then most (otherwise) good power supplies do.


Dave
 
kony said:
Most often I've heard high-pitched sounds from 2 areas, loose
transformers and damaged ball-bearings... Non damaged bearing would
make a more consitent sound but when one or more balls has a flat
spot, very minor temp and RPM change seems to get it moving, or stuck,
and intermittent noise results. The power supply inductors are likely
far too thick and sparsely wound to have a similar resonant effect as
a loose transformer, though there isn't a whole lot left that I'd
expect to make such a noise in that area... it's "possible" that a
capacitor was installed backwards and makes such a high-pitched noise,
I've seen it happen, but it's a rare thing and pretty easy to spot.
I'd always supposed that one backwards would explode but apparently
if installed that way new, it'll reform itself if the voltage is low
enough, work a bit instead of exploding, but never be quite properly
functioning.

It's all low-voltage though, if you can't pinpoint the noise any other
way then it's time for the good 'ole touch-testing.

If you're using onboard sound it could possibly be the power supply...
not usually but it can happen, voltage fluctuations can even make
chips produce noise, or so it seemed. Underclocking the CPU as much
as possible "might" help, but less likley since the CPU is using 12V
and not the other non-related motherboard circuitry.

Thanks again.
I have two near identical systems, and they both make this noise. I'll try
underclocking to see what happens.
 
Back
Top