Jump to content
Corsair Community

H80iV2 hit 40C oddly.


GuruSR
 Share

Recommended Posts

Been using the machine for some light gaming and usually the H80i's fluid temp hits around 37.1C, well, I was away from my desk when I heard a beep and then the machine went off 30 seconds later (as the fan noise vanished).

 

Came back a few minutes later (as I couldn't leave to see what happened), powered the machine on and the fans all hit 100% instantly, in Windows, the fluid was sitting at 39.7C, never seen it that high as I have it set to 37 to be at 100% fan speed, so something has changed. Though today, our ambient temperature is lower due to cooler weather, so the ambient temperature in here is actually far lower than it has been. I put the pump to Performance and am using the CPU temp (so when it reaches 40C, the pump will hit 100% on fans), so now I'm listening to variable fan noise from a low air rush to a jet engine. (I don't need this CPU being damaged, it's a replacement already.)

 

I have the Corsair software now set to 39C to shutdown, the 40 probably wasn't a good idea, the CPU has another software monitoring it and at a safe to T-Case 58C sleep setting, so I know which one happened, the Corsair software shut the machine down. My concern is now, what caused this issue in the first place, can the innards be actually breaking down this early or could there be contaminants in the fluid causing issues with cooling? At power on, the fans went to 100% in windows the fluid temp was 39.7C, it took roughly 1 minute for it to lose 0.1C. Both fans are in a push/pull method around the unit, all hoses are "soft curves" and tension is as low as I could get it to avoid any hoses pulling the connections apart.

 

GuruSR.

Link to comment
Share on other sites

Any particular reason why you set a shutdown to 40 or even 39° on the coolant temp?

 

35 - 45° coolant is very typical on AIOs. yours running at 40 is totally fine.

 

You may want to look at the radiator, and see if it's not clogging with dust, just under the fan doing the push

Link to comment
Share on other sites

Any particular reason why you set a shutdown to 40 or even 39° on the coolant temp?

 

35 - 45° coolant is very typical on AIOs. yours running at 40 is totally fine.

 

You may want to look at the radiator, and see if it's not clogging with dust, just under the fan doing the push

 

Well, it was recommended before not to go past 40 on this model from someone on this forum who knows their stuff.

 

GuruSR.

Link to comment
Share on other sites

And those pump speeds are consistent, regardless of the CPU temperature (not the pump)?

 

What is ambient temp? Is the cooler intake or exhaust?

 

40C isn't a "shutdown" temp but it's a good 'rule of thumb' to keep it under that. I wouldn't start an auto shutdown until it gets closer to 50C. But the recent increase in temps could be indicative of a larger issue and I'd like to see if we can determine if that's the case. We do see increase in coolant temps when there is blockage, pump failure or when the pump isn't powered properly. Your pump speeds would indicate that it is being powered properly so I'd like to rule out the others, if we can. Another challenge with the H80 is that its typical installation (exhaust, rear) leads it to be quite warm as it pulls GPU waste heat through it. But the 1050 isn't that particularly hot (compared to, say, a 1080Ti or a 2080)

Link to comment
Share on other sites

And those pump speeds are consistent, regardless of the CPU temperature (not the pump)?

 

What is ambient temp? Is the cooler intake or exhaust?

 

40C isn't a "shutdown" temp but it's a good 'rule of thumb' to keep it under that. I wouldn't start an auto shutdown until it gets closer to 50C. But the recent increase in temps could be indicative of a larger issue and I'd like to see if we can determine if that's the case. We do see increase in coolant temps when there is blockage, pump failure or when the pump isn't powered properly. Your pump speeds would indicate that it is being powered properly so I'd like to rule out the others, if we can. Another challenge with the H80 is that its typical installation (exhaust, rear) leads it to be quite warm as it pulls GPU waste heat through it. But the 1050 isn't that particularly hot (compared to, say, a 1080Ti or a 2080)

 

The ambient temp was around 22C (was a cool day, air wasn't even on), 40C was what I was told to keep it under, so I have been. Now the case actually has a filter in front of the fans and I do clean it on the 3rd of every month, I looked and there was so little that it wouldn't have caused this instant influx in temp, but, I have the 6700K underclocked to 3600MHz because at 4200MHz, the H80 loses the battle to keep the temp in check (under the T-Case which Intel sent to me in an email circled, telling me to keep it below that). I haven't done anything to the chip because this was a factory replacement after a motherboard "hiccup" that neither party could explain who caused it. The H80 is on the bottom of the front of the case, so it's exhaust actually heats the 1050. I figured the 1050 wouldn't get that much hotter, since it is the OC version and is supposed to cap at around 130C, which it never gets anywhere near, typically it is about 10C hotter than the temp on the H80.

 

What could have happened without me seeing, the very rare occasion, Microsoft's reduction on the CPU doesn't take, I wasn't watching the core speed and it could have gone to 4200MHz, which the H80 won't keep up with, while I wasn't in the room, it is quite possible it pushed the H80 up to 40C and the machine shut down, which is why I couldn't find the issue when I turned it on (aside from it being hot and all fans on full). But that aside, what other means would there be to determine IF there is an issue with the H80, as I don't recall seeing any viewing window anywhere to see if the fluid is indeed running through.

 

GuruSR.

Link to comment
Share on other sites

What vCore are you seeing on the CPU?

 

1.040 Minimum, 1.104 Maximum and right now the coolant is still creeping up past 38C with the fans at 100%, it isn't keeping it down, usually it sits around 36.8-37.1 depending on room temps, but with this room being cooler than normal (~22C), I can see this hitting and going past 40C easily.

 

GuruSR.

Link to comment
Share on other sites

If you are only 1-2C out of your normal range, it could be most anything. The most likely is simple fluctuations in local room temperature. It is not going to be uniformly X degrees everywhere and the case will produce a heat aura.

 

Even if you jack your voltage and frequency to the maximum, there is a clear power limit on the CPU and and it is really hard to see the 6700K going past 150W. It is the watts that matter for the coolant temperature, not the actual CPU temp per se. It's hard for me to imagine a 6700K overwhelming a working H80i.

 

I agree you need to move the 40C shut down point up. 50C sound fine to me. There is nothing magical about 40C, other than that is the fan max programmed into the Corsair curves and for a highly overclocked CPU, 40C coolant/base temp is about the highest you can go before hitting the thermal limit. People with more normal overclocks will not run into the ceiling as easily and nothing happens at 41C.

 

How are things at idle with the coolant temp? If you have some type of building blockage that is reducing performance, you should see it elsewhere too. One of those places is at cold boot and idle. Power on and the coolant should sit just above room temp for a while. If you can watch it quickly tick up +1C every 30 seconds, the fluid may not be getting around. Another thing to test is flip the pump to Extreme/Performance. If the coolant suddenly drops at idle, there probably is a building restriction. Also look at post load behavior. The coolant should start dropping immediately. The last few degrees may take more time, but if the coolant temp seems to hang at the load level when you quit, again signs of low flow rate.

Link to comment
Share on other sites

If you are only 1-2C out of your normal range, it could be most anything. The most likely is simple fluctuations in local room temperature. It is not going to be uniformly X degrees everywhere and the case will produce a heat aura.

 

Even if you jack your voltage and frequency to the maximum, there is a clear power limit on the CPU and and it is really hard to see the 6700K going past 150W. It is the watts that matter for the coolant temperature, not the actual CPU temp per se. It's hard for me to imagine a 6700K overwhelming a working H80i.

 

I agree you need to move the 40C shut down point up. 50C sound fine to me. There is nothing magical about 40C, other than that is the fan max programmed into the Corsair curves and for a highly overclocked CPU, 40C coolant/base temp is about the highest you can go before hitting the thermal limit. People with more normal overclocks will not run into the ceiling as easily and nothing happens at 41C.

 

How are things at idle with the coolant temp? If you have some type of building blockage that is reducing performance, you should see it elsewhere too. One of those places is at cold boot and idle. Power on and the coolant should sit just above room temp for a while. If you can watch it quickly tick up +1C every 30 seconds, the fluid may not be getting around. Another thing to test is flip the pump to Extreme/Performance. If the coolant suddenly drops at idle, there probably is a building restriction. Also look at post load behavior. The coolant should start dropping immediately. The last few degrees may take more time, but if the coolant temp seems to hang at the load level when you quit, again signs of low flow rate.

 

The temps sit around 25c on the cores at idle on boot, but according to the Corsair software 28.1 was the lowest fluid temp it saw (not sure if that is accurate), after "Sims 4" for a few hours, on/off due to the temp climbing past the 37.1 it was staying at and creeping to 38.2, this in itself seems to also take (after the last gaming hit that mark) a good 2 minutes before it drops below 38 and the cores dropped down to 28-30 while it was still sitting at 38.2 with fans at 100%, this isn't an air flow issue, as there is a filter in front of the fans and it is in the corner and the filter covers it completely and I monthly clean out what little it gets. I did leave the pump on Performance, but it was always on Quiet, not knowing the volume per foot movement inside the loop, I can't determine which method would be best suited for cooling the fluid faster while it is in the radiator. Right now the coolant is sitting at 34.1, it usually (and very very slowly) works down to roughly 32-33. The 5 minutes I spent writing this (with interruptions), coolant is sitting at 33.5, which is probably where it will hover until possibly getting down to 33.1, never drops below that once on, seeing 27-29 on the cores. The coolant has never jumped up in temperature very fast when the cpu usage spikes, but it does take a very long time to get back down.

 

I'll give the machine a few hours of idle time to let it settle, then I'll check the pump differences.

 

GuruSR.

Link to comment
Share on other sites

It's tough to make a definitive call at this point without clear evidence either way. Is this new behavior? For instance, 2 weeks ago it would be sitting at 34C coolant temp for gaming and now I am touching 40C?

 

If I didn't know better, I would ask if the radiator was in the rear exhaust slot. We know it's not, but looking at the pictures of the P50, I see a solid front panel, with side inlet airflow, and what appears to be a large honeycomb dust filter. There's no way for me to assess the filter from pictures, but no filter helps airflow. It is possible the combination creates slightly higher coolant temps than we expected. The quick drop of the CPU cores below coolant temp after gaming is typical of environmental heat driving the coolant up vs a a blockage where the cores would stay above coolant almost always.

 

However if that is true, we should probably be looking at this differently. There is nothing wrong with hitting 40C on the cooler. This value is massively environmentally dependent. Someone in a Mediterranean climate will spend most of the year at 40C. If there hasn't been a significant slide in performance (higher temperatures), then we are not likely looking for a problem. Whether or not you use the dust filter is up to you. If you are happy with the end CPU temps, then there isn't an issue at all. CPU and coolant temp have a +-1 relationship. If we can reduce your coolant temp by 4C, that will only reduce your CPU temps by 4C. That may not be of real value given what it would take to reduce the case temp by 4C.

Link to comment
Share on other sites

It's tough to make a definitive call at this point without clear evidence either way. Is this new behavior? For instance, 2 weeks ago it would be sitting at 34C coolant temp for gaming and now I am touching 40C?

 

If I didn't know better, I would ask if the radiator was in the rear exhaust slot. We know it's not, but looking at the pictures of the P50, I see a solid front panel, with side inlet airflow, and what appears to be a large honeycomb dust filter. There's no way for me to assess the filter from pictures, but no filter helps airflow. It is possible the combination creates slightly higher coolant temps than we expected. The quick drop of the CPU cores below coolant temp after gaming is typical of environmental heat driving the coolant up vs a a blockage where the cores would stay above coolant almost always.

 

However if that is true, we should probably be looking at this differently. There is nothing wrong with hitting 40C on the cooler. This value is massively environmentally dependent. Someone in a Mediterranean climate will spend most of the year at 40C. If there hasn't been a significant slide in performance (higher temperatures), then we are not likely looking for a problem. Whether or not you use the dust filter is up to you. If you are happy with the end CPU temps, then there isn't an issue at all. CPU and coolant temp have a +-1 relationship. If we can reduce your coolant temp by 4C, that will only reduce your CPU temps by 4C. That may not be of real value given what it would take to reduce the case temp by 4C.

 

It is not common for sure, since having the unit in place, running that game and other more taxing games, never got past 37.1C, it always slowly creeped up to there and stopped going further, even in warmer days unlike today, though the coolant got down to 31.4C with the cores sitting at 26-28, but that is idle for about 90mins. Typically in more summer weather it is ~33.1 at idle and 37.1 at max, while it being cooler, it's hitting 38.2, the other game (SWTOR) pushes the CPU harder and I am certain it would again pass 39 and hit that limit I have set, have played SWTOR for hours in the past with this unit and it never did this, don't even think I could hit an hour with SWTOR now, as when it happened, SWTOR was on and I was away from the system and it was off in less than 15 minutes.

 

Yes, no filter helps airflow, but dust in a radiator doesn't help it work either, so the filter is a necessary evil to avoid clogging the radiator. As for the cores, when they were air cooled, the chip typically (in the same case) were 23c on idle on warm days and 19C on cool days, so the temps I'm seeing now are lower than the coolant, but while the temp sits at 38C+, the CPU at idle won't go much below 34C, so I am seeing the cores pre-heated by the coolant during that time, they do drop down to the somewhat normal for a bit but bounce back up, only after the coolant gets down to close to 30C (as it never gets actually to 30C or below), the cores sit below it by roughly 4 to 5C, which is typical with a TIM'ed heat spreader, as the TIM as acting as a mild thermal barrier otherwise the coolant's temperature would ramp up faster than I have seen it do so. The H80 is located in the front of the machine (bottom fan spot, in line with the exhaust of the 1050Ti), so it is pulling low room air in directly, did that to remove the recycled air from inside the case and the 1050Ti isn't suffering from it (but is matching or beating the H80 in temps, because it is constantly heating it). Right now, the H80 is at 31.6C and the 1050Ti is at 30C and it isn't doing anything to warrant that temperature, it's being heated by the H80's exhaust which is practically at the end of the card (less than an inch away). My only worry was the fact I couldn't get the hoses out of the way of the exhaust from the H80, so I can imagine that is also helping re-heat the coolant in those lines.

 

GuruSR.

Link to comment
Share on other sites

 Share

×
×
  • Create New...