Sign in with
Sign up | Sign in
Your question

Setting up a push/pull configuration with a Corsair H60

Last response: in Overclocking
Share
December 15, 2011 5:42:15 AM

I have a question about the fans used regarding a push/pull setup. In this particular situation, even more so with the "push" portion, it seems to me that the static pressure of the fan becomes very important. In essence, its like torque. Correct? The ability to overcome resistance. Here are the provided stats for the fan included with the H60.

Fan Speed: Up to 1700 RPM -
Fan Airflow: 74.4 CFM -
Fan dBA: 30.2 dBA -
Fan Static Pressure: 3.2mm/H20 -

Seems like a decent fan. Could find better, but not bad for stock. I haven't found conclusive evidence, but most websites list the included case fan (Antec 300 Illusion case), say that it puts out comparable cfm, but has .1 INCH / H20. ( 3.2 millimeter = 0.125 984 251 97 inch), so not as good. Ok, on to my questions.

#1, should I put the Corsair fan pushing and the Antec fan pulling? Or to clarify, the Corsair first pulling in the air from the back of the case, pushing it through the radiator, and then the Antec fan finishing the job? Also, any suggestions on attaching a filter, if at all, since the rear port was originally intended for exhaust, not intake.

#2, my motherboard (P8Z68-V/Gen3) has 2x 4 pin fan connectors. Should I hook up both fans to the motherboard, and the water pump to the power supply? I'm not familiar with the H60 system yet (arrives Saturday), but any suggestions are welcome. Share your own water cooling solution please. Thanks.
a b K Overclocking
December 15, 2011 9:44:07 AM

Generally its recommended to use two fans that are exactly the same when using a push/pull set up.
m
0
l
a c 324 K Overclocking
December 15, 2011 12:40:44 PM

You don't 'have' to use the exact same fans...just ensure they are moving air in the same direction...flowing from one side of the cooler through the other. If you want motherboard PWM control of your fans, you can use the fan headers, but if you want them running 100%, just connect them to a 3-pin to molex connection and be done. Another option would be a fan controller.

Regardless of what you decide, just make sure that one fan blows air into the radiator, and the other one pulls it out.
m
0
l
Related resources
a b K Overclocking
December 16, 2011 8:44:14 AM

rubix_1011 said:


Regardless of what you decide, just make sure that one fan blows air into the radiator, and the other one pulls it out.



Haha yeah if nothing else, at least make sure the direction of air flow is right :lol: 
m
0
l
a c 324 K Overclocking
December 16, 2011 12:44:42 PM

That really should be the only issue here. Fan speeds or types really isn't that important in push/pull on a rad. Yes, it's 'ideal' if you have fans of similar speed and static pressure, but there will be little to no measurable difference if you use mismatched fans.

We're dealing with case fans, here...not rocket surgery.
m
0
l
December 23, 2011 2:24:46 PM

Oddly enough, I was curious if there was any effect with fan speed vs. direction in the amount of cooling possible and I tested the same.

The best cooling solution I found when still using the thermal MB settings to control fan & pump speed at any given temp was surprising.
I say this because in any other config, temps were "never seen" to be as good as these unless achieved in this manner.

Corsair H60 only. Rear mounted, verticle position, stock cooling paste with a thin damp paste smear.
In a case with a top mounted 140mm fan and using all mainboard connections (thermal control) - Best result achieved is:

- The direction should be to pull cool air into the case; through the radiator; and out the top of the case.
- The push fan should range at speed 900 rpm - 1200 rpm (thermally ramped/controlled).
- The pull fan should range at speed 1200 rpm - 1600 rpm (thermally ramped/controlled) an average idle speed 300-400 rpm higher.
- The pump speed should range 1750 rpm idle - 2100 rpm (thermally ramped/controlled)

This resulted in a best average cooling temp of -10 degrees below the ambient temps at all but the highest temps seen which was -4 & -5 degrees below the ambient temps at full workload. All temps were thermally controlled using the Asus Probe II utility with mainboard connections.
Two identical 120mm front intake fans on a fan controller running at a constant 1400 rpm throughout (quiet).

Surprised me a great deal because I thought the same thing. That it really wasn't important as long as air was moving in one/same direction.
Much better results here and a much more stable thermal range condition throughout the entire range of temps seen.

Since this question was here. I thought you might want to know.

NOTES: As an afterthought I should say the CPU was at 3.564Ghz, a mild O.C. Even at 4.2884Ghz (moderate O.C.) the result was a simular one.
That was approaching room ambient temp and hovered there until fully loaded which climbed +2 & +3 degrees over ambient temp max.

To comment on this, I think this shows that running all fans flat out, can in fact prevent optimized cooling on these CLC's. Most likely due to a combination of less time to both "absorb" and to "cast off" the heat. I'd say giving the cooling medium more time to absorb heat and time to eliminate it does make a big differance in any given result. I also feel the pump's head pressure probably operates best in these RPM ranges. Establishing a slight vacuum between the fans for the radiator probably allows for more heat to be absorbed and dumped into the air more quickly.

NOTE: Edited to reflect more accurate temps NOT under ambient range.
m
0
l
a c 324 K Overclocking
December 23, 2011 3:40:25 PM

Quote:
This resulted in a best average cooling temp of -10 degrees below the ambient temps at all but the highest temps seen which was -4 & -5 degrees below the ambient temps at full workload


You can't cool below ambient with air coolers, closed loop coolers (H60) or with normal watercooling. It's impossible.
m
0
l
December 23, 2011 4:14:58 PM

Well, "the proof is in the pudding I'll say". I did edit the post to include more information.

Even at worst case (say the ASUS utility is flawed) a benchmark is still a benchmark as long as its used for everything.
I'll stand by what I say. My room temps were 77 degrees F. The utility indicated 67 Degrees F. Not C and yes it was configured to display Fahrenheit in the settings. I have no reason to state that which was not seen by my own eyes. Yet along publicy embarass myself.

On what basis do you call me a liar?
Your own expierence? Your qualified opinion?
These were my results. And yes, that room corner is a bit cooler than the rest of the room. But --- I stand by it.
m
0
l
a c 324 K Overclocking
December 23, 2011 4:36:00 PM

It's simply a law of physics- you cannot achieve cooling lower than the ambient temperature in the room unless the cooling agent has the ability to remove more heat than what exists at a normal, rest state at which the cooler would be in (ambient temperature)...you'd have to have the ability to remove more heat than is being produced by the components in question as well as remove additional heat watts contributed by ambient air.

Liquid cooling is similar to your car's cooling system- it does not cool your car's engine below what the actual external air temperature is; regardless if it is summer or winter. All forms of cooling listed above- air, water or combinations of the two cannot cool any component below the air temperature in the surrounding area. Even if you put your PC outside in the winter, and let's say it's -10C, the coolest your PC could be cooled is -10C at 100% efficiency (which is impossible due to several reasons). You also have to take into account that water temperatures inside a liquid cooling solution can also never go below ambient unless you have a component that requires extra energy to remove this additional heat in watts. (phase change, water chiller as examples). Water temperatures in side a loop at equilibrium is different than the temperatures listed by your hardware's thermal sensors. The calculation of what these temperature differences are, as well as the inclusion of ambient temperatures, total heat in watts being produced, flow rates, and total ability of radiator dissipation properties works out to be your delta-T of the loop.

I'm not trying to call you a liar- so please don't take so much offense to my debate. What is likely occurring is flawed temperature readings and/or display of these readings. The very best you could achieve would be 1:1 exactly with ambient air temperature- which is what happens in the best case scenario with a good cooler and with hardware at idle.
m
0
l
December 23, 2011 5:05:04 PM

I'll grant most of that. There are many scenarios to cooling below ambient temps. Swamp coolers and heat exchangers to name but only two.
And yes they do require the input of more energy to go there, just like a radiator requires fans.

I do appreciate your sentiment. Thank You!

I am definately willing to concede flaws in the ASUS utility. I've never know any of these types of things to be "precise wares". However I do wish to highlight the end result - absent of any "real temp accuracy", is still the result. A "best condition" was achieved regardless of what the "true" (as in precise) temps really were and that IS my point. The indices were the indices, so one must conclude that the best condition was achieved, using what is at hand, using the same benchmark, in the same conditions with only variations to the specific speed vs. direction.

I don't feel that any "exempt precision" in any way, changes the results except in definition only.

Do you agree?



m
0
l
December 23, 2011 5:28:03 PM

For the sake of getting to "precise" temps.

I'll replicate the various testing using "certified precise" thermocouples on a Tektronics tester as soon as I can make time.
m
0
l
a c 324 K Overclocking
December 23, 2011 8:11:55 PM

ASUS Probe is very inaccurate. Try RealTemp or CoreTemp to see what values you get reported. I understand your methods- they are correct in what you were trying to achieve. I believe the flaw lies in the actual reporting software, not your approach.
m
0
l
December 27, 2011 2:52:47 PM

And.. indeed it is...

I had a chance over the weekend to spot probe the system with both a laser thermal probe and thermocouples (but only in the best result configuration as it stood). You can add about +12 degrees F to the final result seen. Next, is to try to best that but even those temps are pretty good temps considering. I'll let you know if I do achieve better ones but I suspect to see the same results, now only with more accurate temps in the ranges seen.

I'll still say for now, it's a "best configuration" for the H60 with push/pull fans.
m
0
l
a b K Overclocking
January 26, 2012 12:15:37 PM

i would just put the faster one pulling through the radiator and the slower one pushing it. just seems to make sense to me. you could always do two setups and show us the results :) 
m
0
l
February 1, 2012 9:00:32 AM

The ideas/observations on this thread have given me a few things to think about as I ponder how to improve the performance of my H50.

About 2 years ago I made a newb inexpensive attempt at LC for the first time, after doing a little research of course. I hadn't done a new build in a few years.

Was (remains):

1. Gigabyte 890FXA-UD5 (it has a few cul-de-sacs technologically but it still won out over competing boards 2 of which were RMA'd back to mfr./Amazon b4 (finally) I gave up.
2. Silverstone Strider Plus for power.
3. AMD Phenom II X4 965 3.4 GHz.
4. The Corsair H50 stock configuration for cooling.
5. 2 Dimms (4g) of 'good' RAM. I can't remember. I'd have to grab one and take a look at it. Quick. It could use another one (a shot in the arm), but it doesn't seem to need it.
6. ATI HD5770.
7. An old 128 bit PCI sound card I got on eBay years ago (around the time I bought the case). It drives an equally old Klipsch home theater 4.1 I have, which is never turned up louder than half way as I just don't need to.
8. 2 HDD's. Both SATA, one a 500gb Seagate, the other a 250gb WD (black) which is a little faster. No hardware or software raid--just don't really need it. Striped RAID (even hardware based) doesn't seem to dramatically improve overall system speed/performance (in my experience) so I figured why bother.
9. The hard part: case is an old CM Stacker from several years ago. Steel frame, aluminum panels, heavy. Has a lot of little holes in it, a small (80mm) fan on the ceiling. For reference:
.

This was actually not a *terrible* case at the time. It seemed to offer (more or less) what I was looking for, and still has plenty of room for HDD's and I'd guess about 5 more case fans than I have ever had in it. As it is, it has 1 80mm fan in the middle of the side panel, 2 card/slot fans pushing exhaust out the back (sandwiched strategically between the sound card and video cards), a 120mm up front for intake, and the H50 (configured w/single fan only), and *pulling* air in from outside the back of the case.

Lately, HWMonitorPro and Speedfan are both saying all 4 core temps are evenly at around 41C when idle. It climbs after several hours, and climbs more noticeably if the GPU is having to really do any work. I haven't had time for Fallout or Crysis or anything like that in quite a while, so I hesitate to think what might happen if I really put it to work. On top of all that, AMD's own metrics for the CPU are giving me little warnings (!): "Your CPU fan is not working. You need to turn off your system." Evidently the H50 is not working as well as it once was. I used the single (only?) 120mm PWM fan that came with the H50 when I connected everything up.

Everything in the case accumulates quite a bit of dust. It's heavy and I'm lazy. So I usually wait until I can't put it off any longer (like...now) I find the time, unhook everything, take it outside and blow it out. The build was about 2 years ago now. I do not *think* any of the fans are failing, nor do I have reason to suspect their performance has been impeded significantly enough to explain the unusually high temp. readings. So far the max of those as measured by SpeedFan and HardwareMonitorPro has been about 46C.

Pulling cold in from the outside seem to make a lot more sense to me, intuitively, than exhausting warm air (overcoming the positive pressure gradient is point well taken). This afternoon I ordered 2 new silent hi-air flow, 120mm fans with PWM (a CM and a Nactua), and 2 new 80mm fans (similar). Thinking about this from a pressure and static/dynamics perspective, I wanted to get the most out of it if I was going to (economically, as always) attempt to:
1. Make the case quieter.
2. Get the internal temps (particularly core and GPU) down lower.
3. Make the H50 more efficient, while moving more air through it.
4. Avoid counterproductive turbulence acting against the first 3 priorities.

So thank-you, guys, for giving me some things to think about--very much appreciated. If you have any other ideas or suggestions, that would also be very much appreciated. I am going to test out a few different scenarios. I might wind up putting 2 identical Nactuas on the H50, or not (still thinking about it).

The PSW wound up only working in a kind of heingemacht fashion because the exhaust for it is on the top side of the unit, and the (ceiling) case exhaust is ~6" away from it. Top it all off the PSW does not actually level-fit the enclosure now for it now so as it stands the PSW is just kind of 'resting' on the shelf its in--more gaps introducing dust and causing unwanted noise to emit from the case.

I know I need a new case but I just kind of like the old thing. I'm by no-means disappointed in it.
It doesn't deserve to be relegated to the position of ashtray and I refuse to take it out in the desert and pump 12 gauge slugs into it. It's been reliable if nothing else. It's tough as nails and heavier than a Russian tank. Right now it's also free and a new Lian Li or something like that is going to set me back more than I want to part with at the moment.

Thanks for taking the time to read this guys and please let me know if you have any further thoughts/suggestions/ideas.
m
0
l
February 1, 2012 1:40:25 PM

michaelchute said:
I have a question about the fans used regarding a push/pull setup. In this particular situation, even more so with the "push" portion, it seems to me that the static pressure of the fan becomes very important. In essence, its like torque. Correct? The ability to overcome resistance. Here are the provided stats for the fan included with the H60.

Fan Speed: Up to 1700 RPM -
Fan Airflow: 74.4 CFM -
Fan dBA: 30.2 dBA -
Fan Static Pressure: 3.2mm/H20 -

Seems like a decent fan. Could find better, but not bad for stock. I haven't found conclusive evidence, but most websites list the included case fan (Antec 300 Illusion case), say that it puts out comparable cfm, but has .1 INCH / H20. ( 3.2 millimeter = 0.125 984 251 97 inch), so not as good. Ok, on to my questions.

#1, should I put the Corsair fan pushing and the Antec fan pulling? Or to clarify, the Corsair first pulling in the air from the back of the case, pushing it through the radiator, and then the Antec fan finishing the job? Also, any suggestions on attaching a filter, if at all, since the rear port was originally intended for exhaust, not intake.

#2, my motherboard (P8Z68-V/Gen3) has 2x 4 pin fan connectors. Should I hook up both fans to the motherboard, and the water pump to the power supply? I'm not familiar with the H60 system yet (arrives Saturday), but any suggestions are welcome. Share your own water cooling solution please. Thanks.


I'm not so familiar with Corsair H60 but I've used (2) hybrid liquid coolers in the past (a CoolIt Vantage ALC for a long time & an Antec Kuhler H2O 920 for a short time prior to going full liquid cooling). The H60 is essentially the same as any other liquid coolers in the market now as this hybrid liquid coolers were either made by CoolIt or Asetek & branded by either Corsair or Antec before being sold in the market.

The thing that I've noticed with CPU hybrid liquid coolers is that the fan is CPU controlled. In the case of CoolIt Vantage ALC you set the fan speed level via set bottons on the CPU liquid cooling block (quiet, performance, extreme). With the Antec Kuhler H2O 920, the fan speed is set via USB connection on the motherboard (you do the setting via a downloadable app in the CD that comes with the product. In order for the hybrid liquid cooler to work, it has to be connected to the PWM CPU (4-pin) fan connector. This way if the motherboard senses that the CPU temp is too hot, the set hybrid liquid cooler fan speed kicks in according to user setting.

The maximum fan speed on the fans that comes with the product is around 2,500 rpm. This can get very loud if set at extreme level & the fan runs at full speed. Don't even bother of changing the fan with the main purpose of limiting the noise level. Even those fans advertised as silent fans can be very noisy & in my experienced, I barely noticed any difference.

Yes, you need to connect one of the fans to the PWM CPU liquid cooler wire connected to the CPU for the liquid cooler to work. Otherwised, you'll get beeps error in installation. The 2nd fan you can connect to the other CPU fan connector or you can buy a PWM wire splitter to spit the fan signal to equal the push & pull fan speed.

As what Rubix said as long as both of your fans move air in the same direction, you'll be fine. My take with what some experts said about variable speed is that it's pure BS. Unless you're running fans with extremely high airflow (CFM) then you'll be fine or maybe not at all as you'll get so crazy being anal with fan speeds almost constantly adjusting the push & pull fans if you keep on making adjustments. Just keep the push & pull fans on same speed, period.

The pump inside the hybrid liquid cooling unit gets its power from the 12 volts CPU fan connection so you don't connect it to the PSU but to the motherboard CPU fan connection (same wire the hybrid liquid cooling fan).
m
0
l
February 1, 2012 1:44:41 PM

In terms of attaching filters, you really dont need to attach one if you're going to attach it on fans that move air going outside, not inside. Unless you're moving air going inside (like intake fans) then you're not really introducing dusts but blowing it away outside of your case.
m
0
l
a c 324 K Overclocking
February 1, 2012 1:58:08 PM

Filters just reduce airflow. If you are planning on maintaining a PC, dust clean out should be part of your normal routine. If you forget to clean dust filters, they clog and become far more detrimental to cooling than not having them at all.
m
0
l
March 28, 2014 3:41:57 AM

I use the H60 v2 with a push pull setup (blowing air into the case) and notice temp drop of about 8 °c instead of using just one fan. The push fan is the default corsair h60 fan, the pull fan is a scyte fan running @ 950 rpm (it can't go higher).
m
0
l
!