So the situation is this... NSK 3480 with a Nexus D12SL12 in the rear opening with the Nexus silicone isolators. The rear opening has been de-grilled. Airflow wise the case is otherwise stock (wiring has been moved behind the mobo tray), cpu is an i3-530 cooled by a Scythe Mugen2 passive. The fan is operated by the CHA_FAN header on a Gigabyte H55M-UD2H - an uncontrolled 12V header.

Here's the rub... this should be darned near silent! With the side panel open the fan runs near inaudible levels, as expected. However, once I close the side panel on the case the airflow noise/woosh from the fan increases significantly. Of course by "significant" I mean a little bit, but it's definitely audible to anyone who frequents SPCR and definitely audible past 3'/1m in my suburban apartment.

I am wondering if this is caused by an airflow restriction at the intake side? Not enough air coming into the case... too much negative pressure maybe? What other thoughts are out there?

Or am I just chasing the SPCR ghost... each thing I silence makes the next thing louder.

If I remember correctly this case has a plastic funnel in the side panel. If it's on then perhaps it's amplifying the woosh sound, especially if the case sits on the right side of your desk. Just a quick thought

Actually I need to be more specific...

I've swapped the side panels so that the air doesn't short circuit through the Mugen. I've also closed up the rear expansion covers and rear expansion vents to prevent the same short circuit.

The Mugen is running semi-passive, it does have it's own D12SL12 attached to the CPU_FAN header, the BIOS is running it as voltage regulated (not PWM) and I have the fan profile such that the fan doesn't even come on until the cpu temp gets to 30c. Most of the time the the fan is off.

Your intuition is probably correct. Higher intake resistance when the case is closed is making the fan noisier. At the end of the day, the NSK 3480 has a pretty restrictive front, due to the semi-tortuous path air must take to get in.

Do you need to run the exhaust fan at 12V?

See the link here about the NMB fans and the link to the charts on back pressure.
http://www.silentpcreview.com/forums/vi ... hp?t=58350
(I think it is about the 4th or 5th message down)

It may be that you just need a fan with different performance characteristics that doesn't care how much back pressure it is facing. Consider trying a Gentle Typhoon with the same or even a lesser CFM rating. It probably will not even notice the increased pressure.

Now with the GT I am just guessing about its performance profile. But an Akasa Apache has a high impedance published spec (to the extent you can trust manufacturer's specs) and the Noctua case fan has reasonable back pressure published specifications. You can try one of those as well.

My gut though tells me the GT is the one to experiment with. The fast GT downvolts to about 500 rpm I think. You can try that and experiment to see if that fixes the fan noise problem. Though some people have been complaining about GT bearing noise.

Sounds like you would be better off opening up the air flow in the front/bottom of the case? Running the fan faster will only be noisier, and probably barely make a different in the temps.

Make it quieter and cooler by providing a low resistance intake.

How fast is your fan spinning with your side panel on and off? It's normal for fans to run slower when you open the case. With an open case, more air is being moved per rotation (more work per rotation), so the RPM must drop if the supplied power is constant. Likewise, the fan will speed up when you close it. (It's like how your vacuum cleaner speeds up when the hose gets stopped up.) The effect is usually pretty small, but if your case has lots of restrictions, then maybe the effect is more significant. I notice a small effect when I open and close the door on my P180 case.

If the increase in noise is not due to increase fan speed, then maybe you have some unusually resonance. Is your fan isolated from your case with soft mounts? Maybe the case could use some dampening material?

Jason

It almost sounds like a hdd with the heads constantly moving.

I booted into Windows so I could get to Gigabytes "EasyTune" software and look at fan speed. Unfortunately I have no fan monitoring or control in Linux.

With the case open, fan running in "free air" it reports ~725 rpm. So maybe the CHA_FAN header *is* somehow controlled. It's not controlled directly by the bios though. The fan was running full bore until I opened EasyTune - it then slowed down. So it's software controlled and running at about 7.5v if the reported speed is correct according to this: http://www.silentpcreview.com/article695-page2.html

I tested with closing several panels.

The stock vented panel (Pv)
Stock solid panel (Ps)
The top panel unvented (Pts)
The top panel vented 50% (Ptv)

Free air is ~725 RPM nearly inaudible
Ps ~ 708 RPM
Ps + Pt ~700 RPM, loudest
Pv ~ 715 RPM
Pv + Pt ~710 RPM
Pv + Ptv ~712 RPM

As each RPM level reduces the subjective sound level increases.

Currently I am using the Pv + Ptv though I can still hear some "rumbling" coming from the box.

I used the screwdriver trick to "listen" to the case. All three fans are isolated with silicone and the hdd is suspended and everything is as it should be... no sympathetic vibrations.

If this really is a negative pressure issue then I wonder what might happen if I added a 92mm fan up front to help pressurize the case.

Gigabyte MB's can usually control 2 fans.
check with speedfan if you can change the speed of the case fan.

1. Please explain the "screwdriver trick"

2. Try using a GT or an Apache to replace the Nexus exhaust fan. I believe they will be resistant to impedance generated noise. More so than the Nexus.

IME the NSK3480 is a bit prone to cavity resonance.

Maybe you're experiencing such an effect.

Regards,
Luca

P.S.: wow, a 530 with Mugen2 passive at 30°C!

30C in open air.

With the case closed and only the top door open on the cabinet (last picture) it sits at 36C with the rear fan at 725 RPM. This is running all four threads at 2.93 GHZ with Fold at Home on Ubuntu/Linux.

If I engage the IGP to watch a video (h.264, VLC, full screen) it climbs to 39C.

I haven't seen it break 40C yet.

I did manage to get fan control running under Ubuntu 9.10.

It required lm_sensors 3.1.2 which wasn't available in the 9.10 repositories (it had 3.1.0). I installed it from the 10.04 repos and got positive reporting on all of my sensors and fans. It appears as though fancontrol/pwmconfig can manage both of the fan headers on the motherboard.

Setting up fancontrol is a different matter though. I'm still working on a good profile.

Resonance sounds like the most likely cause. What is the solution then?

Case resonance is helped by the same things that improve air flow through the case and lower noise -- reduce the restrictions on the intake and exhaust, and/or dampen the case (which can change the volume of air inside).

Doesn't resonance have more to do with the cavity shape and size than the other factors you mention? Maybe the solution is just putting a block of acoustic foam in there.

There are two ways to reduce resonance if you think of your case like a large drum...

1. Fill the drum with stuff to keep it from making noise when you bang of the head. The equivelant, is as pointed out, to throw some acoustic foam inside the case. My fear is that by adding material to the inside of the case I begin to restrict the already restriced airflow of the 3480.

2. Don't bang on the drum head. The equivelant is to reduce all amounts of vibrations from getting to the panels, and then making the panels less like drum heads by adding mass to them.

Cavity resonance really never ocurred to me to be an issue, the 3480 is a pretty small case! I'll have to see what I can come up with.

Cavity resonance?! Is that like a whistling / howling sound that changes with the speed of the fans? My Silverstone TJ08 does that, but there's no room for foam in there: http://www.silentpcreview.com/forums/vi ... =30#466639


//

That is a nice tight case. How do you like it, other than the noise?

There is room for foam. Some on the floor and smaller pieces here and there.

Does the sound stop when you open the side of the case? If not, it is unlikely to be resonance.

The howling is there even with the side covers off! Weird huh?

The TJ08 is the perfect size though its quility is poor, (I listed the problems in the linked thread above).

If it continues to howl with the covers off then it's not a pressure or air flow issue. It's also unlikely to be a resonance issue. Have you gone through the standard procedure of stopping components individually to see if/where the sound is reduced?