Temperature way too high after switching to liquid cooling

[MChrisp]

Hi everyone,

 

I have a "slightly" annoying problem. Yesterday I installed custom liquid cooling from EKWB. Right now I am just cooling my Asus IX Fromula Mainboard and my 7700K. I have a tripple 120mm radiator and a D5 Pump with 240 mL reservoir.

According to EKWB I should get water temps of about 25°C. It is summer so it is a little bit warmer than that, but not over 35°C... The thing is, that my package temperature during idling is a toasty 65 to 80°C. Now you gonna say I should check the thermal paste. I did. First time I applied a thin layer of MX4 perfectly even and spread out. Very thin, but not too thin. 65 to 80°C. I tried the included thermal paste and of that a fair amount. 65 to 80°C. I again switched to MX4. 65 to 80°C. So that basically should not be the problem. And I also checked if the pump is running. It is.

 

The cores itself are about 35°C to 40°C with spikes to the 80s. My MoBo on the other hand is also watercooled and is about 40°C.

 

The second strange thing is, that my cpu package temp does not give a damn about load. It just stays at the idling temp. (See attached picture) It also does not care about putting the fans on 100% or 50%. Just stays at the given temperature. (Even after a while). Compared to that, my MoBo Temp went down to 30°C from 40°C doing that.

 

This should not happen at all with this kind of coolin. I expected something like 40°C or lowe during idling. This is realistic, isn't it? I have no clue what to do. Even my 20$ air cooling was way better than this...

 

Does anyone has a clue whats going on? I would be very thankfull if someone can help me. I have no Idea what to do...

 

Maybe it is a faulty sensor, but normally they would be way off, wouldn't it? Or it just does not show the correct temps...

 

PS.: In the bios the cpu temp is about 30°C, and I cuppled one fan to this temp, so I actually hear what my MoBo thinks of my CPU Temp, and it certainly also think it is high, as my fan goes nuts...

Spoiler

PPS.: That weird tube with a 90° Angle before the cpu is for the later implementation of my 1080ti, which I sadly broke during taking of the stock cooler. (Scratched of a SMD Resistor, which I soldered on myself, but than I noticed that I ordered the wrong freaking water block...  So i have to wait for another shipment to even check if my soldering job worked...)

 

[Master Disaster]

Do you have your pump running a full speed or did you fill your loop with lemonade?

 

You need to unplug your PSU from the board, short it and let your pump run until all those bubbles disappear from your res.

 

I'd suggest you get another app (coretemp or hwmonitor) and verify what's being reported is correct.

[MChrisp]

8 minutes ago, Master Disaster said:

Do you have your pump running a full speed or did you fill your loop with lemonade?

 

You need to unplug your PSU from the board, short it and let your pump run until all those bubbles disappear from your res.

 

I'd suggest you get another app (coretemp or hwmonitor) and verify what's being reported is correct.

Thanks for the fast reply. Not Lemonade. EKWB pre-mixed coolant. The bubbles are just in the reservoir. The rest of the loop has no air in it. I tilted the case so all airpockets should be gone. I don't know, why there are so many bubbles in the reservoir. They are not coming from the loop.

[Hiitchy]

Just out of curiosity, what configuration do you have your radiator fans running in? Not that it matters but I don't even see fans on the radiator?

[MChrisp]

2 minutes ago, Hiitchy said:

Just out of curiosity, what configuration do you have your radiator fans running in? Not that it matters but I don't even see fans on the radiator?

They push from the outside. The mount is between the radiator and the fans.

[Hiitchy]

1 minute ago, MChrisp said:

They push from the outside. The mount is between the radiator and the fans.

Understood, you should probably do what @Master Disaster says to get the bubbles out of your system

[For Science!]

9 minutes ago, MChrisp said:

They push from the outside. The mount is between the radiator and the fans.

Just for my record, is this a new system? as in, have you had this system running on air before for comparison?

[MChrisp]

Just now, Hiitchy said:

Understood, you should probably do what @Master Disaster says to get the bubbles out of your system

As I said, they are just in the reservoir. No bubbles at all in the tubes. Don't know why they are in the reservoir... But the system should be flooded very good. Tilted it in a lot of ways, to get the air out.

[MChrisp]

11 minutes ago, For Science! said:

Just for my record, is this a new system? as in, have you had this system running on air before for comparison?

I had a Alpenföhn Air Cooler for 1 month. And it was never noticable loud, so I think everything was fine back than, but I never checked, as it was just a temporary solution and I had no troubles with it... I Could mount it again, but I don't think that helps. Problem should be the liquid cooling...

[For Science!]

2 minutes ago, MChrisp said:

I had a Alpenföhn Air Cooler for 1 month. And it was never noticable loud, so I think everything was fine back than, but I never checked, as it was just a temporary solution and I had no troubles with it... I Could mount it again, but I don't think that helps. Problem should be the liquid cooling...

Can you get a reading of the water temperature? I believe the motherboard has some sensors on it. Also could you just confirm the BIOS version number?

[MChrisp]

2 hours ago, For Science! said:

Can you get a reading of the water temperature? I believe the motherboard has some sensors on it. Also could you just confirm the BIOS version number?

MoBo Temp is 40 right now. (I guess that's the temp of the watercooled part) Other MoBo 76°C. Temp is Package is 60°C. Bios should be the latest, but I check. BRB

Bios is 0906 acc. to CAM. Can't check if that's the latest. Asus Website seems to be down...

[For Science!]

23 minutes ago, MChrisp said:

-snip-

No immediately obvious solutions but here are some thoughts and comments:

 

- Those bubbles in the reservoir are just some adherence via surface area. To remove them you will need to rotate the reservoir and let a bigger air bubble merge with the little ones, and then they will travel to the top of the res as one big air bubble.

 

- The high idling temps to me indicate a dodgy mount onto the CPU, since it sounds like you pre-spread the thermal paste before the mounting block, it would be hard to know if you are mounting the block with enough pressure. Perhaps you can try once with the dot method just to confirm the mounting pressure is adequate

 

- I noticed you are pushing quite a lot of voltage (1.35/1.38V) for quite a modest overclock. Did you set this yourself? While not completely responsible for the bad idling temps, it would play a factor. I was rock solid stable at 1.3 V for 4.8 GHz (and many others would quote even better voltages)  if you want to use that as a reference.

 

- I was mistaken about your motherboard having sensors for water temperature, you only have headers that could be used if you plugged in a sensor. The 40C is likelyjust a random sensor on the board and is probably a better representation of your case ambient + components close to that sensor. The 76C is probably your chipset, quite hot, but not alarmingly so. You can expect idling temperatures of about 5C more than the water temperature for an 7700K.d

 

- More of a thought than a suggestion; although we typically say that loop order does not matter, I noticed that most CPU monoblocks cool the CPUs before the VRMs. This could just be design, but perhaps worth taking into consideration. VRMs are the hottest things on the board afterall. But I would only tinker with this after establishing that the mounting pressure is good.

[Arika]

why is your entire case upside down?

[For Science!]

38 minutes ago, MChrisp said:

-snip-

Also could you take a look using CPUID HWMonitor? I'm intrigued by your core temps vs package temps.

 

1 minute ago, Sierra Fox said:

why is your entire case upside down?

Inverted motherboard layout cases are fairly common from be quiet!

[MChrisp]

2 minutes ago, Sierra Fox said:

why is your entire case upside down?

Because it is a be quiet dark base pro and i inverted it, as my pc stands on the left side of my desk and i wanted to look at it.

[MChrisp]

4 minutes ago, For Science! said:

Also could you take a look using CPUID HWMonitor? I'm intrigued by your core temps vs package temps.

 

Inverted motherboard layout cases are fairly common from be quiet!

The OC is a standard setting of the MoBo. But I did have the same problem without OC. Just wanted to check if it affects the temps. it gets a bit hotter. About 5°C.

I have chosen the loop order just on aesthetics, as I thought I would have more than evough cooling headroom... BTW thanks for your help.

[For Science!]

1 hour ago, MChrisp said:

-snip-

Is this a delidded CPU? 40 degree delta on one core is not normal at all. 

 

Is it always Core #1 that is hot? or does the hot core alternate between the 4 cores?

[MChrisp]

17 minutes ago, For Science! said:

Is it always Core #1 that is hot? or does the hot core alternate between the 4 cores?

No, it varies. Overall it varies quite a lot. not sure if that is normal. See attached picture. (on line = 5°C)

It is not delidded. Didn't know this is possible

[For Science!]

4 minutes ago, MChrisp said:

No, it varies. Overall it varies quite a lot. not sure if that is normal. See attached picture. (on line = 5°C)

It is not delidded. Didn't know this is possible

Very weird, indeed at this point I can only recommend changing back to the air cooler and seeing whether the same behaviour is present. If so, it is just down to the TIM between the die and IHS. If not, then mounting pressure of the EVO waterblock.

 

Hope you are successful

[MChrisp]

42 minutes ago, For Science! said:

Very weird, indeed at this point I can only recommend changing back to the air cooler and seeing whether the same behaviour is present. If so, it is just down to the TIM between the die and IHS. If not, then mounting pressure of the EVO waterblock.

 

Hope you are successful

I changed to my old air-cooler... Exact Same behavior. just a bit warmer... cores go up to 100°C. But no thermal throtling. Shouldn't that happen at this temperatures? Also the Heatsink is not even remotely hot. It is just a bit warm. like 35°C or so... This is not normal if the package is at 90°C, is it?

[MChrisp]

52 minutes ago, For Science! said:

Very weird, indeed at this point I can only recommend changing back to the air cooler and seeing whether the same behaviour is present. If so, it is just down to the TIM between the die and IHS. If not, then mounting pressure of the EVO waterblock.

 

Hope you are successful

I correct myself. if i stop the fan for like 20s it does thermal throttle and get's fairly warm. Not hot though...

[MChrisp]

1 hour ago, For Science! said:

Very weird, indeed at this point I can only recommend changing back to the air cooler and seeing whether the same behaviour is present. If so, it is just down to the TIM between the die and IHS. If not, then mounting pressure of the EVO waterblock.

 

Hope you are successful

Now that I hear how the cpu fan goes nuts and fluctuates in speed I am pretty sure, I did not have the Problem before. I would have noticed. So basically I have to buy a new CPU I guess... Maybe I get a refund...

[For Science!]

1 hour ago, MChrisp said:

-snip-..

Did you apply the thermal compount using the dot method? Can you if possible post photos of:

 

1. The dot of thermal compound you applied before putting the cooler on

 

2. How the thermal compound looked like on the CPU after putting the cooler on and taking it off

 

3. The same as #2 but on the CPU cooler side

 

 

[FrankV]

Not sure it's been mentioned but did you remove the sheet of plastic that is on the water block? 

[MChrisp]

28 minutes ago, FrankV said:

Not sure it's been mentioned but did you remove the sheet of plastic that is on the water block? 

Sure. It is kind of obvius. Besides, it is written on it  But Thanks.