Which fan headers to use for which components when watercooling

[WithAlligators]

Hey guys,
first off, here's my system:
Ncase M1 case
Maximus VII Impact
i7 4790K lapped / delidded
apogee drive II
ST30 240mm radiator
2 x Cougar Vortex CF-V12HP fans
Samsung 850 Evo 256GB
16GB Hyper X 1866

If you've seen the Ncase M1, the 2 fans sit on the side. Mine are intake fans, forcing air through the rad and onto the MB and then out the top.

I am wondering what is the best config for efficiency. Right now I have the pump hooked up to the CPU fan header, and each fan hooked up to a port on the Asus Coolhub. I can control each port through the BIOS.

Which is more important to cool the CPU during waterclocking? Increasing the speed of the pump as it gets hotter or increasing the speed of the fans on the rad?

The motherboard is getting kind of hot. During stress testing in Aida 64, it is getting up to 45C after 10 minutes. One of the fans pushes air right onto the motherboard, should I increase the speed of that one?

How much quieter are the Gentle Typhoon AP-15s? The cougars aren't too loud at 800rpm (the lowest they'll go with Q-fan) but they are far from silent.

Also, how hot is too hot for an SSD? Mine is sitting at 40C right next to the PSU.

 

[BonzaiDuck]

Which is more important to cool the CPU during waterclocking? Increasing the speed of the pump as it gets hotter or increasing the speed of the fans on the rad?

I don't think the answer needs to be "mutually exclusive" in its consideration of which RPM to increase.

AP-15 fans are the more acceptable GT fans for the noise factor compared to something like an AP-30 (which I myself use -- selectively).

I would then ask whether you've fully explored the thermal-control features of the motherboard and fan-plugs. Stress-testing doesn't yield "real-world" noise or temperature behavior. I would be more tolerant of noise while stress-testing than would I be for regular gaming or other uses. And the truth is, after you finish the stress-testing, you can further tweak fan settings to reduce noise for regular usage.

I'd hope the Maximus VII board -- even with a limited number of fan ports -- implements ASUS' thermal control fan features with the proprietary software for PWM fans.

But you really only need a single PWM CPU_FAN port, although two (including a CHA_FAN plug) would be better.

Use a device like the Swiftech 8W-PWM-SPL-ST splitter:

http://www.performance-pcs.com/swiftech-8-way-pwm-splitter-box-sata.html

Akasa and some other manufacturers also provide a similar product, and they are all cheap.

You could run both a pump(s) and fan(s) off the same splitter, powering them from the PSU directly but controlling them thermally from the PWM signal of a single mobo fan port. Since the control is provided as a matter of "duty-cycle-%" instead of RPM, both types of devices would increase speed relative to that %, whatever their absolute RPM range.

There is only a minor drawback: you can only MONITOR one device connected to the splitter -- which controls up to eight devices (fans and/or pumps). I'd suggest that you might want to monitor at least one pump in preference to fans connected to the same splitter.

With two such splitters and two PWM motherboard fan-ports, you can control fans/pumps in groups to define an intake-exhaust strategy for your case. But if using more than one pump, I'd control them all from the same splitter unless they are in different, isolated water-loops.

You don't mention graphics card(s). I WOULD suggest that you need higher fan RPMs pushing more CFM over the motherboard (and graphics card(s)?) to keep that assembly (mobo and graphics) cooler.

And given the limited size of the case (very nice, by the way), I'd focus initially on desired, achievable temperatures first, and worry about noise afterward.

 

[PhIlLy ChEeSe]

Word of advice, don't plug multiple big fans into one fan header on the MB. She will become a smoker....................:'(

 

[BonzaiDuck]

Word of advice, don't plug multiple big fans into one fan header on the MB. She will become a smoker....................:'(

Abso-freakin-LUTELY!

But he doesn't need to power ANY fans from the mobo with those $10 splitters. The splitters only use the PWM signal to control up to 8 fans. The power comes directly from a PSU SATA-plug.

 

[WithAlligators]

I admit, I actually haven't installed any of the Asus software. I have only been using the BIOS and the Q-Fan option. Q-fan spools the fans and such up and then determines the lowest settings they can be at, etc.

While stress testing at 100% (4.4GHz stock turbo) the cores hover in the sixties with the occasional spike into the high sixties and the fans are louder, certainly, but not super loud.

At 800 RPM on the fans, and the pump on minimum (which is silent indeed), at idle, the fans are a little more audible than I'd like. CPU temps are low to mid thirties. I have AS5 between the IHS and Apogee, though I may just put CLU in there as well.

The GPU I have now is a stand in R9 280X until the new AMD cards come out next month. I will then probably buy one of those or pick up a last generation card at a discount since the new gen will be out. I will also add another rad and fans and put the GPU in the loop as well.

Anyway, my question was more of a theoretical one than anything. When water cooling, what is more important, the fans blowing faster to push heat off the fins, or the pump moving faster to push heat off the CPU and into the water loop?

And the reason I mentioned the stress testing in the original post, is because it showed my motherboard getting quite hot. One of the fans blows through the rad and directly onto the board, and I thought increasing the speed of that might help cool it. Is there any way to tie one of the fan headers to that temp sensor? So that it scales with that one? Or should I just set that one a bit higher so it's always blowing a bit more onto the board.

And speaking of that, does anyone know what the Q-fan in the BIOS scales with? I assume the CPU temp but I don't know for sure.

Cheers

 

[BonzaiDuck]

I admit, I actually haven't installed any of the Asus software. I have only been using the BIOS and the Q-Fan option. Q-fan spools the fans and such up and then determines the lowest settings they can be at, etc.

While stress testing at 100% (4.4GHz stock turbo) the cores hover in the sixties with the occasional spike into the high sixties and the fans are louder, certainly, but not super loud.

At 800 RPM on the fans, and the pump on minimum (which is silent indeed), at idle, the fans are a little more audible than I'd like. CPU temps are low to mid thirties. I have AS5 between the IHS and Apogee, though I may just put CLU in there as well.

The GPU I have now is a stand in R9 280X until the new AMD cards come out next month. I will then probably buy one of those or pick up a last generation card at a discount since the new gen will be out. I will also add another rad and fans and put the GPU in the loop as well.

Anyway, my question was more of a theoretical one than anything. When water cooling, what is more important, the fans blowing faster to push heat off the fins, or the pump moving faster to push heat off the CPU and into the water loop?

And the reason I mentioned the stress testing in the original post, is because it showed my motherboard getting quite hot. One of the fans blows through the rad and directly onto the board, and I thought increasing the speed of that might help cool it. Is there any way to tie one of the fan headers to that temp sensor? So that it scales with that one? Or should I just set that one a bit higher so it's always blowing a bit more onto the board.

And speaking of that, does anyone know what the Q-fan in the BIOS scales with? I assume the CPU temp but I don't know for sure.

Cheers

Likely all the thermal fan-control references the CPU temperature. This may have changed since my sig-rig ASUS board, but I've also got a Z77 board and it seemed to be the same.

That's why -- just off top of my head -- I'd suggest using the CPU_FAN header (advisedly with a Swiftech splitter) and a CHA_FAN -- separately. You can then define two different fan curves for thermal control, which might give you flexibility to set the floor on intake fan speeds blowing on the motherboard a little higher. And of course, if you can use bigger fans for that latter purpose, they won't be as noisy even for spinning faster than their minimum.

Your only other option -- to link fan speeds to motherboard temperature -- would be use of an off-board fan controller equipped with thermal sensors that can be taped to board components. For me, this is the least favorable choice if a solution otherwise can be found with splitters and motherboard fan control: it adds complexity, it costs more -- especially if you it is capable of reading mobo sensor readings through a USB connection.

Also, I'd say your 45C stress-test motherboard temperature is only about 10C higher than what other folks have with bigger cases and better case airflow. If it never exceeds the mid-40's C, I don't see it as a major problem.

 

[aigomorla]

Right now I have the pump hooked up to the CPU fan header, and each fan hooked up to a port on the Asus Coolhub. I can control each port through the BIOS..

please tell me you have the PWM sense header connected to the board, while the power draw is on 4 pin cable...

The Apogee uses a pump which can draw too much on the fan header alone...
That is why it has 2 cables.

Anyway, my question was more of a theoretical one than anything. When water cooling, what is more important, the fans blowing faster to push heat off the fins, or the pump moving faster to push heat off the CPU and into the water loop?

lol... thats like asking... what gives me more Torque? Having the engine rev really high, or having a better gear / power ratio in your transmission.

The answer... both....
Faster fans means more static on the fins, which in terms means the coolant inside the rad gets cooler until it hits ambient.
So having the most static on the fans will indeed help with coolant.
Water pump moving faster, has to do more with FLOW.
The greater the FLOW in the system, the higher holding potential water has, however the returns on this value can get to dimished values very quickly.
Typically one would require a really small loop and very powerful pumps to reach area, so your still in the golden zone.

But what is more important in a LC system? id say both is... you cant cut one without drastically reducing the overall efficiency on the entire closed loop system.
The pump can be downvolted when the cpu and gpu are in idle.... because heat values are low, hence not requiring the holding capacity, which translate to a quieter system.

However if ur gaming, burning all 4 cores and GPU, you would want to maximize both fans on radiator, and pump rpm to work with the increased system, and overall boost efficiency in the LC system.

 

[WithAlligators]

Whoa, I missed this. I know it's both, but I was wondering which is MORE important. Anyway, I've been playing with settings for a bit. I can keep the pump on minimum duty (like 18%) and only scale it up when the CPU gets over 75C, and it'll never get there. Aida64 for an hour and it only hit 70 with average temps in the high 50s. My mainboard will hit 50C though if I'm stressing the GPU, so that's a little worrying.

I feel like flow isn't as important, as flow too fast pushes the water back through the block before it's had sufficient time to cool in the rad? Especially in a setup with a reservoir and 240mm rad, low flow seems fine. I also have my fans set a little lower, but scaling up a bit earlier. Still pretty quiet.

So, if I were to shell out for some Gentle Typhoons, should I go AP-15s or 14s? Quiet is my goal. Also, the only header that the BIOS lets me control via DC and not PWM is the CPU header, as far as I can tell. So, I need to use a Fan Splitter with all fans to the CPU header, and power from the PSU, right? Then I can use the PWM off the pump to one of the chassis headers? Cause it all scales from the CPU temp, correct?

Cheers
Alex

 

[BonzaiDuck]

You seem to be going in the right direction. You'll want to fiddle with the fan-control etc. until you find the right combination.

Personally, I think it's not a bad idea to moderate the CPU idle temperature at 40C, if you're trying to mitigate noise. I think last year we discussed how to mount pumps for acoustic advantage, so it might not matter as much running them at a higher speed.

 

[WithAlligators]

Well, I'm using the Apogee Drive, which is mounted to the CPU block, so I don't have options there. Anyway, I have everything running as minimally as possible during idle (and even load, as it never gets to the point where it scales up), to mitigate noise, and idle varies between 28 and 35ish depending.

So here's a question about the Gentle Typhoons, as it's something I've never dealt with: If I use a splitter, with power drawing from the PSU, how do I control the fan speed, since they aren't PWM fans? Are there splitters that support DC control and somehow modulate power from the PSU? I ask this because I'll be putting a total of 4 fans plus the pump in there, and there are only 4 headers in total. Maybe I can run one off the GPU if the header isn't covered by the waterblock

Or can you guys suggest quiet and powerful PWM fans with high static pressure?

Cheers

 

[BonzaiDuck]

Well, I'm using the Apogee Drive, which is mounted to the CPU block, so I don't have options there. Anyway, I have everything running as minimally as possible during idle (and even load, as it never gets to the point where it scales up), to mitigate noise, and idle varies between 28 and 35ish depending.

So here's a question about the Gentle Typhoons, as it's something I've never dealt with: If I use a splitter, with power drawing from the PSU, how do I control the fan speed, since they aren't PWM fans? Are there splitters that support DC control and somehow modulate power from the PSU? I ask this because I'll be putting a total of 4 fans plus the pump in there, and there are only 4 headers in total. Maybe I can run one off the GPU if the header isn't covered by the waterblock

Or can you guys suggest quiet and powerful PWM fans with high static pressure?

Cheers

The Gentle Typhoons -- AP-30 (hard to find anymore), AP-15 and similar models come as either PWM or 3-pin. While I can only confirm this with certainty for the AP-30, it and likely any other AP- -- look at AP-15 -- have the solder pads exposed which can be modded -- wired -- to take advantage of a built-in but unused PWM capability. I'm only extrapolating this from what I know about the AP-30, and someone else can correct me -- I will "stand corrected."

Otherwise, only order PWM fans. If you can't get the GT AP fan you want, look at the Noctua iPPC 120 and 140mm fans. I had no hesitation about getting a 3,000 RPM iPPC PWM and using it as a CPU pusher fan.

The only quick and dirty way to split a 3-way fan other than to use an aftermarket controller would be to wire the fans together in parallel to a single 3-pin plug to the motherboard. You would only use a tach wire from one of them -- leave the other disconnected, or put it on the tach-pin of a different unused motherboard fan port. I had done this in the past, it "works" but I was less than happy with it. Especially, the total amperage of the two fans cannot exceed what would likely be a 1.0A limit per port.

Instead, buy PWM fans with some thought toward a strategy of "group fan control," wire them to a Swiftech 8W-PWM-SPL-ST 8-port splitter connected to (a) the CPU_FAN or CHA_FAN header (if 4-pin PWM) for the signal only, and (b) with power from an SATA plug from the PSU. You can only monitor the "lead" fan -- wired to a port initially capped with a red or white plastic protector. (the rest are black). Monitoring is one thing; control is another -- and all the fans (different sizes, speeds and amperage as desired) or pumps will be controlled according to the duty-cycle % determined from the PWM signal wire of the motherboard.

Any 3-pin fans in the mix I would hope there are enough motherboard 3-pin ports to avoid the soldering and parallel wiring kloodge. I suppose the last resort with no more PWM or 3-pin plugs available could be a simple wiring to the PSU via Molex or SATA, and then you'd want those particular fans to have top-end performance at relatively low RPM and airflow. Of course, the larger the fan, the lower the top-end RPM and the less noise, but your case will dictate what size of fans you use for intake or exhaust purposes.

If you wire any fans together, make clean, solid solder connections.

 

[WithAlligators]

The PWM mod only works with the AP-30s and one of the other high RPM fans as far as I know. It doesn't work with the lower RPM ones. And from what I've read, the AP-30s aren't very good at low RPMs.

Here is what I think I will do. Each AP-15 is rated at very low amperage, but has a .36 startup surge. I only need 4 fans, so I'll use 1 Y splitter for the bottom fans. That will keep even the startup surge below 1A. How does that sound?

Cheers

 

[BonzaiDuck]

The PWM mod only works with the AP-30s and one of the other high RPM fans as far as I know. It doesn't work with the lower RPM ones. And from what I've read, the AP-30s aren't very good at low RPMs.

Here is what I think I will do. Each AP-15 is rated at very low amperage, but has a .36 startup surge. I only need 4 fans, so I'll use 1 Y splitter for the bottom fans. That will keep even the startup surge below 1A. How does that sound?

Cheers

== MY PUZZLEMENT ==
I may hold unconventional views about fans.

I'm running a PWM AP-30 in each of two systems. So I'm curious -- what does it mean that "AP-30s aren't very good at low RPMs?"

In my fan-control configuration, I think I've had them set to a %-duty-cycle floor that would run them at anywhere from 1,100 to 1,700RPM. They are noiseless in my systems at somewhat higher speeds than that, and the acoustic mods I've made don't show any noise into the low 3,000's.

So . . . what is "not very good?"
===================

Thanks for the clarification on the PWM mod, and I can see how such might be the case. A lot of folks here have chosen AP-15's for this and that.

Is this Y-splitter designed for the usual 3-pin fans? If the AP-15's are PWM, the Swiftech and similar splitters are <= ~$10.00/each. Then you won't have to worry about motherboard fan-port amperage limits.

Here's a 3-pin fan-splitter I found at the Egg -- didn't know they even made these. It actually looks as though the tach-wires for two fans are both connected to the same mobo pin.

http://www.newegg.com/Product/Produ...gclid=CL7y94vr38UCFQSVfgod_xUAAg&gclsrc=aw.ds

I've done it -- for two and three-fan configurations to one port respectively. I'm just not very enthusiastic for it.

 

[WithAlligators]

I honestly have no experience with Gentle Typhoons at all. I couldn't find PWM versions anywhere, except for the mod, so I assumed they didn't make them. Did they?

I also have no evidence that the modded 30s aren't very good at low RPMs. Are you saying they are? I've heard they are quiet, but because they lack the same amount of blades as the low RPM ones, they simply can't push as much air or generate as much pressure as the low RPM ones when running at low RPM. Do you disagree?

As far as that Y splitter, they do make them with only one tach wire from one fan going to the mobo:
http://www.amazon.com/Splitter-Exte...8164&sr=8-8&keywords=3+pin+fan+cable+splitter
which is what you would want, I imagine. The mobo would have trouble reading the speed if there where two signals mixed on the same pin. No?

 

[BonzaiDuck]

I honestly have no experience with Gentle Typhoons at all. I couldn't find PWM versions anywhere, except for the mod, so I assumed they didn't make them. Did they?

I picked up two AP-30 PWM fans two years ago. They were getting hard to find. The OEM manufacturer is Nidec Servo, and I have no idea why they're scarce now -- they just are. I can't find them anymore.

But they were sold as 3-pin as well as PWM, and the 3-pin version was PWM-ready for the minor soldering mod that would enable it.

I also have no evidence that the modded 30s aren't very good at low RPMs. Are you saying they are? I've heard they are quiet, but because they lack the same amount of blades as the low RPM ones, they simply can't push as much air or generate as much pressure as the low RPM ones when running at low RPM. Do you disagree?

Like I said, I picked up a pair because I was making enhancements to my heatpipe cooling. I never saw any difference in the shape or number of the blades, and they don't give me any airflow problems at the lower RPMs.

As far as that Y splitter, they do make them with only one tach wire from one fan going to the mobo:
http://www.amazon.com/Splitter-Exte...8164&sr=8-8&keywords=3+pin+fan+cable+splitter
which is what you would want, I imagine. The mobo would have trouble reading the speed if there where two signals mixed on the same pin. No?

That's what I would've assumed without further information. If they're selling those units the way they are and there aren't any customer-review compliants suggesting a problem, there must not be any. But if I were to make my own splitter, I'd pick one fan to monitor, use its tach-wire exclusively, and only wire the red and black wires in parallel. Or I believe you could run tach wires from other fans on that string to unused fan ports.

But for splitting a single 3-pin mobo port or using a port that swings either way for 3-pin fans, I'd use identical fans - or no cigar. Since they'd be identical, you'd only need to monitor one of them because you'd casually inspect them to see they're all running properly during the periodic computer-case cleaning exercise.