Trouble with heat... Need some ideas.

[Crescendo]

I have a Phantek Evol ATX case with 2 x 140mm SP fans in the front sucking in and 1 x 140mm in the rear sucking out. I have a 120mm radiator uptop with 2 Corsair SP fans blowing air out through the radiator (push configuration.) the cooling solution is a Corsair H110i GTX AIO cooler. I run a GTX 980TI Hybrid from EVGA.

GPU stays around 50-55C under load, but the CPU is the main concern, it goes as high as 60c stock while gaming. If I overclock to 4.5GHZ @ 1.275v it goes as high as 80-85c. I know for a fact my cooling should be doing much better than this, I didnt spend extra on a case/motherboard/fans/liquid cooling and a K processor to NOT be able to overclock 500mhz respectively.

So I have my 120mm radiator mounted uptop with a push configuration, a rear 140mm ran exhausting and 2 x 140mm intakes in the front. I also have my Hybrid radiator mounted as an INTAKE in the front. Should I swap my Hybrid radiator to the rear and use it as a outtake where the 140mm one is? I should not be seeing 80-85c on water cooling @ 1.275v, my 6600K on air cooling in a crappier case was able to do 4.5 @ 1.270 and stay in the 70's so what gives?

I know overclocks arent gauranteed but when I spent an extra $50 on my motherboard, $50 on my CPU, $100 on cooling and I can't even overclock to a measily 4.5ghz that's pathetic...

Also I noticed my fans don't seem to go as fast as they used to, it used to sound like a jet engine when my PC would kick up so I turned the fans down... but now no matter what I do in BIOS the fans won't go any faster, the CPU fans are @ 1300RPM and the case fans are around 900-1000RPM does this seem like normal speeds for my case fans and cooler fans?

I don't know pretty upset right now and if this is what overclocking has become I will not be opting for K processors and special motherboards/cooling because its quite pathetic I can't even hit 4.5GHZ on liquid cooling. Hell even at stock my temps hit 60+ which is also pathetic.

Ontop of all of that my BIOS reads my voltage correct (1.275v) but my desktop, every program reads my voltage @ .5-1v (hwmonitor,cpuz,aida64 etc.) which is weird because ive never had this issue before. Like honestly I bought all liquid cooling to OC and keep things cool and quiet and while it's quiet my PC is hot as hell. I'm wondering if the fans arent going the proper speed? cause like I said at one point my case sounded like a jet taking off and I have never beena ble to get the fans to go that fast since I changed the settings.

I wonder if anyone can offer some insight and help me start clearing up these issues. Should I place the hybrid radiator as an outake in the rear instead of an intake in the front? How come my fans won't adjust speed when I adjust them in BIOS? how come my voltage isn't properly reading? and why the heck are my temps so damn high when I invested an extra $300 in my system just to properly overclock and keep it cool? It ticks me off when I see some kid with a 6600K and a 212 EVO and he manages to hit 4.7GHZ and keep temps under 70c it's like seriously? I've never had voltage reading issues and temps like this with such little overclock.

SYSTEM:

6700K @ Stock - 60C load
980TI hybrid 1480 - 53C load (radiator is set to intake from the front)
Phantek Evolv ATX (2 x 140mm intake SP fans in front, 1 x 140mm outtake in rear, 2 x 120mm SP fans upto push configuration outtake.)
2 x 2400mhz 8GB corsair LPX vengeance
1 x 500GB samsung SSD
1x 1TB WD black HDD
850w EVGA GQ PSU
EVGA Z170 FTW motherboard


Heres a picture of my case/configuration:

As you can see CPU radiator uptop in push config, 2 x 140mm in front sucking in, hybrid radiator mounted infront sucking in, 140mm in read pushing air out. All fans are SP series fans so it's not that. I was thinking of moving the GPU radiator to the REAR and having it push the air out and then have 3 x 120mm fans sucking in in the front. But I really like the way it looks right now and I'm not sure if that will be better considering I have so much air coming in but not really alot coming out. The fans on the top have barely any room to breathe because of the case. I feel like removing the rear 140mm will make temps even worse because the air will just sit inside the case.

 

[BonzaiDuck]

I've spent a lot of time identifying "last grains of rice" in cooling improvement, and I have yet to use an AiO cooler -- many of which are quite good.

The airflow strategy can be somewhat different between air and water. There would likely only be a slight difference between radiator at intake versus exhaust.

Your concerns are really quite common for the i7-6700K. One has to choose whether your goal is simply a lower CPU temperature, or what you might surmise to be "best" between an air and water solution. I knew this for my own Skylake project (just completed this week after two-months' building and tuning), after "research" over maybe 6 months' time. Extend that time for maybe 18 months, since I was also looking at the 4790K Haswell.

I spent more on the processor than you did, but it's the same processor. I bought it "binned" to run at 4.8 and ~ 1.42V, and for $50 I had it de-lidded and then re-lidded with CLU Liquid Ultra. The air-cooler is a TR LG Macho, which shows to be 5C warmer than an EKWB Predator 240 on the same test-bed with overclocked loading. The Silicon Lottery re-lid makes it perform 7C cooler than the EKWB.

If your 4.5 Ghz voltage is 1.275V, I'd try actually raising it to ~ 1.29V to see if you see an actual drop in temperatures at same room ambient. But those temperatures are just a tad on the high side of what you might expect.

Someone else here reported ~83C under load for 4.5 with a CM 212 EVO. So, I, too, might wonder why your temperatures aren't a bit lower than what you report.

It may just be that you didn't win the real silicon lottery, or that the polymer TIM and gap between IHS and processor die on that particular CPU makes it run warmer.

The 6600K doesn't have hyper-threading, and this actually accounts for better overclocking and lower thermals -- up to a point.

For myself, I'm not going to run my Skylake at 4.8Ghz, and while 4.7 for 24/7 works, I don't like the voltage around 1.41V. It's either 4.5 @ ~ 1.30V, or 4.6 @ 1.35V. And those voltages provide a margin to assure an ability to run LinX for 30 iterations using my cooling solution and temperatures around 70 to 73C (4.6Ghz)(package temperature maximum as trapped in HWMonitor or HWInfo.)

For the fans, there are some more options, but they involve a trade-off between noise and speed/CFM. My 140mm fans top out at maybe 1,600 RPM to 1,800 RPM pushing air relative to a 110 CFM spec. Filtering at intake considerably reduces the possibility of that level of airflow. There is no noise, because they're good fans and they are totally noise-isolated from the case: no contact between hard surfaces and no metal screws. If you can do this for the radiator, you can reduce the noise and push more air through the fins at lower dBA. Radiators are worse noise-transmitters than case-metal. And noise-isolation for radiators is more of a challenge.

I hope someone else offers more information or insight for you about this.

PS You may be showing common-sense judgment about your thoughts concerning money and parts. This may be the last machine I ever build for "the Joys of Over-Clocking." As far as I can see, it rivals or exceeds my Sandy Bridge machines, just for being "good computers" as opposed to CPUs that are dated technology.

PSS Don't mistake this for my Trumpian moment of gloating. This is a fairly cold day in So-Cal, with room-ambient around 73C. I thought I'd see if what I said about my thermals was accurate. After six LinX iterations (affinitized), I have 69C at 4.6 Ghz. VCORE registers at 1.355V. I am quite sure without the CLU-re-lid, my results would be approximately the same as yours.

Try working on the fans, even if you need to spend some chump-change. Use rubber noise isolators if you can. If you're not put off by the white-noise of "AC-vent" airflow or "Whoosh," if the fans don't have a noisy motor whine at top-end, you may improve your cooling a few degrees C that way.

 

[Valantar]

The fans on the top have barely any room to breathe because of the case.

That sounds like it might be your issue right there, especially if you say your fans don't run as fast as they used to. Radiators already introduce significant resistance to airflow - a case that further restricts this might more or less kill it. I take it you've checked all fan profile settings in BIOS/UEFI and any applicable software? Outside of that, there's no reason why they shouldn't be running the same as before - a 100% PWM signal/12V input is the same regardless of source.

As a test, I'd try removing the radiator from the top and just letting it hang loose with the side panel off. If that lowers temps significantly, the case+radiator combo is restricting your fans too much to sustain sufficient airflow. If that's the case, better/higher static pressure fans or a push-pull setup ought to help.

 

[Tim]

I am running into the same issue as the OP, but with the itx version of the same case. I'm convinced beyond anything else that the top panel having such poor venting is the cause of my Skylake temperature woes.

 

[Crescendo]

My radiator fans are SP. I can't even see my cpu fans show up in software of BIOS but the case ones do. I would like to make my CPU fans spin faster but they won't do it. I did what the manual said and I have my radiator fans hooked into the pump splitter and then that's connected into CPU_2 on the motherboard. Yet the fans are not showing up in BIOS and I cannot adjust them... They run extremely slow, if I plug the fans directly into the hub on the case they spin in accordance and I can set the fans with PWN (case hub has PWM.)

I think its a combination of the case and the fans spinning slow. I googled my case and a ton of people are reporting the top has no air flow and that most have to remove it to overclock, put that with the fact my fans are only @ 40-50% I think thats my heat problem. I don't think 4.5 @ 1.275v is bad at all, I can actually do 4.7 @ 1.375v but then my temps go as high as 80-85c so I backed it off.

That still doesnt explain why my voltages are wrong in all software and my CPU fans arent showing up. Do I need corsair link to adjust fans and have them show up? do I need the plug the pump header into CPU_1 not 2? idk I followed the manual with my cooler and watched videos and I dont see what I'm doing wrong.

As of now my PC is sitting at 74c @ 100% 4.5ghz 1.275v which I guess isn't terrible. but I was hoping I could stay under 65c while gaming.

If your 4.5 Ghz voltage is 1.275V, I'd try actually raising it to ~ 1.29V to see if you see an actual drop in temperatures at same room ambient. But those temperatures are just a tad on the high side of what you might expect.

this statement confuses me how will raising my voltage drop temperatures?

But bottom line I still feel my temps are a bit high, I'd like to be under 65 while gaming and under 72 while stress testing. If I can't handle that then during the summer I won't even be able to overclock because Ill be seeing atleast a 5c gain. and I'm not running 70-80c 24/7.

I also want to know why my voltages are showing so incorrectly and why my CPU fans won't go faster or slower since they are hooked up via PWM the way corsairs manual told me to do it. do I need Corsair LINK or?

Normally my voltages show up in software no problem, it's really hard to overclock and test things when I can't see my voltage unless I'm in the BIOS.

 

[BonzaiDuck]

My radiator fans are SP. I can't even see my cpu fans show up in software of BIOS but the case ones do. I would like to make my CPU fans spin faster but they won't do it. I did what the manual said and I have my radiator fans hooked into the pump splitter and then that's connected into CPU_2 on the motherboard. Yet the fans are not showing up in BIOS and I cannot adjust them... They run extremely slow, if I plug the fans directly into the hub on the case they spin in accordance and I can set the fans with PWN (case hub has PWM.)

I think its a combination of the case and the fans spinning slow. I googled my case and a ton of people are reporting the top has no air flow and that most have to remove it to overclock, put that with the fact my fans are only @ 40-50% I think thats my heat problem. I don't think 4.5 @ 1.275v is bad at all, I can actually do 4.7 @ 1.375v but then my temps go as high as 80-85c so I backed it off.

That still doesnt explain why my voltages are wrong in all software and my CPU fans arent showing up. Do I need corsair link to adjust fans and have them show up? do I need the plug the pump header into CPU_1 not 2? idk I followed the manual with my cooler and watched videos and I dont see what I'm doing wrong.

As of now my PC is sitting at 74c @ 100% 4.5ghz 1.275v which I guess isn't terrible. but I was hoping I could stay under 65c while gaming.

If your 4.5 Ghz voltage is 1.275V, I'd try actually raising it to ~ 1.29V to see if you see an actual drop in temperatures at same room ambient. But those temperatures are just a tad on the high side of what you might expect.

this statement confuses me how will raising my voltage drop temperatures?

But bottom line I still feel my temps are a bit high, I'd like to be under 65 while gaming and under 72 while stress testing. If I can't handle that then during the summer I won't even be able to overclock because Ill be seeing atleast a 5c gain. and I'm not running 70-80c 24/7.

I also want to know why my voltages are showing so incorrectly and why my CPU fans won't go faster or slower since they are hooked up via PWM the way corsairs manual told me to do it. do I need Corsair LINK or?

Normally my voltages show up in software no problem, it's really hard to overclock and test things when I can't see my voltage unless I'm in the BIOS.

Some AiOs supposedly have their own PWM controller or hub. That's what I learned researching the EKWB Predator.

For the load turbo speed, the CPU will push more power in watts. Too few volts means more amps. V x A = W. I can't speak with authority, but I believe this is what they mean by a "sweet spot."

A rather contentious member augmented the fact of Intel discontinuing publication of a "safe voltage" spec at Gen 3, noting that Skylake "board partners" of Intel (the board makers) were the only ones informed on this.

My Z170 board and apparently others with "auto" settings will push voltage at turbo to 1.4V without overclocking the multiplier. That may apparently be the information known "only to board partners." I suggest the 1.290V measured VCORE (as opposed to VID, which should be a bit higher) because the expected voltage to run at 4.5 is about 1.282V. Some in the lottery might run at lower, others higher. But I'd suspect the flaws in each and every chip would likely show up at higher speeds and voltages more than at the lower end of the scale. I say "suspect."

Also, other "skinny" shows that Intel described Skylake as "more resilient" to electro-migration degradation, which could support higher voltages (but I wouldn't go beyond 1.4V and I like 4.6 @ 1.35V better). I just suspect that the processor is more thermally sensitive for two reasons: the TCASE spec is lower than some prior gen chips or 68.5C, and the reports by enthusiasts suggest that temperatures around 85C begin to generate noise and instability. All the reviews I'd seen for speeds between 4.7 and 4.8 suggest that it becomes impossible to achieve a stable 24/7 overclock

But if you cool it so you might abuse it, you could tune it so you could just use it, and it would then be cooler than it needed to be.

You'll most likely have to experiment with the fans. With a Swiftech 8-way PWM splitter, the signal comes from the CPU_FAN port, the power comes from the PSU. If the pump shows itself as properly thermally controlled, then control the fans separately from the mobo. With that idea, you can only monitor the speed of the lead fan. And they all must be PWM fans, or the 3-pin variety will only run at 100%.

Higher static pressure, avoid sleeve-bearing fans and pay for the best (like Mag-Lev), get the highest airflow spec that would keep the dBA below 40 at top-end, and isolate the fans from contact with the radiator. Once you can control the fans, if they are a tad in excess of what's needed to approach diminishing returns of heat-removal by air, you can reduce the prospect of that noise.

I just don't know what the limitations are for fan control on your motherboard. But I remember way back when -- with a Northwood Pentium and Springdale motherboard, I could control a 3-pin fan from the CPU_FAN port. That was before PWM. Whether you have to spend some chump change, you should be able to sort this out.

 

[Crescendo]

The weird thing is my fans on the CPU don't run at 100% the ones in my case fluctuate based off thermals. I can't see the fans on my radiator in software monitors or the BIOS. My radiator fans run at around 50% at all times and never above or below, I'd like them to go faster as it gets hotter and I'd like to be able to monitor the fans. I can't do this with my current setup and I don't know why. I literally followed corsairs instructions to a T and I'm not using Corsair Link so I'm wondering if that's the problem, or should I still be able to monitor and adjust fans without it?

 

[Crescendo]

I mean does 75C @ 1.275v with a H110i GTX and Phantek Evol ATX case a bad temp? The CPU fan is running @ 50%.

 

[BonzaiDuck]

I'd bet some chump change that you need to use Corsair Link. This was one of the prospective hassles I observed with AiO's, but this matter of the Corsair H1xx series and the Link software has appeared in threads before.

If you mean "75C @ 1.275V under stress and 4.5 Ghz," no -- that's not bad, or it would not appear so, but it depends on your choice of stress-test. I'd only worry about what stress-test you use to validate your overclock settings. IF you actually used Prime95 to get that temperature, you might not approach my following suggestion with as much caution. There is a good test with Prime95:

Prime95 "1344" in-place short test

With this test, if you are close to absolute stability but not close enough, it will bork within 15 minutes to a half hour. Usually, if you can run it for an hour, it should pass just about every other test including LinX-affinitized, IntelBurnTest, certainly OCCT (either CPU for 3 hours or Linpack). These are all the tests people have been more leery of since Haswell. I would say that if ANY of them push your system to 85C, then don't use them until you drop the clock enough so it runs cooler. Or apply a cooling solution that allows you to scale the clock.

The guideline of 1.282 for 4.5 and other voltage settings for higher multipliers were in a graph I found and posted in an older but recent thread. It attracted criticism, because I hadn't been clearer about a probability distribution around each voltage value as a mean. But it's a good starting point for tweaking above or below a given value.

If you're "close" but not close enough, there's a better chance that any of the strenuous tests will only terminate itself without throwing a BSOD. When I OC'd my own Skylake, I think I had three or four BSODs during the first preliminary hour or so of tweaking, and after that, the stress-tests would simply throw an error without throwing Windows.

You have the option to experiment a little with the Corsair. But I have no firsthand experience.

 

[Valantar]

My radiator fans are SP. I can't even see my cpu fans show up in software of BIOS but the case ones do. I would like to make my CPU fans spin faster but they won't do it. I did what the manual said and I have my radiator fans hooked into the pump splitter and then that's connected into CPU_2 on the motherboard. Yet the fans are not showing up in BIOS and I cannot adjust them... They run extremely slow, if I plug the fans directly into the hub on the case they spin in accordance and I can set the fans with PWN (case hub has PWM.)

I think its a combination of the case and the fans spinning slow. I googled my case and a ton of people are reporting the top has no air flow and that most have to remove it to overclock, put that with the fact my fans are only @ 40-50% I think thats my heat problem. I don't think 4.5 @ 1.275v is bad at all, I can actually do 4.7 @ 1.375v but then my temps go as high as 80-85c so I backed it off.

That still doesnt explain why my voltages are wrong in all software and my CPU fans arent showing up. Do I need corsair link to adjust fans and have them show up? do I need the plug the pump header into CPU_1 not 2? idk I followed the manual with my cooler and watched videos and I dont see what I'm doing wrong.

As of now my PC is sitting at 74c @ 100% 4.5ghz 1.275v which I guess isn't terrible. but I was hoping I could stay under 65c while gaming.

If your 4.5 Ghz voltage is 1.275V, I'd try actually raising it to ~ 1.29V to see if you see an actual drop in temperatures at same room ambient. But those temperatures are just a tad on the high side of what you might expect.

this statement confuses me how will raising my voltage drop temperatures?

But bottom line I still feel my temps are a bit high, I'd like to be under 65 while gaming and under 72 while stress testing. If I can't handle that then during the summer I won't even be able to overclock because Ill be seeing atleast a 5c gain. and I'm not running 70-80c 24/7.

I also want to know why my voltages are showing so incorrectly and why my CPU fans won't go faster or slower since they are hooked up via PWM the way corsairs manual told me to do it. do I need Corsair LINK or?

Normally my voltages show up in software no problem, it's really hard to overclock and test things when I can't see my voltage unless I'm in the BIOS.

So let me check that I got this right: the fans are connected not to the motherboard, but a splitter connected to the pump, and only the pump is connected to the motherboard? And you're surprised that the fans aren't visible in the BIOS. I think we've found your problem.

1) A single fan/pump header can only read/control a single device. If you connect a (PWM if applicable) splitter, it can connect several fans at the same "speed" (voltage or PWM %, not necessarily the same rpm). RPM will only be read from one connected fan/pump. Running a pump and two fans off one header means only one of these can be controlled, at best.

2) Your AiO is a H110i. It's meant to be controlled by software. Do you even have the USB header connected? And why on earth have you not tried using the included software?

3) Sure, your fans are "SP" fans. That does not mean they are the highest static pressure fans around. Corsair makes decent fans, not great ones. And even great fans struggle against a rad + a restrictive case if they don't run flat-out. The more restrictive the setup, the higher speeds are necessary, and the more noise you produce.

But yes, that would explain why your fans are running too slow. Use Corsair Link, speed them up. That should fix everything. RTFM.

 

[Crescendo]

So let me check that I got this right: the fans are connected not to the motherboard, but a splitter connected to the pump, and only the pump is connected to the motherboard? And you're surprised that the fans aren't visible in the BIOS. I think we've found your problem.

1) A single fan/pump header can only read/control a single device. If you connect a (PWM if applicable) splitter, it can connect several fans at the same "speed" (voltage or PWM %, not necessarily the same rpm). RPM will only be read from one connected fan/pump. Running a pump and two fans off one header means only one of these can be controlled, at best.

2) Your AiO is a H110i. It's meant to be controlled by software. Do you even have the USB header connected? And why on earth have you not tried using the included software?

3) Sure, your fans are "SP" fans. That does not mean they are the highest static pressure fans around. Corsair makes decent fans, not great ones. And even great fans struggle against a rad + a restrictive case if they don't run flat-out. The more restrictive the setup, the higher speeds are necessary, and the more noise you produce.

But yes, that would explain why your fans are running too slow. Use Corsair Link, speed them up. That should fix everything. RTFM.

I dont have the Corsair Link cable bought my AIO used didnt come with it.

so what are you saying I should hook the pump and fans to a different header? corsairs instructions say to plug the fans into the splitter then the pump into the CPU_1 header.

/edit

also I moved my GPU radiator to the rear as exhaust I am now running 4.7GHZ @ 1.35v and 75c Max Load. So I fixed my temp issues now to fix my voltage readings and fan RPM readings because they still arent listed and I still cant control the CPU fan speed.

 

[XavierMace]

Generally I prefer to run my radiators as intakes with push+pull fans then the remaining fans as exhaust. With your radiator as an exhaust you're pulling warmer air across it then further hampering it by having it mounted in a restrictive location.

 

[Crescendo]

Generally I prefer to run my radiators as intakes with push+pull fans then the remaining fans as exhaust. With your radiator as an exhaust you're pulling warmer air across it then further hampering it by having it mounted in a restrictive location.

Then why did my temps drop dramatically when I changed my GPU radiator to exhaust? Isn't the point of a radiator to put the heat somewhere else other than your case? so what sense does it make to suck hot air into the case from the front radiator? Wouldn't it make more sense to put all radiators as exhaust so the head goes directly out of the case? atleast that's my understanding. what I don't get is if your method is correct why I get much lower temps now?

Also what warm air will be pulled through it? the CPU's radiator is top mounted exhaust and GPU radiator is rear exhaust the rest of the system barely puts out any heat.

I was sitting at 75-85C @ 1.275v 4.5GHZ before, now I sit at 75-80 1.35v 4.7GHZ so obviously the swap made a big difference, for the best.

The hottest thing in my case is the GPU radiator, so with it now exhausting directly out of the case wouldn't that make more sense on keeping the rest of the case cooler? I mean thats what radiators are for I though was the ability to mount it and get the air out without the case getting hotter.

I mean my general consensus is to have the GPU radiator sucking air in and the rear exhausting that's how I originally had it, but most people top mount their CPU rads on the top as exhaust and not on the front as intake unless they have to. So it only makes sense to do the same with the GPU radiator.

Am I wrong or?? and if so why are my temps better this way?

 

[BonzaiDuck]

. . . . .
I was sitting at 75-85C @ 1.275v 4.5GHZ before, now I sit at 75-80 1.35v 4.7GHZ so obviously the swap made a big difference, for the best.

The hottest thing in my case is the GPU radiator, so with it now exhausting directly out of the case wouldn't that make more sense on keeping the rest of the case cooler? I mean thats what radiators are for I though was the ability to mount it and get the air out without the case getting hotter.

I mean my general consensus is to have the GPU radiator sucking air in and the rear exhausting that's how I originally had it, but most people top mount their CPU rads on the top as exhaust and not on the front as intake unless they have to. So it only makes sense to do the same with the GPU radiator.

Am I wrong or?? and if so why are my temps better this way?

Did you try the special Prime 95 stability test I posted earlier in this thread?

I don't pretend to know everything, but there's a probability frequency distribution of 6700K's for overclocking to a higher multiplier and there is a probability distribution of voltages around a mean value for those multipliers.

Even for a binned chip, the mean voltage value for reaching a stable 4.7 is around 1.40V.

For instance, I might be able to boot at 4.6 with a voltage as low as 1.30V, but I won't pass that Prime95 test until I'm closer to 1.34, so my setting assures VCore voltage under load at above 1.35V. I can tell if there's the slightest insufficiency of voltage by looking at the GFLOPS in affinitized LinX as a probability distribution and looking for outliers. A stable system will have a narrow variation of GFLOPS with a range (min to max) of less than 2 GFLOPS, and I'd wager it should look more like 1.5. Sometimes the outliers can be thrown by background processes that are identifiable, or some AV or system process that needs to run. But you can identify those things.

I suppose it's possible to get such a winner in the chip lottery that you could be totally rock-stable at 4.7 and 1.35V, but it's just hard to believe. Who knows?

Actually I wouldn't be fiddling with clock settings until you're sure you can control both the fans and the pump on that water-cooler. If Corsair provides a splitter, you'd follow the instructions. Alternately, you could use a Swiftech PWM splitter for the fans such that the tach of a single fan is reported and all the rest spin up or down according to the PWM signal. [If they're all the same spec fan, you'd mostly only need to monitor one.] You could use another port for the pump, or simply use a port designated for "pump."

you would then only need to assure that the monitoring software shows those devices being monitored to spin up at warmer temperatures and slow down at lower ones.

 

[Crescendo]

You agreeing with me or what? haha >confused.

I just want to make sense of this all.

 

[BonzaiDuck]

I'm just saying for your future reference and because I've followed similar detours before and learned from them -- you might have been better off attending to proper operation of the Corsair first before fiddling with the OC.

The second thing I'm saying, as long as your overclocks are in the mix here, you should run that Prime95 "1344" test to see if you're really "stable" at those settings of 1.35V and 4.7. The recommendations for that test are "15 minutes" but even a half-hour might not turn up a failure. An hour seems more practical.

Personally, from what I've seen in posts, online whitepapers, serious lab test reviews and so forth, even the top 25% of 6700Ks would need more voltage than that to achieve that clock.

 

[Valantar]

I dont have the Corsair Link cable bought my AIO used didnt come with it.

so what are you saying I should hook the pump and fans to a different header? corsairs instructions say to plug the fans into the splitter then the pump into the CPU_1 header.

/edit

also I moved my GPU radiator to the rear as exhaust I am now running 4.7GHZ @ 1.35v and 75c Max Load. So I fixed my temp issues now to fix my voltage readings and fan RPM readings because they still arent listed and I still cant control the CPU fan speed.

So you don't have the cable that lets you control the AIO, yet you expect to be able to control it while following the manual otherwise. See the flaw in your logic?

First off, what do you mean when you say you can't control the CPU fan speed? Do you mean the fans, or the pump connected to CPU_1? If it's the pump, have you checked if it supports speed adjustments? If so, how? PWM or voltage, or only through software? If you're actually talking about the fans, and they're connected through a single splitter: Are they PWM fans? Is it a PWM header? Is the splitter a proper PWM splitter that carries the RPM signal from only one fan to the motherboard? If it's some sort of werid splitter that Corsair made special, it might not work. Try another PWM splitter. If the header they're connected to isn't PWM capable (i.e. 3-pin), does it support voltage adjustments for fan speed control? If not, then ... find another one? Buy a fan controller?

 

[BonzaiDuck]

So you don't have the cable that lets you control the AIO, yet you expect to be able to control it while following the manual otherwise. See the flaw in your logic?

First off, what do you mean when you say you can't control the CPU fan speed? Do you mean the fans, or the pump connected to CPU_1? If it's the pump, have you checked if it supports speed adjustments? If so, how? PWM or voltage, or only through software? If you're actually talking about the fans, and they're connected through a single splitter: Are they PWM fans? Is it a PWM header? Is the splitter a proper PWM splitter that carries the RPM signal from only one fan to the motherboard? If it's some sort of werid splitter that Corsair made special, it might not work. Try another PWM splitter. If the header they're connected to isn't PWM capable (i.e. 3-pin), does it support voltage adjustments for fan speed control? If not, then ... find another one? Buy a fan controller?

Yup! Yup! Yeah! And I'll second all those motions.

The Swiftech 8-port PWM splitter had been a $10 item until last year some time, but the Egg has it for $14. Still -- chump-change. Last time I bought one, I purchased three. Those things are handy and allow you to define an air-handling strategy with as few as two PWM ports and a 3-pin port

If you had to, you could still do it with as little as a single PWM port. You can mix pumps and fans on a single splitter, but you'd want to monitor the pump and one fan -- hopefully one of a few identical units. I use it to control the speed of three Akasa Viper 140's, one round and two square and ~ same CFM. My rear exhaust connects to the CPU_FAN port.

The reason I bought my Sabertooth Z170 S was a wonderful compromise. Sure -- I only get a single M.2 port. But I can control fans with curves defined in BIOS on 6 CHA-FAN ports, 2 CPU ports, 1 PUMP port, 3 "Assistant Fan" ports, and a few more "extension" fan ports that I don't want to bother learning about.

For any port, I can choose in BIOS between PWM and Voltage control. If I have simply a string of 3-pin fans wired in parallel with total amperage acceptable to mobo limits, I can use any one of those ports to control the string, monitoring one.

But that's not the icing on the cake. The icing on the cake is the option to tie each or any of those fan ports to temperature on a) CPU, b) mobo "VCORE" VRM1, c) VRM2, c) the USB3 controller, d) a point on the board nearest the x16_1 PCI-E slot, e) the RAM, and f) the Northbridge/chipset. There may be more sensors for this thermal control --there's the commonly-known "motherboard" -- but I'm most interest in two for my purposes.

Yes, there are boards more and less expensive than mine, some with a stronger feature set, some boards like the AsRock which throw in features Asus makes you pay for. But for me -- this is "the One."

 

[Crescendo]

So you don't have the cable that lets you control the AIO, yet you expect to be able to control it while following the manual otherwise. See the flaw in your logic?

First off, what do you mean when you say you can't control the CPU fan speed? Do you mean the fans, or the pump connected to CPU_1? If it's the pump, have you checked if it supports speed adjustments? If so, how? PWM or voltage, or only through software? If you're actually talking about the fans, and they're connected through a single splitter: Are they PWM fans? Is it a PWM header? Is the splitter a proper PWM splitter that carries the RPM signal from only one fan to the motherboard? If it's some sort of werid splitter that Corsair made special, it might not work. Try another PWM splitter. If the header they're connected to isn't PWM capable (i.e. 3-pin), does it support voltage adjustments for fan speed control? If not, then ... find another one? Buy a fan controller?

People supposidly have adjusting fan speed without the proprietary cable. But that's why I was asking if I need Corsair Link to adjust fan speed, I'm referring to CPU fans not the pump, I know I need Corsair Link to control the pump. Fans are all PWM and supposidly the splitter corsair provides allows PWM fan control as it's 4 pin.

 

[Valantar]

People supposidly have adjusting fan speed without the proprietary cable. But that's why I was asking if I need Corsair Link to adjust fan speed, I'm referring to CPU fans not the pump, I know I need Corsair Link to control the pump. Fans are all PWM and supposidly the splitter corsair provides allows PWM fan control as it's 4 pin.

The manual for the H110i GTX talks about the 3-pin connector going into the motherboard as the "Pump RPM connector." I.e. it has no relation to the fans whatsoever, it simply conveys an RPM signal to the motherboard so that it doesn't freak out and give you CPU_FAN_ERROR messages and refuses to boot. I would have assumed that the fans, in the absence of a Link connector, simply ran at 12V flat-out. However, it seems that Corsair has gone the sensible(?) route to avoid noise and defaults the fan to a mid-range speed.

In other words, you need to connect the fans directly to the motherboard to have any hope of controlling them without the Link cable. Otherwise, how do you propose the motherboard communicate with the fans? From reading the manual it sounds like the splitter is fixed to the pump. Is this true? If so, get another PWM splitter, connect the fans to separate motherboard headers, or get a Link cable. Btw, the cable can be found cheap on ebay.

 

[BonzaiDuck]

Who knows about the Corsair Pump? Is it a three-pin or PWM device?

If you could simply control the pump from the motherboard with either PWM or voltage, control the fans from another port perhaps powered directly from PSU with the Swiftech splitter, you might not need Link and the hardware you missed in the used purchase.

I can't speak with any authority about this. There are some threads around going back months which explain my "custom" AiO water plan, and then I craw-fished and bought a TR LG Macho.

And I had my 6700K "CLU-re-lidded." No need for water here.

But I know enough about some pumps and fans and splitters. It could all be done through the motherboard if the BIOS or software defines % RPM control of the ports -- PWM -- allowing one to set at least three points on a "graph" or curve. No less for the voltage-controlled ports.

 

[XavierMace]

Then why did my temps drop dramatically when I changed my GPU radiator to exhaust? Isn't the point of a radiator to put the heat somewhere else other than your case? so what sense does it make to suck hot air into the case from the front radiator? Wouldn't it make more sense to put all radiators as exhaust so the head goes directly out of the case? atleast that's my understanding. what I don't get is if your method is correct why I get much lower temps now?

No the point of a radiator is to cool the liquid that's cooling your components, therefore you want the air flowing through the radiator as cool as possible and you want as much air flowing through it as you can stand the noise of. Just like the radiator on a car. That's why they are up front with a nice big grill in front of them even though that means air is going to be (marginally) hotter when it gets to everything else in the engine bay.

Then you want as much unrestricted exhaust as possible to get that hot air out as quickly as you can. Any "data" you're getting right now is moot because of your bad fan setup. The Corsair fans as others mentioned are meh fans in general. You have SP fans on non-radiator mount points and you have more exhaust than intake. IMO, that's all less than ideal.

My current setup is an H440 with an X61 radiator (280mm) in a push/pull intake, an X31 on the rear as a push/pull intake, then 3x 140mm up top as exhaust. Last I checked, I sit about 62c running full load (3770k @ 4.4Ghz, GTX970).

 

[BonzaiDuck]

No the point of a radiator is to cool the liquid that's cooling your components, therefore you want the air flowing through the radiator as cool as possible and you want as much air flowing through it as you can stand the noise of. Just like the radiator on a car. That's why they are up front with a nice big grill in front of them even though that means air is going to be (marginally) hotter when it gets to everything else in the engine bay.

Then you want as much unrestricted exhaust as possible to get that hot air out as quickly as you can. Any "data" you're getting right now is moot because of your bad fan setup. The Corsair fans as others mentioned are meh fans in general. You have SP fans on non-radiator mount points and you have more exhaust than intake. IMO, that's all less than ideal.

My current setup is an H440 with an X61 radiator (280mm) in a push/pull intake, an X31 on the rear as a push/pull intake, then 3x 140mm up top as exhaust. Last I checked, I sit about 62c running full load (3770k @ 4.4Ghz, GTX970).

I labored over these issues before I "craw-fished" and killed the idea of using the Predator 240 over the TR LG Macho (heatpipe). Since I took so long only to arrive at that decision, I prepared my Stacker 830 case for two 140mm intake fans on the front panel to assure myself I could get a 280mm radiator in there. I could make a switch again at a moment's notice. But I probably won't.

I DID have an idea, though. If you chose to top-mount the radiator as case exhaust, it might be possible (anything is possible) to change the exhaust fan to intake, duct it to a motherboard plate, and duct the motherboard plate to half the radiator. If one used double the number of case-intake fans for a total of >= 3, the case wouldn't be starved of pressure.

The problem there is the restriction of the duct-plate. Just thinking . . .

 

[Valantar]

In a single-radiator or single-loop watercooled setup, radiator placement has very little effect on cooling (given equal access to air for all mounting points). In this case, though, the case has two loops. Ironically, AIO cooling on GPUs (which are far more power hungry than CPUs) are almost universally limited to 120mm rads. In other words, those radiators get hot, and the air coming off of them is far hotter than that from, say a 240mm rad cooling a CPU. In a case like this, having the GPU rad as an intake delivering air to the CPU rad would be a bad idea, as you're effectively limiting the cooling potential of the CPU rad by supplying it with hotter air. If they were on a single loop with the same radiators, this wouldn't be much of an issue as the heat would be evenly spread across all rads and thus the air from the intake rad wouldn't be as hot.

In other words: make your GPU radiator an exhaust. Then fix your fan control issues. I would get a Link cable, that would be the easiest way (and you'd be getting the most out of the AIO you already paid for).