Jump to content
Corsair Community

Time to upgrade/replace my H80i??


Hocknose

Recommended Posts

I have one of the original H80i's, had it for about 3-4 years.

 

I've recently noticed that my CPU has started to run very hot while under load.

 

My idle temps are <=30 degrees, but when I play quite an intensive game such as Destiny 2 or AC:Origins I've noticed that the temps can go as high as 85+ degrees and regulary sits at about 75 degrees.

 

I have an I7 7700k (not OC'd)

 

Things I've done to try and remedy it:

  • Cleaned all dust from radiator
  • Reseated the cooler and reapplied thermal paste
  • Have fan speed run at max rpm

 

Am i being over cautious here or do you think the cooler may not be working as good as it used to?

 

Is it time to upgrade or replace the cooler?

 

Thanks

Link to comment
Share on other sites

Take a look at the H80i Temp in Link. This is your coolant temperature and a better indicator of where the problem might be. While case and load specific, coolant temperatures will range from 4-7C over your room temperature and a minimum up to about 40C. If you mount a H80i in the rear slot, you might see near 40C during GPU load during warm months.

 

Other indicators the H80i is not behaving include very slow to drop CPU (and coolant) temperatures after a load. So, it goes to 80C when gaming, drops to 50C when you quit, then takes 2 hours to come back down to 40C.

Link to comment
Share on other sites

Thanks for the suggestions...

 

Just did a few tests...

 

My room temperature is 18.5 degrees and my coolant was 30 degress idle, it is mounted in the rear slot.

 

I just got the cpu up to 84 degrees, quit the application and it went to below 40 degrees in less than 10mins. In this test the coolant went to 42 degrees and is now sitting at 32 degrees after about 10 mins.

 

So apart from the quite high temp of the coolant at idle the temps all came down quite quickly...

 

maybe its my case that is the issue...

Link to comment
Share on other sites

OK, then lets look at this from the other side. Kaby Lake is notoriously temp jumpy and voltage heavy when on the auto setting. In Link or any other monitoring program, what is your peak Vcore recorded? What do you see value when on the desktop? Your coolant to CPU temp delta is about 45C. That is too much for stock clocks. The cause can be poor contact, bad TIM, but I think most likely we are looking at an overly aggressive auto voltage response. If it were either of the first two, you would not see the corresponding rise in coolant temperature.

 

I have been playing Origins all morning. On my 5930K it eats up 60-80% CPU usage, although my actual CPU core temps are very good in the 40-50C range. If your voltage is climbing up into the 1.30-1.40 range, that would explain a lot.

Link to comment
Share on other sites

The highest temp I remember seeing recently is 94 degrees, this was reflected in both Link and an application called RealTemp 3.7. The temps in both prgrams match always.

 

OK I'l look into the voltage of my cpu...what should the voltage be for stock i7 7700k...1.2v?

 

edit...just noticed that my core voltage in CPU-Z is going up and down constantly...does this mean its set to auto?

Link to comment
Share on other sites

There is no exact voltage for a 7700K. Each sample will have it's own auto voltage table it follows. Yes, for the stock 4.5 Turbo frequency, something in the 1.25 region would be enough. What has changed is how it handles loads and the voltage it provides accordingly. For Skylake and even more so on Kaby Lake, that Auto setting is far too adaptive. Intel's purpose is make sure you are stable in all circumstances and the easy way to do that is to pour on some voltage. However, if you combine that with unanticipated requests or in an nontraditional manner, the voltage seems to really scale upward. This is why most board makers give you a warning about running synthetic stress test with an adaptive voltage and this includes the Auto adaptive setting. More specific to games, the way in which they load their shader models when you first launch often creates a huge near maximal CPU load. (AC Origins, ME: Andromeda, etc.). Whether this is necessary or sloppy programming is another matter, but if you have something that keeps a line graph of core temps, you may often notice your high marks are in the first few seconds after launch and not during actual gameplay. That should ease your mind somewhat, however I still think you should set a specific voltage in the BIOS, even if you do not intend to overclock.

 

I don't know what board model you are on, but most have some similarities. This overclock guide is for Asus and it is technical, but you can probably get the gist of it. You need to set some voltage, no matter what you intend to do with clocks and most boards have a setting that will tighten the overshoot on voltage, whether this is a IA Load Line calibration or some other named variant.

Next time you play a game, take a look at the peak Vcore number after a few minutes. Also, keep some kind of running temp line graph up if you have one. Corsair Link provides an OK one. You may be able to see more by letting Intel XTU or AIDA run in the background. XTU is free and if you click the wrench on the line graph it can be configured to show individual core temps. It can track Vcore as well.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...