Passive CPU cooling not necessarily the quietest solution

[nomoon]

(Forgive me if this has already been covered here)

I think that an argument can be made that a “passivelyâ€

[Wayne Redpath]

I use only one fan in the upper chamber of my P180 case. The top fan position is blocked off and air enters through the filter on the lower front and exits through the upper rear fan position. A 1150 rpm Scythe Gentle Typhoon does double duty by both getting the heat away from the CPU and moving room air through the chassis. The fan is mounted on the Scythe Ninja on the side adjacent to the rear chassis fan location and blowing towards the rear of the chassis. A 2.5 cm (1 inch) duct made out of a 120mm fan with the guts cut out is mounted on the rear chassis fan location. The CPU fan and the duct align well enough for most of the air to exit the chassis. The fan is connected to the CPU fan connector and the ASUS fan control is set for DC fan type and Silent profile. When the system is in idle the CPU fan runs at 700 rpm and speeds up to 1200 rpm at full load. Idle temperature is around 20'C above ambient and full load temperature is 45'C above ambient.

This cooling setup compared to normal configurations reduces noise in these ways:
=> There is only one fan making noise.
=> The Scythe Gentle Typhoon has the most non-annoying noise signature that I know of.
=> The fan makes less noise sucking air from the heatsink than blasting turbulent air at the heatsink.
=> The less turbulent air from the Gentle Typhoon has time to moderate before hitting that nasty cheap steel stammping rear grill.

This computer is almost silent during the day and when things quiet down in the late evening I can here a gentle, quiet whirr.

[audiojar]

With a CPU fan you are still going to want a case fan. The exception is a setup like Wayne's. If you aren't ducting the exhaust of the CPU fan out of the case then you are likely creating a situation where the air is circulating around inside the case, with little actual outside airflow.

Also, I'm not sure you'd really notice the additional noise of a quality case fan running at low RPM, over the noise of the CPU fan, PSU fan, and drives.

Additionally, the case fan could allow the CPU fan to run slower as the temp inside the case would be lower due to increased airflow. Two slower fans is probably better than one faster fan (even if that fan is well inside the case).

[Wayne Redpath]

Another problem I often see is a "fan fight" situation where the chassis fan and the PS fan are trying to suck air from the same place that the stock Intel CPU fan is trying for.

It is even possible for the PS to be completely starved for air by a very strong chassis fan.

[josephclemente]

Assuming a 100% negative pressure case fan setup, if a fan is ducted to do the job of both case exhaust and cpu cooling, would air entering and exiting the case be the same as having a separate exhaust/cpu fan?

I would guess that the ducted fan would need to run faster due to the resistance added by the CPU heatsink, in order for case airflow to be the same as a non-ducted setup. Am I correct?

[nomoon]

With a CPU fan you are still going to want a case fan.

I certainly agree. I was only suggesting that the use of a CPU fan would allow the case fan to run at a lower speed.

Also, I'm not sure you'd really notice the additional noise of a quality case fan running at low RPM, over the noise of the CPU fan, PSU fan, and drives.

Two slower fans is probably better than one faster fan (even if that fan is well inside the case).

I agree with this as well.

Another problem I often see is a "fan fight" situation where the chassis fan and the PS fan are trying to suck air from the same place that the CPU fan is trying for. It is even possible for the PS to be completely starved for air by a very strong chassis fan.

That's not a problem with my P180 case.

A 1150 rpm Scythe Gentle Typhoon does double duty by both getting the heat away from the CPU and moving room air through the chassis.

1150 rpm sounds like it would be fairly loud. I find 900 rpm is where I start to really notice my fans

I have an Intel E8500 inside a P180 case with (gasp) six fans (including the one inside the PSU). It idles at 34C. The CPU fan idles at 515 rpm on a Thermalright HR-01 Plus. All of the case fans are running at 715 rpm. It's extremely quiet. The CPU fan is controlled with Speedfan and the case fans run at a fixed speed and are controlled with a Fanmate2. I'd run the case fans even slower if I could. The four hard drives are really the only thing that is audible, and this will be getting quieter once I move three of them to a NAS in the next room.

I can't find the article, but I think that it was Mike Chin who wrote about his results of measuring noise from multiple fans, and found that it wasn't as loud as theory predicted. Noise theory predicts that noise should increase by 3 db if the noise power is doubled. However, Mike saw an increase in noise of only 1 db when two fans were running compared to one fan. If this is true, then it may further support the idea that running multiple slow speed fans is much better than a single fan at high speed.

Jason

[wwenze]

Well, the 3dB assumption is true only for fully constructive interference between two point sources. Think somewhere in mid-high education w'all have been taught about waves.

From that we know that the doubling of loudness will only occur at specific points, and at other points destructive interference cancels out the sound.

[Wayne Redpath]

need to run faster due to the resistance added by the CPU heatsink

Heat sinks like the Scythe Ninja, Thermalright HR-01 Plus and Noctua NH-U12P have very little air flow resistance.

[Wibla]

I'll agree with the OP - to a certain point. I dont have fan control on the cpufan in my P182, so its a nexus running at a fixed 1000rpm on a TRUE. But that doesnt really bother me, I cant hear it outside the case anyway. What that fan does is allow me to run all the other case fans at a much lower speed, and still keep good themal performance. I ran this setup with a Q6600 oc'ed to 3.2GHz doing folding@home, and the case fans still only ran at like 4-500rpm. Tho, it isnt a "one or two fan"-system by any stretch, with two nexus intake fans (one kama bay), a papst top exhaust and nexus rear exhaust to keep the airflow going. The only thing I can hear from this machine is the annoying rumbling sound from the GTS250 cooler even at 35%, and some seeks from the F1 640GB if I pay attention.

However, there are many ways to nirvana, and mine isnt the perfect one, it just works quite well for my setup.

[nomoon]

Well, the 3dB assumption is true only for fully constructive interference between two point sources. Think somewhere in mid-high education w'all have been taught about waves.

From that we know that the doubling of loudness will only occur at specific points, and at other points destructive interference cancels out the sound.

The 3 db assumption assumes a noise that is spread across a spectrum of wavelengths, where some wavelengths interfere constructively and others destructively at a fixed point in space. If I recall correctly, the 3 db increase will occur if the the noise power is spread across the spectrum in a white noise distribution. If you integrate power over the spectrum, the net increase will be 3 db. Here is a discussion by Mike Chin where he made measurements. I suspect that the apparent discrepancy with Mike's measurements is because of the spectral distribution of noise for the fan is not the same as for white noise. If fans produced white noise, then he might see the 3 db increase in noise when a second fan was added. The end result is that running multiple fans will often provide a quieter solution than running a single fan.

Jason

[Fayd]

while a fan on the heatsink will be more effective at cooling the heatsink, it wont be as effective at exhausting the air from the case.

given that, there's no real reason to not mount an exhaust fan..

and if your exhaust fan pulls enough air to passively cool the CPU, then by all means.

people are way overly critical of whether or not their CPU is getting cool enough.

they're designed to function well north of 70C, but most people seem to think that that's the end of the world.

if you're using a tower heatsink that is reputed to function well passively, you're running it passively, and you're having heating problems, then you're doing something wrong.




btw, i've tested this: w/ fan or fanless, doesnt make any noticeable difference on my scythe ninja in my case (p182) with my processor (Phenom II X4 945)

last: i need to do fanswaps for the gentle typhoons. i wonder if i could get away with the 500 RPM models...

probably not. 800 RPM is a nice number.

[ces]

while a fan on the heatsink will be more effective at cooling the heatsink, it wont be as effective at exhausting the air from the case.

You can have your cake and eat it too if you use the full HR-01 setup.

Thermalright makes the HR-01 plus. It was originally designed as a passive heatsink. It is a relatively older design, but even today with slow fans it performs just as well as the Prolimatech and even with faster fans it performs almost as well as the Prolimatech (per SPCR testing).

Thermalright makes two compatible accordion ducts that clip right onto the the HR-01. It permits one fan to act as both the case fan and as the CPU cooler fan.

But that is not all, if you are willing to flout convention (I love doing that) you can run your case fan to push air into your case, providing the CPU cooler with the fresh cool outside air first. This air is generally going to be about 15C cooler than the air from inside your case. A 15C advantage is a heck of a cheat.

If you bring in a cubic inch of cool air, somehow a cubic inch of warm air will get pushed out of the case somewhere. If the case is reasonably well ventilated you don't need to make any other changes. If your other fans all all pushing air into the case, you just need to reverse one or more of them to reduce the pressure you your new case/cpu fan.

In the end, you get one fan doing the work of two.

[nomoon]

while a fan on the heatsink will be more effective at cooling the heatsink, it wont be as effective at exhausting the air from the case. given that, there's no real reason to not mount an exhaust fan..

I agree. I've never advocated not having a case fan.

and if your exhaust fan pulls enough air to passively cool the CPU, then by all means.

I would argue that:
1. Your case fan(s) would be able to run slower if there is an active fan running on the CPU heat sink.
2. A fan mounted directly on the CPU heat sink will cool the CPU more efficiently than a case fan trying to cool a passively mounted CPU heat sink. The directly mounted fan would require a lower RPM than a case mounted fan to give the same cooling ability.

The case fans on my P180 run at a fixed 625 rpm and my Speedfan controlled CPU fan idles at 535 rpm.

Jason

[ces]

2. A fan mounted directly on the CPU heat sink will cool the CPU more efficiently than a case fan trying to cool a passively mounted CPU heat sink. The directly mounted fan would require a lower RPM than a case mounted fan to give the same cooling ability.

My personal experimentation confirms that in spades.

I have come to the conclusion that there are only two functions for fans.

ONE: Moving cool air into the case (if you do that, the hot air will be pushed out).

TWO: taking that cool case air and pushing it past whatever you want to cool. The closer you place the fan to whatever you are trying to cool, the less air you have to move and the less noise you have to make in moving it.

Oh and THREE: That first whiff of air movement that moves the stagnant warmed up air along, replacing it with cooler air, generates the most cooling per CFM or per DB. Each additional CFM is less efficient at cooling than the prior CFM.

[Fayd]

while a fan on the heatsink will be more effective at cooling the heatsink, it wont be as effective at exhausting the air from the case. given that, there's no real reason to not mount an exhaust fan..

I agree. I've never advocated not having a case fan.

and if your exhaust fan pulls enough air to passively cool the CPU, then by all means.

I would argue that:
1. Your case fan(s) would be able to run slower if there is an active fan running on the CPU heat sink.
2. A fan mounted directly on the CPU heat sink will cool the CPU more efficiently than a case fan trying to cool a passively mounted CPU heat sink. The directly mounted fan would require a lower RPM than a case mounted fan to give the same cooling ability.

The case fans on my P180 run at a fixed 625 rpm and my Speedfan controlled CPU fan idles at 535 rpm.

Jason

and there you're wrong.

whether or not you have the CPU being actively cooled, you still need to exhaust the same amount of air from the case, because you're still putting roughly the same amount of heat into the case.

[Fayd]

while a fan on the heatsink will be more effective at cooling the heatsink, it wont be as effective at exhausting the air from the case.

You can have your cake and eat it too if you use the full HR-01 setup.

Thermalright makes the HR-01 plus. It was originally designed as a passive heatsink. It is a relatively older design, but even today with slow fans it performs just as well as the Prolimatech and even with faster fans it performs almost as well as the Prolimatech (per SPCR testing).

Thermalright makes two compatible accordion ducts that clip right onto the the HR-01. It permits one fan to act as both the case fan and as the CPU cooler fan.

But that is not all, if you are willing to flout convention (I love doing that) you can run your case fan to push air into your case, providing the CPU cooler with the fresh cool outside air first. This air is generally going to be about 15C cooler than the air from inside your case. A 15C advantage is a heck of a cheat.

If you bring in a cubic inch of cool air, somehow a cubic inch of warm air will get pushed out of the case somewhere. If the case is reasonably well ventilated you don't need to make any other changes. If your other fans all all pushing air into the case, you just need to reverse one or more of them to reduce the pressure you your new case/cpu fan.

In the end, you get one fan doing the work of two.

"flouting convention" in this case means forgoing air filtering. not something i'd be willing to do.

[ces]

"flouting convention" in this case means forgoing air filtering. not something i'd be willing to do.

I agree with you. I am a firm believer in positive case pressure with all incoming air being run through a filter.

But there is no need to compromise.

Cut out whatever grating is in the way, or not, and use a FilterRight filter. They are the best that I have been able to find. See:
http://www.performance-pcs.com/catalog/ ... ts_id=2275

Once you use one of these, the others just aren't good enough any longer. They are just the size of filtering. They catch all the dust and they are so easy to clean.

[nomoon]

---- snip --------
I would argue that:
1. Your case fan(s) would be able to run slower if there is an active fan running on the CPU heat sink.
2. A fan mounted directly on the CPU heat sink will cool the CPU more efficiently than a case fan trying to cool a passively mounted CPU heat sink. The directly mounted fan would require a lower RPM than a case mounted fan to give the same cooling ability.

The case fans on my P180 run at a fixed 625 rpm and my Speedfan controlled CPU fan idles at 535 rpm.

Jason

and there you're wrong.

whether or not you have the CPU being actively cooled, you still need to exhaust the same amount of air from the case, because you're still putting roughly the same amount of heat into the case.

I don't see how you arrived at your conclusion. Here is my reasoning:

• With a slower case fan, the average air temperature inside the case will be warmer, and the heat flow rate out of the case will remain constant. Edit: (Also, the temperature of the air leaving the case will be hotter. )
  
  Heat flow = (delta Temperature) x (mass flow) x (heat capacity)
• A heat sink fan will increase the rate of airflow through the heat sink.
• Lots of warm air flowing through the heat sink will have the same cooling as a little bit of cold air

My original conclusion was that an active fan running on the CPU heat sink would allow the case fans to run slower. This still seems correct to me. I don’t see a problem until you take it to an extreme. Situations where the case fan speed is reduced by 30%-60% sound very feasible and beneficial in typical scenarios.

In most cases, the CPU heat sink is located near the case fan, so much of the air heated by the CPU won't be mixed with air in the rest of the case. Therefore, the air leaving the case may be hotter (and slower) with a slow case fan, though the air entering the CPU heat sink won't be significantly hotter.

Maybe there's something here that I'm missing? I appreciate polite scrutinizing and constructive comments.

[ntavlas]

I have to agree with nomoon here. Very often a 500rpm exhaust is enough to keep the case reasonably cool. However, such a fan will have difficulty properly cooling even a low resistance heatsink like the ninja at such low speeds (in my case the exhaust has to run at 800rpm to keep my q8400 below 70c at stock speeds). Having a fan sit directly on the cooler allows me to keep the exhaust at 500rpm with no adverse effect on the rest of my system temps.

Of course there are circumstances where you could remove the fan from the cooler and still keep your exhaust at a very low speed, when running a low power cpu for example. Ducting could also achieve the same thing as long as the rest of the components don`t rely too much on the exhaust fan to stay cool.