Jump to content
Corsair Community

Concerned about temps with pump in quiet mode.


Astral85
 Share

Recommended Posts

Apart from my 100% fan issue which some of you here are aware of I'm also not too happy about the temps I have been seeing lately or rather over the last few months.

 

It is the pump in quiet mode that seems to be struggling, I'm seeing 5C+ in load temps than I have in the past. I have also noted how rapidly the coolant temp is rising under load, really a matter of minutes from 30-40°

 

Now in spring coming into summer I would expect some temp increases but I have been concerned about the overall performance for at least the last few months while still in winter after I clean installed Windows 10.

 

A month or so ago I was getting 68° in OCCT load test. Today I get 75° and very quickly. I stopped it there because that is warm enough for me.

 

Maybe it is just an increase in ambient temp but the rapid rise in coolant temp doesn't seem right...

Edited by Astral85
Title change
Link to comment
Share on other sites

  • Replies 83
  • Created
  • Last Reply

Top Posters In This Topic

Current gaming temps are:

 

Coolant temp: 40.5C , CPU: average 53-55C, max 61C

 

Core clock 4.0Ghz, 1.120v

 

40 mins gameplay.

 

Is this rather average? Is 40.5C quite on the high side? What concerns me is that I've dropped my CPU voltage considerably over the last few months, from 1.220v down to 1.120.

 

It might sound silly but I have seen better temps than this when I was using the higher voltage. To be fair it might have been the middle of winter and now it is spring.

 

Idle coolant temp is 30-31C. Is a 10C rise in coolant temp under these circumstances a bit extreme?

Edited by Astral85
Link to comment
Share on other sites

I don't think a 10C rise is extreme for an H80i in the rear slot as exhaust. When gaming, it isn't just CPU waste heat increasing the water temperature, but also GPU waste heat raising the temperature of the surrounding area. Now a 970 isn't a real burner, but it will still increase the local area temps and your water temp by a few degrees. Your CPU temps are still well below the concern point, so I don' think this is a real limitation until you start raise the voltage with a higher overclock.

 

We often dismiss differences between Quiet and Performance mode on some of the other coolers, but the smaller the unit and the lower the volume of water inside, the more the flow rate matters. On an H80i, you may see a benefit from using the performance mode when gaming. Obviously that is simple enough to test out.

 

If you want to see how much your GPU waste heat contributes to the overall water temp, run a CPU only benchmark and compare. For this particular test, I would use Intel XTU. It's a mild stress test that also has very even loading. You don't need to run it for 40 minutes. 10-20 min should be enough to replicate the conditions. Mark your starting and ending water temperature. Water temp delta should be consistent between seasons, although the starting water temp will vary with room temperature.

 

Of course, the real irritant at the moment is the jump to 100% fan speed when it hits 40C, even on Quiet mode. I don't have any further answers for that yet, but perhaps someone else does. If Performance Pump mode will keep you under 40C, that might be a good compromise for the time being.

Link to comment
Share on other sites

Do you think it's more likely a rise in ambient/room temperature that's contributing to or causing the higher temps? I probably should have named this thread something more like "Am I getting the best possible cooling/temps". I know people will always say things like those temps are fine, but are they the best? I also really should be monitoring and documenting my coolers performance and temps throughout the seasons a lot better!

 

I notice that Intel XTU offers tuning facilities, are these still better applied in the BIOS? How does the stress test in XTU compare to others? You say it's quite mild, what sort of temps would you generally expect from this test? I've briefly ran a 5 min stress test with a max of 72C. Sound ok or too high for given test/program? You also said that this could give me an idea of cpu temps without the added gpu's heat, is there a particular test for that because 72C is not my normal gaming temps.

 

As a temporary mitigation against the 100% fan triggering I have set a custom profile for the max coolant temp to be 50C, therefore under extended gaming periods where I seem to maxing at 41-42C coolant temp the fans aren't being triggered. I'll update this in the other thread. I will also contact Corsair for their explanation on this.

Edited by Astral85
Link to comment
Share on other sites

Ambient temperature increases/decreases will be additive to the base water temperature. So, when it's 20C in your room and the H80iGT Water temp is 26C, then when it's 25C room temp it should idle at 31C. Weather is turning colder on my end and I can now appreciate the significance of a 5C drop in room temp.

 

Unfortunately, I would need both a 4690K and a H80iGT/v2 to give you some meaningful feedback on the data. However, your cooler is not malfunctioning. I am a little surprised your reached 72C on XTU and 75C on OCCT. I would have expected a larger spread between them. However, sometimes peak numbers can be deceiving. There is a big difference between 72C for a split second in time and averaging 72C. If you configure the XTU display to show you the line graphs for your individual core temps, you can see a nice fluid sine wave like graph. I suspect this median was much lower than 72C.

 

Don't use XTU to overclock. Use the BIOS. It may work, but software based tuning is always a bit iffy and it requires XTU and your BIOS to have a little chit chat. Things may get lost in conversation.

 

I think the custom curve is a good idea and ultimately what most users need to do anyway. I usually advise the same thing for people with big watt GPUs running 4K who are heating up their cases quite a bit. They also run into the 40C water temp issue and don't really need their fans at 100%. Different hardware, but really the same issue.

Link to comment
Share on other sites

Hi c-attack,

 

A couple of things stand out in your reply, you being surprised in my XTU and OCCT temps and people running into the 40C water temps with big watt/multi GPUs which I don't have.

 

Why are you surprised by those temps would you expect lower? I will run the test again but from what I remember it was averaging 65C and peaking 72C.

 

Are you suggesting that 40C water temps for my hardware configuration is out of proportion?

Link to comment
Share on other sites

No, I think 40C water temp on a H80i/GT/v2 is pretty common when in the rear exhaust. After gaming, and perhaps more so on short CPU stress tests, does the H80iGT Temp return back to near the 30C mark? It may take a few minutes, but it should. A long gaming session may take longer to drop, particularly if the case ambient temp has increase during the session. That was the connection with "other large GPUs". In those examples they are using 280mm radiators in the roof, but the connection is their water temp is increasing not from CPU load or lack of effectiveness, but because of a general increase in case temperature. When this happens, the water temp will increase right along with it --- load or no load. You make less GPU heat, but the radiator is in closer proximity to the source.

 

I expected XTU to come in lower because on my two set-ups it is a good 5-10C below AIDA64 on average and AIDA is a good 5C lower than OCCT. Now, we certainly have different models CPUs and the design elements of the 4690K are pretty well documented in relation to heat exchange. This is why I would like another 4690K owner with a H80i. However, my first impression is related to CPU design and not cooler performance. One way to check would be to compare the water temp delta between the two tests. Take a start and end reading for XTU and OCCT around a 10 minute run (or however long). The water temp change will give you an approximation of how much heat is dumped into the water stream. The fans remove heat from the water, but at the same speeds the removal should be equal.

 

I don't think you need to do this, but if you are hunting for more information this may be of interest. The water temp change will assess cooler function. The difference in end CPU core temps versus peak water temp gives you an idea of how much thermal stress is placed on the CPU. Not all 100% loads are created equal. I don't see a cooler problem (other than the fan thing) and I think if you had one, your gaming temps and everything else would suffer too. At 61C you still have 20C of headroom at a minimum. That's quite a bit, but I am willing to continue analysis if you want.

Link to comment
Share on other sites

I found this today, I hope it will remind people that maintenance of *all* computer parts is necessary. This is about 19 months of never taking the radiator apart to clean it!

 

Unfortunately and surprisingly? my temps aren't amazingly better after the cleaning, maybe 3C, so I'm still with the thinking that I've still got an issue elsewhere.

 

So XTU should come in lower than both AIDA64 and OCCT? I've done a brief re-run of XTU after the cleaning with maybe a 3C change but still a peak of 70C.

 

I'll get some readings with the two tests you've suggested. In the mean time what ideas do you have where we can head with this? I really appreciate your time for this.

 

20161016_152939.jpg

20161016_152846.jpg

Edited by Astral85
Link to comment
Share on other sites

Yay some better luck with XTU this time: (I would think that cleaning should do something!)

 

Intel XTU Stress test 5 Mins.

 

Water temp start: 29.2C Stop: 36.5C

65C Peak, Average (HWiNFO64) 48-50C

Fan curve custom, Pump - Quiet mode.

 

Is HWiNfO ok to use for a reliable average temp?

Edited by Astral85
Link to comment
Share on other sites

Oh dear, just as things get better they suddenly get worse :(:

 

Twice now on reboots the idle has been at 40C! compared to 30C despite normal water temp. A 5 min game test of 65C!

 

One of the tubes is *very* warm at the block end and the water temp is *not* coming down.

 

Have set the pump to performance which has brought the water temp back down.

 

Hope I haven't damaged anything while cleaning it, I didn't dismount the block.

Edited by Astral85
Link to comment
Share on other sites

Current idle water temp of 34.3 and CPU 40C, not good :(:

 

I really can't think what's happened, handled the radiator well while I cleaned it, didn't touch or remove the block, did tighten the thumbsrew's a fraction...

Edited by Astral85
Link to comment
Share on other sites

I guess I need to remember to ask is the radiator is obstructed or blocked in some way. A thick layer of dust will do that. Stress test results now more like what I expected and it is likely airflow across the fins was the limiting factor in the first runs.

 

But on to the new problem... The delta between idle water temp and CPU temp is still good, so I don't think you pulled the block off the CPU. It may be some particulate inside the cooler was disturbed when you took the radiator off. It has now settled at the "strainer" at the pump and is restricting flow. You can:

 

1) Let it run and don't do anything overly strenuous. No more stress tests. Gaming and normal use should be OK. Hopefully, it will work out the restriction on it's own.

 

2) Take the cooler out, including the pump off the block. Give the block end some careful shakes to see if you can dislodge whatever is stuck at that end. You will need to clean the CPU and block, then put on new TIM before seating it again. Be prepared before you take it off.

 

3) I suppose a bubble could also do this, but usually you get some pump noise with that and it would be very unusual for a bubble to restrict flow. Still, lying the case on its side for a few hours (while powered on) might give 'whatever' a chance to come loose. This is not my favorite troubleshooting method, but it is less invasive than #2.

Link to comment
Share on other sites

Hi c-attack,

 

Later last night the temps stabilized *again* with a nice little gaming run and a peak of 52-55C, average 45C. I haven't seen temps this good under quiet mode of the pump, that's about a 7C difference that has to be attributed to the cleaning... I also noted the water temp about 5C cooler than before for the same game so looks good...

 

I will have to monitor the temps as it's quite likely something like what you describe happened. So far this morning idle looks good at 30C.

Edited by Astral85
Link to comment
Share on other sites

I think there is definitely something stuck at the pump, or least that's all I hope it is. Just done another short gaming test and the temps look fine but upon exiting the water temp is coming down ever so slowly keeping the idle temps up.

 

I'm happy to just keep running the machine for a bit before considering option #3. It's obvious the pump *is* working and I don't think it's unsafe to be running it in this state. What do you suggest?

Link to comment
Share on other sites

I've started up a gaming test again despite my current concerns but interestingly the CPU *load* temps are remaining stable. The issue seems to be with the idle temps, I had a couple of good moments with 30C idles but for the last 6+ hours it has consistently been idling at 40C. I also note that the water temp sometimes rises to 40C while idling but sometimes falls back to 30C. Despite both high (unusal) and low *idling* water temps the CPU idling temp remains high.
Link to comment
Share on other sites

I've ordered some thermal compound (Noctua NT-H1). If the idles don't clear up before the TIM arrives I suppose there's nothing to lose in remounting the block.

 

If everything fails would you suggest RMA'ing? Unit is 16 months old, I believe these carry a 5 year warranty...

Link to comment
Share on other sites

You are not in a dangerous position yet, but you need to keep an eye on temperatures. No more stress tests -- they serve no purpose. The water temp is the base CPU temperature, so you would need to apply a very strong full core load to hit 80C.

 

A 10C change in idle water temp (H80iGT Temp) is concerning. That and should not happen and indicates there is some kind of flow problem. When it happened once and went away quickly, I had hopes it was unlikely to occur again. However, the re-occurrence suggests the problem is circulating rather then settled at a low point and that means all this shaking just buys more time, until the next time. Obviously, there is not supposed to be a bunch of material floating around inside. I think you should file a support ticket with Corsair directly. This is the first step in an RMA. This is not a binding path, so if it clears up or you decide to go another direction, you don't have to follow through. Also, the number of world wide RMA centers is low and limited to Western Europe and California. Do the advance RMA if you can. If you happen to live in Australia there is going to be some shipping time and expense. You may want consider alternatives as well.

 

I would not take the block off before getting more TIM. That only adds another layer of complication and potentially even worse CPU temps. No margin for that right now. I would tip the case back and forth or over before trying that.

Link to comment
Share on other sites

Yes, I was going to ask you about that, what would this material be? I was also going to ask, where would it eventually settle, but it's now possible as you suggest that it's circulating.

 

It is definitely intermittent behavior (the high and low idle water temps) but what surprises me is how it's not affecting the load temps and that they're actually better than before. Would you expect this to affect the load temps as well?

 

I think I will set up an RMA and continue monitoring. Are you suggesting removing the block to shake it is not worth it anymore or is it still worth a shot?

 

Not quite Australia but not far from it. I live in New Zealand actually. I've RMA'd a mouse with Corsair and it was replaced by the store in NZ that I originally brought it from based on Corsair RMA granting. Maybe I could have that same luck.

Edited by Astral85
Link to comment
Share on other sites

Well, at this point I am clearly speculating... It's not like someone left a hunk of plastic loose in the unit. It is most likely some combination of anti-corrosive and the sealant adhesive that managed to congeal or coagulate into a mass. It then gets stuck around the pump intake/outflow. If the specific gravity is high enough, it will settle at the low point and be relatively harmless. If it is light, it's going to float around and cause re-occurring blockages and strange temperatures.

 

This is not your cooler or even a Corsair product, but it does show an extreme example of the issue.

http://en.community.dell.com/resized-image/__size/550x0/__key/communityserver-discussions-components-files/3746/photo.JPG

 

 

It may be that is how Corsair handles RMAs for Oceania. The shipping cost would be bad for everyone.

Edited by c-attack
Link to comment
Share on other sites

That sounds feasible. Temps look good first thing this morning with a cold boot water temp of 20C rising to 25.8C in 15-20 mins from cold boot. Idle temps are what I like to see at 30C. I suppose it's matter of just continuing to monitor.

 

I've had a quick reply from Corsair requesting the original invoice for the unit.

Link to comment
Share on other sites

Is the buildup that yellowish stuff on the block? What exactly is the low point? The radiator reservoir?

 

Whatever is the lowest point of gravity in the system -- radiator, tube droop, pump block. I suppose you could bend the tubes slightly to drop below the pump. However, it seems more likely the particulate or whatever is not going to sit still. I wouldn't put a lot of faith in that as a preventative measure.

 

Sounds like you are back to normal today. No more tipping or shaking.

Link to comment
Share on other sites

Things are so far normal today and I took some notes starting at the cold boot this morning. The temps this afternoon may need analyzing.

 

Cold boot water temp: 11:12 AM 20C, CPU idle 26-30C

 

I can imagine it's quite normal for the water temp to be at it's lowest on the first cold boot after sitting unpowered all night...

 

 

12:00 AM water temp: stable at 26.5C, CPU idle 28-33C

 

1:20 PM water temp: 27.4C, CPU idle same.

 

I then put the machine to sleep and have just woke it at 5:00 PM (3.5 hours sleep)

 

5:00 PM water temp: 32.1C, CPU now idling between 34-38C

 

We are at the start of summer here in NZ and it's starting to get quite warm, probably 20C or so. My room in particular really gets hit by the sun in the afternoon and can get very warm, in fact my phone is reading 24.1C (5:30 PM) Would you say the higher water temp and idle this afternoon look to be in line with ambient/room temperature? Or is it idling too high now?

Edited by Astral85
Link to comment
Share on other sites

 Share


×
×
  • Create New...