New x99 rig, EK Supremacy Evo bad temps at stock

[Unframed]

Hey Guys

So i just finished my new rig last week and have been using it at stock speeds for the last couple of days.
Info about rig/WC setup.
I7 5930k
MSI X99S Gaming 9 Ack
2x Asus Strix 980

CPU Block: EK Supremacy Evo Nickel
GPU Block: EK-FC980 GTX Strix - Acetal+Nickel
Rads: 1x Alphacool XT45 360mm 2x Alphacool UT60 240mm


With the info out of the way, lets get to the problem.
So last night I wanted to start overclocking my system.
Before I began I wanted to do some benchmarks to compare stock vs OC.

I ran Cinebench and immediately got worried since the temps on the CPU started hitting 69-71°C and this is at stock speeds with a voltage of 1.025-1.035.

I saw this guy on Youtube getting his CPU to around 45°C at stock with a Corsair H100i doing the same test, and I would think my WC setup should at least be able to match that.

I stopped testing the CPU and went for the GPU's instead.
I ran FurMark for about an hour or so and both cards hit a max temp of 43°C which I'm pretty happy about.
So with that in mind I'm thinking it has something to do with the CPU block and not my loop.

I'm thinking I might need to remount the CPU block. I am using Coollaboratory Liquid Ultra on both the CPU and GPU's and it was also the TIM I used on my last rig, with great results.
Could it be the layer between the CPU and the block is to thin? I was using the last drop of the Liquid Ultra on the CPU but I got the whole surface covered.
If I do the remount should I try with the TIM that came with the Block or just try the Ultra again?
Also when i mounted the block I did it in the Goofy position so 90° CCW, but from what I have read that should only make a couple of degrees difference, and not 20-30°C(compared to the H100i).

Hope you guys are able to help me with my problem and feel free to come with other suggestions than what is on my mind. Also if there is any info missing please tell me and Ill post it. Thanks in advance

-Unframed

 

[Warchild]

Is the voltage set to auto?

Could be the reason why the temp ramps up high. I know you state 1.025 V but is that what you are seeing at idle? or is that what you set it to?

Other thing could be that your CPU block is not seated correctly. I have a problem with my EVO at the moment that there is virtually 1mm clearance from the 3 capacitors on the left side of the socket. I had to carefully bend them away to allow the EVO block to sit correctly ontop of the CPU.

as for Cool lab paste. I use it and you only need a tiny drop to spread it completely.

edit* did you flush your rads before installing them? Shine a light into your EVO block and check the jet plate. It might be clogged. Bad/low flow= high temps

mine at stock with the 4790K went up to 60C with WC but that is because volts were on auto and jumped to 1.35V when at full load.

 

[tinytomlogan]

45c wouldnt have been full load on the H100i mate.

Even with water youll probably see 60c on load.

You can check our heatsink and CPU reviews to check reliable temps

 

[Unframed]

Yes the voltage is set to auto, but is between 1.025-1.035 during the Cinebench, according to CPU-Z.

Not really anything around the socket that should curse a problem for it to sit correctly on top of the CPU.

The rads I'm using is from my last build and I did do a good flushing before using them then and as well now before using them in my new build.
It's a bit hard to look inside the block since I got the Original CSQ model with a frosted look to it. But from what I can see it doesn't look clogged. But shouldn't this also effect the GPU temps if there was bad flow through the system?

 

[Warchild]

if its auto then your volts will rocket at full load. There is your reason.

 

[Unframed]

During his Cinebench run his CPU hit a max of 45°C. I know its a short test but when i do it mine it reaches 70°C, both being the 5930k. He is saying he is using the H100I in the video but you are not seeing it at any point, so I guess I have to take his word on him using that. Also he is using auto voltage.

Ill try and do a run with my voltage set to manual and see what happens.

 

[Warchild]

During his Cinebench run his CPU hit a max of 45°C. I know its a short test but when i do it mine it reaches 70°C, both being the 5930k. He is saying he is using the H100I in the video but you are not seeing it at any point, so I guess I have to take his word on him using that. Also he is using auto voltage.

Ill try and do a run with my voltage set to manual and see what happens.

I very very very much doubt his is auto. My pc is is cooled by a ridiculous amount of fans and as I said, even at stock speed and auto setting I hit 60C in cinebench.

There is a nice guide on here how to undervolt your CPU. I suggest you give it a go and then try cinebench again. Once you are satisfied and confident you can start playing about with overclocking your CPU.

 

[Unframed]

So setting the Core voltage to 1.035 the cores range from 68-74°C during the Cinebench run which is a little higher than on auto, and I guess that's because the voltage goes down to 1.025 during the run on auto.
So I'm guessing it's something else than the voltage.

btw this is the video I'm referring to /watch?v=yePdgt9Tmuw maybe you are able to see/hear something I'm not.
2 min into the video he starts talking about auto voltage and his H100

 

[tinytomlogan]

So setting the Core voltage to 1.035 the cores range from 68-74°C during the Cinebench run which is a little higher than on auto, and I guess that's because the voltage goes down to 1.025 during the run on auto.
So I'm guessing it's something else than the voltage.

btw this is the video I'm referring to /watch?v=yePdgt9Tmuw maybe you are able to see/hear something I'm not.
2 min into the video he starts talking about auto voltage and his H100

Stop comparing to that YouTube video when you have no idea about all of his settings.

Cinebench is also not a program to bother running for CPU stress testing.

Follow the guide below and then post back with screen shots and lists of settings dude - try and help us to help you

http://www.overclock3d.net/articles/cpu_mainboard/how_to_manually_set_and_undervolt_your_cpu_vcore/1

 

[Kushiro]

OP download OCCT and use that instead, its free and quite good too. Nonetheless you should probably check your block seating as others have said, as you shouldn't be getting those kind of temps with the amount of rad space and fans you say you have.

Also next time just use a simply TIM like MX-4 or something, it wont make a massive difference.

 

[Unframed]

Okay I'll go have a look and try it out, thanks

 

[Unframed]

So after some testing I've now found the lowest voltage my CPU will run at and still run OCCT.

I started with all settings on auto and the core voltage was stable at 1.020 V during the 30 min test.
I then went into the Bios and set it manually to 1.020 V and ran the test again.
Here is the result after 30 min http://imgur.com/dC1USAJ

After this i went on and decreased the voltage .020 at a time, all the way to 0.940 V.
With 0.940V it did the 30 min run as well, here are the result http://i.imgur.com/6aCZ4uo.png

I then tried with 0.930V but just after booting into windows it came with BSOD.
Went back up to 0.940V but this time setting the test time to 2 hours.
after 10-20 min it came up with BSOD.

So i went back into BIOS and upped the voltage to 0.950V
Got into Windows ran the OCCT test for 2 hours and here are the result http://imgur.com/hqVrIoC

So comparing there is a small decrease in temps as there should be, but shouldn't I be able to get better temps with that low Voltage together with the amount of rad space I've got?

 

[Unframed]

I did a remount of the water block, and when I removed the thumb screws i noticed that a PVC washer fell out from one of the screws. I then inspected all the screws and found 1 more inside another screw. The washer is supposed to be used together with another socket, and i guess it had found a way inside the screw during shipping.

So having removed the washers and mounting the block again, put everything on stock and did a OCCT run.
Max temp after 30 mins is 42°C, now that's an improvement

I guess the washers made the block sit uneven on the CPU or something.

I've been OC'ing the chip today and so far I've got it running at 4.6GHz at 1.3V and a max temp of 67°C for 1 hour. Going to do some more testing and make sure it's stable.

Thanks for your suggestions guys, much appreciated.