Question CPU heat issues

[Praetorian349]

Hi guys, in desperate need of some help and hoping you might be able to shed some light on this.

I’m rather new to the whole PC tinkering thing and trying to pick it up. I recently got a new PSU and opened up my PC to install it, as well as give it a quick clean inside. I went from a 600W bronze certified PSU that was 4 years old to a 850W HX850 Corsair platinum certified PSU (need it for a new GPU I plan on installing). Swap went by without a hitch, I thought.

However now my CPU (Intel 8700k) will start up at incredibly high temperatures (80-90 degrees C) and keep climbing. I have an H100i Corsair water cooler, checking the coolant temperature in this in iCue, it starts at 25 degrees C and steadily climbs. The CPU temperature also climbs to 100 degrees C. This is all while the system is idle, mind you. As the temperature climbs the fans on the H100i obviously kick into high gear, up to 2700 RPM, and the pump fan reads around 2000 RPM. The pump appears active as one coolant pipe is significantly hotter than the other. But strangely the case fans never start (plugged into the MB).

I can’t seem to stabilise the temperatures or find a reason for why this is happening? I have tried replacing the thermal paste on the CPU (which was also done as recently as November last year), resetting the BIOS and have swapped the old PSU back to test, to no avail. I’m at a loss. As far as I am aware the temperatures were not an issue before the PSU swap, I didn’t check them but I didn’t hear ungodly amounts of fan activity either.

I am trying to contact Corsair as the H100i is the most likely faulty part, from my perspective, not that I can see where it is faulty. It is 4 years old so should still fall under warranty and I may be able to replace it. However I don’t have a spare CPU cooler to test if it is at fault or not. :-/

Any insight on further troubleshooting I could do? I will provide any further info you think could be helpful, but I am hesitant to have the computer on too long to provide screenshots, etc as I don’t want the components to be damaged by excessive heat.

 

[Praetorian349]

Someone responded just before with a link for tips. This reply appears to have been removed.

However, most of these tips I had already completed. One which I hadn’t was under clocking the CPU, I have just attempted this to no avail, temperatures remain around 80 degrees on start up for the CPU, with the H100i coolant getting up to 50 degrees. :-/

 

[Praetorian349]

I have tried another CPU cooler, the Corsair H60 from my wife's computer, in the system and that works, so the H100i was at fault and I have sent this away to be replaced under warranty.

While I am here, would anyone have any insight into a GPU that functions flawlessly while underclocked as far as MSI will underclock it, but when no underclock the GPU appears to crash games to desktop?

1080ti, 4 years old, temperatures don't go beyond 75 degrees celsius, doesn't appear to be under heavy load, however performance tab of task manager indicates it spikes in load to 100% while crashing to desktop... I thought the PSU was at fault, but have replaced this and it still happens. :-/

 

[Nighthawk117]

I have tried another CPU cooler, the Corsair H60 from my wife's computer, in the system and that works, so the H100i was at fault and I have sent this away to be replaced under warranty.

While I am here, would anyone have any insight into a GPU that functions flawlessly while underclocked as far as MSI will underclock it, but when no underclock the GPU appears to crash games to desktop?

1080ti, 4 years old, temperatures don't go beyond 75 degrees celsius, doesn't appear to be under heavy load, however performance tab of task manager indicates it spikes in load to 100% while crashing to desktop... I thought the PSU was at fault, but have replaced this and it still happens. :-/

I'd normally say dodgy pump with AIO's, you shouldn't need to underclock an 8700K with a 240mm Corsair AIO, it's not that difficult to cool. You say the case fans don't start, do you have any fan control software running. Have you checked the fan curves in the bios just in case they are linked to CPU temp or something?

As for the GPU I assume your saying the GPU is fine at idle but not when under load? Insufficient power can often cause that but if you've already tried another PSU. It's 4 years old so I take it this is a recent development?

Did it do it when you had the H60 in? A CPU hitting crazy temps could also cause instability.

 

[Praetorian349]

I'd normally say dodgy pump with AIO's, you shouldn't need to underclock an 8700K with a 240mm Corsair AIO, it's not that difficult to cool. You say the case fans don't start, do you have any fan control software running. Have you checked the fan curves in the bios just in case they are linked to CPU temp or something?

As for the GPU I assume your saying the GPU is fine at idle but not when under load? Insufficient power can often cause that but if you've already tried another PSU. It's 4 years old so I take it this is a recent development?

Did it do it when you had the H60 in? A CPU hitting crazy temps could also cause instability.

Yeah, was definitely the H100i at fault. Case fan issue has been resolved, the Molex cable for the fan controller wasn't plugged in.

The GPU was having issues beforehand starting around 3 months ago, crashing in games or failing to render, screen going black which seemed to be the GPU restarting or something before it returned to rendering the game as normal. This was unrelated to the CPU overheating. I replaced the PSU in preparation for getting a bigger GPU that needs more power, which led to my H100i and CPU temperature woes. 😛

 

[Karadjgne]

Don't rely on gpu temp as if it was Gospel. A 1080ti is a lengthy card that contains several heat producing sources, the gpu, the vram and the VRM's being the major players. Out of them, the only place that gets a temp reading is the gpu itself. Gpu being the graphics processing unit, not the generic name of the entire card.

So if the gpu is hitting 75°C, that does not mention the temp of the VRM's, nor the temp of the vram both of which can easily hit 100°C or more. If you've noticed artifacts before the shutdowns, pretty much that's vram overheat. If it's sudden shutdown, that's generally a VRM issue. Both of which have nothing between them and the heatsink except a thermal pad made of sponge material.

Thermal pads should be replaced with any repaste and vice-versa.

What undervolting the card does is lower voltage demands on the VRM's, giving them a break. So instead of hitting 100°C, they hit 90's instead, which doesn't shut the card down.

The 1080ti was released in March of 2017, so regardless of your actual purchase date, it's possible that card is almost 5 years old.

Meaning it's possible the card is due for a maintenance upgrade with new paste and new pads, only failing now due to pad deterioration.

 

[Praetorian349]

Okay, I might get a mate over and have a crack at reapplying the thermal paste in the 1080ti. If we're successful then it'll at least be worth a bit of cash or I can chuck it in my wife's computer. 😀

An unrelated question that I hope I can get some insight on, I've now installed my new GPU, an EVGA FTW3 3080ti, and since installing this I am hearing coil whine under any form of load. However, the sound is originating from the HX850 PSU I installed for the card. This PSU didn't make this noise with the 1080ti (although I didn't play anything with it in as I was struggling with the CPU temperature issue). Both the PSU and GPU are under warranty, so now I am faced with which component is it that I claim the warranty on?

Could the GPU's high power demands, or some fault in the GPU, be causing the PSU to emit coil whine, or is the PSU at fault as it is where the noise is coming from? Very irritating. 🙁

For the record, I have tried undervolting the GPU as well, to see if that did away with the coil whine, but no dice.

 

[Karadjgne]

Read this applies to motherboards, gpus, psus.

 

[Praetorian349]

Read this applies to motherboards, gpus, psus.

Righto, thanks. So what I take from that is the best way to find out would be to put my 3080ti in my mates computer, put his 3080 in my computer and see whether the 3080ti in his system generates coil whine, or the PSU in mine does with his GPU.

 

[Karadjgne]

Yep. Coil whine can be adjusted for slightly with clock speeds and voltage changes, you change the amount in the inductors, but it's harmless otherwise, just slightly annoying to those without earphones.

 

[Praetorian349]

Yeah, I can hear this through my headphones. xD It's really unacceptable.

And I'm normally not too bothered with noise. My wife hated regular fan noise from my computer before, when she gets back from her secondment she will not be having this coil whine. 😉

 

[Praetorian349]

With the 3080 in my system it was still generating coil whine from the PSU. I'm just wanting a second opinion that that means the PSU is what is at fault and would need replaced?

He doesn't get coil whine with his 3080 in his system, and the noise is coming directly from the coils in the PSU. I'm just anxious I get a new PSU back and still get coil whine in it when paired with my GPU! 😛

 

[Karadjgne]

Coil whine is harmless, and it's evident in every inductor, whether that's gpu or psu, even some motherboards. The difference is its either so quiet you don't hear it outside of the case over the fan/ambient noise or it's such a high frequency it's beyond your hearing range. Psu manufacturers will even slop on some rubberized epoxy, gooping it on the inductors to reduce harmonic vibrations and silence the whine. Many opening electronics for the first time see that goop and think 'what a crap messy glue job' when it actually serves a purpose.

Kinda hard to do on a slim-line gpu board squished upto a heatsink.