Xc7 , 3 radiators and 13900k overheating?

[Sir Beregond]

6 minutes ago, NorKris said:

but i wonder if the "take" is that there is a temp diff between waterblocks... i dont think there is

I'd venture a guess and say there might be a mounting/contact issue here. 32-33c water temp is pretty good under load.

[FUSER]

5 hours ago, Sir Beregond said:

I'd venture a guess and say there might be a mounting/contact issue here. 32-33c water temp is pretty good under load.

 

On 1/19/2023 at 6:37 AM, Selle said:

If the water temp tops out at 32c your cooling capacity is good, and the issue has to be related to getting the heat from the cpu to the water. It could be the block not being good enough, a bad mount, bad thermal paste application or a issue with the tim in between the cpu did and the heat spreader

 

On 1/18/2023 at 2:36 AM, NorKris said:

intake is better for temps than exhaust. 4in, 6 out.. 

 

rads should be intake 

 

with that said, my 13900k also reaches 100c in stresstests, do i care that much, no.

why? cuz 70-75c in gaming is fine

 

another bit of info we need is: how much power does ur cpu take/use 

Ok so the following changes have been made:

- all 9 radiator fans (3x3) are intake

- 1 case exhaust fan

- replaced intel locking mechanism with thermal grizzly frame

- removed all paste from xc7 and processor and reapplied one pea-sized kryonaut paste

- reservoir filled to the top

 

results:

 

- water temp can now range from 23-24 at idle up to 37-38 at continuous maxed out occt stress test, with cpu temp never above 99C. This suggests that at least the new frame and/or the paste revision helped significantly.

- case temp never goes above 34 so i guess no bottlenecking there

- needless to say gpu temp maxes at mid 50s with hotspot low 80s, at 3ghz ocer @ 1440p ( asus tuf 4900 oc)

 

my biggest bottleneck is my mediocre silicon (sp 95 combined p/e core), maxing out for now score of 38000 with cinebench r23

 

[andrewmp6]

On 2/6/2023 at 5:28 PM, NorKris said:

low end waterblock?

A heatkiller cpu block is 80 us and will be 5 to 8 degrees cooler then that corsair one.So yes for the price and what you get the corsair stuff is low end.

[NorKris]

3 hours ago, andrewmp6 said:

A heatkiller cpu block is 80 us and will be 5 to 8 degrees cooler then that corsair one.So yes for the price and what you get the corsair stuff is low end.

how do u know this?

[NorKris]

8 hours ago, FUSER said:

Ok so the following changes have been made:

- all 9 radiator fans (3x3) are intake

- 1 case exhaust fan

- replaced intel locking mechanism with thermal grizzly frame

- removed all paste from xc7 and processor and reapplied one pea-sized kryonaut paste

- reservoir filled to the top

 

results:

- water temp can now range from 23-24 at idle up to 37-38 at continuous maxed out occt stress test, with cpu temp never above 99C. This suggests that at least the new frame and/or the paste revision helped significantly.

- case temp never goes above 34 so i guess no bottlenecking there

- needless to say gpu temp maxes at mid 50s with hotspot low 80s, at 3ghz ocer @ 1440p ( asus tuf 4900 oc)

 

my biggest bottleneck is my mediocre silicon (sp 95 combined p/e core), maxing out for now score of 38000 with cinebench r23

 

ok few things i would like to point out here,

1. Since we are dealing with thermal limits and we would like to squize out every bit of performance,  thermalpaste should be spread out by hand so that you cant see any of the top of the cpu(IHS). 

2. No1 cares about idle anything (unless they are crazy high...) 

3. Abit confused about the +5c in water temp BUT this can be cuz ur CPU is now boosting higher and /or using more W now. you did not tell us what ur Wats was before the change (my bad for not pointing that out). 

4. What is case temp?

5. The Temp you told us before was a "vague" 90+, but now ur telling us "never over 99c", 

[Sir Beregond]

On 2/7/2023 at 11:13 PM, NorKris said:

how do u know this?

Been around the block enough to see people who have used both and what the results were. In general the Corsair block was a more middle of the pack block. I wouldn't say "low end", but its definitely a lower performer vs something like a Heatkiller IV Pro.

 

And really most of Corsair's water-cooling products are either rebrands with a jacked up price (the Corsair radiators are HWLabs L-series with the Corsair sailboat logo tax), or else just really low quality materials compared to comparably priced competitors. For example, put a Corsair reservoir up against the Aquacomputer Ultitube and tell me which one is better quality of materials, yet the Corsair costs more.

 

Corsair definitely did a good job of creating a line of products that makes it easy to get into watercooling, but you are definitely paying the Corsair tax for them. I do like their fittings though, as they are Bitspower rebrands, but often cheaper than actual Bitspower fittings. Kind of the exact opposite of what they did with the radiator rebrands where they are more expensive than actual HWLabs L-series and often costing the same as HWLabs Nemesis series which are just plain and simple better radiators than the L-series.

[NorKris]

4 hours ago, Sir Beregond said:

Been around the block enough to see people who have used both and what the results were. In general the Corsair block was a more middle of the pack block. I wouldn't say "low end", but its definitely a lower performer vs something like a Heatkiller IV Pro.

 

And really most of Corsair's water-cooling products are either rebrands with a jacked up price (the Corsair radiators are HWLabs L-series with the Corsair sailboat logo tax), or else just really low quality materials compared to comparably priced competitors. For example, put a Corsair reservoir up against the Aquacomputer Ultitube and tell me which one is better quality of materials, yet the Corsair costs more.

 

Corsair definitely did a good job of creating a line of products that makes it easy to get into watercooling, but you are definitely paying the Corsair tax for them. I do like their fittings though, as they are Bitspower rebrands, but often cheaper than actual Bitspower fittings. Kind of the exact opposite of what they did with the radiator rebrands where they are more expensive than actual HWLabs L-series and often costing the same as HWLabs Nemesis series which are just plain and simple better radiators than the L-series.

ok i wil call misinformation on this.

 

LTT did a test where they compaird a H100 vs a custom loop(i think it was EK) with same size rad. if ur statement is true the temps diff should have been many c.  

 

[FUSER]

On 2/8/2023 at 5:20 AM, NorKris said:

ok few things i would like to point out here,

1. Since we are dealing with thermal limits and we would like to squize out every bit of performance,  thermalpaste should be spread out by hand so that you cant see any of the top of the cpu(IHS). 

2. No1 cares about idle anything (unless they are crazy high...) 

3. Abit confused about the +5c in water temp BUT this can be cuz ur CPU is now boosting higher and /or using more W now. you did not tell us what ur Wats was before the change (my bad for not pointing that out). 

4. What is case temp?

5. The Temp you told us before was a "vague" 90+, but now ur telling us "never over 99c", 

Cpu is consuming around 250 watts max

case temp is the air inside the case

the temp i mentioned earlier was cpu package temps 

 

[NorKris]

1 hour ago, FUSER said:

Cpu is consuming around 250 watts max

case temp is the air inside the case

the temp i mentioned earlier was cpu package temps 

 

what is 250 in both tests?

 

i know what "Case temp" is,  but what i want to know is, how do u know what it is

[FUSER]

20 hours ago, NorKris said:

what is 250 in both tests?

 

i know what "Case temp" is,  but what i want to know is, how do u know what it is

240 watts on OCCT all-core maxed out test.

 

I have a temp probe hovering between the cpu , raised GPU  and the reservoir (equidistant from all), to capture the air temperature of the case. That's the air that has moved through the radiators from outside in. It tells me basically if my exhaust fan can remove any accumulating hot air fast enough (observing the delta between reservoir temp and case air temp).

 

[NorKris]

5 hours ago, FUSER said:

240 watts on OCCT all-core maxed out test.

 

I have a temp probe hovering between the cpu , raised GPU  and the reservoir (equidistant from all), to capture the air temperature of the case. That's the air that has moved through the radiators from outside in. It tells me basically if my exhaust fan can remove any accumulating hot air fast enough (observing the delta between reservoir temp and case air temp).

 

Now:

240W gives you 99c+

before:
xxxW gives you 90c+ ?

 

[FUSER]

6 hours ago, NorKris said:

Now:

240W gives you 99c+

before:
xxxW gives you 90c+ ?

 

Correction: 240W before the changes.
295-300W after the changes. Apparently it can pull more power now, Temps before were 99+ requiring thermal throttling.

 

Temps now are 90-98, with fans+punp @ 100%, without thermal throttling, scoring  roughly 2000 points higher on cinebench

 

vcores at 1.25-1.26v on p-cores

[NorKris]

19 minutes ago, FUSER said:

Correction: 240W before the changes.
295-300W after the changes. Apparently it can pull more power now, Temps before were 99+ requiring thermal throttling.

 

Temps now are 90-98, with fans+punp @ 100%, without thermal throttling, scoring  roughly 2000 points higher on cinebench

 

vcores at 1.25-1.26v on p-cores

Ah nice! 300w  > 100c is about the same as my setup looks good 

[Sir Beregond]

On 2/9/2023 at 1:47 PM, NorKris said:

ok i wil call misinformation on this.

 

LTT did a test where they compaird a H100 vs a custom loop(i think it was EK) with same size rad. if ur statement is true the temps diff should have been many c.  

 

H100...How long ago was that lol. I outperformed an original H100 with an XSPC Raystorm back in the day. But keep in mind custom loop also has other variables like number of rads, flow rates, restrictions, pump speeds, etc.

[NorKris]

10 minutes ago, Sir Beregond said:

H100...How long ago was that lol. I outperformed an original H100 with an XSPC Raystorm back in the day. But keep in mind custom loop also has other variables like number of rads, flow rates, restrictions, pump speeds, etc.

XSPC Raystorm and 240 rad?

 

when linus and others tested the EK kit vs h100  the EK won with 1c... 
it had better pump, copper Rad, better fans. better coldplate.  

 

if u have seen the inside of a h100 coldplate, its KEKW compared to the one in the EK set.   

 

 

[Sir Beregond]

1 hour ago, NorKris said:

XSPC Raystorm and 240 rad?

 

when linus and others tested the EK kit vs h100  the EK won with 1c... 
it had better pump, copper Rad, better fans. better coldplate.  

 

if u have seen the inside of a h100 coldplate, its KEKW compared to the one in the EK set.   

 

 

I'm not a big fan of EK, so I'm not surprised.

 

But again, how long ago was this. H100 released nearly 12 years ago. I'm really not sure what point you are making with it. This thread we are talking about an XC7 water block.

[NorKris]

31 minutes ago, Sir Beregond said:

I'm not a big fan of EK, so I'm not surprised.

 

But again, how long ago was this. H100 released nearly 12 years ago. I'm really not sure what point you are making with it. This thread we are talking about an XC7 water block.

when you come with statements like XC7 is a lowend coldplate and you can get alot cooler temps by changing ...

"Block A" ( H100)  lost by 1c to "Block B" (EKs kit) and block B had better pump and rad...

so where is all the temp improvements your talking about? 

[Sir Beregond]

7 hours ago, NorKris said:

when you come with statements like XC7 is a lowend coldplate and you can get alot cooler temps by changing ...

"Block A" ( H100)  lost by 1c to "Block B" (EKs kit) and block B had better pump and rad...

so where is all the temp improvements your talking about? 

I didn't call the XC7 low-end, someone else did. I did call it middle of the pack. Also, still fail to see why you keep bringing up the H100 vs 12 year old waterblocks.

[andrewmp6]

On 2/8/2023 at 1:13 AM, NorKris said:

how do u know this?

https://www.guru3d.com/articles-pages/review-liquid-cooling-with-a-techn-waterblock,5.html

The corsair block in other reviews runs the same as the ek velocity in temps 

[andrewmp6]

On 2/15/2023 at 12:06 AM, Sir Beregond said:

I didn't call the XC7 low-end, someone else did. I did call it middle of the pack. Also, still fail to see why you keep bringing up the H100 vs 12 year old waterblocks.

I called it low end,For the price you can buy better cooling parts.Some cases a lot better parts.