Overclock.net banner

1 - 20 of 24 Posts

·
GPU Enthusiast
Joined
·
2,560 Posts
Discussion Starter #1
It's been a few weeks since I've finished installing my first loop (was busy with school/life/etc...), and I've had a serious problem bleeding this loop, and also getting a decent delta T.

Current OC: 7800x (delidded) @ 4.6ghz 1.83 vin, 1.15vcore, 1.05vccio, 1.0v vccsa

Parts:

- EK-XRES 100 SPC-60 MX PWM
- Alphacool fittings
- EK 360 SE slim rad
- 3x EK Vardar @ ~900 rpm
- EK ASRock x299 Monoblock
- 3/8 ID, 5/8 OD tubing
- Distilled mixed with EK Cryofuel

I currently have some very small micro bubbles left in the loops, and with a current ambient (20c) my idle temps sit around 2-3c higher than room temp. When I run x264 for a few loops on normal priority across all 12 threads, my temps will peak at roughly 72c, an almost 50c detla?! I'm currently running the pump at a fixed 2k rpm and fans at ~900-1000rpm. Kicking the fans up to max speed (~1800rpm) will only drop the loop temps about 2-3c. I am expecting too much out of a single 360 slim rad, or is there a serious issue I'm missing?

I have another 240mm slim rad I could install, but I'm leaning more towards getting an Alphacool Nexxos UT60 or something similar. What would be the best option to keep noise down and get a more tame delta.


 

·
Registered
Joined
·
180 Posts
What is your pump speed at?

Have you checked your block mounting?

Is your rad sitting with the fittings up top? If so, you'll always have bubbles from near as I can tell from the way your res is sitting lower than the top. Air is going to end up at the high points.
 

·
GPU Enthusiast
Joined
·
2,560 Posts
Discussion Starter #3
What is your pump speed at?

Have you checked your block mounting?

Is your rad sitting with the fittings up top? If so, you'll always have bubbles from near as I can tell from the way your res is sitting lower than the top. Air is going to end up at the high points.
Set at ~2000 rpm pump speed currently. Pretty sure the max is ~2400rpm; is it bad to run a pump at max speed 24/7?

I'll double check the mount after I try and get the remainder of what bubbles I can out tonight.

Would flipping the rad to have the fittings at the bottom work/bleed better?
 

·
Registered
Joined
·
180 Posts
They are rated for their lifespan at full speed. Is it a DDC style pump? If so the more high strung versions wind out to 3500+ rpm.

Putting the fittings at the bottom will force the air trapped in the rad to the top, and should stop the mini bubbles from coming from there. Fittings at top means it's working like a fountain, capturing air in the coolant. That's if there is air in there, and unless you can move the res to the very top of everything for bleeding, you should get most of the air out. From what I've read, it may take a few tries to finally get all of the air out.
 

·
GPU Enthusiast
Joined
·
2,560 Posts
Discussion Starter #5
They are rated for their lifespan at full speed. Is it a DDC style pump? If so the more high strung versions wind out to 3500+ rpm.

Putting the fittings at the bottom will force the air trapped in the rad to the top, and should stop the mini bubbles from coming from there. Fittings at top means it's working like a fountain, capturing air in the coolant. That's if there is air in there, and unless you can move the res to the very top of everything for bleeding, you should get most of the air out. From what I've read, it may take a few tries to finally get all of the air out.
I cut the tubing to fit, so everything is pretty fixed unfortunately. The pump is a PWM pump, peak flowrate at just under 2gpm and peak rpm of ~2500rpm according to what I can find; so not the strongest pump, but should be enough to get the job done.

I'm really leaning towards replacing the rad with a thicker, lower FPI rad. Opinions?
 

·
Registered
Joined
·
1,168 Posts
Correct me if I'm wrong, but air is always going to be at the top of a vertically mounted rad whether your fittings are high or low. If your fittings are at the top where the air is, it is easier to bleed by leaning your case and moving bubble towards the out fitting. If the fittings are at the bottom you may have to invert the case and bubbles are much more difficult to maneuver out.
 

·
Registered
Joined
·
180 Posts
With the Swiftech rad, it specifically says it will keep air out of the coolant as long as the fittings aren't at the top. I would try doing the rotating the case slowly with the system off, to see if you can get air to collect and see if you can work it towards the res. Can you tip the case on it's side, and stand up the pump and res, so It's vertical to the horizontal case, that would force it into the highest position and should get the trapped air to move. It may take some time for the air to move, and the pump has to be off.


I would seriously lift the mono block and check to see if it's making proper contact. That's getting toasty, and a 360 slim rad should be able to handle that CPU, without that kind of temperature spike.

Can you stick a temp probe into the rad fins? Does the radiator get really hot when the cpu temperature is spiking like that? How does the air feel coming out of the radiator? Does it feel like a blow dryer hot? If the stuff is making good contact, that system should be getting noticeably hot to the touch.
 

·
GPU Enthusiast
Joined
·
2,560 Posts
Discussion Starter #8
With the Swiftech rad, it specifically says it will keep air out of the coolant as long as the fittings aren't at the top. I would try doing the rotating the case slowly with the system off, to see if you can get air to collect and see if you can work it towards the res. Can you tip the case on it's side, and stand up the pump and res, so It's vertical to the horizontal case, that would force it into the highest position and should get the trapped air to move. It may take some time for the air to move, and the pump has to be off.


I would seriously lift the mono block and check to see if it's making proper contact. That's getting toasty, and a 360 slim rad should be able to handle that CPU, without that kind of temperature spike.

Can you stick a temp probe into the rad fins? Does the radiator get really hot when the cpu temperature is spiking like that? How does the air feel coming out of the radiator? Does it feel like a blow dryer hot? If the stuff is making good contact, that system should be getting noticeably hot to the touch.
The radiator is warm and the water in the loop feels warm through the tubes, but probably only 10c warmer than ambient (85-90f). Going to go check the mount now.
 

·
Registered
Joined
·
2,099 Posts
Hi there

Sadly LGA2066 chips or Skylake X are running bit too hot and deliding can help but in many cases this simply doesn't help and monoblock usually introduce higher temperatures by 3-10°C against the normal CPU block,because they usually cools VRM although X299 boards needs to have good VRM cooling, EK blocks are not best performers if you are comparing them against the Heatkiller or Aquacomputer or XSPC etc

On other hand, pump itself should be enough for yours loop, but not sure if pump will be enough if you add to loop GPU and another radiator

SE radiators are not best radiators, I would say they're poorest radiators from EK, you will be better getting Magicool radiator which is cheaper than this SE

You shouldn't care about the delta between the idle and load CPU temperatures, yours main concern should be water delta T not CPU delta

Get water temperature sensor and check yours water temperature and check ambient temperature, if its in range 5-10°C then yours loop performs good

If you like quietness then I would suggest get extra radiator to have better water delta and therefore you can run fans slower

If you have at home spare waterblock then try spare waterblock and check what temperatures you will have with this waterblock than monoblock

If you are really want low temperatures then chiller is only option

Hope this helps

Thanks, Jura
 

·
Registered
Joined
·
184 Posts
Just to add on with Jura comments. You should be more concern on your idle & load individual core temp. Cpu temp doesn't really give an accurate figures. My 6600k delided on Maximus IX Formula cpu temp is idling/load @ 48deg/57deg with an ambient of 33deg. But my individual core stays frosty 35deg/63deg in a stress test. Water temp idle/load is @ 36deg/40deg. And all this is cooling with overclocked Asus Strix Gtx 1080 Advanced inlcuded on a 480 SR2 Hwlad rads.
 

·
Registered
Joined
·
2,302 Posts
You're delidded and running a mono block. If you didn't reglue the IHS then the surface of the IHS can sit lower than it would have stock, but the mono block sits at stock height because it's on the VRMs as well as the chip.

Investigate the mount and see what the spread of your tim is like. You might just need a re mount, or you may need to experiment with thinner thermal pads etc.
 

·
Shooting down fallacies
Joined
·
8,197 Posts
Just to add on with Jura comments. You should be more concern on your idle & load individual core temp. Cpu temp doesn't really give an accurate figures. My 6600k delided on Maximus IX Formula cpu temp is idling/load @ 47deg/56deg with an ambient of 33deg. But my individual core stays frosty 35deg/64deg in a stress test. Water temp idle/load is @ 31deg/36deg. And all this is cooling with overclocked Asus Strix Gtx 1080 Advanced inlcuded on a 480 SR2 Hwlad rads.
You have a mistype here....as written, you are stating that your liquid is below ambient. Or your liquid temp sensor is skewed.
 

·
Banned
Joined
·
3,726 Posts
https://www.techspot.com/review/1433-intel-core-i9-core-i7-skylake-x/page4.html

However, the frequency isn't my greatest concern right now. Using just 1.2 volts, which is all I needed for a stable 4.6GHz overclock on all cores, the system consumption went from the 259 watts just seen to an insane 402 watts! That might not even be the worst part, the deal breaker is probably the operating temps. Chilling the 7900X was Corsair's H100i v2 and despite being a premium 240mm AIO liquid cooler, temps skyrocketed as the CPU was place under load, reaching 90 degrees instantly before climbing further towards 100 degrees. Unless you have an amazing cooler, I'm not sure overclocking is going to be worth it.

so yes they are using a AIO, but this one preform as well, as any of the best after market air coolers. still there CPU that was not dilided, with the same OC, reached 100C. so I would say that this setup is functioning as well as possible.
 

·
Registered
Joined
·
626 Posts
It's been a few weeks since I've finished installing my first loop (was busy with school/life/etc...), and I've had a serious problem bleeding this loop, and also getting a decent delta T.

Current OC: 7800x (delidded) @ 4.6ghz 1.83 vin, 1.15vcore, 1.05vccio, 1.0v vccsa

Parts:

- EK-XRES 100 SPC-60 MX PWM
- Alphacool fittings
- EK 360 SE slim rad
- 3x EK Vardar @ ~900 rpm
- EK ASRock x299 Monoblock
- 3/8 ID, 5/8 OD tubing
- Distilled mixed with EK Cryofuel

I currently have some very small micro bubbles left in the loops, and with a current ambient (20c) my idle temps sit around 2-3c higher than room temp. When I run x264 for a few loops on normal priority across all 12 threads, my temps will peak at roughly 72c, an almost 50c detla?! I'm currently running the pump at a fixed 2k rpm and fans at ~900-1000rpm. Kicking the fans up to max speed (~1800rpm) will only drop the loop temps about 2-3c. I am expecting too much out of a single 360 slim rad, or is there a serious issue I'm missing?

I have another 240mm slim rad I could install, but I'm leaning more towards getting an Alphacool Nexxos UT60 or something similar. What would be the best option to keep noise down and get a more tame delta.


Honestly, I don't think your temps, when compared to ambient, are bad at all considering you are using a monoblock. I would check to see if your block is making good contact with the IHS as someone else has suggested. If you cleaned off all the silicone on the die before putting the IHS back on chances are that the IHS is sitting slightly lower than it would at stock. I'm running a delidded 7900X OC'ed to 4.7 on all cores at 1.24V and my temps are in the low 70's under load and can peak at ~81 when running Realbench or Cinebench. I am using 2 EK PE 240 radiators and a D5 with a full nickle EK Supermacy block. I have a DerBauer Die Frame ready to go on in a couple days in hopes of dropping temps a little more.
 

·
Registered
Joined
·
184 Posts
You have a mistype here....as written, you are stating that your liquid is below ambient. Or your liquid temp sensor is skewed.
Thanks for pointing that out.. couldn't even figure it out what the heck i'm typing. Anyway mine is as follows, quite outdated by 2 months o so though

i5 6600k Delided (4.7ghz vcore: 1.38v)
Temp as follows;

Ambient Temp: 31 deg C
Water Sensor Temp: 34.3 deg C

Idle Core Temp: 33,35,32,34 deg
Idle Package Temp: 48 deg

ROG Benchmark
Stress Core Temp: 64,62,60,59
Stress Package Temp: 74 deg
 

·
Registered
Joined
·
626 Posts
Thanks for pointing that out.. couldn't even figure it out what the heck i'm typing. Anyway mine is as follows, quite outdated by 2 months o so though

i5 6600k Delided (4.7ghz vcore: 1.38v)
Temp as follows;

Ambient Temp: 31 deg C
Water Sensor Temp: 34.3 deg C

Idle Core Temp: 33,35,32,34 deg
Idle Package Temp: 48 deg

ROG Benchmark
Stress Core Temp: 64,62,60,59
Stress Package Temp: 74 deg
Not bad at all. 3 degrees over ambient is great and your stress load temps are fine.
 

·
Registered
Joined
·
2,099 Posts
https://www.techspot.com/review/1433-intel-core-i9-core-i7-skylake-x/page4.html

However, the frequency isn't my greatest concern right now. Using just 1.2 volts, which is all I needed for a stable 4.6GHz overclock on all cores, the system consumption went from the 259 watts just seen to an insane 402 watts! That might not even be the worst part, the deal breaker is probably the operating temps. Chilling the 7900X was Corsair's H100i v2 and despite being a premium 240mm AIO liquid cooler, temps skyrocketed as the CPU was place under load, reaching 90 degrees instantly before climbing further towards 100 degrees. Unless you have an amazing cooler, I'm not sure overclocking is going to be worth it.

so yes they are using a AIO, but this one preform as well, as any of the best after market air coolers. still there CPU that was not dilided, with the same OC, reached 100C. so I would say that this setup is functioning as well as possible.
Hi there

I suspect in this review they've used non delided 7900X, with delidding you should see 10-15°C temperature drop per core maybe bit more

Comparing H100i to custom loop, you shouldn't compare that, this H100i in many cases is no better than good air cooler or cools sometimes lot worse than water cooler

OP temperatures are not bad, but still I would suspect bit better temperatures too,monoblock still would cools bit worse than normal CPU waterblock

Hope this helps

Thanks, Jura
 

·
GPU Enthusiast
Joined
·
2,560 Posts
Discussion Starter #18 (Edited)
You're delidded and running a mono block. If you didn't reglue the IHS then the surface of the IHS can sit lower than it would have stock, but the mono block sits at stock height because it's on the VRMs as well as the chip.

Investigate the mount and see what the spread of your tim is like. You might just need a re mount, or you may need to experiment with thinner thermal pads etc.
The IHS was reglued, I had Silicon Lottery delid the chip; I'm not entirely sure where the IHS sits in comparison to stock at this point. I do believe I may have slightly over tightened the two screws for the VRM portion which may be skewing the block; at points during x264 loops I have up to a 12c delta between cores, average core to core delta seems to be around 8c.

I went to remove the block but unfortunately the rear portion of my case blocks one of the VRM screws so I can't easily remove the block without disconnecting the loop at this point. I also think the socket design on the OC formula is slightly different from some of ASRocks other x299 boards and I can't mount the block as in the directions, so I may need some thinner thermal pads for the vrm.

Honestly, I don't think your temps, when compared to ambient, are bad at all considering you are using a monoblock. I would check to see if your block is making good contact with the IHS as someone else has suggested. If you cleaned off all the silicone on the die before putting the IHS back on chances are that the IHS is sitting slightly lower than it would at stock. I'm running a delidded 7900X OC'ed to 4.7 on all cores at 1.24V and my temps are in the low 70's under load and can peak at ~81 when running Realbench or Cinebench. I am using 2 EK PE 240 radiators and a D5 with a full nickle EK Supermacy block. I have a DerBauer Die Frame ready to go on in a couple days in hopes of dropping temps a little more.
As mentioned earlier I did go to remove the block to check the spread but can't because the motherboard tray is blocking one of the vrm screws, but this did lead me to recheck the thumbscrews on the cpu portion of the block, 2 of 4 which were able to be turned a few more times. It didn't seem to help. Definitely think there might be a slight issue with mounting (but not as major as I originally thought).

During the retest (x264 loops) after finishing bleeding and tightening the loose block screw, I ended up running the pump at max speed (~2400rpm), and fans at max (~1800rpm), and temps were about 10c lower; peak of 63c on 1 core instead of 72c. At this point I have to take apart the loop to check the thermal pads and spreads and figure it's best to replace the radiator with something more optimized for low rpm fans. I'd like to find a 360mm rad around 45mm thick (should leave me extra space I need between my gpu and the tubes coming out of the pump). I was originally thinking a UT60 but I think that would cause mounting issues with my pump and GPU.

Any recommendations for an low rpm 360mm rad? Top contenders I see for price vs performance @ lower fan speeds (~1000rpm) put the XSPC 360 V3 ($90), EK Coolstream PE ($90), and the HWLabs Nemesis 360 GTS ($63) as the best buys as far as I can tell.
 

·
Overclocker in training
Joined
·
13,929 Posts
Hi,
Just had to take apart a new 3 week old x99 mono block clogged from ek cryofuel
Navy blue premix
Although your temps aren't bad from the pump speed and fan speed was it ever better ?

Go by the cpu package temps or are you already ?
https://www.hwinfo.com/download.php
 

·
GPU Enthusiast
Joined
·
2,560 Posts
Discussion Starter #20 (Edited)
Hi,
Just had to take apart a new 3 week old x99 mono block clogged from ek cryofuel
Navy blue premix
Although your temps aren't bad from the pump speed and fan speed was it ever better ?

Go by the cpu package temps or are you already ?
https://www.hwinfo.com/download.php
The temps have pretty much stayed the same. Not until I attempted full speed on the pump and fans was I able to keep the highest core to 63c which is decent enough. Just to clarify I only have the EK corrosion inhibitor in the loop with distilled water (so I may have misnamed it in the op).

I'm going by peak core which is normally +/- 1c from package temp in hwinfo.

Before I take apart the loop to check the block I'm going to go and replace the rad, the EK 360 SE is sadly the worst performing 360 rad (I originally got most of my parts from an EK kit and thought the rad would be ok). Currently the XSPC RX360 V3 looks like the best all around option for lower fan speeds at it's price, I like the price/performance of the HWLabs BI 360 GTS, but I think the weaker pump I have from the EK kit may be a problem since that rad is almost as restrictive as a cpu block.
 
1 - 20 of 24 Posts
Top