Jump to content
Corsair Community

Better Cooling fans needed


Highroller59
 Share

Recommended Posts

Hello All,

 

I have recently built a new Home theater system. The system is working great but I'm having some cooling issues. The configuration is in my signature. I used the Carbide Air 240 case, configured with the window at the top. The CPU is not overclocked and is cooled by Corsair H100i V2 Hydro system with the stock fans (loud) mounted in the front of the case (power switch/audio/usb ports front). The H100i V2 fans are sucking air from the outside, through the Radiator and into the case. The 3 stock fans the came with the case are installed and configured as follows. Looking from the front of the case, I have 2 installed on the left side sucking air into the case, on the right side I have the one remaining fan blowing air out of the case. The power supply fan is facing the bottom of the Mother board right under the M2 ssd and sucks air out of the case and heat away from the M2 seeing 13*C on the M2 in Corsair link which is fine with me.

 

Now to the temp issues. The motherboard only supports PWM fans. The stock case fans are three pin and run 100% all the time (12 to 1400 Rpms) The H100i V2 is connected to the CPU_OP w/pump Header and is working as designed. The temps I'm seeing during Plex transcoding are in the for the motherboard are 27*C idle and under load 20 to 40% 35 to 45*c. The CPU during idle is running 27 to 29*c and during load runs 40 to 50*c. I would like to cool this down a bit if possible with better and hopefully quieter fans.

 

My motherboard other than the Water Pump header only has two fan ports CPU_Fan and Chassis Fan and only one USB 2.0 header which for monitoring CPU cooler and power supply I need two. So for the sake of expandability I'm looking at installing a Commander Pro. for fans I am looking at 4 AF120's two sucking in and two blowing out and two SP120's to replace the hydo cooler fans.

 

Looking for suggestions are the selection of fans really that much better than the stock fans.

without going for the best Noctua fans do you suggest better fans than my selections. I need the case fans to move alot of air and the Hydro fans to have an enhancement of pressure.

 

Thank you for your time.

Link to comment
Share on other sites

Your motherboard fans may be 4-pin and support PWM, but you should be able to set them to DC control mode either in the BIOS or on AsRock's Windows utility.

 

CPU transcoding is going to heat up your processor. This will be true whether you are using Intel QuickSync (HW Transcoding in Plex) or not, though QuickSync won't heat it up as much. Your temps during transcoding sound like you are using the hardware transcoding and are, honestly, pretty much in line with what I'd expect for hardware transcoding. Fans won't help you here either; the thing that cools your processor is the liquid in your cooler ... and since you have your radiator on intake, the waste heat from your processor is actually dumped back into the system via the liquid.

 

Ideally, if you can, configure the radiator as exhaust. You don't have a discrete GPU in there to heat up the internal case temperature so, without the radiator dumping waste heat into the case, those internal temps won't rise too much. If I'm understanding your specs and install correctly, you should be able to mount this on the "bottom" (left side of your installation).

 

With the radiator as exhaust, you likely won't need any additional fans. You can mount the ones included as intake, giving you positive pressure to keep dust at bay.

 

Now ... for the noise. Yeah, the stock fans are loud and you'll want quiet. I would recommend using the ML series fans. You can get the non-RGB fans in a 2 pack from Amazon for a reasonable price. Set it to quiet profile in Link and that should do the job pretty well. If you get standard SP fans, they can't be controlled from the pump header (it's PWM only). If you get the High Performance PWM SP fans ... well, they really aren't any different from the stock fans except for the colored ring and some vibration reduction.

 

Finally ... the challenge will be how you control the speed of the intake fans. You will likely want them to ramp up as the cooler fans ramp up (to keep a balance). You can do this loosely based on CPU temperature (there will be a delay, however) or you can do it with the CoPro. Some caveats and things to consider. IF you base the intake fans on the H100i v2 temperature (like the radiator fans), you'll need to have Link running all the time. IF, however, you base the fan temps on one of the temperatures from one of the CoPro's sensors, you won't need to run Link all the time; the fan curve will run in the hardware.

 

IF, however, you want to replace the fans, I would, again, go with the ML series fans. They will give you excellent airflow and, most importantly for a home theater system, be quiet.

 

The only thing left to consider would be cooling any mechanical drives that you have. If you have 3 of them stacked up, they could heat up. If possible, it'd be ideal to have a fan blowing over them to keep them at a reasonable temperature.

Link to comment
Share on other sites

As above, no need to buy more SP120s. That is the same fan you already have but without the decorative ring. The ML120 Pro would be the best choice in the Corsair line.

 

So your motherboard temp sensor ranges from 27 to 45C when loaded? That is a lot of internal heat. Presumably this is from CPU waste heat, but I would be interested in your H100i v2 Temp to compare. The air coming out the back of the radiator should be more or less the same as the coolant temp. If the coolant temp is substantially lower than 45C, that may suggest heat from other sources are contributing.

Link to comment
Share on other sites

Your motherboard fans may be 4-pin and support PWM, but you should be able to set them to DC control mode either in the BIOS or on AsRock's Windows utility.

 

CPU transcoding is going to heat up your processor. This will be true whether you are using Intel QuickSync (HW Transcoding in Plex) or not, though QuickSync won't heat it up as much. Your temps during transcoding sound like you are using the hardware transcoding and are, honestly, pretty much in line with what I'd expect for hardware transcoding. Fans won't help you here either; the thing that cools your processor is the liquid in your cooler ... and since you have your radiator on intake, the waste heat from your processor is actually dumped back into the system via the liquid.

 

Ideally, if you can, configure the radiator as exhaust. You don't have a discrete GPU in there to heat up the internal case temperature so, without the radiator dumping waste heat into the case, those internal temps won't rise too much. If I'm understanding your specs and install correctly, you should be able to mount this on the "bottom" (left side of your installation).

 

With the radiator as exhaust, you likely won't need any additional fans. You can mount the ones included as intake, giving you positive pressure to keep dust at bay.

 

Now ... for the noise. Yeah, the stock fans are loud and you'll want quiet. I would recommend using the ML series fans. You can get the non-RGB fans in a 2 pack from Amazon for a reasonable price. Set it to quiet profile in Link and that should do the job pretty well. If you get standard SP fans, they can't be controlled from the pump header (it's PWM only). If you get the High Performance PWM SP fans ... well, they really aren't any different from the stock fans except for the colored ring and some vibration reduction.

 

Finally ... the challenge will be how you control the speed of the intake fans. You will likely want them to ramp up as the cooler fans ramp up (to keep a balance). You can do this loosely based on CPU temperature (there will be a delay, however) or you can do it with the CoPro. Some caveats and things to consider. IF you base the intake fans on the H100i v2 temperature (like the radiator fans), you'll need to have Link running all the time. IF, however, you base the fan temps on one of the temperatures from one of the CoPro's sensors, you won't need to run Link all the time; the fan curve will run in the hardware.

 

IF, however, you want to replace the fans, I would, again, go with the ML series fans. They will give you excellent airflow and, most importantly for a home theater system, be quiet.

 

The only thing left to consider would be cooling any mechanical drives that you have. If you have 3 of them stacked up, they could heat up. If possible, it'd be ideal to have a fan blowing over them to keep them at a reasonable temperature.

 

Thank you, but the motherboard Fan header is set to DC in the bios. The only fan header that has this option is the CPU_OPT W/PUMP which is where the H100i is connected and running correctly. Seems Asrock have changed things abit, from what I've read. Their 4 pin fan headers will run 3 pin fans but only at 100% due to the fact they headers are sending a constant voltage when using 3 pin fans. The fans that come with the case are 1400 rpm fans and barely move enough air to move cigarette smoke, I've done this in a test. I'm not as worried as much about the CPU temps as I am the motherboard. I think better (Move more air) fans will lower the motherboard temps down to acceptable levels. I am going to reverse the fans (blowing out of the case) on the H110i as you suggested, but I feel the trade-off will be lower Motherboard temps at the cost of higher CPU temps. I'll let you know how it goes.

Link to comment
Share on other sites

As above, no need to buy more SP120s. That is the same fan you already have but without the decorative ring. The ML120 Pro would be the best choice in the Corsair line.

 

So your motherboard temp sensor ranges from 27 to 45C when loaded? That is a lot of internal heat. Presumably this is from CPU waste heat, but I would be interested in your H100i v2 Temp to compare. The air coming out the back of the radiator should be more or less the same as the coolant temp. If the coolant temp is substantially lower than 45C, that may suggest heat from other sources are contributing.

 

Yes I read in the FAQ the ML fans are the way to go. I let you know the coolant temps the next time the system is under load.

Link to comment
Share on other sites

Thank you, but the motherboard Fan header is set to DC in the bios. The only fan header that has this option is the CPU_OPT W/PUMP which is where the H100i is connected and running correctly. Seems Asrock have changed things abit, from what I've read. Their 4 pin fan headers will run 3 pin fans but only at 100% due to the fact they headers are sending a constant voltage when using 3 pin fans.

Interesting. That's somewhat unusual.

 

The fans that come with the case are 1400 rpm fans and barely move enough air to move cigarette smoke, I've done this in a test.

So that sounds like the AF Quiet edition fans. You also had a pretty unbalanced system, with a good deal more intake than exhaust. A better balance would probably move air better. You do want positive pressure but you don't want to overload the positive pressure ... any intake needs to overcome the internal case pressure.

 

I'm not as worried as much about the CPU temps as I am the motherboard. I think better (Move more air) fans will lower the motherboard temps down to acceptable levels. I am going to reverse the fans (blowing out of the case) on the H110i as you suggested, but I feel the trade-off will be lower Motherboard temps at the cost of higher CPU temps. I'll let you know how it goes.

Yes, more airflow will cool the motherboard. And better balance. Since you don't have a significant source of heat inside the case (no GPU), as long as you have good airflow, you'll have good internal temperatures. This internal case air will then be used to cool the radiator and shouldn't be to much higher than ambient.

Link to comment
Share on other sites

OK I have run some stress tests on the system monitoring with Corsair link and Hardware info both match so I will go with Corsair link.

The Motherboard has 5 temp sensors reads in corsair link. under 100% load for 20 minutes these are the readings.

Temp #1 = 29*c, Temp #2 41*c, temp #3 29*c, Temp#4 29*c, temp#5 41*c,

CPU temps = Core #2 41*c, Core #3 41*c, Core #4 41*c, Core#5 41*c, Package 40*c

H100I v2 Water temp 26.5*c Fan 2640rpm, Pump 2850rpm

Corsair link set to Performance Mode

Fan #1 1370Rpm, Fan #2 1277rpm, Fan #3 1400<<<< Not sure what Fan 3 is I only have 2 case fans.

Memory temps #1 31*c #2 31*c

SSD Temp 9*c

 

It looks like Motherboard temps #2 & 5 must benear the CPU and picking up Heat from the CPU.

How do I lower the CPU temps or can I! The longer the stress test runs the hotter the system will get.

 

Forgot to mention I have corsair link to ramp up to 100% at 38*c

 

Here are the idle temps in Corsair link.

Motherboard

Temp #1 = 27*c, Temp #2 29*c, temp #3 25*c, Temp#4 27*c, temp#5 29*c, VCPU .69

CPU temps = Core #2 27*c, Core #3 26*c, Core #4 28.5*c, Core#5 26*c, Package 28*c

H100I v2 Water temp 23.5*c Fan 1500rpm, Pump 2850rpm

Corsair link set to Performance Mode

Fan #1 1370Rpm, Fan #2 1277rpm, Fan #3 1400<<<< Not sure what Fan 3 is I only have 2 case fans.

Memory temps #1 25*c #2 25*c

SSD Temp 9*c

 

These temps are about 20 minutes after the stress test. I shut down the stress test before the temps went much over 40*c if I continue the system temps will keep rising, I have seen 55*c and will continue to rise. I'm not sure how hot it will get but I don't won't to go into meltdown.

Edited by Highroller59
Link to comment
Share on other sites

A lot of the Link temp readings are garbage and I have no idea where they come from. Same problem with the fans and duplicate readings. You mentioned your motherboard temp sensor ranging from 27 to 45C. If true, that is a lot and the clear place to attack to reduce the temperature of every thing in the case. What I was curious about is the H100i v2 temp under load. That does correspond with the motherboard temp sensor and at cool idle those two values should be similar. In effect, the local ambient temperature is setting the minimum possible coolant temperature, which is also the min. possible CPU temp. The same thing would hold true if the local ambient temp is 45C. Then your baseline temp for the CPU is 45C as well and that is the place to make gains.

 

One of the ideas mentioned was turning the H100i to exhaust instead of intake. To determine whether that might benefit, I was wondering about the coolant temp under load. If the local case temp is 45C, then the coolant temp will wind up there as well. On the other hand, if the motherboad temp sensor is near a heat source and giving a misleading reading and the coolant temp is 33C (or something similar), then you might be better off as you are but with more exhaust.

 

The mesh on the back of the case is limiting. You can put 80mm fans back there, but they are buzzy and annoying and this should not be necessary without a discrete GPU. Adding 2x120 to both top and bottom should help. I think my preference would be for radiator intake, then bottom intake and top exhaust. This should get heat coming off VRM modules out (likely the second biggest source of heat) and keep a general flow out the top and mesh back. The missing piece is your coolant temp. If you really are getting up to near 40C, then dumping that out may be better for the overall case ambient.

Link to comment
Share on other sites

What did you use for a stress test? Different stress tests have different thermal characteristics. Just saying "I ran a stress test" doesn't tell us that. There's are pretty significant differences between Aida and RealBench and Prime95, for example.

 

The motherboard temps you are seeing aren't picking up heat from the CPU. The CPU doesn't heat up the rest of the system; it's liquid cooled. They are likely picking up heat either from the VRM or the chipset. In fact, one of these is likely a temperature sensor on the chipset itself.

 

Unless your system is completely enclosed, your temps should, eventually, level out. They typically don't run away. This is where your thermal balance is and, on a liquid cooled system especially, you'll need to run a stress test for an extended period of time to actually hit it. Personally, I have to run RealBench, which stresses the CPU and the GPU simultaneously, for about 20-30 minutes or so before I hit this. CPU temps are the least stable but you'll typically hit your max temp pretty quickly.

 

Your vCPU readings indicate idle. What is the max value for this?

 

Also, I, personally, find it more valuable to record the value during a test run and then analyze them in Excel. I'll look for the max and average ... and also add rolling average trendlines. Looking at just a single temperature at a single point in time isn't really telling.

Link to comment
Share on other sites

A lot of the Link temp readings are garbage and I have no idea where they come from. Same problem with the fans and duplicate readings. You mentioned your motherboard temp sensor ranging from 27 to 45C. If true, that is a lot and the clear place to attack to reduce the temperature of every thing in the case. What I was curious about is the H100i v2 temp under load. That does correspond with the motherboard temp sensor and at cool idle those two values should be similar. In effect, the local ambient temperature is setting the minimum possible coolant temperature, which is also the min. possible CPU temp. The same thing would hold true if the local ambient temp is 45C. Then your baseline temp for the CPU is 45C as well and that is the place to make gains.

 

One of the ideas mentioned was turning the H100i to exhaust instead of intake. To determine whether that might benefit, I was wondering about the coolant temp under load. If the local case temp is 45C, then the coolant temp will wind up there as well. On the other hand, if the motherboad temp sensor is near a heat source and giving a misleading reading and the coolant temp is 33C (or something similar), then you might be better off as you are but with more exhaust.

 

The mesh on the back of the case is limiting. You can put 80mm fans back there, but they are buzzy and annoying and this should not be necessary without a discrete GPU. Adding 2x120 to both top and bottom should help. I think my preference would be for radiator intake, then bottom intake and top exhaust. This should get heat coming off VRM modules out (likely the second biggest source of heat) and keep a general flow out the top and mesh back. The missing piece is your coolant temp. If you really are getting up to near 40C, then dumping that out may be better for the overall case ambient.

 

I understand where your coming from. The thing that has me scratching my head is when you say fans in the bottom and top. The way I have to case configured is with the clear window on top, with that configuration you only have the option of placing up to two 120 fans on each side of the case, with the possibillty of two 80mm in the rear, the radiator is up front. Is there a basement fan location I have missed in the case, somewhere down where the PS lives?

Link to comment
Share on other sites

No, I wrote from the original orientation and in relation to the motherboard. It gets too wacky to talk about left/right when you can position it in any direction. Normal heat convection should be meaningless in a case this small and with multiple fans. There is no local temp advantage for bottom vs top, etc. You can compare this to a larger case like the 740 that has a 3C difference between top and bottom. That difference is muted when side lying.
Link to comment
Share on other sites

No, I wrote from the original orientation and in relation to the motherboard. It gets too wacky to talk about left/right when you can position it in any direction. Normal heat convection should be meaningless in a case this small and with multiple fans. There is no local temp advantage for bottom vs top, etc. You can compare this to a larger case like the 740 that has a 3C difference between top and bottom. That difference is muted when side lying.

 

I understand what your saying. Heat rising off the motherboard versus air flowing across and out. I had to configure the case with the clear window on the top due to space between two shelves in my entertainment system my TV sets on. I have thought about modifying the clear windows to accept a fan or two so heat would be vented through the top. This idea was a last resort but I am willing to do it, if you think it would help.

Link to comment
Share on other sites

I understand what your saying. Heat rising off the motherboard versus air flowing across and out. I had to configure the case with the clear window on the top due to space between two shelves in my entertainment system my TV sets on. I have thought about modifying the clear windows to accept a fan or two so heat would be vented through the top. This idea was a last resort but I am willing to do it, if you think it would help.

 

With good, balanced airflow, I doubt it'll make much difference. However ... on thing to note is that in an enclosed area, the local ambient temperature (the "source" of cool air for the system) can also increase. That's a challenge with most entertainment systems; they aren't necessarily well designed to keep airflow to the components. I don't know the design of your entertainment system so it may not be an issue in your case but it's certainly something to consider.

Link to comment
Share on other sites

I understand what your saying. Heat rising off the motherboard versus air flowing across and out. I had to configure the case with the clear window on the top due to space between two shelves in my entertainment system my TV sets on. I have thought about modifying the clear windows to accept a fan or two so heat would be vented through the top. This idea was a last resort but I am willing to do it, if you think it would help.

 

No, just the opposite. There is no need for that and in the small case you are emptying the entire internal volume at high frequency. There is nothing wrong with putting the 240 on its side, it was part of the design, and I even turn over the 540 and 740 versions that were not expressly fashioned that way. You are not taking a penalty by using it in this position.

Link to comment
Share on other sites

What did you use for a stress test? Different stress tests have different thermal characteristics. Just saying "I ran a stress test" doesn't tell us that. There's are pretty significant differences between Aida and RealBench and Prime95, for example.

 

The motherboard temps you are seeing aren't picking up heat from the CPU. The CPU doesn't heat up the rest of the system; it's liquid cooled. They are likely picking up heat either from the VRM or the chipset. In fact, one of these is likely a temperature sensor on the chipset itself.

 

Unless your system is completely enclosed, your temps should, eventually, level out. They typically don't run away. This is where your thermal balance is and, on a liquid cooled system especially, you'll need to run a stress test for an extended period of time to actually hit it. Personally, I have to run RealBench, which stresses the CPU and the GPU simultaneously, for about 20-30 minutes or so before I hit this. CPU temps are the least stable but you'll typically hit your max temp pretty quickly.

 

Your vCPU readings indicate idle. What is the max value for this?

 

Also, I, personally, find it more valuable to record the value during a test run and then analyze them in Excel. I'll look for the max and average ... and also add rolling average trendlines. Looking at just a single temperature at a single point in time isn't really telling.

 

I used Intel Burn Test & Heavy load 64 was looking for loads on cpu only. The CPU seemed to stabilize around 56*c the hot areas of the motherboard where nearly exactly the same temps from sensor 2 & 5. The idle of the VCPU is 68 to 70, under load this tops out at 1.09 with a quick 1.10 here and there. This is all through the full 10 minute intel test. Max CPU temp was 56*c max motherboard temp 56*c

Edited by Highroller59
Link to comment
Share on other sites

OK, whatever that is cannot be a the standard motherboard temp sensor. If I were guessing, it is a copy of CPU temp. Link seems to name everything "motherboard temp #(1-99). Maybe fully exit Link and use HWinFO, HWMon, or whatever you were using before. Those should list some motherboard specific values like VRM, PCH, and the actual board temp sensor. CPU temps are so load dependent. We need to look at the coolant temp delta to understand the amount of heat moving around and how much fan speed affects it. Some of those other programs can see your cooler as well, although not likely if both are running. Link is currently not compatible with just about everything else.
Link to comment
Share on other sites

OK, whatever that is cannot be a the standard motherboard temp sensor. If I were guessing, it is a copy of CPU temp. Link seems to name everything "motherboard temp #(1-99). Maybe fully exit Link and use HWinFO, HWMon, or whatever you were using before. Those should list some motherboard specific values like VRM, PCH, and the actual board temp sensor. CPU temps are so load dependent. We need to look at the coolant temp delta to understand the amount of heat moving around and how much fan speed affects it. Some of those other programs can see your cooler as well, although not likely if both are running. Link is currently not compatible with just about everything else.

 

I have hardware info the latest version and the temps match, but I didn't look for ghost fans or temp sensors. I'll try that and see what I come up with. For the Coolant temp what do you want to do?

Link to comment
Share on other sites

To all that helped I thought I would give you an update.

 

After every ones input I had some ideas, not exactly as suggested but it worked.

 

As noted above I have a Corsair Air 240 case configured with the clear window on top..I was seeing temps into the 40's and sometime 50's with minor to major loads.

 

I replaced all the stock case fans and the two H100i V2 fans with ML120 fans all connected to a Commander pro.

 

Two Ml 120's on the Radiator sucking in

two Ml 120's on the right side sucking in

Two Ml 120's on the left side blowing out

The side fans are directly across from each other and move air across the motherboards

 

With all the fans running at around 900rpm's I get the same cooling levels in quite mode as I was getting in performance mode and I must say so quite you can tell the fans are running.

 

Temp Probe 1 is the Radiator Intake

Temp Probe 2 is the Radiator exhaust

Temp Probe #3 is side case intake air temp

Temp Probe #4 is side case Exhaust air temp

 

I'm seeing an over-all reduction in temps, they where running when playing a movie in PLEX 40 to the 50's*c, now temps are running 29 to 31*c.

 

I'm seeing a 1 to 2*c temp difference in intake to exhaust temps on the Radiator and almost equal on the side fans.

 

These upgrade where in the $175.00 but well worth the piece of mind. If ever your upgrading your fans look into the ML120 fans and Commander pro you can't go wrong over stock.

 

I'm a happy camper!

Link to comment
Share on other sites

 Share

×
×
  • Create New...