i7-5820k running super hot idling

[shrapnel_indie]

First, Current setup:

CPU: Intel Core i7-5820K 3.3GHz 6-Core Processor (Currently NOT overclocked)
CPU Cooler: CRYORIG A40 83.0 CFM Liquid CPU Cooler (lines are not kinked)
Motherboard: Gigabyte GA-X99P-SLI ATX LGA2011-3 Motherboard
Memory: ]G.Skill Ripjaws V Series 32GB (4 x 8GB) DDR4-3200 Memory - f4-3200c15q32-gvk (Currently at default 2133 speed and timings)
Storage: Samsung 850 EVO 500GB M.2-2280 Solid State Drive
Other: Flashmen® Black B+M key socket 2 M.2 NGFF (SATA) SSD to 2.5 SATA adapter card with case (MoBo doesn't support SATA on the M.2 slot)
Storage: Hitachi Ultrastar 7K3000 2TB 3.5" 7200RPM Internal Hard Drive
Video Card: Sapphire Radeon R9 380X 4GB NITRO Video Card
Case: Rosewill BLACKHAWK-BLUE ATX Mid Tower Case (radiator mounted inside, top outer panel modded to allow for AiO radiator fans)
Power Supply: Corsair HX750 Professional 750W 80+ Gold Certified Semi-Modular ATX Power Supply
Optical Drive: LG WH14NS40 Blu-Ray/DVD/CD Writer
Operating System: Microsoft Windows 10

Have had a history of aggravation with this build. (Initially post issues.)

Now... Motherboard has been RMA'd 2x Gigabyte says nothing wrong it. RMA'd the CPU, Intel sent me a new CPU.

Reassembled system, and it is thermal limiting to under 1GHz and temps 90-93ºC Idle... (WAY too hot, and considerably hotter than it was ever before, even when previously overclocked and under load) The system is unstable (obviously) too.

I'm fairly sure the connection between CPU and cooler is not the issue as one line gets really hot (and the other line stays cool.) The radiator stays cool too though. Does this sound like the pump of the AiO is at issue? Or should I be looking elsewhere for the culprit?

The AiO gets its power from a SATA connector and connects to the MoBo so the MoBo can monitor speeds... Fan Profile is set for maximum, 100% of the time. I've disconnected and reconnected it a couple of times already.

 

[jossrik]

Either the pump or some other fault in the cooler. Sounds like it's not making good contact with the heat spreader. Barring that, faulty pump would be the first place to look. If you can verify the pump is working make sure you have enough TIM and that the block is making good contact.

 

[shrapnel_indie]

Update: Tightened down the waterblock and achieved lower temperatures. Unfortunately not good enough temps. (Idle was full speed @ around 60º C b(Room was 22ºC.) Attempting to just run Firefox with a few tabs open was enough to drive temps to 94º and throttling. Shut down the browser and temps dropped back down along with a return to normal rated speed on CPU. So... It does seem to be cooling... some at least...

 

[1LiquidPC]

Bad pump by the sounds of it. You mention one line being hot, and the other being cool and radiator being cool. It sounds like your AIO isn't pumping water, or very slowly. Water temps level out after some time of PC being on, just like a car. You shouldn't have one line hotter than the other. Your CPU, especially at stock, should idle around 30C or so I'm guessing depending on ambient. My 5820k is watercooled and I'm idling around 31C right now avg. across 6 cores.

 

[shrapnel_indie]

Bad pump by the sounds of it. You mention one line being hot, and the other being cool and radiator being cool. It sounds like your AIO isn't pumping water, or very slowly. Water temps level out after some time of PC being on, just like a car. You shouldn't have one line hotter than the other. Your CPU, especially at stock, should idle around 30C or so I'm guessing depending on ambient. My 5820k is watercooled and I'm idling around 31C right now avg. across 6 cores.

Somehow I figured that might be the case. I guess time to figure out if they want just the Radiator & waterblock-pump assy alone or to include the fans. Wish it didn't die.. would have been nice to see how this chip overclocks. (The one that died didn't have an issue hitting 4.5Ghz under load.) <sigh> That leaves one small issue where the BIOS and Windows only sees 24GB RAM (Windows with a "reserve" on a small part) and CPUID sees all four RAM sticks for 32GB... but that is an inconvenience and best suited for a different forum.

 

[1LiquidPC]

Not sure about the RMA on your kit. They may not need the fans. May be a block mounting issue causing the RAM issue. Uneven pressure can cause issues reading correct memory installed. A flaky stick or reseating may help. You had 4.5Ghz on previous chip? What kind of voltage did you use?

 

[shrapnel_indie]

Not sure about the RMA on your kit. They may not need the fans. May be a block mounting issue causing the RAM issue. Uneven pressure can cause issues reading correct memory installed. A flaky stick or reseating may help. You had 4.5Ghz on previous chip? What kind of voltage did you use?

Well, the RAM in question as been shuffled around (in the slots, so it's not the RAM)... which does kind of lead to the socket (and slower post as the BIOS tries to figure it out)... EDIT: I should note that the shuffling involved a single stick at a time to test them and the earlier post issues.

I believe I was at 1.25-1.27V @ 4.5GHz on the provision Core Temp and CPU-z were not misinforming me... I think the BIOS agreed with the voltage too.

Temps generated with Prim95 v25.11, build 2. using 12 threads and testing with small FFTs..... no Core hotter than 84ºC after an hour. Idled around 44-45. Realbench gave me a peak temp of around 80.

I hope this one can do just as good... but first.... gotta get it staying cool and stable at stock before even thinking of it.

 

[1LiquidPC]

You had a great chip What actually killed it if I may ask as your voltage looked fine. I've had mine since X99 launch, though I've seen newer steppings seem to OC better with lower voltage. I've been running at 4.4Ghz with 1.325 since Sept. 2014 or so. I could boot at 4.5Ghz, but benchmarks weren't stable. I didn't want to go any higher on voltage.

 

[shrapnel_indie]

You had a great chip What actually killed it if I may ask as your voltage looked fine. I've had mine since X99 launch, though I've seen newer steppings seem to OC better with lower voltage. I've been running at 4.4Ghz with 1.325 since Sept. 2014 or so. I could boot at 4.5Ghz, but benchmarks weren't stable. I didn't want to go any higher on voltage.

I'm not 100% certain what did it in, other than the possibility of while I had it shut down due to utilities working on the mains power (scheduled) potentially a small spike sneaked through the system and got it. Everything else seems to be fine, or at least as fine as it was before hand.... well maybe other than the cooler.

 

[1LiquidPC]

Ouch. Hopefully thats all it was and the replacement is fine, along with RMA cooler. I think that will fix your problem.

 

[shrapnel_indie]

Ouch. Hopefully thats all it was and the replacement is fine, along with RMA cooler. I think that will fix your problem.

I'm working on the RMA now, and in the process I went and double checked the pump twice. It was noisier than I remembered, especially the first time of the two I checked for noise.

UPDATE: Cryorig has my replacement on the way. I'm assuming they'll want the old one shipped back soon after... Nice to see some trust if they do. (I'll of course know more when the replacement gets here.)

UPDATE 2: Replacement arrived. Installed. Stock speed (I may have the fans on max in the BIOS)... Temps so far have not exceeded 45ºC peak. (Room temp at 27ºC) Lowest core temp currently 31-32ºC