Quiet mid-tower case supporting Corsair H105 cooler

Bill wrote:
Paul wrote:
Bill wrote:

My personal preference is for fixed fan speed, even if a little
noisy. I tend to forget about it after a while. When a fan
wanders around speed wise, it tends to attract unwanted
attention.

If the PWM signal is not connected to anything (it is left
floating), the fan would run at 100% speed. So if you were
to connect the four pin fan to a three pin header, the fourth
signal which is PWM, hangs in the air and is not connected. And
that's one way to get 100% fixed fan speed.
On my current Gigabyte GA-P55A-UD4P motherboard, I ended up
having to disable the "CPU Smart Fan Control", because otherwise the
audio was detecting my mouse movement and making very annoying noises
(sort of "following me around").
I assume that this feature is the same as PWM. If there is anyone
here who could
possibly explain that phenonema, I expect it is you or Flasherly! If
it happens again,
then PWM is just a waste of money... FWIW, I have a Logitech corded
mouse, but, IIRC, I tried another mouse too.

Bill


PWM (Intel style) operates at 25KHz. The pulse width of
the square wave signal varies, as an indicator of how fast
the fan is to go.

The 25KHz was selected specifically tn not be in the audio band.
While it is possible to alias an out of band signal down into
the audible range, you would need a suitable signal to do that.

Similarly, the CCFL tubes on your LCD monitor, run at 25KHz. And
that's another attempt to prevent things like coil noise from
an inverter, being heard by the user.

I've not read any posts, tracing noise to PWM. But of course
stranger things have happened.

*******

Motherboard audio doesn't have a particularly clean noise floor.
The quickest fix, is reach for a $10 PCI sound card and use that
in a PCI slot. Alternately, you can also get a USB dongle that
looks like a flash stick, with a couple 1/8" connectors on the
bottom, and the green connector on that is Line Out. That's another
way to potentially replace motherboard audio.

If the motherboard audio (HDAudio or AC'97) runs in parallel with
NIC drive wires (to the motherboard side of the Ethernet
isolation transformer0, the audio can pick that up. That gives
noises in response to packets on the NIC.

The motherboard audio must be powered with something. If it
were to run off +5VSB say, and the mouse runs off +5VSB, perhaps
the sound of the mouse leaks through that way. You may notice
a three terminal regulator on the motherboard, next to the
HDAudio chip, which is an attempt to provide "clean" unshared
power to the HDAudio chip. That might take +12V and make
+5V for the audio. No other circuit should use that power.

Some motherboards now, have decent audio performance. Older
motherboards, audio was an afterthought. It was the appearance
of RightMark, and audio testing, that perhaps alerted motherboard
makers to the need to clean up their act (bad PR). Now, you can find
pretty strange attempts at shielding. Some of which are mainly
for show (random strips of metal with branding printed on them).
Some of the effective solutions, are just layout changes in the
PCB, to route audio signals away from noise sources. But bulk
shielding has been known to buy 10dB in noise floor, so there
is some evidence that such helps. My current audio solution
(came with motherboard), is in a shielded box. Presumably so
it wouldn't look like an unshielded solution.

Another mechanism you might look into, is the CDROM audio cable.
Some people still connect the CDROM audio cable to the white
four pin header on their motherboard. The cable has four wires,
and the intent was for differential transmission and reception.
But on the motherboard end, the reception might be single ended,
and do a poor job at common mode noise removal.

So for that one, there are two possibilities. You listen to a
CD, and pick up all manner of "digital" beeps and boops in
the background. Removing the audio cable, and switching to
digital audio extraction (packet audio from the CD drive),
eliminated that as a source.

But even if you weren't listening to a CD, that white connector
could pick up noise. On the "Recording" audio interface,
remember to "mute" all recording inputs you are not using. For
example, any path with sensitive input (microphone in), might be
a source of noise pickup. Right now, I'm using the
Line In (1V RMS levels) from an external source. Which
doesn't amplify much. And my background audio is quiet.
My audio likely doesn't have -110dB performance figure,
but at least it isn't beeping and booping while I work.
Everything but Line In is muted.

HTH,
Paul
Your post was very interesting. However recall that disabling "CPU
Smart Fan Control"
completely eliminated the problem (I got the solution from an online
forum). Does my problem suggest that the circuit which helped provide PWM
created some sort of interfering electric field on the board, somehow
allowing the audio to detect mouse action events?
Obviously, I am not an EE but I am still interested in how the problem
manifested itself.

Bill

As a collector of trivia, I'm interested too :-)

The PWM signal itself is inaudible. If I coupled that
signal into the audio, now maybe the audio doesn't have a sharp
enough prefilter. Maybe what you're hearing is 22.050KHz sampling
minus 25KHz signal. The difference is 2.9KHz or so (high pitched).

The fan motor probably has some current ripple from the
brushless DC commutation. Would that be the noise ?

You would start with capturing a sample with Audacity,
and Audacity has an FFT module. You could look for
the frequencies involved in there. And maybe come up
with a theory as to where it comes from.

Paul

Paul wrote:
Bill wrote:
Paul wrote:
Bill wrote:

My personal preference is for fixed fan speed, even if a little
noisy. I tend to forget about it after a while. When a fan
wanders around speed wise, it tends to attract unwanted
attention.

If the PWM signal is not connected to anything (it is left
floating), the fan would run at 100% speed. So if you were
to connect the four pin fan to a three pin header, the fourth
signal which is PWM, hangs in the air and is not connected. And
that's one way to get 100% fixed fan speed.
On my current Gigabyte GA-P55A-UD4P motherboard, I ended up
having to disable the "CPU Smart Fan Control", because otherwise the
audio was detecting my mouse movement and making very annoying
noises (sort of "following me around").
I assume that this feature is the same as PWM. If there is anyone
here who could
possibly explain that phenonema, I expect it is you or Flasherly!
If it happens again,
then PWM is just a waste of money... FWIW, I have a Logitech
corded mouse, but, IIRC, I tried another mouse too.

Bill


PWM (Intel style) operates at 25KHz. The pulse width of
the square wave signal varies, as an indicator of how fast
the fan is to go.

The 25KHz was selected specifically tn not be in the audio band.
While it is possible to alias an out of band signal down into
the audible range, you would need a suitable signal to do that.

Similarly, the CCFL tubes on your LCD monitor, run at 25KHz. And
that's another attempt to prevent things like coil noise from
an inverter, being heard by the user.

I've not read any posts, tracing noise to PWM. But of course
stranger things have happened.

*******

Motherboard audio doesn't have a particularly clean noise floor.
The quickest fix, is reach for a $10 PCI sound card and use that
in a PCI slot. Alternately, you can also get a USB dongle that
looks like a flash stick, with a couple 1/8" connectors on the
bottom, and the green connector on that is Line Out. That's another
way to potentially replace motherboard audio.

If the motherboard audio (HDAudio or AC'97) runs in parallel with
NIC drive wires (to the motherboard side of the Ethernet
isolation transformer0, the audio can pick that up. That gives
noises in response to packets on the NIC.

The motherboard audio must be powered with something. If it
were to run off +5VSB say, and the mouse runs off +5VSB, perhaps
the sound of the mouse leaks through that way. You may notice
a three terminal regulator on the motherboard, next to the
HDAudio chip, which is an attempt to provide "clean" unshared
power to the HDAudio chip. That might take +12V and make
+5V for the audio. No other circuit should use that power.

Some motherboards now, have decent audio performance. Older
motherboards, audio was an afterthought. It was the appearance
of RightMark, and audio testing, that perhaps alerted motherboard
makers to the need to clean up their act (bad PR). Now, you can find
pretty strange attempts at shielding. Some of which are mainly
for show (random strips of metal with branding printed on them).
Some of the effective solutions, are just layout changes in the
PCB, to route audio signals away from noise sources. But bulk
shielding has been known to buy 10dB in noise floor, so there
is some evidence that such helps. My current audio solution
(came with motherboard), is in a shielded box. Presumably so
it wouldn't look like an unshielded solution.

Another mechanism you might look into, is the CDROM audio cable.
Some people still connect the CDROM audio cable to the white
four pin header on their motherboard. The cable has four wires,
and the intent was for differential transmission and reception.
But on the motherboard end, the reception might be single ended,
and do a poor job at common mode noise removal.

So for that one, there are two possibilities. You listen to a
CD, and pick up all manner of "digital" beeps and boops in
the background. Removing the audio cable, and switching to
digital audio extraction (packet audio from the CD drive),
eliminated that as a source.

But even if you weren't listening to a CD, that white connector
could pick up noise. On the "Recording" audio interface,
remember to "mute" all recording inputs you are not using. For
example, any path with sensitive input (microphone in), might be
a source of noise pickup. Right now, I'm using the
Line In (1V RMS levels) from an external source. Which
doesn't amplify much. And my background audio is quiet.
My audio likely doesn't have -110dB performance figure,
but at least it isn't beeping and booping while I work.
Everything but Line In is muted.

HTH,
Paul
Your post was very interesting. However recall that disabling "CPU
Smart Fan Control"
completely eliminated the problem (I got the solution from an online
forum). Does my problem suggest that the circuit which helped
provide PWM
created some sort of interfering electric field on the board, somehow
allowing the audio to detect mouse action events?
Obviously, I am not an EE but I am still interested in how the
problem manifested itself.

Bill

As a collector of trivia, I'm interested too :-)

The PWM signal itself is inaudible. If I coupled that
signal into the audio, now maybe the audio doesn't have a sharp
enough prefilter. Maybe what you're hearing is 22.050KHz sampling
minus 25KHz signal. The difference is 2.9KHz or so (high pitched).

The fan motor probably has some current ripple from the
brushless DC commutation. Would that be the noise ?
The noise was from the "rolling" of the mouse. No move, no noise.
No Smart Fan Control, no noise too. Go figure.

Go on to more important work! I'll buy some new computer components and
bring you some real questions!

Thanks,
Bill



You would start with capturing a sample with Audacity,
and Audacity has an FFT module. You could look for
the frequencies involved in there. And maybe come up
with a theory as to where it comes from.

Paul

Bill wrote:
Paul wrote:

snip

Go on to more important work! I'll buy some new computer components
and bring you some real questions!

Thanks,
Bill



Paul

Okay, I'm ready now! : )

Am considering Consair 650D case with H105 cooler.

This case comes with 200mm fan directing air IN on the FRONT and another
200mm fan directing air OUT on the TOP,
and a 60mm fan directing air out the back.

If installed as directed (in the directions), the H105 cooler, with its
two 60mm fans would replace the fan on top.
Furthermore, they would be directing air IN to the case. My intuition
suggests installing these fans so they direct
air OUT the top of the case.

2 questions:

1. What do you think of this configuration?

2. What does anyone care about "positive air pressure in the case"? I
don't understand the significance of this.

Thank you,
Bill

On Sun, 30 Nov 2014 06:12:01 -0500, Bill
wrote:

2. What does anyone care about "positive air pressure in the case"? I
don't understand the significance of this.

What it implies, (to me w/out hunting it down), is you're dealing with
powerful enough fans to overcome a design factor of the case: either
a) too much air, or b) not enough air is being circulated, as based on
the fan's rating for cubic/ft of air moved (insideoutside ||
outsideinside);- It's irrespective of existing thermal conditions
inside the case, as well, and will continue to be until such
condition(s a&b) is brought into balance, whereupon the case is
"tuned" for efficiency (to the fans' rated measure for air movement)
and its inside temperature may be measured for some sort of
coefficient common to cases. Although I've doubts such a rating
practically should exist;- I see no reason to as well derive
(extrapolate for) a rating for a measure cases and their filtering
efficiency, in part based on fans, and how filthy and quickly they
turn the case and all inside fuzzy brown and black from airborne
contaminants.

Bill wrote:
Bill wrote:
Paul wrote:

snip

Go on to more important work! I'll buy some new computer components
and bring you some real questions!

Thanks,
Bill



Paul

Okay, I'm ready now! : )

Am considering Consair 650D case with H105 cooler.

This case comes with 200mm fan directing air IN on the FRONT and another
200mm fan directing air OUT on the TOP,
and a 60mm fan directing air out the back.

If installed as directed (in the directions), the H105 cooler, with its
two 60mm fans would replace the fan on top.
Furthermore, they would be directing air IN to the case. My intuition
suggests installing these fans so they direct
air OUT the top of the case.

2 questions:

1. What do you think of this configuration?

2. What does anyone care about "positive air pressure in the case"? I
don't understand the significance of this.

Thank you,
Bill

Airflow is a lively topic.

http://www.overclock.net/t/1062965/p...e-or-less-dust

I have positive pressure on my current case. It's still dirty in
there, because I absolutely refuse to use filters. Filters require
attention, and the last thing I need it "continuous maintenance".

And the new system, with the monster fan on the back, it's too early
to tell how bad it'll be (negative pressure). It should get pretty dirty,
but the old motherboard in there wasn't that bad.

As the thread above mentions, some sort of filter would help, in terms
of the quality of air feeding the case. But if you "do it right",
say a hepafilter like the first one I bought for my new gas
furnace, there will be high arrestance, and you'll need a good
fan to drive it, and then you're going to have a "noisy vacuum cleaner
case design". I choose to accept a bit of dust, and use Medium speed
fans (somewhere around 35CFM a piece). Yes, my case is still noise,
it's still slightly dirty, but it's not super noisy + super clean.
That would drive me nuts (the super noisy part).

*******

The purpose of the instructions saying to arrange the fans
on the H105 to blow *into* the case, that's an attempt to
give the water cooling system impressive statistics. If you
use room air, the inlet side of the H105 is 25C. The water
experiences better cooling. Now, if you grabbed case air
and blew it out of the case, the inlet air to the H105
would be 35C. And the water could never be cooled below
35C, and Corsair "wouldn't look as good". So they only
specified such an approach, for their own interests.

You put cooling where it is needed. You use enough cooling
for the hard drives, so they stay under 50C. My current
two drives are running at 31C and 25C.

You remove enough heat from the CPU, so it doesn't throttle.

You remove enough heat from the GPU, so it doesn't hit 105C
or whatever the top temperature is. There have been some
gamer cards, that would run at 90C, even with the fan on
the card running at a noisy speed. In such an event, you'd
want to "sweep warm air" from around the card, and that means
adjusting the case fan configuration to achieve that objective.
I've had one computer case, where there was a pretty bad
cloud of warm air right around the CPU cooler, and it was
pretty hard to move that out without a lot of noise.

*******

OK, if I aim for a slightly positive air pressure,
it looks like this. That's an attempt to balance
the 200 with the three other fans. (I didn't check
CFM numbers, instead using pie R squared.)

^ ^
| |
+------- 120 120 -----------------+
H105
-60

|
| Video
| HDD
200-
- PSU HDD

+-------------------------------------+

If instead, I arranged them the following way, now I've
got way too much positive pressure. The 120's and the
200 are working against backpressure. I would need to
remove more slot plates in the back, and 5.25" tray
covers in the front, to let the air out. I expect
with this setup, more air will leave via the front
than the back, and the lower left corner will be
"warm cloud" country. The video card temps might be
a bit higher.

| |
v v
+------- 120 120 -----------------+
H105
-60 Remove --
Covers --
| Here --
| Video
| HDD
200-
- PSU HDD

+-------------------------------------+

I'm going to have to go with the first config. It's going to
make the H105 a bit warmer. But the lower diagram, the
air pattern looks like it could be a bit more stagnant.
I really really like a defined airflow (in on one side,
out on the other). It's hard to predict how much
worse the bottom diagram will be. The H105 will be
heroic, and the CPU will be at 26-27C when idle, because
room air at 25C will be hitting the radiator. I'm willing
to trade a little CPU temp, at least until it hits
the throttle point. If the CPU starts to throttle,
then I have to switch to some other configuration.

So yeah, there are some compromises involved. And
some test cases to run.

Paul

Paul wrote:
Bill wrote:
Bill wrote:
Paul wrote:

snip

Go on to more important work! I'll buy some new computer components
and bring you some real questions!

Thanks,
Bill



Paul

Okay, I'm ready now! : )

Am considering Consair 650D case with H105 cooler.

This case comes with 200mm fan directing air IN on the FRONT and
another 200mm fan directing air OUT on the TOP,
and a 60mm fan directing air out the back.

If installed as directed (in the directions), the H105 cooler, with
its two 60mm fans would replace the fan on top.
Furthermore, they would be directing air IN to the case. My
intuition suggests installing these fans so they direct
air OUT the top of the case.

2 questions:

1. What do you think of this configuration?

2. What does anyone care about "positive air pressure in the case"?
I don't understand the significance of this.

Thank you,
Bill

Airflow is a lively topic.

http://www.overclock.net/t/1062965/p...e-or-less-dust


I have positive pressure on my current case. It's still dirty in
there, because I absolutely refuse to use filters. Filters require
attention, and the last thing I need it "continuous maintenance".

And the new system, with the monster fan on the back, it's too early
to tell how bad it'll be (negative pressure). It should get pretty dirty,
but the old motherboard in there wasn't that bad.

As the thread above mentions, some sort of filter would help, in terms
of the quality of air feeding the case. But if you "do it right",
say a hepafilter like the first one I bought for my new gas
furnace, there will be high arrestance, and you'll need a good
fan to drive it, and then you're going to have a "noisy vacuum cleaner
case design". I choose to accept a bit of dust, and use Medium speed
fans (somewhere around 35CFM a piece). Yes, my case is still noise,
it's still slightly dirty, but it's not super noisy + super clean.
That would drive me nuts (the super noisy part).

*******

The purpose of the instructions saying to arrange the fans
on the H105 to blow *into* the case, that's an attempt to
give the water cooling system impressive statistics. If you
use room air, the inlet side of the H105 is 25C. The water
experiences better cooling. Now, if you grabbed case air
and blew it out of the case, the inlet air to the H105
would be 35C. And the water could never be cooled below
35C, and Corsair "wouldn't look as good". So they only
specified such an approach, for their own interests.

You put cooling where it is needed. You use enough cooling
for the hard drives, so they stay under 50C. My current
two drives are running at 31C and 25C.

You remove enough heat from the CPU, so it doesn't throttle.

You remove enough heat from the GPU, so it doesn't hit 105C
or whatever the top temperature is. There have been some
gamer cards, that would run at 90C, even with the fan on
the card running at a noisy speed. In such an event, you'd
want to "sweep warm air" from around the card, and that means
adjusting the case fan configuration to achieve that objective.
I've had one computer case, where there was a pretty bad
cloud of warm air right around the CPU cooler, and it was
pretty hard to move that out without a lot of noise.

*******

OK, if I aim for a slightly positive air pressure,
it looks like this. That's an attempt to balance
the 200 with the three other fans. (I didn't check
CFM numbers, instead using pie R squared.)

^ ^
| |
+------- 120 120 -----------------+
H105
-60

|
| Video
| HDD
200-
- PSU HDD

+-------------------------------------+

If instead, I arranged them the following way, now I've
got way too much positive pressure. The 120's and the
200 are working against backpressure. I would need to
remove more slot plates in the back, and 5.25" tray
covers in the front, to let the air out. I expect
with this setup, more air will leave via the front
than the back, and the lower left corner will be
"warm cloud" country. The video card temps might be
a bit higher.

| |
v v
+------- 120 120 -----------------+
H105
-60 Remove --
Covers --
| Here --
| Video
| HDD
200-
- PSU HDD

+-------------------------------------+

I'm going to have to go with the first config. It's going to
make the H105 a bit warmer. But the lower diagram, the
air pattern looks like it could be a bit more stagnant.
I really really like a defined airflow (in on one side,
out on the other). It's hard to predict how much
worse the bottom diagram will be. The H105 will be
heroic, and the CPU will be at 26-27C when idle, because
room air at 25C will be hitting the radiator. I'm willing
to trade a little CPU temp, at least until it hits
the throttle point. If the CPU starts to throttle,
then I have to switch to some other configuration.

So yeah, there are some compromises involved. And
some test cases to run.

Paul

Thank you for your thoughtful and insightful answer! I (strive to)
choose every component optimized
for peace and quiet. The vast majority of the time, my computer is not
working hard (like as I type this message).

Besides for the H105 CPU fan, other components on my list include the
(new) Gigabyte Strix GTX750ti GPU
and Corsair RM750 power supply. I have been using their HX power supplies,
but I recently learned about the "RM" series, optimized for quiet. For
some reason,
the RM750 has a much better track record than the RM650 (which
power-wise) probably
would have been quite adequate for me.

Cheers,
Bill