[A] Little help pl[ea]s[e?]

[Caddish]

Hi I just finished building a CPU+GPU water cooling loop in a mini p180 for fun (not really into over-clocking). I have the 2 stock fun running at low and 2 Scythe S-FLEX SFF21F 120MM 1600RPM 28DBA 63CFM on the rad. When I'm playing the noise is tolerable but on iddle it is very noisy and the CPU/GPU are iddling at 30C which I find pointless.

I'm torn between :

• slowing down the 2 scythe to 1000-1200 when the computer is iddling but I'm a bit lost with all this PWM gibberish and exactly how to slow down a fan (if you could just point me to a guide explaining PWM/voltage control/fan controller for noob that would be awesome)
• adding another 120.2 radiator and replacing the 2x1600rpm with 4x1200rpm and slowing them down to 800rpm when iddling

So basically will 4x800rpm (best ratio CFM/noise fan) be less noisy than 2x1200rpm of the scythe I have and if yes is it easy to link them to CPU temp ?

Any other suggestion to lower noise is welcome

[ces]

PWM is controlled by a fan header on your motherboard. That in turn is controlled either through your bios and/or a utility supplied by the motherboard manufacturer.

The nature and extent of that control is determined by the motherboard manufacturer.

Almost always this control is linked back or associated in some way with the temperature the board or chip is reporting for the CPU and/or its cores.

Some low end boards have a PWM header but don't really attempt to control the speed of the fan.

The PWM header looks similar to a 3 pin fan connector but has 4 pins.

I don't believe there are many motherboards, if any, that have more than one PWM fan header. But some will have extra 3 pin headers that will use voltage to control the speed of non PWM fans/

Some people use the utility speedfan to control fan speed. I find it a pain in the rear because you have to set it every time you boot.

SOLUTION 1:

Get 2 Akasa Apache fans. Get their special cable that permits you to hook two of them up to a single 4 pin PWM header. If things are tooling along at a low temperature they will lope along at 500 rpm without making any noise. If things get hot, they will kick it up to 1500 rpms.

The Apache handles high impedence situatios (like your radiator) substantially way better BY A LOT, than the S-Flexes.

What that means is that when they are getting high levels of pushback, they don't strain. That straining generates noise. The ability to handle that impedance is what distinguishes a good CPU fan from a good case fan. The tradeoff is that a good CPU fan is likely to be slightly louder than a good case fan in low impedance situations (all things being equal - which they never are).

The Apache's impedance rating (as supplied by the manufacturer) is higher than that of the Noctua CPU fans. Actually about double. I recollect seeing benchmark performance figures comparing S-flex impedance handling to Noctua's impedance handling. Noctua kicked the S-Flex's a*s.

Apache (per manufacturer published ratings) kicks Noctua's a*s in this particular category of performance. The apache also appears to be manufactured by the same manufacture of the Noctua, only with fan blades that appear to be optimized for impedance handling.

Someone on this board purchased two Noctuas and complained that they started making noise after a few months. I recollect that they use real bearings (the S-Flex use fancy sleave type Sony bearings... but not old fashioned honest to god bearings). Real bearings tend not to ship well. The points of contact are smaller than sleave bearings and they can be damaged with lower g forces.

SOLUTION 2:

Nanoxia has an adapter that you connect to the 4 pin PWM header on your motherboard and to the 3 pin fan connector on your regular fan. It permits you to control your regular fan as if it were a PWM fan.

You have two fans, so you might need to use the Apache 2 fan PWM cable in conjunction with the Nanoxia PWM adapter.

SOLUTION 3:

Scythe has traditionally had two PWM fans. A Kama flex one and a Slipstream one. They have just come out with a second version of the Slipstream PWM fan that allows you to adjust the band of RPM performance that is adjustable by your motherboard through the PWM header.

SOLUTION 4:

Scythe Gentle Typhoons may work. They likely handle high impedance better than any of the above fans. If you just look at their fan blades you can sort of see why. They have to run at faster speeds to generate the same CFM as most other fans. But they are quieter at those faster speeds than most other fans.

The problem with them is that they have resonance points where they can make noise. You are generally safe if you run them at 12 volts. But if you play around with their speed, you can end up running them at a resonance point and they then make more noise (but even then they are quieter than most other fans)

[Caddish]

Awesome reply. I wish somebody would have mentionned this before. Choosing a fan has been quite random I must say except the full CPU/GPU load CFM recquired which I calculated.

I should have mentionned that I have a gigabyte p55m-ud4 mobo which has 2x4pin but I think only the cpu header is bios controlled although most people seem to complain that there is very little control on the fan in the bios

Solution 1 : If I understand well I plug them both in the CPU header and control them through the bios. Would love that but there is pretty much no control for the fan in the bios. Also 57CFM won't be enough for full load

Solution 2 : I can't seem to find nanoxia in north america

Solution 3 : interesting but can I plug 2 of them in the CPU header

Solution 4 : CFM is not enough for full load but I might consider them if I had a radiator

Solution 5 : NZXT sentry 2 fan controller http: // ww w.newegg.ca/Product/Product.aspx?Item=N82E16811992005&Tpk=nzxt%20sentry%202
What do you think. It's cheap, little bit of eye-candy and would have much more control than my bios I guess altough much less accurate CPU temp reading. Although people have mentioneed that it make cliking noise

or bigNG which if I understand correctly has the advantage of being controled by software through internal usb ?

the aerocool touch 1000 is nice but it does not seem to change fan rpm in relation to temp

Also I checked and 4x800rpm with 2 rad would perform similarly to my current 2x1600rpm but what about noise ?

(By the way money is not really a concern as long as it stop doing that much noise on iddle/movie watching)

[Caddish]

I made some more research and since the auto rpm mode of the nzxt sentry 2 is not configurable I'll probably go with the bigNG. But there is something I wonder will the fan be running at a 100% until the software is loaded or the setting are stored in the bigNG ?

[ces]

Solution 1 : If I understand well I plug them both in the CPU header and control them through the bios. Would love that but there is pretty much no control for the fan in the bios. Also 57CFM won't be enough for full load

57CFM is with no impedance. I am highly confident that if the S-flex are rated for 63CFM (which is also under no load) that they apaches under load will deliver substantially more actual CFM than the S-Flex. Say if the S-flex drop to 30 CFM and the Apaches drop to 45 CFM, they are delivering 50% more CFM.

Solution 2 : I can't seem to find nanoxia in north america

Aerocooler caries it.
http://www.aerocooler.com/shop.cart?act ... d=FANNXPWM
SVC
http://www.svc.com/pwmx.html
Frozencpu
http://www.frozencpu.com/products/9081/ ... oller.html

Solution 3 : interesting but can I plug 2 of them in the CPU header

Rosewill 12" PWM Splitter Model RCW-FPS-401
http://www.newegg.com/Product/Product.a ... 6812119248

Solution 4 : CFM is not enough for full load but I might consider them if I had a radiator

"I just finished building a CPU+GPU water cooling loop" doesn't that mean you have a radiator?

Take another look at the GT blades. Then go take a look at the Noctua case fans blades (not their CPU fans). Then take a look at the S-flex blades.

Which fan do you think can push air best against high impedance. The GT blades provide no place for the air to leak backwards. The Free air CFM ratings are one thing. Actual delivered CFMs against impedance is another.

In high impedance situations I don't believe any other quiet fan can begin to compete against the GT. I just don't see anything out there.

Its like assuming that a 500 HP ferarri can pull as much weight as a 400 hp diesel truck. It can't. It's not geared for it. It can race faster around the track, but not if it is pulling a 20 ton trailer.

It will stall, just like these fans stall.

Solution 5 : NZXT sentry 2 fan controller http: // ww w.newegg.ca/Product/Product.aspx?Item=N82E16811992005&Tpk=nzxt%20sentry%202
What do you think. It's cheap, little bit of eye-candy and would have much more control than my bios I guess altough much less accurate CPU temp reading. Although people have mentioneed that it make cliking noise

or bigNG which if I understand correctly has the advantage of being controled by software through internal usb ?

the aerocool touch 1000 is nice but it does not seem to change fan rpm in relation to temp

I need to think about it. Can you give me a more complete description of your system and what your intentions are for it, and why.

Also I checked and 4x800rpm with 2 rad would perform similarly to my current 2x1600rpm but what about noise ?

That doesn't sound right. Can you tell me where you are getting your data?

Why are you even messing around with water cooling? What exactly are the constraints you are dealing with that caused you to go water?

[Caddish]

57CFM is with no impedance. I am highly confident that if the S-flex are rated for 63CFM (which is also under no load) that they apaches under load will deliver substantially more actual CFM than the S-Flex. Say if the S-flex drop to 30 CFM and the Apaches drop to 45 CFM, they are delivering 50% more CFM.

I did not understand when you talked about impedance that it could have so much impact on CFM but that makes sense

Aerocooler caries it.
http://www.aerocooler.com/shop.cart?act ... d=FANNXPWM
SVC
http://www.svc.com/pwmx.html
Frozencpu
http://www.frozencpu.com/products/9081/ ... roller.htm

Rosewill 12" PWM Splitter Model RCW-FPS-401
http://www.newegg.com/Product/Product.a ... 6812119248

Thanks

"I just finished building a CPU+GPU water cooling loop" doesn't that mean you have a radiator?

Take another look at the GT blades. Then go take a look at the Noctua case fans blades (not their CPU fans). Then take a look at the S-flex blades.

Which fan do you think can push air best against high impedance. The GT blades provide no place for the air to leak backwards. The Free air CFM ratings are one thing. Actual delivered CFMs against impedance is another.

In high impedance situations I don't believe any other quiet fan can begin to compete against the GT. I just don't see anything out there.

Its like assuming that a 500 HP ferarri can pull as much weight as a 400 hp diesel truck. It can't. It's not geared for it. It can race faster around the track, but not if it is pulling a 20 ton trailer.

It will stall, just like these fans stall.

Good to know like I said I did not realise the impact of the rad on the airflow. I just tough I need about 65CFM and I didn't know some fan where better in these condition. But heck if a apache is much less noisy (17DBA VS 28DBA) and has better airflow on a rad I'm all good with that

I need to think about it. Can you give me a more complete description of your system and what your intentions are for it, and why.

I'll put the spec in my sig. Here is a resume of how it is right now :

Full load/gaming : Temp are correct, noise is tolerable
Iddle/movie/internet : Temp are unneccessarly cold , noise is giving me headache

So basically my primary goal is to lower iddle noise by a lot with a fan controller regardless of temp (I'm ok with 50C iddling I change CPU often)
Secondary goal lower full load temp and noise a little

That doesn't sound right. Can you tell me where you are getting your data?

There is a spreadsheet I found on internet that estimate CPU/water temp in a water cooling loop according to the number of block, CPU/system wattage number of rad a fan rpm. Which is how I found that with a 120.2 rad I needed approximately 1600rpm fan but it was not taking into account that some fan perform better on rad like you said. I could find the link again if you are interested

Why are you even messing around with water cooling? What exactly are the constraints you are dealing with that caused you to go water?

My contraint were I had too much money and a buttload of time to waste and I must say it was a quite fun project

Thanks for helping

[ces]

You didn't give all the details of your system, such as the case, CPU etc.

Consider just getting the Noctua 3 radiator heatsink and using three Nexus fans at 7 or 8 volts.

Or a Prolimatech or Thermalright Venemous with two Nexus fans running at 7 or 8 volts.

Or a Thermalright HR-01 plus with with a Thermalright Duct bringing in outside air (probably 15C lower than your inside the case air) with the 1145rpm GT running at 12 v or the apache running on pwm, pushing the air into the duct.

Or if you have space problems either the Thermolab BADA or the new Prolimatech Armageddon heat sinks.

[Caddish]

Don't get me wrong I'm very satisphied with my water cooling and it was very fun to build, I just want to lower a bit the iddle noise. With your help I narrowed down the solution to either 2 apache fan bios controled or adding another rad + 4 apache and a bigNG fan controller but let's just say another rad would be very thigh

Spec are in my sig. Here is a picture if you want to see
http://img96.imageshack.us/img96/8572/dsc03731m.jpg

[ces]

Looks pretty.

Yeah the simplest approach is to try some new fans. I would buy one Apache, one Noctua and one 1900rpm GT.

Then get this:
Akust Power Splitter PCI Card with 16 Separate Outputs!
http://www.performance-pcs.com/catalog/ ... s_id=25981

Experiment, fan by fan, speed by speed. and see what happens.

Come back and post the results of your experiments.

[Caddish]

That's a great card but it's not temperature controlled right? if I go with fixed voltage it will be silent on iddle but warm on full load

[QuietOC]

I don't believe there are many motherboards, if any, that have more than one PWM fan header. But some will have extra 3 pin headers that will use voltage to control the speed of non PWM fans/

My Gigabyte GA-MA785GM-US2H has 2 PWM headers. Luckily both headers also support DC voltage control since I only have one PWM fan, and I am using it on another motherboard that only offers PWM control.

[Caddish]

How do you tell if it's PWM. I have 2x4pin on my mobo but in the bios there is only option to control CPU fan. Could it be because nothing is pluged in the other 4pin that the option does not appear in the bios

[ces]

How do you tell if it's PWM. I have 2x4pin on my mobo but in the bios there is only option to control CPU fan. Could it be because nothing is pluged in the other 4pin that the option does not appear in the bios

I doubt that. Probably one of them just runs at full bore. But try it and see what happens.

[Caddish]

According to the manual the other 4xPin fan header is suppose to be speed control but when I plugged my non-pwm s-flex 1600rpm it lowered it to 500rpm but did not increase with temp. So my father soldered both fan together and now they are voltage regulated by the CPU header. Careful as always he cut the fan wire in the middle so now I can't hide the cable, actually half an inch shorter and I would not have been able to plug the fan. But at least my PC is now silent enough that I can hear the pump and HDD on idle . I'm tempted to put the pump on a miniNG and maybe I'll add another rad and better rad fan. I can't find any comparison between the apache akasa and the gentle typhoon