Hydro X High CPU Temps

I've recently just put together my first custom loop and I believe I've done everything right but my CPU temps are pretty high when gaming. They would consistently stay around 70-80 C when playing any game. I have my pump set to 4000 RPM and all my fans on my 360 Rad set to max RPM but my rad would get hot to the touch. When I had an h150i, my CPU temps would never go above 60 C when gaming. Is this normal?

 

My loop order is GPU->CPU->Rad->Pump

 

I7 8700k OC'ed to 5 GHZ

 

RTX 2080 Ti

What's your coolant temp?

What's your coolant temp?

 

It's Idle at around 30 C and goes up to around 36 C when gaming

With that coolant temp, that's a bit warmer than I'd expect. My gut would be a contact issue.

How are the GPU temps?

With that coolant temp, that's a bit warmer than I'd expect. My gut would be a contact issue.

How are the GPU temps?

 

My GPU temps are pretty good, doesnt go above 50 in gaming. Do you think it would be safe for me to unmount the CPU block without draining and undoing the tubing?

It looks like you have enough flexibility with your hoses to do so. You may also be able to detect a CPU contact issue from desktop observation without additional load. The CPU temps should get rather spiky with simple tasks like opening folder or launching basic programs. For 5.0 (delid?) I would expect normal task CPU temps to go +10-15C over idle and immediately drop. If you see anything over 60C on the desktop, you probably don't have good contact.

It looks like you have enough flexibility with your hoses to do so. You may also be able to detect a CPU contact issue from desktop observation without additional load. The CPU temps should get rather spiky with simple tasks like opening folder or launching basic programs. For 5.0 (delid?) I would expect normal task CPU temps to go +10-15C over idle and immediately drop. If you see anything over 60C on the desktop, you probably don't have good contact.

 

My Idle temps are pretty good 39-45 C when just doing normal tasks. Its Just when i play games. My water temperatures would go up to 36 and my CPU would consistently be at around 70-80. I also just tried to run a benchmark and then press my finger onto the CPU block but it didnt affect temperatures at all. Maybe the preapplied thermal paste isnt full spread?

odd to see games really driving your cpu that hard relative to your gpu. I also agree sounds like a potential contact/seating issue.

I just reapplied the thermal paste on my cpu, still no improvement

Have you checked the heat spreader and the XC7 with a straightedge for cupping/dishing? Maybe one or both are curved and preventing proper contact?

i dont see why it would dish or cup when i just bought it two days ago

Not saying it did. Eliminating possibilities, because what you're describing sounds and smells like bad contact.

 

These may seem like insultingly simple questions but you never know.

 

When you installed the block, you tensioned it down gradually in a crisscross pattern like when you bolt on a tire?

 

You are using the proper intake port?

 

There wasn't any crud in the block preventing flow across most of the fins?

 

Struggling to throw possibilities past you for consideration. Give you enough wrong ideas maybe one will turn out right.

Oh no, i didnt take any insult to ur question, sorry if my reply seemed that way but anyway

Yes i did tension it down using a crisscross pattern. is it possible i tightened it too hard?

 

I just found out that my cpu block is in reverse flow but would it really affect temperatures that much? On the manual it says i can run it reverse for a performance drop but with my temps, it almost seems necessary to run it correctly. I dont see why corsair would even mention that i could run it reverse if that were the case.

 

No there isnt any gunk preventing the flow

 

Maybe my overclock is too much for this waterloop compared to the h150i?

Reverse flow is fairly bad yeah, but your temps are extremely high. You're running +20C compared to your old AIO.

 

The 360 rad should perform about equivalently to the H150i, since rad capacity mostly determines the cooling capability of your loop and those're the same. Yeah better cpu block and flow rate do help, and so does airflow, but not all that much.

 

I'm running an i9 all cores 5Ghz and I'm down about 5C on my XD5+XC7+XR5-420 compared to the H150i with its 360 rad. A little of that I chalk up to the Kryonaut vs Arctic Silver 5, because the XR5-420 is only marginally more radiating capability. The major improvement for me will come when I get my 2x XR7-480 installed.

 

Given you know your flow is backwards that'd be the first thing to fix just to ensure it's fixed and not contributing, and from there I don't know - if you think you're mounted properly I can't think of great ways to confirm that.

Typically going against the designed CPU block flow costs you 2-3C, not the much larger difference you are looking for. It’s up to you if you want flip around anyway. The block cap can be taken off and the ports rotated in order to maintain your current lengths.

 

Maybe we need to go back to basics. What kind of voltage are you running for 5.0 GHz? Is your 8700K delidded or standard issue? Try using CPU-Z to do a basic coolant temp to cpu assessment. There is a flat load linpack test in cpu-z under the Bench tab. As soon as you start it up the cpu temp should jump and hold steady. You only need to run it for 15-30 seconds. Record the cpu temp and also the coolant temp from when you started. We can make some inferences from the difference between the coolant and load cpu temp.

Typically going against the designed CPU block flow costs you 2-3C, not the much larger difference you are looking for. It’s up to you if you want flip around anyway. The block cap can be taken off and the ports rotated in order to maintain your current lengths.

 

Maybe we need to go back to basics. What kind of voltage are you running for 5.0 GHz? Is your 8700K delidded or standard issue? Try using CPU-Z to do a basic coolant temp to cpu assessment. There is a flat load linpack test in cpu-z under the Bench tab. As soon as you start it up the cpu temp should jump and hold steady. You only need to run it for 15-30 seconds. Record the cpu temp and also the coolant temp from when you started. We can make some inferences from the difference between the coolant and load cpu temp.

 

I ran 5Ghz with 1.33 Volts, but I tried downclocking it to 4.8 GHZ with 1.3 Volts last night but I’m still getting the same temperatures when gaming. My 8700k is standard issue. I’ll do the CPU-Z thing when I get home.

Reverse flow is fairly bad yeah, but your temps are extremely high. You're running +20C compared to your old AIO.

 

The 360 rad should perform about equivalently to the H150i, since rad capacity mostly determines the cooling capability of your loop and those're the same. Yeah better cpu block and flow rate do help, and so does airflow, but not all that much.

 

I'm running an i9 all cores 5Ghz and I'm down about 5C on my XD5+XC7+XR5-420 compared to the H150i with its 360 rad. A little of that I chalk up to the Kryonaut vs Arctic Silver 5, because the XR5-420 is only marginally more radiating capability. The major improvement for me will come when I get my 2x XR7-480 installed.

 

Given you know your flow is backwards that'd be the first thing to fix just to ensure it's fixed and not contributing, and from there I don't know - if you think you're mounted properly I can't think of great ways to confirm that.

 

I’m going to try to do correct my water flow tonight. I can’t find any information online about how much temperatures are affected by reverse flow on the XC7, probably cuz how it’s still kinda new.

 

Ive just tried remounting the XC7 again last night, still no improvements. I also tried pressing down on the block while it’s running to see if harder contact will cause the temperatures to drop, again still no improvements.

 

Do you think one 360 radiator is maybe not enough to run a 2080ti and 8700k together?

I have the 8700K same temps as mine coolant hits 40c while gaming, I think best bet is delid. 8700k hottest cpu I have ever had the spikes are insane.

 

Only time I had low temps was turning off the xmp profile in the motherboard and setting the cpu 4.3 with 1.1 V

I get that it’s a hot CPU but again, I was able to have very decent temperatures with the h150i, so i know it’s capable of running fine without delidding. Right now I’m guessing one 360mm isn’t enough to cool my GPU + CPU. I’m going to try to fix up my flow, and if that doesn’t help, I may add another 360 Rad to my build

1 x 360 for 400-500W may be cutting it a little thin. However, this leads to higher coolant temp which leads to higher component temps. The coolant temp data you gave us was only +6C and that about as low as you go. This also another reason to try the cpu-z test and eliminate the GPU from the equation.

I doubt addding another radiator will lower it down alot, AIO was using a 360 radiator just for the CPU now you are cooling two high end parts in the same loop plus overclocking. I would consider to delid it.

My setup 360 + 240 temps we got about the same temps.

Edited September 24, 2019 by Antixs

I doubt addding another radiator will lower it down alot, AIO was using a 360 radiator just for the CPU now you are cooling two high end parts in the same loop plus overclocking. I would consider to delid it.

My setup 360 + 240 temps we got about the same temps.

 

wait, so youre getting the same temps as me on the cpu?

I get that it’s a hot CPU but again, I was able to have very decent temperatures with the h150i, so i know it’s capable of running fine without delidding. Right now I’m guessing one 360mm isn’t enough to cool my GPU + CPU. I’m going to try to fix up my flow, and if that doesn’t help, I may add another 360 Rad to my build

 

Yeah that's probably a wise approach. I could saturate a 360 with just my i9 alone with higher overclocks during prime or cinebench testing, without adding in a 2080ti.

 

I'm still troubled by the amount of heat your CPU is failing to get out into the water in the first place though. Cooling the additional component should mean youd' saturate your loop sooner, not so much that your CPU temp alone would shoot for the stratosphere.