AIO Cooling - Performance Degrades Over Time?

Hurin

2[H]4U
Joined
Oct 8, 2003
Messages
2,410
Hi All,

Just wanted to see if anyone else has noticed the same thing over the years. I've had two Corsair AIO cooling units. The H100i and the H115i.

In both cases, I've noticed that a year or two after initial install temperatures aren't as good as they were initially. Just FYI, ambient temperature is cooler today (while I investigate this) than it was when these systems were originally set up. We're having a bit of a cool day today. =)

---------------------------------------------------------

Corsair H100i
On the H100i and my 4770K, at initial install and OC (to 4.5GHz), temps were 81c / 79c / 87c /81c for each core. Under load in Prime95 w/ 512k min FFT

Three years later, I note that temps are at 98c / 93c / 97c / 87c

I wash the radiator to clear some dust build-up that compressed air couldn't remove and remount the block. . . 86c / 89c / 89c / 85c

-----------------------------------------------------------

Corsair H115i
Initial install was a year ago with a 6700K at 4.6GHz. Temps were 76c / 80c / 68c / 74c

A year later, I noted today: 89c / 90c / 86c / 87c

After some dust blowing and a remount of the block, I was able to get this down to: 81c / 82c / 75c / 79c

Those are respectable. But nowhere near what I originally got with the H115i.

-----------------------------------------------------------

Does anyone see this gradual degradation of AIO performance over time. There doesn't seem to be much else for me to do to try to rejuvenate these coolers and get them back to where they were when new.

Is this due to the AIO units aging? The CPUs aging? I tried to rule out changes in the testing environment by using older editions of Windows 10 (1607), Prime95, and HWMonitor.

--H
 
I've never heard of them degrading. My h100 has been running without degradation to temps for 4 years. But...You had a huge drop in temp after you cleaned the radiator. How much crap was in that thing?


Maybe the block shifted. Thermal paste tried up. I know with my h100 I had to use plastic washers on the mounting system to mount the block to the cpu with enough pressure.
 
I used a Corsair H100 for over five years with no degradation in performance. If I remember right I used MX2 with it. I don't remember ever cleaning the radiator either (and it was nasty when I removed it). This was with a 2500k at 4.5Ghz most of the time.
 
Seems odd that even after cleaning and reapplication of thermal paste I still wasn't able to get either system back to where they were at initial install.

Perhaps the mounting bracket bends over time, lessening contact? May try the washer method eventually. Though I'd think "contact is contact" after a certain point and it does seem to be on there pretty darn firmly. For now, my 6700K is still workable and the OC is stable. The 4770K is no longer stable but isn't my primary rig anymore. So it can be at stock until I have more time.
 
I've been running an H100, which I bought used, for nearly two years now and I haven't noticed any deterioration. My only problem is one side has 6 stripped mounting holes. Hot glue has been my savior.
 
Conclusion: The H115i is failing.

The Noctua NH-D15 performs better than the H115i ever did even when the H115i was new.

Noctua NH-D15 - Test #1
6700K @ 4.6GHz adaptive voltage at 1.290v
Prime95 v29.4 - 45 minutes - Custom/Minimum FFT set to 512 in 74.5f room

75c / 74c / 72c / 72c

vs

H115i when new at those settings

76c / 80c / 68c / 74c


Noctua NH-D15 - Test #2
6700K @ 4.6GHz adaptive voltage at 1.290v
Prime95 v29.4 w/ AVX disabled in local.txt - 15 minutes - small FFTs in 74.5f room

71c / 70c / 62c / 68c

vs

(apparently) failing H115i at those same settings -- for only four minutes

87c / 93c / 75c / 85c

I can't say whether it's the pump on the H115i that is failing, whether there's an osbstruction, or whether something has gone wrong with the mount which is making for uneven contact (even after trying adding washers as a last resort). All I know for sure is that I can't get good performance out of it anymore and I'll see about warranty service on it (and then sell what I get back as a refurb).

Incidentally, the Noctua NH-D15 is a dream! It installs so easily (except needing to remove my video card to access the fan clips), performs spectacularly, and is even much more quiet. Could not be happier to have this solved and I'm giddy about the performance. Might even "reopen the books" on my overclock and try to exceed 4.6GHz at reasonable temps again.

I have yet to replace the H100i in my other rig (currently in use by the kids). But I suspect I'll get the same findings.

Best Regards,

Hurin

P.S. Full disclosure: There is a tiny chance that the problem was that I got a bum tube of MX-4. I can't 100% rule that out because I used a brand new tube that I ordered along with the Noctua NH-D15. So, really, performance was awesome after I changed two things: The cooler, and the paste vintage. Prior, I was using a tube of MX-4 that was only a few weeks old. But, well, I hate changing more than one variable at a time when tracking down a problem. But having done so, I figure I should note it.
 
I'm moving away from my AIO (old Cooler Master Seidon 120M) to a Cryorig H5 air cooler. My AIO is getting a little long in the tooth, and I've decided I'm just more comfortable with air cooling.
 
I'm moving away from my AIO (old Cooler Master Seidon 120M) to a Cryorig H5 air cooler. My AIO is getting a little long in the tooth, and I've decided I'm just more comfortable with air cooling.
Only real downside is aesthetics. My system certainly looks less modern and fancy than it did with the Corsair LED blazing white and just a couple of black tubes running to a largely hidden radiator.

But. . . I only peer into the side of my case about three times a year. So. . .
 
The only degradation I have noticed on any Corsair H-series I have used or mounted for others. Is due to the crappy grey goop that is pre-applied when you buy them. If you don't remove that and use something better you will see a significant degradation in performance within 1-3 years as that goop will expire. Replace it with something better, I use Noctua NT-H1, and you will see consistent performance for several years.
 
The only degradation I have noticed on any Corsair H-series I have used or mounted for others. Is due to the crappy grey goop that is pre-applied when you buy them. If you don't remove that and use something better you will see a significant degradation in performance within 1-3 years as that goop will expire. Replace it with something better, I use Noctua NT-H1, and you will see consistent performance for several years.
I used MX-4 from the get-go. Sadly, my H115i seems to just be a dud after about a year.
 
I've had the same problem with an h100i went from 63C to 68 in 2 years.
 
My Deepcool Captain 360 degraded over 1 year (pump) and leaked coolant on $1200 Titan XP, bricking it.

I suggest to not risk it, and get a good air cooler if you can fit it in.
 
Conclusion: The H115i is failing.

The Noctua NH-D15 performs better than the H115i ever did even when the H115i was new.

Noctua NH-D15 - Test #1
6700K @ 4.6GHz adaptive voltage at 1.290v
Prime95 v29.4 - 45 minutes - Custom/Minimum FFT set to 512 in 74.5f room

75c / 74c / 72c / 72c

vs

H115i when new at those settings

76c / 80c / 68c / 74c


Noctua NH-D15 - Test #2
6700K @ 4.6GHz adaptive voltage at 1.290v
Prime95 v29.4 w/ AVX disabled in local.txt - 15 minutes - small FFTs in 74.5f room

71c / 70c / 62c / 68c

vs

(apparently) failing H115i at those same settings -- for only four minutes

87c / 93c / 75c / 85c

I can't say whether it's the pump on the H115i that is failing, whether there's an osbstruction, or whether something has gone wrong with the mount which is making for uneven contact (even after trying adding washers as a last resort). All I know for sure is that I can't get good performance out of it anymore and I'll see about warranty service on it (and then sell what I get back as a refurb).

Incidentally, the Noctua NH-D15 is a dream! It installs so easily (except needing to remove my video card to access the fan clips), performs spectacularly, and is even much more quiet. Could not be happier to have this solved and I'm giddy about the performance. Might even "reopen the books" on my overclock and try to exceed 4.6GHz at reasonable temps again.

I have yet to replace the H100i in my other rig (currently in use by the kids). But I suspect I'll get the same findings.

Best Regards,

Hurin

P.S. Full disclosure: There is a tiny chance that the problem was that I got a bum tube of MX-4. I can't 100% rule that out because I used a brand new tube that I ordered along with the Noctua NH-D15. So, really, performance was awesome after I changed two things: The cooler, and the paste vintage. Prior, I was using a tube of MX-4 that was only a few weeks old. But, well, I hate changing more than one variable at a time when tracking down a problem. But having done so, I figure I should note it.
Many of us have found top tier air coolers to be as good or better than CLCs and quieter too .. not to mention only costing about 2/3rds as much and almost never wearing out.

I would be extremely surprised if it was a 'bum tube' of MX-4.

Noctua has the Chromax cooler covers and Chromax black fans out now
LL

http://noctua.at/en/products/product-line-chromax
 
I have two systems with Corsair 240mm units. The rest are air cooled. I now only use AIOs when there is a specific thermal limitation in case design. Otherwise, I grab the biggest Noctua or Thermalright heatsink I can cram into whatever case I'm using. AIOs wowed people with initial temps and the exoticness of liquid cooling. However, when people learned about pump failures, heat saturation,and the amount of noise they generate compared to a top tier heatsink, they're celebrity seems to be fading.
 
Back
Top