Dead H100i v2 Pump?

I have a CPU cooling problem and I wanted to cover all my bases before I run out and get a new cooler. I was using a H100i V2 cooler in my case, however I now suspect it maybe dead.

I am using a newly built computer I put together a week ago however the 100i is from my previous build and now about a year and a half old. As I was leaving my apartment started installing Asus AI Suite 3, a program that required restarting my computer. I was not here for the process but when I got back I noticed that my cpu temps were in the 80's at idle. First thought was that AI Suite messed up the sensors, so I uninstalled the program and then re-flashed the bios, however this did not solve the problem. Finally I removed the cooler from my cpu and installed an old Intel stock cooler I had laying around. With the stock cooler I am now getting temperatures around 32 degrees, so the problem is seemingly solved.

 

My issue comes from the fact that in bios my motherboard was rpm values from the pump. The Strix 270e motherboard has an AIO pump header so my assumption is that if the pump was not running there would have been no rpm value for that header in bios. Finally when taking off the cooler I noticed that the tubes on the cooler were warm near the block but room temperature near the radiator and that the radiator itself was cool.

 

What do you guys think, is this a dead pump or is there a blockage in the tubing I may be able to clean out?

 

Thanks.

If I was you, I would use the Support Ticket System link above your post and take it up with Corsair. You will need to register there as your details from here won't work. Corsair coolers have a 5 year warrantee, so upload a copy of your receipt.

That is part of my problem I don't have the receipt so I figured I was out of luck on that.

Normally, the way analyze a cooler failure (pump fail/blockage) is by looking at the H100i v2 Temp (coolant temperature) in Link. When the liquid doesn't flow, either through restriction or non-working pump, the coolant temperature will climb slow enough for you to watch it go up/ 40-45-50-55 and continue on. Your CPU temps do the same. How high it goes depends on the severity of the blockage or failure.

 

It would be unusual for the system to suddenly fail because you installed software, even something notorious like AI Suite. It seems more probably the initial configuration did not allow the pump to run. On first boot AI Suite should incorporate the existing BIOS fan curves into its settings -- at least until you run the tuning protocol.

 

Can you describe how the H100i v2 was connected to the motherboard? Which header? What settings did you use in the BIOS Q-fan for that header?

The header on the motherboard is under the io and is labeled AIO_Pump. In bios it also shows up as a pump running about 1400 rpm with a Q-fan curve that is set to 100% full full time. I have the H100i v2 plugged in and just ran Prime95 for about 5 minutes (This is a new CPU and I am hesitant on how long I should run it at these temps) and Core Temp pegged at 100 degrees almost instantly. Using corsair link the registered temp from the cooler climbed from 29 to 45.1 over that time and the CPU throttled from 4.20 to 4.01 over that time.

 

I agree that installing AISuit would make it fail is very unlikely but I am at a loss to describe what happened.

 

EDIT

 

Idle temps do come down to 40 quite quickly when Prime95 was stopped but then the Link and CPU temp took 10 minutes to down to 30 degrees which seems to be its lowest value.

 

I do have another H100i v2 that I just bought yesterday but I am leaving it sealed in case I can return it.

Don't use Prime for Kaby Lake. It has some value for binning a bunch of CPUs, but it is not a good test for the cooling system. No cooler is going to handle it well on recent Intel CPUs. +16C coolant delta in 5 minutes sounds high to me, but I would like to see a normal stress test program.

 

I would suggest trying Intel XTU for a 5 or 10 minute run. It's a cupcake stress test, so if temps are high we know something is wrong. It is also mild enough you won't fry your CPU in figuring this out. Click the wrench on the line graph and set it so your 4 cores show. In should be a fairly smooth sine wave like pattern on load. If you see lots of jumps or spikes, that may suggest a contact problem. Again, take note of start/end coolant temp. I would expect something in the 6-9C delta range depending on length, but we'll see. You can also use Aida or OCCT (large data set), but those are more dynamic in nature and the cores will jump around by design. If you see 80C+ with XTU, stop the rest and report back.

 

The 1400 rpm looks like the normal one-half actual pump speed reported in the BIOS for that model. You can see the actual speed in Link and it should be near 2800-3000 while in Performance Pump mode. Presumably you found a way to deal with the CPU boot error (either by ignoring or placing other fans on CPU_FAN). There is nothing special about the AIO/W_PUMP header on Asus boards, other than it is pre-configured to run at 100%. You can do the same thing with the CPU fan header and that would be the normal place to put the H100i v2 lead, however at this point I don't see a connection and you can leave it as is. The AIO header should be immune from AI Suite control, even after tuning.

 

Don't open the other H100i v2 yet. Still some normal possibilities to check. Also, how is the H100i v2 positioned in the case. For now, I am assuming top exhaust.

 

 

EDIT: Also take note of the peak Vcore value during the test. If you are still running in stock configuration with no refinement, Prime may have pulled in excess of 1.40v. There are some setting you can tweak to limit this, but lets see the results first.

Well, I gave in. I was still getting better temps with a 2600k stock cooler so I decided to try the other cooler and Prime95 right away gave me temps in the 70's. I still want to find out what is wrong with the old cooler and use it in my old 2600k rig but I am not sure how likely a fix will be.

 

I know the AIO_PUMP header should be no different and I even plugged in the old pump to a different header to see if anything changed but it did not.

 

The radiator is outside the top of my case with the fans inside the case pulling air through the radiator into the case.

 

I plan to get some numbers with the old cooler tomorrow to see what is going on.

 

EDIT: Under Prime95 the CPU was pulling around 1.2 volts also.

Since you have a pump speed we can assume the pump is electrically functional. The other two probabilities are a mechanical failure (pump physically cannot move coolant) or more likely there is some sort of obstruction blocking the flow.

 

If for whatever reason you cannot obtain another copy of the receipt from the vendor, I would still contact Corsair and see what they say. I don't know how long you've had it, but return to the vendor is always quicker and you already have a replacement. If none of that pans out and this thing becomes your lab toy, the very unscientific place to start is by giving it a thorough shaking. You may be able to dislodge the obstruction. Hit or miss to be sure and this is not a real solution. You would then need to install it, then let it sit for at a little while before powering on. Hopefully the debris settles down and does not become mobile. When it does, it collects at the pump grate and that is what really blocks things up. If this is a mechanical failing, you would likely hear a higher frequency pump sound and see odd surges or generally higher pump speeds than expected. This is a low probability and unfortunately there is no clear fix. Taking these AIO units apart is tough. I would say they were deliberately made to prevent you from doing it. I have tried to swap cold plates between units and the screws are unbelievably difficult to get out - certainly 12 of them. Surgery tends to be terminal, so keep that in mind before starting.

 

70's for Prime95 sounds better and it appears the second unit is functional.