Jump to content
Corsair Community

RPM and temperature difference between two-H110i GT -Help!


theotorm

Recommended Posts

Good morning

 

I wanted to submit this strange thing that happens to me.

I am owner of two H110i GT, one bought in February 2015 and the other bought last month (2016)

I noticed several differences using them on two identical processors (I7-4790K 4.7 Ghz)

Here the test with AIDA64 for 3 hours.

 

H110i GT-2015

 

 

http://i68.tinypic.com/wqx28h.jpg

 

H110i GT-2016

 

http://i66.tinypic.com/28suex0.jpg

 

I noticed these differences

The speed 'of the pump in the one of 2015 turns to 3000 rpm while the new turn in 2850

The liquid temperature is more 'high in that of 2015

The fans run more 'fast in the one of 2015.

Same room temperature

 

As you can see temperatures in the new one in 2016 they are more 'high of 3-4 degrees.

Could you give me an opinion? They are good temperatures in AIO again?

I tried to take it apart and reassemble it more 'times putting the pasta but the situation is still the same.

 

I await your expert opinion

Thanks so much

Link to comment
Share on other sites

First, no two processors are necessarily identical. Even if you can create perfectly identical environmental conditions, right down to duplicate voltages, cases, and fan speeds, you may still see some small temperature differences under full load. You also have nearly 1 hour and 45 minutes difference between the two tests. When you start getting into hours run, other things besides your voltage can play a role, particularly heat retention in the case. Both of the temperature line graphs in view looks smooth and consistent. The problem with using peak values to evaluate a 3 hour run is it would only take a single hiccup, a voltage overshoot, a momentary hesitation in voltage regulation to create that peak number and it may not be indicative of the overall test.

 

I would argue your system is preforming quite well. Those are excellent water temperatures for a test that lasts that length of time. It means the heat is being effectively removed from the radiator and it seems to be doing it with relatively moderate fan speed. With motherboard temps in the 19-21C range, it must be fairly cool in your room, but nevertheless those are good results. There is always more fan speed if you really want to scrub off another 1-3C.

Link to comment
Share on other sites

First, no two processors are necessarily identical. Even if you can create perfectly identical environmental conditions, right down to duplicate voltages, cases, and fan speeds, you may still see some small temperature differences under full load. You also have nearly 1 hour and 45 minutes difference between the two tests. When you start getting into hours run, other things besides your voltage can play a role, particularly heat retention in the case. Both of the temperature line graphs in view looks smooth and consistent. The problem with using peak values to evaluate a 3 hour run is it would only take a single hiccup, a voltage overshoot, a momentary hesitation in voltage regulation to create that peak number and it may not be indicative of the overall test.

 

I would argue your system is preforming quite well. Those are excellent water temperatures for a test that lasts that length of time. It means the heat is being effectively removed from the radiator and it seems to be doing it with relatively moderate fan speed. With motherboard temps in the 19-21C range, it must be fairly cool in your room, but nevertheless those are good results. There is always more fan speed if you really want to scrub off another 1-3C.

 

Meanwhile, thank you so much

I worried about this new h110i-gt that I just bought.

I was afraid that the pump might not touch well the surface of the processor 'cause the old H110i-GT (2015) reached much more' high of 'water temperatures and the fans are more.

I thought of 'water temperature did not increase because' the pump did not touch well the processor. Anyway max 82 ° are good for Aida 64 right?

It does not pass in game 61° the fans to 700 rpm.

So I can feel comfortable?

Link to comment
Share on other sites

If you had a contact error between the pump and CPU, you would idle at high temperatures and under any stress test you would overheat within seconds. Any other problem with the cooler would result in drastic temperature increases and you would not be able to run AIDA for hours. If you want to drive yourself crazy, you can switch the two coolers and see if there is a change in temperature, but that doesn't seem like a good use of your time. You still have a lot of fan speed unused, if you need to bring the limit temperatures down a little. I wouldn't put too much emphasis on a single max temperature value that could have occurred at any point during that long test run. Your average numbers appear to be in the low to mid 60's after hours of running. That is good and shows you have an efficient system.
Link to comment
Share on other sites

If you had a contact error between the pump and CPU, you would idle at high temperatures and under any stress test you would overheat within seconds. Any other problem with the cooler would result in drastic temperature increases and you would not be able to run AIDA for hours. If you want to drive yourself crazy, you can switch the two coolers and see if there is a change in temperature, but that doesn't seem like a good use of your time. You still have a lot of fan speed unused, if you need to bring the limit temperatures down a little. I wouldn't put too much emphasis on a single max temperature value that could have occurred at any point during that long test run. Your average numbers appear to be in the low to mid 60's after hours of running. That is good and shows you have an efficient system.

 

ok. however I can feel comfortable with these temperatures? so 'I put the soul in peace) also because' at least that H110i is much more 'quiet compared to other template :) If you increase the fans of the core temperature drops ...

Link to comment
Share on other sites

I can't say for sure, however my first guess is this is a result of the separate control protocols for CPU vs Water temperature (H110i GT Cooler Temp). The fans speeds will increase as the water temperature increases. I do not know what kind of delay is programmed into the software, but it is likely there is some amount of time between adjustments to keep the fans from fluctuating up and down. There may be a change in the AIDA program that occurs at that point or the fans may finally speed up at that point and reduce the temperatures.

 

One way to test this is use a fixed fan speed for your testing. Pick something you can stand --- 1200, 1400, etc. and set the fans to that level before initiating the AIDA test. This should remove the fan curve and any "delays" as a factor in the testing. You don't need to run for hours. 20 minutes should be enough to see if the pattern continues. I usually test this way all the time so I can make more accurate comparisons between various runs. The room temperature will affect your water temperature and alter the fan speed. This makes it hard to duplicate test runs exactly.

Link to comment
Share on other sites

I can't say for sure, however my first guess is this is a result of the separate control protocols for CPU vs Water temperature (H110i GT Cooler Temp). The fans speeds will increase as the water temperature increases. I do not know what kind of delay is programmed into the software, but it is likely there is some amount of time between adjustments to keep the fans from fluctuating up and down. There may be a change in the AIDA program that occurs at that point or the fans may finally speed up at that point and reduce the temperatures.

 

One way to test this is use a fixed fan speed for your testing. Pick something you can stand --- 1200, 1400, etc. and set the fans to that level before initiating the AIDA test. This should remove the fan curve and any "delays" as a factor in the testing. You don't need to run for hours. 20 minutes should be enough to see if the pattern continues. I usually test this way all the time so I can make more accurate comparisons between various runs. The room temperature will affect your water temperature and alter the fan speed. This makes it hard to duplicate test runs exactly.

 

ok. however I can feel comfortable with these temperatures? so 'I put the soul in peace) also because' at least that H110i is much more 'quiet compared to other template :) If you increase the fans of the core temperature drops ...

Link to comment
Share on other sites

The Corsair fans are PWM (pulse width modulation) ones rated at 2400 rpm max.

 

For most motherboard BIOS that I have seen, the PWM control will be at 100% when the CPU temperature reaches 60-70°C, which mean 2400 rpm. It may be noisier than you like, but is necessary for cooling. The idea is to use the CPU loads to control the fan speed. So the cooler fans must be connected to the CPUFAN 4-pins header. Corsair provided a Y-cable for connecting the 2 fans to one header.

 

In typical non-gaming condition, the CPU temperature will stay below 40°C or so, which will control the PWM fan to run around 800-1200 rpm. It is very quiet.

 

Your biggest concern is why the fans run so slow in both cases after over an hour of the AIDA torture test. Did the fan speed go up as the AIDA test progress? Are the fan connected to the 4-pin CPUFAN header on the motherboard? Is the BIOS PWM control disabled? If you connect the cooler fan to the SYSFAN headers, the rpm of the fans may be set according to the ambient temperature which is much lower than the CPU temperature. You are not taking advantage of what the H110i is capable of cooling.

 

The pump should be always at full speed. The 2850 rpm is good. The 3000 rpm is good too. They are in spec of the Corsair design. The pump can be connected to any MB 3-pin header. It does not accept PWM signal and cannot and should not be slowed down.

 

Something in the HWMonitor screen show that your may have messed up fan connections. The first (2015 cooler) has no CPU rpm. The second (2016 cooler) shows 2800+ CPU rpm. The pictorial CPU rpm shows the same. It is impossible and higher than the Corsair PWM fan spec.

 

Good luck and let us know how it works out for you.

Link to comment
Share on other sites

The Corsair fans are PWM (pulse width modulation) ones rated at 2400 rpm max.

 

For most motherboard BIOS that I have seen, the PWM control will be at 100% when the CPU temperature reaches 60-70°C, which mean 2400 rpm. It may be noisier than you like, but is necessary for cooling. The idea is to use the CPU loads to control the fan speed. So the cooler fans must be connected to the CPUFAN 4-pins header. Corsair provided a Y-cable for connecting the 2 fans to one header.

 

In typical non-gaming condition, the CPU temperature will stay below 40°C or so, which will control the PWM fan to run around 800-1200 rpm. It is very quiet.

 

Your biggest concern is why the fans run so slow in both cases after over an hour of the AIDA torture test. Did the fan speed go up as the AIDA test progress? Are the fan connected to the 4-pin CPUFAN header on the motherboard? Is the BIOS PWM control disabled? If you connect the cooler fan to the SYSFAN headers, the rpm of the fans may be set according to the ambient temperature which is much lower than the CPU temperature. You are not taking advantage of what the H110i is capable of cooling.

 

The pump should be always at full speed. The 2850 rpm is good. The 3000 rpm is good too. They are in spec of the Corsair design. The pump can be connected to any MB 3-pin header. It does not accept PWM signal and cannot and should not be slowed down.

 

Something in the HWMonitor screen show that your may have messed up fan connections. The first (2015 cooler) has no CPU rpm. The second (2016 cooler) shows 2800+ CPU rpm. The pictorial CPU rpm shows the same. It is impossible and higher than the Corsair PWM fan spec.

 

Good luck and let us know how it works out for you.

 

Hello ... h110i in 2015 was connected to cpu opt 'cause in the Cosmos II case I had before the CPU_FAN connector was covered by the radiator ... now with new homes obsidian 900d is properly connected on cpu fan ... if comuqnue you tell me that my temperature are fine and normal I keep it so '. how about? Meanwhile, thank you for your help. :)

Link to comment
Share on other sites

Hello ... h110i in 2015 was connected to cpu opt 'cause in the Cosmos II case I had before the CPU_FAN connector was covered by the radiator ... now with new homes obsidian 900d is properly connected on cpu fan ... if comuqnue you tell me that my temperature are fine and normal I keep it so '. how about? Meanwhile, thank you for your help. :)

 

I have been corrected that the H110i is a Cool-It design where the water block pump is controlled by the USB connection. The important issue is whether your radiator fans increase rpm when the CPU temperature goes up.

 

I have the H75 which is a Asetek design and uses constant pump speed. I have no insight on how the pump speed should be controlled on the H110i. I suspect the pump speed may adjust according to the CPU temperature too. Any expert can enlighten us on the Cool-It pump control scheme?

 

You may want to check some of the H110i reviews on the internet. Your temperatures seem to be high, by about 15°C when compared to review test under torture conditions. But your temperatures should still be safe for the Intel processor. After spending all those money on the cooler and case, you would like to get the most out of them, right? Good luck.

Link to comment
Share on other sites

You should be fine. Your fan speeds are low because you also have a relatively low water temperature of 28-30C in those two test. You can alter the curve or take manual control if you want it. That 28C is a very low temperature for a multi-hour stress test. I am guessing at your ambient temperature, but a water temp 10C over room temp with a sustained load of 100% is excellent.
Link to comment
Share on other sites

You should be fine. Your fan speeds are low because you also have a relatively low water temperature of 28-30C in those two test. You can alter the curve or take manual control if you want it. That 28C is a very low temperature for a multi-hour stress test. I am guessing at your ambient temperature, but a water temp 10C over room temp with a sustained load of 100% is excellent.

 

This is a curious question. On the H110i, how do you connect the radiator fans so that their RPM is controlled by the water temperature instead of the CPU temperature?

 

If the CPU temperature is as much as 40-50°C higher than the water block temperature as in theotorm's case, which means very poor heat transfer to the cooling water. I would suspect improper water block installation. I expect the delta temperature should be in the 20-30°C range. Just my guess.

Link to comment
Share on other sites

This is a curious question. On the H110i, how do you connect the radiator fans so that their RPM is controlled by the water temperature instead of the CPU temperature?

 

This is the standard operation for the "i" series coolers. On some of them, that is the only method available without moving the fans to another control source.

 

If the CPU temperature is as much as 40-50°C higher than the water block temperature as in theotorm's case, which means very poor heat transfer to the cooling water. I would suspect improper water block installation. I expect the delta temperature should be in the 20-30°C range. Just my guess.

 

Your 20-30C delta above water temperature is reflected in the graphs above on both coolers. That 82C peak value could have happened at any one millisecond in time during the 4+ hour test. In all probability, the water temperature was not 30C at the moment that value was recorded. All we can see is the last 20-30 minutes of the test, and the temperatures are very stable and consistent. That is a good sign after hours of testing.

 

It is exceptionally difficult to predict what the delta should be. On the positive side you have room temperature + hardware waste heat + voltage added temperature. On the negative side you have the inherent capability of the water to carry the heat + flow rate + fin density/surface area + fan speed. Several of those variables are dynamic and not all them have a direct correlation to each other.

Link to comment
Share on other sites

This is a curious question. On the H110i, how do you connect the radiator fans so that their RPM is controlled by the water temperature instead of the CPU temperature?

 

If the CPU temperature is as much as 40-50°C higher than the water block temperature as in theotorm's case, which means very poor heat transfer to the cooling water. I would suspect improper water block installation. I expect the delta temperature should be in the 20-30°C range. Just my guess.

 

My I7-4790K CPU runs at 4.7 with vcore 1.29.

I've tried to take it apart and put back the dough 5 times .... nothing changes and the pump is mounted correctly I'm sure ...

I also tried to put shims behind the plate but the result is always the same ...

Pero 'on' another h110i (2015) of the 'water temperature reaches 30 ° while in this (2016) comes to 28.8 maximum.

Clear ... if you increase the fan the temperature stays more 'low with AIDA64.

Instead these are the temperature after an hour of play in Assassin's Creed Syndicate:

 

http://i68.tinypic.com/2z6s9zd.jpg

 

Instead these are the temperatures at idle:

 

http://i65.tinypic.com/24nqxjd.jpg

 

on the other h110i (2015) are a bit more 'low 3-4 degrees ...

So what should I do? I tried and tried but nothing changes ...

I can stay tranquyillo or should I worry? :(

Link to comment
Share on other sites

This is the standard operation for the "i" series coolers. On some of them, that is the only method available without moving the fans to another control source.

 

 

 

Your 20-30C delta above water temperature is reflected in the graphs above on both coolers. That 82C peak value could have happened at any one millisecond in time during the 4+ hour test. In all probability, the water temperature was not 30C at the moment that value was recorded. All we can see is the last 20-30 minutes of the test, and the temperatures are very stable and consistent. That is a good sign after hours of testing.

 

It is exceptionally difficult to predict what the delta should be. On the positive side you have room temperature + hardware waste heat + voltage added temperature. On the negative side you have the inherent capability of the water to carry the heat + flow rate + fin density/surface area + fan speed. Several of those variables are dynamic and not all them have a direct correlation to each other.

 

So what should I do? do I have to worry?

Link to comment
Share on other sites

This is a 3 hour test of AIDA64 with the closed case (before it was open)

I noticed that of 'water temperature is more' high of 1-2 degrees.

I have already removed 5 times the pump and put the dough ...

the pump is mounted properly ... I have to do?

with the old h110i gt (2015) temperatures were more 'low of 4-5 degrees ...

I do not know what to think ... if you tell me that I can be so 'without danger ....

 

http://i63.tinypic.com/2rz5nwi.jpg

 

They are in your hands :(

Link to comment
Share on other sites

You are not in danger. You are also giving to much weight to small differences in water temperature. As a base, the water temperature will always start with your room temperature, then both system waste heat and CPU heat will be added to it. So, if it is 1 single degree warmer in your room and everything else is equal, your water temperature will also be 1 degree warmer. Based on your drive temperatures, I am guessing your room temperature is in the 17-20C range. A +10C over ambient water temperature after a stress test is a good result. There is nothing wrong with the system or your mount. You can stop.

 

I suspect what you are really concerned about is the peak core temperatures. Those numbers aren't helpful without some sort of context. If it concerns you, what you really need to do is run AIDA64 with it's line graph of individual core temperatures. You only need to run for 10-15 minutes, but you need to watch it or expand the graph window so it will show the entire test in one screen. My gut feeling is your spike occurs right at the beginning of the test. Again, this is one of the reasons to test with a fixed fan speed set prior to initiating the test. You also still have a lot of fan speed left in reserve. That fans are capable of reducing that peak number by a few degrees and bringing it under 80C. I am guessing this is the number than bothers you. It looks like you are using a custom profile for fan speed vs. Temperature (Water? Or did you get it to CPU temperature?). One solution may be to raise the higher end of the curve by a few degrees to give you a slight increase in fan speed when temperatures are higher. However, there is nothing wrong with your H110iGT or the installation. This is just the tweaking phase to make it perfect for your use. Keep in mind that if you are running from water temperature, you will likely need a different curve for the warm Italian Summer when your base water temperature will be higher. Otherwise, your fan speeds will be much higher at normal levels.

 

Of course, chances are the limiting factor is the 4790K itself. You have what looks like two excellent and efficient overclocks. That is a good frequency for that voltage. However, there is no doubt the weak point of Haswell is in the chip construction itself. It is certainly possible no amount of cooling or fan speed will reduce the on die temperatures any more without delidding the CPU and that is one step too far for most people, including me.

Link to comment
Share on other sites

You are not in danger. You are also giving to much weight to small differences in water temperature. As a base, the water temperature will always start with your room temperature, then both system waste heat and CPU heat will be added to it. So, if it is 1 single degree warmer in your room and everything else is equal, your water temperature will also be 1 degree warmer. Based on your drive temperatures, I am guessing your room temperature is in the 17-20C range. A +10C over ambient water temperature after a stress test is a good result. There is nothing wrong with the system or your mount. You can stop.

 

I suspect what you are really concerned about is the peak core temperatures. Those numbers aren't helpful without some sort of context. If it concerns you, what you really need to do is run AIDA64 with it's line graph of individual core temperatures. You only need to run for 10-15 minutes, but you need to watch it or expand the graph window so it will show the entire test in one screen. My gut feeling is your spike occurs right at the beginning of the test. Again, this is one of the reasons to test with a fixed fan speed set prior to initiating the test. You also still have a lot of fan speed left in reserve. That fans are capable of reducing that peak number by a few degrees and bringing it under 80C. I am guessing this is the number than bothers you. It looks like you are using a custom profile for fan speed vs. Temperature (Water? Or did you get it to CPU temperature?). One solution may be to raise the higher end of the curve by a few degrees to give you a slight increase in fan speed when temperatures are higher. However, there is nothing wrong with your H110iGT or the installation. This is just the tweaking phase to make it perfect for your use. Keep in mind that if you are running from water temperature, you will likely need a different curve for the warm Italian Summer when your base water temperature will be higher. Otherwise, your fan speeds will be much higher at normal levels.

 

Of course, chances are the limiting factor is the 4790K itself. You have what looks like two excellent and efficient overclocks. That is a good frequency for that voltage. However, there is no doubt the weak point of Haswell is in the chip construction itself. It is certainly possible no amount of cooling or fan speed will reduce the on die temperatures any more without delidding the CPU and that is one step too far for most people, including me.

 

Of course I understand ... my thought was' cause the old h110i-gt (2015) was of the 'more' high water temperature and of course the fans.

I did not understand why 'are not equal :)

In this new fans follow the water temps with profile "Default Mode"

In fact, when will come 'summer I will see' how to adjust the fans.

Try 'to do testing with a personal curvature.

Meanwhile, I calmed down a bit 'with your help ...

I keep you updated.

Link to comment
Share on other sites

You are not in danger. You are also giving to much weight to small differences in water temperature. As a base, the water temperature will always start with your room temperature, then both system waste heat and CPU heat will be added to it. So, if it is 1 single degree warmer in your room and everything else is equal, your water temperature will also be 1 degree warmer. Based on your drive temperatures, I am guessing your room temperature is in the 17-20C range. A +10C over ambient water temperature after a stress test is a good result. There is nothing wrong with the system or your mount. You can stop.

 

I suspect what you are really concerned about is the peak core temperatures. Those numbers aren't helpful without some sort of context. If it concerns you, what you really need to do is run AIDA64 with it's line graph of individual core temperatures. You only need to run for 10-15 minutes, but you need to watch it or expand the graph window so it will show the entire test in one screen. My gut feeling is your spike occurs right at the beginning of the test. Again, this is one of the reasons to test with a fixed fan speed set prior to initiating the test. You also still have a lot of fan speed left in reserve. That fans are capable of reducing that peak number by a few degrees and bringing it under 80C. I am guessing this is the number than bothers you. It looks like you are using a custom profile for fan speed vs. Temperature (Water? Or did you get it to CPU temperature?). One solution may be to raise the higher end of the curve by a few degrees to give you a slight increase in fan speed when temperatures are higher. However, there is nothing wrong with your H110iGT or the installation. This is just the tweaking phase to make it perfect for your use. Keep in mind that if you are running from water temperature, you will likely need a different curve for the warm Italian Summer when your base water temperature will be higher. Otherwise, your fan speeds will be much higher at normal levels.

 

Of course, chances are the limiting factor is the 4790K itself. You have what looks like two excellent and efficient overclocks. That is a good frequency for that voltage. However, there is no doubt the weak point of Haswell is in the chip construction itself. It is certainly possible no amount of cooling or fan speed will reduce the on die temperatures any more without delidding the CPU and that is one step too far for most people, including me.

 

Hello, I lowered the vcore from 1,290 to 1,275 and is very stable after 4 hours of AIDA64 tests. I also gained 2-3 degrees less. In the game it does not pass the 64/66 °. You say you can 'be good now? I remain so 'and I enjoy the pc? I await your opinion .Thanks

 

http://i67.tinypic.com/73pc1t.jpg

Link to comment
Share on other sites

I think you can stop testing and enjoy. However, the question of when the peak core temperatures are occurring is still unanswered. What might help answer this is a short AIDA test (10-15 minutes) with the System Stability Test window maximized (or as wide as possible) so I can see the whole test on the line graph. I am wondering if your peak values occur in the first few moments before the system increases to it's present speeds.
Link to comment
Share on other sites

I think you can stop testing and enjoy. However, the question of when the peak core temperatures are occurring is still unanswered. What might help answer this is a short AIDA test (10-15 minutes) with the System Stability Test window maximized (or as wide as possible) so I can see the whole test on the line graph. I am wondering if your peak values occur in the first few moments before the system increases to it's present speeds.

 

I made a 10-minute test. HWMonitor marks as the maximum temperature of the core 79, while Aida 78 (reached after about 6 minutes) .All 'beginning are 70/73.

Even with the old h110i gt (2015) I was around these temperatures.

I also tried to put shims behind the backplate but does not change anything, in fact it is worse.

That's the test. what do you think about it?

 

http://i67.tinypic.com/t6739k.jpg

 

AIDA TEMPERATURE:

 

http://i64.tinypic.com/21c8tc4.jpg

 

HW MONITOR + LINK + REAL TEMP

 

http://i64.tinypic.com/5x1v02.jpg

 

I repeat, even with the old H110i gt (2015) I was around these temperatures, with this (2016) are higher by 2-3 degrees.

 

Thank you very much await a reply.

Link to comment
Share on other sites

Known issue, devs are investigating the issue.

 

I have this problem with CL 4. The case fans are shown in the "home" but not in the "configure" screen. This is normal? Another problem: Even if I widen the window and give more 'great when you restart the always find the right and most' small ... can you help me please?

Link to comment
Share on other sites

You don't need shims and this has nothing to do with the installation. You CPU temps are warmer in the first 10 minutes than they are at 4 hours. This is most likely because the initial low fan speeds cannot quite contain the instant 100% load. As I mentioned earlier, you can rule this out of your testing by setting a constant fan speed prior to the start of the test. It will also provide more consistent results. While average CPU temps may be somewhat consistent between tests on different days, individual peak core values may not unless you have a way to maintain consistent room temperature, case temperature, and fan speed. I don't think you need to take it that far.
Link to comment
Share on other sites

Archived

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

×
×
  • Create New...