The Corsair User Forums

The Corsair User Forums (http://www.asktheramguy.com/forums/index.php)
-   Cooling (http://www.asktheramguy.com/forums/forumdisplay.php?f=155)
-   -   HELP - H100i GTX + I7 4790K (http://www.asktheramguy.com/forums/showthread.php?t=181564)

Lautaro46 10-04-2018 12:59 PM

HELP - H100i GTX + I7 4790K
 
Hi people! First of all I apologize for my bad English xD, I do this post because of a doubt I have for a long time and I do not have the opportunity to talk to whoever happens something similar and I've been with the same problem for years.

The theme is the following, I have an Intel I7 4790k mounted on an Asus Maximus VII Hero (with the latest update of Bios)

Currently I have the processor running at stock frequencies, with the turbo mode activated at 4.4Ghz with 1.195v voltage, and the average temperatures it gives me in games are between 68 and 70 with maximum peaks of 76 (the thermal paste which has placed is Artic Cooling MX4)

Clearly with these temperatures it is impossible to think of any type of OC, since when I raise the voltages to more than 1.25v, I get to have peaks over 90 and I am forced to cut the tests.

Now the query is the following, do you think these temperatures are normal using this type of refrigeration? I've been with this problem for years, and I never find any kind of solution to this issue, as a detail I could say that when placing the heatsink, I feel as if the back base was half loose at the time of mounting the screws, although I saw in many videos that this is so, I also tried to use rubber grommets to leave it completely rigid and still I still have the same problems ....

On the other hand, from the Corsair Link software, I can not establish personalized profiles, since I put the values ​​in the table but it seems not to respect the climbs according to the marked temperatures, however if I put any of the 3 profiles by default (Relaxed, Balanced or Performance) if it respects these values, but I would like you to recognize the values ​​in the personalized load, since even using the Performance profile the difference that I get with the Relaxed is less than 3 .

I am waiting for your answers and from now I thank you in advance!:laughing:

DevBiker 10-04-2018 01:08 PM

The temps don't sound unreasonable for a non-delidded processor.

When you create your profiles, what is the temperature that you use for the source of the fan curve? Typically this should be your coolant temperature (and that's the default) so you won't see the fans speed up as the processor heats up, only as the coolant heats up. Note that this is the expected behavior.

Lautaro46 10-04-2018 11:55 PM

Quote:

Originally Posted by DevBiker (Post 975163)
The temps don't sound unreasonable for a non-delidded processor.

When you create your profiles, what is the temperature that you use for the source of the fan curve? Typically this should be your coolant temperature (and that's the default) so you won't see the fans speed up as the processor heats up, only as the coolant heats up. Note that this is the expected behavior.

First of all, thanks for the quick response! Do you really think it's normal? because I look at many reviews on this model and in all temperatures with the processor to more than 4.6Ghz gives them in OCCT no more than 62/64 and in my case I get above 90 only in 4.4Ghz ....

Do you think it is possible to improve my temperatures by changing the factory cooler for others with better flow rates? I discard what to do delidded since I do not want to risk losing it and never did it before

DevBiker 10-05-2018 01:04 AM

What are your coolant temperatures?

Lautaro46 10-05-2018 06:26 PM

Quote:

Originally Posted by DevBiker (Post 975228)
What are your coolant temperatures?

After a test of 5min with OCCT with the processor in its Stock frequency (4.4Ghz to 1.198v) it marks me a minimum of 22.5 and a maximum of 33.3 , this keeps going up if it is continuous, at that point the processor set the temperature of core at 88

c-attack 10-05-2018 08:05 PM

You had a 10C swing in coolant temp in 5 min of OCCT? That probably should not happen, however the root cause may be a little trickier to pin down.

A failing cooler is terrible in all aspects, load or not. If there is a flow problem, the coolant will continually rise on the desktop with little to no load. In a normal cooler, the instant power on coolant temp is the same as the room/case temperature and then warms up +4-7C over the next 10-20 minutes before leveling off. A cooler past its prime will continue to creep up while using your web browser.

A cooler that is terrible at load but appears normal at idle may be restricted in some way. This could be placement, a case panel blockage, dust filter, or a carpet layer of dust itself. Temps are worse than they should be, but steady and not continually escalating as the paragraph above.

Coolant temp should start dropping the moment you cut the load, often several degrees in the first 1-2 minutes before a slow decline. A cooler that cannot shed it;s heat will retain a lot of heat even after the load stops. This can be from either of the above scenarios.

Lautaro46 10-08-2018 07:07 PM

Quote:

Originally Posted by c-attack (Post 975318)
You had a 10C swing in coolant temp in 5 min of OCCT? That probably should not happen, however the root cause may be a little trickier to pin down.

A failing cooler is terrible in all aspects, load or not. If there is a flow problem, the coolant will continually rise on the desktop with little to no load. In a normal cooler, the instant power on coolant temp is the same as the room/case temperature and then warms up +4-7C over the next 10-20 minutes before leveling off. A cooler past its prime will continue to creep up while using your web browser.

A cooler that is terrible at load but appears normal at idle may be restricted in some way. This could be placement, a case panel blockage, dust filter, or a carpet layer of dust itself. Temps are worse than they should be, but steady and not continually escalating as the paragraph above.

Coolant temp should start dropping the moment you cut the load, often several degrees in the first 1-2 minutes before a slow decline. A cooler that cannot shed it;s heat will retain a lot of heat even after the load stops. This can be from either of the above scenarios.

Here I am again! using the equipment in a normal way, in desk tasks and others, the temperature of the coolant does not rise to more than 34 , but in games, after more than 1 hour, it marks me average of 45-50 , translating in temperatures of processor of about 76 , always in stock frequencies and vcore at 1.19v, I can not think of more power to do to test possible problems ...

What would be the way to prove if the pump is defective? since problems of dirt and placement, I would discard them, since I verified it several times, like the change in the thermal paste, using artic cooling mx-4, changing the coolers that come from stock would be a possible option?

c-attack 10-08-2018 07:16 PM

Quote:

Originally Posted by c-attack (Post 975318)
You had a 10C swing in coolant temp in 5 min of OCCT?


That is how you prove the cooler is not working effectively. Coolant temperature is the primary data for determining if your issue arises from the cooler or something else. Most people will average about +4-7C over their room temperature when at idle. CPU only loads (at 100%) another +6C or so. Gaming is the mystery value and can be +10C as the GPU heats the case up. Your very large differential when gaming could suggest a cooler problem or a general heat management issue related to the GPU. The way to tell the difference is to cut off the GPU aspect and run a CPU only bench for 5-10 min. That takes me back to the original question: Did your coolant temp go up +10C in 5 min of OCCT? That is cause for concern. If you re-run the test, set your fans to a medium fixed speed prior to starting - something like 1350-1500 rpm to eliminate the fan curve as a bias.

Warning signs to watch for are a coolant temp that goes up really fast, but then takes literally an hour to come back down. A healthy cooler should be slow to rise (maybe +1C per min) and come down at the same rate or faster. Also watch for continual creep upwards when on the desktop. If it's 22C in the room and the coolant keeps drifting upwards 26-28-30-32-34-36... something is wrong.

Lautaro46 10-09-2018 12:44 AM

Quote:

Originally Posted by c-attack (Post 975616)
That is how you prove the cooler is not working effectively. Coolant temperature is the primary data for determining if your issue arises from the cooler or something else. Most people will average about +4-7C over their room temperature when at idle. CPU only loads (at 100%) another +6C or so. Gaming is the mystery value and can be +10C as the GPU heats the case up. Your very large differential when gaming could suggest a cooler problem or a general heat management issue related to the GPU. The way to tell the difference is to cut off the GPU aspect and run a CPU only bench for 5-10 min. That takes me back to the original question: Did your coolant temp go up +10C in 5 min of OCCT? That is cause for concern. If you re-run the test, set your fans to a medium fixed speed prior to starting - something like 1350-1500 rpm to eliminate the fan curve as a bias.

Warning signs to watch for are a coolant temp that goes up really fast, but then takes literally an hour to come back down. A healthy cooler should be slow to rise (maybe +1C per min) and come down at the same rate or faster. Also watch for continual creep upwards when on the desktop. If it's 22C in the room and the coolant keeps drifting upwards 26-28-30-32-34-36... something is wrong.

First of all, thanks for the prompt response, I was just able to do a test again using OCCT, configure the fan profile, through the Corsair Link in Balanced (1600-1850rpm)

Pass a test in automatic mode, at 64bits, with 8 threads, which lasted less than 6 minutes, because the processor reached temperatures of 94 and I was forced to stop the test, the temperature of the liquid in the pump, monitored through the Corsair Link, I throw a maximum peak of 42.4 C, and started with an average of 36.2 C, that is to say that in an average lapse of just over 5min, I change by 6.2 C.

I understand that those would not be good values, is not it?

DevBiker 10-09-2018 07:55 AM

No, that sounds too warm.
What is your ambient/room temperature?
How/where is your cooler configured in the case?
And have you tried running tests with the case open? This can help eliminate the possibility of internal case heat buildup actually causing the issue.

c-attack 10-09-2018 12:27 PM

Yeah, the +6C rise is ok, but the 36C starting point is pretty warm unless it is 30C in the room.

Lautaro46 10-11-2018 11:21 AM

Good! Here I am again, do the tests this time with the box open, and the difference was about 2 , I do not change anything, so I guess airflow problems do not think they are

As for the attached placement photo of how I have it placed, will it change into something if I put the hose on the right side instead of the left one? I notice that many people use it that way, but I do not know if this is a case of problems ... even so the hose on the side that is placed exerts a certain price on the lid of the box

The ambient temperature at the time of the tests was on average about 22 , and I throw the same temperatures as mentioned in the previous post + -2

https://ibb.co/h8hzDp

DevBiker 10-11-2018 11:57 AM

There will be no difference with the hose or pump orientation.

And while it doesn't indicate that you have an issue with airflow, if you are starting at over 30C when your ambient is 22C ... something isn't adding up. At all.


All times are GMT -4. The time now is 05:15 PM.

Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2018, vBulletin Solutions, Inc.