Jump to content
Corsair Community

Pump Fluctuations on H80i (2015) Question


Recommended Posts

I've been wanted to say its been a year since I made a topic about CPU temps in stress testing, but I have decided to plug the usb dongle onto my h80i v1, and looking at the pump speed. In the bios, it was 2319rpm on cpu fan 1, but on Windows, the rpm was at 2219rpm. Sometimes the pump fluctuates like the rpm dipped at 2123 rpm, then backs up at 2200rpm. Most of the time it was around 2221 rpm. This is being read in Corsair version 4 as ICUE doesn't detect my cooler, even when using legacy hardware version of Icue. I'd also discovered that HWInfo64 detected my h80i and decided to keep an eye on the pump and coolant temp. I'd noticed something alarmingly as the pump's rpm dipped around 1918rpm, then a second later it went back to 2200rpm. I'm more worried about that drop as I had my h80i for like almost 7 years from now. I know HWinfo64 may have inaccurate sensors, but to my hunch, I decided to switch to Corsair Link 4 and keeping an eye on the pump's rpm for that 1918rpm. As such, I've been keeping an eye on this ever since I've plugged that dongle in. Despite such drop, does this seems kinda normal for a 7 year old h80i or is it time to die? I can try checking back into the bios for the pumps rpm and see if it drips there.

I still haven't decided about reapplying the thermal paste since the temps in normal usage are in the acceptable range and no coolant problems as when I'm playing GTA 5, the coolant temp rose to like 31.4-34c as well as cpu temps are in the late 50s - early 60s range. If the cooler fails, I'll look for a new cpu cooler. I'm just wondering if fluctations like I have are normal for a 7 year old AIO. ><

Edited by RyougaLolakie
Link to comment
Share on other sites

34 minutes ago, RyougaLolakie said:

the coolant temp rose to like 31.4-34c as well as cpu temps are in the late 50s - early 60s range.

You definitely do not have a cooling problem at this time.  Your room/environment temp is the base or lowest possible level of coolant temp, so if you are in the low 30s at load everything is working well.  Hard to say on the pump RPM.  The motherboard and Link may poll at two different times or it might be wobbling a bit.  However, most high speed devices are going to have some variance and it won't read 2300 flat all the time.  What will be more worrisome is if you star to see low and high speed fluctuations (2300-1900-2500-1900, etc).  That's usually a sign the pump is losing its grip.  More often its the flow path that becomes the limitation and you'll see elevated base temperatures with no load.  That is a definite sign its time to start looking for a replacement.  

Link to comment
Share on other sites

10 minutes ago, c-attack said:

You definitely do not have a cooling problem at this time.  Your room/environment temp is the base or lowest possible level of coolant temp, so if you are in the low 30s at load everything is working well.  Hard to say on the pump RPM.  The motherboard and Link may poll at two different times or it might be wobbling a bit.  However, most high speed devices are going to have some variance and it won't read 2300 flat all the time.  What will be more worrisome is if you star to see low and high speed fluctuations (2300-1900-2500-1900, etc).  That's usually a sign the pump is losing its grip.  More often its the flow path that becomes the limitation and you'll see elevated base temperatures with no load.  That is a definite sign its time to start looking for a replacement.  

What you mean is that the high speed fluctuations such as 2300-1900-2500-1900 etc are like in 1 second intervals or does it happen in a few seconds or few minutes at a time? Like, I did see the 1918 rpm fluctuation on HWinfo64 for one bit then it went back up to 2200s and it stayed there most of the time. Right now, I don't see that happening on corsair link 4 yet but will kept an eye out. The coolant temp right now is 26.5 as I'm typing to reply this thread and when that pump rpm fluctuation happened, the coolant temp didn't increase one bit. 

 

Link to comment
Share on other sites

13 minutes ago, RyougaLolakie said:

uch as 2300-1900-2500-1900 etc are like in 1 second intervals

Yes, like each time it polls.  I can't remember for the H80i and Link, but it's likely to be every 1-2 seconds for updates.  If you see a dramatically different number every 2 seconds, things are starting to wear out.  It's not uncommon to see the pump speed drop 100 rpm at uneven intervals, but return to a normal baseline.  It could be a bubble, pressure change, or something fairly ordinary.  However, pump overspeed mixed with drops is a definitely warning sign.  When users show up and their pump is running several hundred RPM too fast for the setting, you don't have a lot of time before a critical error occurs.  

 

In terms of consistency, I would stick with the Link reading.  It is the native sensor and software.  Most likely HWiNFO or the MB is reading it at a different time and it's taking a number of rotations for the 2 seconds (or whatever) and multiplying out by 30 to make it revolutions per minute.  That means fractional rev count changes are multiplied by 30 as well, so anything +-100 rpm is normal.  Also, be careful running HWiNFO at the same time as Link.  It is not as devastating as when paired with CUE 4, but they still can mess each other up trying to access the device at the same time.  If you need HWiNFO for other reasons, engage the Asetek protection in the safety tab or right click on the H80i and tell HWiNFO not to monitor the Corsair internal devices.  Wacky RPM reporting is a common byproduct of the two software programs stepping on each other.  

Link to comment
Share on other sites

12 minutes ago, c-attack said:

Yes, like each time it polls.  I can't remember for the H80i and Link, but it's likely to be every 1-2 seconds for updates.  If you see a dramatically different number every 2 seconds, things are starting to wear out.  It's not uncommon to see the pump speed drop 100 rpm at uneven intervals, but return to a normal baseline.  It could be a bubble, pressure change, or something fairly ordinary.  However, pump overspeed mixed with drops is a definitely warning sign.  When users show up and their pump is running several hundred RPM too fast for the setting, you don't have a lot of time before a critical error occurs.  

 

In terms of consistency, I would stick with the Link reading.  It is the native sensor and software.  Most likely HWiNFO or the MB is reading it at a different time and it's taking a number of rotations for the 2 seconds (or whatever) and multiplying out by 30 to make it revolutions per minute.  That means fractional rev count changes are multiplied by 30 as well, so anything +-100 rpm is normal.  Also, be careful running HWiNFO at the same time as Link.  It is not as devastating as when paired with CUE 4, but they still can mess each other up trying to access the device at the same time.  If you need HWiNFO for other reasons, engage the Asetek protection in the safety tab or right click on the H80i and tell HWiNFO not to monitor the Corsair internal devices.  Wacky RPM reporting is a common byproduct of the two software programs stepping on each other.  

I'll disable the monitoring of h80i coolant temp and pump in hwinfo64 as well as uncheck CorsairLink and Asetek Support from Corsair's site about hwinfo64/corsairlink conflcts. Here's the image of the coolant temp and pump speed while running RetroArch in the background and Edge. 

Reading on what you said about the overspeed pump rpms mixed with drops sounds really scary. And Looking at Link right now, I did see it dropped to 2068rpm and that kinda worries me, even though the coolant temp stays the same and it was just a drop. Does this seem normal at this time? I'd apologized for repeating the same questions.  . 

 

Clipboard02clinkpumpcoolant.png

Link to comment
Share on other sites

This looks pretty steady to me.  If you are experiencing the over/under speed condition, the jumps are going to be large.  Something like 2200->1900->2000->2500->2300->1900 every 2 seconds.  It's a roller coaster ride and most users hear the pump shifting up and down before they visually notice in monitoring.  Most dying pumps will audibly let you know there is trouble.  Also, this over-revving condition is most often found in relatively new pumps and is indicative of a production issue.  If your pump has lasted 7 years, it's obviously a very good sample and I think this type of behavior in highly unlikely for your unit. 

 

Most closed loop AIO units will come to an end because the the flow path becomes clogged over a period of years.  It is a 'zero maintenance' product and anything like that has a definite shelf life.  As the flow path narrows or become restricted in a certain spot, the flow rate slows as you will see elevated coolant temperatures, even when there is minimal load. After heating things up a bit through normal load, it will take an extremely long time for the coolant levels to come down -- possibly not until you shut down and reboot hours later.  For you this would look something like 34C coolant at idle and it keeps slowly creeping up.  Then maybe mid 40s when under load and it stays there after you quit.  Your temps look good and if the coolant temp quickly drops over 2-3 minutes after load stops, you have a fully functional, healthy system.  

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

Posted (edited)
39 minutes ago, c-attack said:

This looks pretty steady to me.  If you are experiencing the over/under speed condition, the jumps are going to be large.  Something like 2200->1900->2000->2500->2300->1900 every 2 seconds.  It's a roller coaster ride and most users hear the pump shifting up and down before they visually notice in monitoring.  Most dying pumps will audibly let you know there is trouble.  Also, this over-revving condition is most often found in relatively new pumps and is indicative of a production issue.  If your pump has lasted 7 years, it's obviously a very good sample and I think this type of behavior in highly unlikely for your unit. 

 

Most closed loop AIO units will come to an end because the the flow path becomes clogged over a period of years.  It is a 'zero maintenance' product and anything like that has a definite shelf life.  As the flow path narrows or become restricted in a certain spot, the flow rate slows as you will see elevated coolant temperatures, even when there is minimal load. After heating things up a bit through normal load, it will take an extremely long time for the coolant levels to come down -- possibly not until you shut down and reboot hours later.  For you this would look something like 34C coolant at idle and it keeps slowly creeping up.  Then maybe mid 40s when under load and it stays there after you quit.  Your temps look good and if the coolant temp quickly drops over 2-3 minutes after load stops, you have a fully functional, healthy system.  

Oh, I haven't encountered the over/under speed condition. The pump's rpm has been steady 2200 rpm and the drop happen for like a second then it gets back up to 2200 rpm and there's no rpm increase. Only slight drops. It's like you're riding in a car in a steady road and you encountered a huge speed bump for a second, then you resume to ride in a steady road again.  You say that over 100 rpm +/- may consider it normal for the 1 second drops, then as I'm replying to this thread, the pump spead rpm drops to 1921 rpm for one second then it goes back to the steady 2200 rpm. So the reading from the hwinfo64 earlier was true. 

Despite all of this, after I exited out on Retroarch and right now the coolant temp is at 25.9c. It seems the one second drops doesn't seem to affect the coolant temp at the slightest. Still, would these 1 second rpm drops be considered normal for this now especially when it drop to like 1921rpm for one second? It seems the system looks functionally healthy, but the drops like that seem to worry me at the slightest. :X

I'll definitely keep an eye out for possible future cooler failures in the meantime and that means I'll look for a different cooler since my case is HAF 922 Cooler Master case and I'll decide whether a new AIO or a decent Air cooler would fit better, except for the G.Skill Ripjaws Z 8GBx4 1600mhz ram I had installed in my mobo.

Edited by RyougaLolakie
Link to comment
Share on other sites

Alright, this is what the drop looks like and today I saw an rpm drop to 1908 rpm. Most of the time its still steady at 2200 rpm. Though the max of coolant temp today was 30.1c and temps are still good. I dunno if that drop's gonna affect the pump's life that way, even for a 7 year old h80i. I'm ocd'ing too much about this. >< 

Thank you for the information about the overspeed/drops and more info about the pump's life and such, C-Attack! If the 1908rpm drop still counts as normal, then I guess I just have to accept this until the thing fails out after 7 long years. 🙂

Clipboard01clinkpumpcoolantpumprpmdrop.png

Link to comment
Share on other sites

I think it’s just a “hiccup” until it starts happening on a regular per minute basis. A pump is a high speed device and the sample rate is taken over 1 or 2 seconds before it is multiplied out. This makes a tiny variance look larger than it really is. Your pump would need to drop 1000 rpm for 5-10 seconds consecutively to impact your temperatures. So I don’t see any kind of performance issue on that end and your coolant temp is excellent, so I think you are still in good shape. 

Link to comment
Share on other sites

Posted (edited)
6 hours ago, c-attack said:

I think it’s just a “hiccup” until it starts happening on a regular per minute basis. A pump is a high speed device and the sample rate is taken over 1 or 2 seconds before it is multiplied out. This makes a tiny variance look larger than it really is. Your pump would need to drop 1000 rpm for 5-10 seconds consecutively to impact your temperatures. So I don’t see any kind of performance issue on that end and your coolant temp is excellent, so I think you are still in good shape. 

That's really good to know. 🙂 

I tested out CPU-Z stress test and seeing core 1 top it off at 72c (last time it was at steady 68c) while rest of the cores top off at 68, 66, and 60. The coolant temperature tops it off at 30.5c and pump speed's chugging at 2217rpm. To tell you the truth, when I got this pc built by Ironside Computers, the radiator fans themselves were hooked on the motherboard instead of being hooked on the cooler as the fans were at 2k rpm. I think its why the coolant temp maxes out at 30.5c during cpu-z cpu stress testing. I haven't bothered to switch the fans to the cpu cooler, but I was wondering if those fans at 2k speed affected the coolant temp that much.

After a couple of minutes later after I stopped the CPU-Z stress test, the coolant temp drops back to 26.6c

Edited by RyougaLolakie
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share

×
×
  • Create New...