Processors
Intel® Processors, Tools, and Utilities
14405 Discussions

Question i7 6850k Temps High After H110i Pump Failure

RKenn11
Beginner
1,197 Views

So, I was running an Intel i7 6850k OC'd to 4.3ghz / 1.35v on a ASUS Rampage V Edition 10 and a Corsair H110i for about 3 years with no issues, idle temps around 25-28c, full load under OCCT around 79c after 4 hours. Anyway, the other night the pump on my H110i failed, causing my system to perform a thermal shutdown when the CPU overheated. I replaced the H110i with a H115i, and since replacing it my idle and load temps are much higher than before...

 

When I installed the new Corsair H115i, I went into BIOS on my first boot and reset BIOS to stock settings, no OC or anything. Temps were pretty normal on the first boot, around 32c idle, a little higher than the H110i, but I figured it was just cause it had just been installed. Anyway, rather than running any stress tests at that time, I went ahead and restarted, enabled my XMP profile, a modest OC to 4.0ghz. I manually set VCORE to 1.22, and then set all other voltages to stock manually, I don't leave any of my voltages on auto. Upon booting into windows my idle temps were now over 50c, and they have been since then. I've tried reseating the pump head. I've reapplied TIM a few times and can confirm a thin even spread, there is no contact issue. I've gone back to stock clocks. I'm not sure what else I can do to fix this at this point. Running a stress test in OCCT produces a 90c+ temp in less than 3 minutes. Pump speed is set to max, 2080 RPM, fans are running at 1100RPM on the radiator. Liquid temp gradually increases over time under load from 24.5c to about 31-32c, so there is heat transfer happening. Temps are relatively safe when gaming, editing and rendering video, etc... they sit at about 60-63c. CPU Package temp is always a bit higher than Core temps. When idling there is about a 20c difference, my cores idle between 26c-33c, while my Package temp never drops below 50c.

 

What I am wondering is what could be causing the heat increase? I would assume this newer model h115i would produce better results than the h110i I had on before, but surely that isn't the case. Also the fact that I used to be able to run OCCT for 4 hours with a higher VCORE and never hit 80c, but now I hit over 90c within 3 minutes, is pretty alarming. Sure I can perform daily tasks and game, but I won't be happy until I can figure out what the issue here is... was something damaged because of the thermal shutdown? I have been told more than likely not the case by a few people on reddit/other forums, as that is the point of the TjMAX. Someone suggested that the contact between my IHS and CPU could be poor because of the overheating, but I highly doubt it. The i7 6850k was one of the last Intel CPUs to be soldered, so delidding is pointless, I also don't understand why I'd have to do it, as nothing has changed but the AIO, which I can confirm is working. No sign of blockage or bubbles, no leaks, etc. Is it possible the motherboard or CPU was damaged or is degrading and causing these heating issue? I built the system in Jan 2017, so it's only 3 years old. I have a 1600w EVGA PSU, I have also tested with a 1200w and that didn't change anything, so I know it isnt the PSU.

 

Has anyone had similar issues with 68XX processor or my motherboard? I received a lot of support on the Corsair forums, but we couldn't come up with an answer. We just came to the conclusion that since the CPU is operating at a reasonable temp when performing my regular work/gaming that I should be happy. Hard for me to be happy with it though as there is obviously something going on. A helpful user suggested I try some other forums, so I figured I'd post here as well as the Intel official forums, I am going to hit the ASUS forums, I have a post on the ROG forum about this, and I am going to look for some more Intel/68XX specific forums.

 

Any help is greatly appreciated. I've included some images for you guys to check out.

 

hwinfotemps4rog.jpg

OCCT.JPG

Thank you in advance for any help!

0 Kudos
5 Replies
n_scott_pearson
Super User
1,026 Views

Everyone seems to think that the throttling that occurs at Tjmax is this magical thing that will completely prevent the system (note I said 'system', not 'processor') from being thermally damaged. This is utterly not true! Any significant amount of time spend in the vicinity of the Tjmax threshold is going to be damaging to the system. Think about it: all of that heat from the processor is not being dissipated in the normal fashion. Where does it go? Well, instead, it is dissipating into and across the motherboard, via board traces, grounding planes, etc., and can, if there for any significant amount of time, damage other components on the motherboard. At the same time, if in this temperature range for any significant length of time, it is definitely going to have a detrimental effect on the lifetime of the processor, of not immediately (relatively speaking) and permanently damaging the processor. Think of throttling as being only a temporary protection. In your system design, you likely are exasperating the situation because you are using a liquid cooling solution. Because you do not have any fans working in the vicinity of the processor, you likely will not have great airflow in this area to begin with, let alone the situation when your liquid cooling system failed.

 

Bottom line, to answer your question, yes, you have likely damaged the processor and perhaps the motherboard as well. With their proximity, if you do not have sufficient airflow in this area, you could even see damage in the DIMMs.

 

Sorry, reality bites,

...S

RKenn11
Beginner
1,026 Views

I think your missing the point, I never spent any time over 60c, even 50c. The only time I was ever 70-90c was when running OC benchmarks. Never had an overheating issue. The pump failed and almost immediately my system had a thermal shutdown, so we are talking seconds. Motherboard, RAM, and everything else appear to be working fine. I am going to try swinging the waterblock 90 degrees to see if it gets better coverage. Airflow in the case is actually very good, I run 7 noctua fans inside the case besides for the AIO, and the CPU is surrounded by 4 fans, all intake. Appreciate your response though.

0 Kudos
RKenn11
Beginner
1,026 Views

Sorry, I totally forgot to list all my specs:

 

Motherboard - Rampage V Edition 10 BIOS 2101 (7.10.19)

Processor - Intel Core i7 6850k

Memory (part number) - CMD64GX4M4C3200C16

Graphics Card - ASUS STRIX ROG 1080 GTX

Monitor #1 - ASUS ROG Swift PG279Q

Monitor #2 - BenQ BL2711U

Storage #1 - Samsung 960 PRO m.2 SSD

Storage #2 - Samsung 950 PRO ssd

CPU Cooler - Corsair H115i

Case - CM Stormtrooper

Power Supply - 1600W EVGA Titanium

OS - Windows 10 Pro (10.0.18363)

RAID Array - 4x 8tb Seagate Enterprise RAID0

0 Kudos
n_scott_pearson
Super User
1,026 Views

No, I didn't miss the point - but I did miss the most-applicable situation. Any time that you see a THERMTRIP, your processor has been heated well beyond its junction temperature. That can potentially cause permanent damage almost immediately. THERMTRIP, like Throttling, is a protection mechanism that saves the processor from meltdown, but only most of the time.

 

What I suggest you do is find another processor to try in your setup. This will let you know if the problem is isolated to your processor or is more systemic.

...S

0 Kudos
RKenn11
Beginner
1,026 Views

Welp, I feel like a dummy... I could of swore I swapped the standoff screws when I installed the new pump, but apparently I didn't. After replacing them I an hitting a max temp of 51c after 3 hours in OCCT. Thanks for the help everyone.

0 Kudos
Reply