Jump to content
Corsair Community

Recommended Posts

Posted

So i've been running this Hydro X loop for awhile and something's not right.

The temps seem way off what i'd expect, maybe there's something up with the loop config?

 

I'd appreciate any help.

 

The setup is Pump/Res>GPU block>280mm Rad (top)>CPU>280mm Rad (Front)>pump/Res

 

Both Rads have 2xNoctua NF-A14 fans on them

My CPU is a i9 9900k at stock config and the GPU is an RTX2080Ti, also stock.

 

Idle Temps

Ambient 28c (its warm here at the moment)

Liquid 34c

CPU 36c

GPU 34c

 

Max Load

Liquid 42c

CPU 90c

GPU 55c

 

GPU seems pretty okay but wow that CPU is running hot. I've also noticed that my top mounted rad seems way hotter than the front, something inefficient in my layout maybe?

 

Let me know if you need more info, loop photo attached, and yes that photo was taken before I added the coolant trust me i'm not that stupid!

IMG_6547.thumb.jpg.2776c4b7a607993e76f2154d211d3801.jpg

Posted

Your coolant temps, both relevant to room temp and load seem on the mark for me. We can talk about efficiency and fan direction and possible using a reverse flow for that layout (back intake, both rads exhaust), but the differences would be a couple of degrees on coolant. I don't think that is what you are after.

 

What I do notice is the 90C CPU and 55C GPU. If the 55C GPU occurs when the coolant is 42C, then that is about right. Most 2080 Ti will have a GPU diode to coolant delta of +12-15C. Just like with a CPU is varies a little bit from chip to chip, but also most certainly from GPU power. That leaves the CPU. At 42C coolant, 90C is a +48C differential. That is not out of bounds, but this value is also directly tied to your Vcore level and individual chip characteristics.

 

Are you running the typical 5.0GHz OC in the 1.25-1.35v range? Do you see this in CPU only load situation where the coolant is not as warm and the differential even greater?

Posted

Thanks guys.

@Infin1tum, so it's the first rad that's hot which I reckon is odd, it's after then GPU and before the CPU. As for it being a mounting issue with the CPU maybe but I actually built two of these machines at the same time, they're identical and both CPUs hit the same temps.. Id doubt I've messed up the mounting twice.

 

@c-attack I agree regarding the GPU temps, the feel about right given my environment but the CPU getting close to thermal limits seems odd. Yes on cooler days its a few degrees lower and no, there's no OC at all, stock config.

Posted

Both rads should be roughly at the same temperature so there is a problem.

If the top one is hot then you have a serious flow issue i would say.

At idle, coolant temperature should be very close to ambient. 34°C is what i'd expect under load without overclock at 28 ambient.

 

I am curious what your pump speed is ?

Posted

The range I give most people is coolant averages +4-7C above ambient for idle. There are a lot of factors involved, but case layout and design and idle CPU wattage are the heaviest. Nobody with 10980X or anything else using 100W at idle will be at ambient. Case and radiator placement the other. Glass cases will be higher. The all mesh blow through model on the low end. Regardless, I don't think +6C over ambient automatically indicates a flow problem.

 

It seems like the CPU is main point of concern and it does seem like the odd value in the data. When do you see 90C? Stress test? Gaming? If it is just in something like Prime 95 or OCCT, I would not be overly concerned. If this is during gaming, that is more alarming. Even if you don't overclock, you may want to see what peak VCore the processor gets up to. "Auto" voltage is almost unusable for most applications. Even if you want to keep things at default settings, giving it a specific target value is one of the largest improvements you can make. We can stuff your room full of 360mm radaitors and not take 10-15C off the top end. You can do that with voltage refinement in a few minutes. I would expect the Vcore to stay under 1.25v for the normal peak/turbo mode settings.

Posted (edited)

My concern was that the top rad was "way hotter" than the front one. That's what makes me think about a pump that's running way too slow. No matter what the loop order is, there shouldn't be temperature differences you could feel that sharply from one rad to the other.

 

If pump speed is really low, coolant temp will also rise. Those NZXT cases are not that hot usually.

Duno if it could be related, but the pipe bend on pump outlet looks really kinked. Could be a flash artifact though

Edited by LeDoyen
Posted

@c-attack yes the 90c is during stress tests, i'm going to go get some proper gaming numbers tonight. I'll also get you the voltage numbers since your curious :)

 

@LeDoyen, the pump is managed with the hydro-x profile max speeds I see under full load are in the 4000-4400rpm range.

Posted

hmm it's almost full speed for a D5, but that hot spot bugs me ^^'

I am curious too, could you take a closer picture of the pipe bend going from the pump to the RTX? it really looks kinked, i wonder if there's like a big flow restriction because of it.

Posted (edited)

my 9900k runs hot in benchmarking- hits high 70s but is o/c to 5Ghz all core @ 1.28v- seems to be the sweetspot for this one, have tried lower volts but bluescreens anything less than 1.28

I have a 240 x 360 rad in my corsair 680x so a little more rad space than yours- I wouldnt be too concerned with it tbh as I believe 105c is throttling temp-please feel free to correct me:D:

If your machine is mostly for gaming then as long as the temps are not stupidly high then it will be fine

Edited by knacker40
×
×
  • Create New...