Overclock.net banner

1 - 20 of 21 Posts

·
Registered
Joined
·
263 Posts
Discussion Starter #1
In waiting for the rest of my wc supplies I decided to put a loop on my GPU to play with it for a bit. Initially my 2080 super ftw3 was getting super hot right away with the Hydrocopper wc block. I reseated it a few times with no luck until evga support advised to put extra thermal paste and so I did and temps improved. Now since I'm new at this wc stuff I'm not sure what temps to expect. I have an alphacool nexxos XT45 280mm radiator at the front with two arctic pst 140mm fans in push config. The flow meter shows around 5-6L/min at 40% pump, fans around 1200 rpm. Ambient is around 25C and with my pc at idle the fluid temp is 33-34 with the GPU at 36. During a two hr gaming session fluid will get as high as 44C and gpu anywhere between 55-65C. Is this ok? The rest of the system is a 3700X inside a BeQuiet Dark base 700 mid tower. I will be adding the cpu when the stuff comes in as well as two more rads, one 140 at the back and one 280 at the top, both 30mm thick. Airflow will be in at the front, out thru the top and rear.
 

Attachments

·
Overclocker in training
Joined
·
13,062 Posts
Hi,
Haven't got a 20 series so don't know but seems high
10 series tops out at 45c.

Sure you got the flow going the designed direction ?
Unless the pump makes noise no reason to slow the flow down.
 

·
Registered
Joined
·
1,623 Posts
Too high, I'd say. Unless your fans spin at 1000-1200RPM under load! I have a 2080Ti Kingpin with 240mm AiO that maxes out <50*C at 2115MHz and 2 NF-12x25 at full speed.
 

·
Registered
Joined
·
1,539 Posts
In waiting for the rest of my wc supplies I decided to put a loop on my GPU to play with it for a bit. Initially my 2080 super ftw3 was getting super hot right away with the Hydrocopper wc block. I reseated it a few times with no luck until evga support advised to put extra thermal paste and so I did and temps improved. Now since I'm new at this wc stuff I'm not sure what temps to expect. I have an alphacool nexxos XT45 280mm radiator at the front with two arctic pst 140mm fans in push config. The flow meter shows around 5-6L/min at 40% pump, fans around 1200 rpm. Ambient is around 25C and with my pc at idle the fluid temp is 33-34 with the GPU at 36. During a two hr gaming session fluid will get as high as 44C and gpu anywhere between 55-65C. Is this ok? The rest of the system is a 3700X inside a BeQuiet Dark base 700 mid tower. I will be adding the cpu when the stuff comes in as well as two more rads, one 140 at the back and one 280 at the top, both 30mm thick. Airflow will be in at the front, out thru the top and rear.
Seems like you have some room to improve. I have a hard time believing your flow meter since I have 2 pumps at 100% and am barely hitting 1 gpm. Then again my loop might be more restrictive. Try increasing pump speed.

Too high, I'd say. Unless your fans spin at 1000-1200RPM under load! I have a 2080Ti Kingpin with 240mm AiO that maxes out <50*C at 2115MHz and 2 NF-12x25 at full speed.
My Kingpin has the stock cooler and runs at 2130/2145 and approaches 60 C at 23C ambient. I have one of the voltage dip switches flipped. How fast are you running those aftermarket fans?
 

·
Registered
Joined
·
1,623 Posts
Should be around 1900RPM but these Noctuas are still pretty quiet and perfectly usable at full RPM. The rad is at the bottom of the case as it's an inverted layout. The stock fans are rubbish imho.
With 30-32*C ambient and fans at full speed I'm at 52-53*C when gaming and 100% GPU load.
 

·
Rig Advisor
Joined
·
2,010 Posts
Should be around 1900RPM but these Noctuas are still pretty quiet and perfectly usable at full RPM. The rad is at the bottom of the case as it's an inverted layout. The stock fans are rubbish imho.
With 30-32*C ambient and fans at full speed I'm at 52-53*C when gaming and 100% GPU load.
Danger zone for coolant temp. Is that a DDC or D5 boxfish? I know that when your loop temps hit high 40's and 50's it causes damage to the pump.

*Edit, I know your board is inverted, but if i'm not mistaken, you have the inlet/outlet backwards on your GPU.

With just a GPU in the loop, your temps are ridiculously high for gaming. Change inlet/outlet and watch it all get better. Consider replacing with better thermal pads >11 W/MK
 

·
Registered
Joined
·
263 Posts
Discussion Starter #7
Too high, I'd say. Unless your fans spin at 1000-1200RPM under load! I have a 2080Ti Kingpin with 240mm AiO that maxes out <50*C at 2115MHz and 2 NF-12x25 at full speed.
Yeah I had the fans just at that speed. I got them at 90% now which is about 1500RPM according the Asus Fan Xpert fan tuning.

Danger zone for coolant temp. Is that a DDC or D5 boxfish? I know that when your loop temps hit high 40's and 50's it causes damage to the pump.

*Edit, I know your board is inverted, but if i'm not mistaken, you have the inlet/outlet backwards on your GPU.

With just a GPU in the loop, your temps are ridiculously high for gaming. Change inlet/outlet and watch it all get better. Consider replacing with better thermal pads >11 W/MK
It's a D5 Boxfish but some ppl seem to think this is a knock off D5 so who knows. The inlet and outlet on the GPU is as in the picture(yellow lines), once I get the rest of the loop parts it will be going from pump/res to cpu>140mm rad>gpu>280 thin rad>280 thick rad>res...see red lines. I read the manual for the Hydrocopper and it states that either left or right can be used for inlet or outlet so I don't know. It is my first time watercooling so bare with me.
Right now the pump is at 60% giving approx 8-9L/min, fans at 90% which is 1500RPM(arctic fans) and the fluid temp as per the display monitor is showing 33C, GPU is sitting idle at 35C, ambient I would say 26-27 but I can't tell for sure as I don't have a thermometer. I haven't played any games since a couple of days ago so. The higher temps were produced during Escape from Tarkov gaming session. Battlefield V did not yield such temps, it was more in the 50s.

I did have issues with the cold plate and die contact on the hydrocopper which led to taking it apart and putting it back together like five times, frustrated AF. But evga support suggested to slap on more thermal paste which led to better temps. The thermal pads are the ones that came with the block. Which other pads are better?

I'm getting new screws for the gpu block soon so I'll try getting a better thermal paste spread(maybe get Thermal Grizzly Kryonaut too)
 

Attachments

·
Overclocker in training
Joined
·
13,062 Posts

·
Rig Advisor
Joined
·
2,010 Posts
Yeah I had the fans just at that speed. I got them at 90% now which is about 1500RPM according the Asus Fan Xpert fan tuning.



It's a D5 Boxfish but some ppl seem to think this is a knock off D5 so who knows. The inlet and outlet on the GPU is as in the picture(yellow lines), once I get the rest of the loop parts it will be going from pump/res to cpu>140mm rad>gpu>280 thin rad>280 thick rad>res...see red lines. I read the manual for the Hydrocopper and it states that either left or right can be used for inlet or outlet so I don't know. It is my first time watercooling so bare with me.
Right now the pump is at 60% giving approx 8-9L/min, fans at 90% which is 1500RPM(arctic fans) and the fluid temp as per the display monitor is showing 33C, GPU is sitting idle at 35C, ambient I would say 26-27 but I can't tell for sure as I don't have a thermometer. I haven't played any games since a couple of days ago so. The higher temps were produced during Escape from Tarkov gaming session. Battlefield V did not yield such temps, it was more in the 50s.

I did have issues with the cold plate and die contact on the hydrocopper which led to taking it apart and putting it back together like five times, frustrated AF. But evga support suggested to slap on more thermal paste which led to better temps. The thermal pads are the ones that came with the block. Which other pads are better?

I'm getting new screws for the gpu block soon so I'll try getting a better thermal paste spread(maybe get Thermal Grizzly Kryonaut too)
Totally fine and totally here to bear with you.

I use Fuji poly, well made and good thermals. I use the 11 W/mK, the 17 is just too expensive for what you get: https://www.amazon.com/s?k=thermal+pads+fujipoly+11&ref=nb_sb_noss_2

Thrash appears to be right about flow with the EVGA copper, though switching them to test wont hurt.

Your Delta T is just too high, I am about 3c Meaning ambient right now is 24c, loop temp 27c, both GPUs at 29/30 and my highest core at 33/34
 

·
Overclocker in training
Joined
·
13,062 Posts
Totally fine and totally here to bear with you.

I use Fuji poly, well made and good thermals. I use the 11 W/mK, the 17 is just too expensive for what you get: https://www.amazon.com/s?k=thermal+pads+fujipoly+11&ref=nb_sb_noss_2

Thrash appears to be right about flow with the EVGA copper, though switching them to test wont hurt.

Your Delta T is just too high, I am about 3c Meaning ambient right now is 24c, loop temp 27c, both GPUs at 29/30 and my highest core at 33/34
Hi,
Those are unforgiving pads way too easy to tear/.... and one off no good after disassembly.
I just use arctic pads all sizes.
 

·
Registered
Joined
·
263 Posts
Discussion Starter #11
Totally fine and totally here to bear with you.

I use Fuji poly, well made and good thermals. I use the 11 W/mK, the 17 is just too expensive for what you get: https://www.amazon.com/s?k=thermal+pads+fujipoly+11&ref=nb_sb_noss_2

Thrash appears to be right about flow with the EVGA copper, though switching them to test wont hurt.

Your Delta T is just too high, I am about 3c Meaning ambient right now is 24c, loop temp 27c, both GPUs at 29/30 and my highest core at 33/34
Yeah I don’t know what could be causing DeltaT to be so high. Maybe ambient is higher than 27. I’ll see if I can get a thermometer off a friend or something. I also have read people have had issues with contact between cold plate and gpu die with these hydrocopper blocks. I bought it in May thinking of EVGA’s great reputation and based on my good experience with their products, and now I can’t return it. Oh well, I’ll see what happens when I throw in the extra two rads and the cpu into the loop and also re paste the gpu block with better/newer TIM. The paste I’m using now is ek, from a friend who built his loop a year ago so it might be old now.
 

·
Registered
Joined
·
89 Posts
Yeah I don’t know what could be causing DeltaT to be so high. Maybe ambient is higher than 27. I’ll see if I can get a thermometer off a friend or something. I also have read people have had issues with contact between cold plate and gpu die with these hydrocopper blocks. I bought it in May thinking of EVGA’s great reputation and based on my good experience with their products, and now I can’t return it. Oh well, I’ll see what happens when I throw in the extra two rads and the cpu into the loop and also re paste the gpu block with better/newer TIM. The paste I’m using now is ek, from a friend who built his loop a year ago so it might be old now.
how bout a copper shim? thermal pads on vram and mosfet too thick? is it anything like the other guy that had a flow indicator blocking flow?
 

·
Registered
Joined
·
263 Posts
Discussion Starter #13 (Edited)
how bout a copper shim? thermal pads on vram and mosfet too thick? is it anything like the other guy that had a flow indicator blocking flow?
How is this shim? https://www.amazon.ca/dp/B07ZSWDJ7M/ref=cm_sw_r_cp_apa_i_E5QmFbSTC0WFQ
I can also get thinner shims, what thickness is recommended? I have liquid metal from collaboratory labs that I used when I delidded my old i7; however, I don't know how I feel about putting liquid metal between die and shim or shim and cold plate. Another guy just filed down the standoffs and apparently that worked great, tho this would probably void the warranty should a leak arise.
I used the 0.5mm thermal pads on VRMs and mosfets but they were too thin so I wen back to the original. That ended up being a waste of money.
Who's that person you're referring to with regards to the flowmeter?
I'll see if I can try it without the flow meter.
The bottom part of the rad is pretty restricted after the rad because of the pump/res being placed there and it's bracket. I don't know if that does anything other than mess the air flow into the case tbh.
 

·
Village Idiot
Joined
·
2,368 Posts
Some of these replies are so ridiculous they make my head spin. Let's break your situation down.

with my pc at idle the fluid temp is 33-34 with the GPU at 36

Your idle GPU temp relative to water temp is completely normal.

During a two hr gaming session fluid will get as high as 44C and gpu anywhere between 55-65C. Is this ok?

Yes, it's OK, but it's not exactly spectacular either. 44C water temps aren't going to hurt anything, but it leaves a lot to be desired in terms of performance. We'll get to that in a minute.

Water cooled Turing cards typically run around 10-15*C above water temperature, but I don't understand your comment about GPU temps being anywhere between 55-65C. Are you saying the card runs at 65C when the water temperature is 44C? What is the water temperature when the GPU temps are 55C? What happens to GPU temps if you run the pump at 100% speed?

If you're seeing fluctuations between 55-65C on your GPU at the same water temperatures, you have something wrong, probably a contact issue. Even 65C load with 44C water temp is kind of high and I would suspect a mounting issue, especially if EVGA's "solution" was telling you to "slap on more thermal paste". I'm assuming you mean more thermal paste on the die... how did you spread the paste? How much did you use? Did it look like the scene out of the infamous Verge PC build video?

I have an alphacool nexxos XT45 280mm radiator at the front with two arctic pst 140mm fans in push config fans around 1200 rpm.

Let's look at some test data for your radiator:
https://www.xtremerigs.net/2016/02/25/hardware-labs-nemesis-280-gtx-radiator-review/5/
(they don't have an article published specifically on the XT45 280mm, but the data is included in this review)

So in open air, with 1300 RPM NF-A14 fans (similar performance to the P14s based on the minimal amount of data out there) they determined that the XT45 280mm can dissipate 254.6 watts of heat while maintaining a 10c delta (temperature rise of the water above ambient). Stuff the radiator inside your case with a restrictive looking closed front panel, with slightly slower fans, and you can expect a lot worse performance. Let's just assume 200 watts, but even that is probably being overly optimistic based on my experience with having the front door open/closed and intake filters on/off with my Fractal R6 and Define 7 XL.

Going back to your water temperature of 44C and an unknown ambient you estimate is 25-27C... Your water temp is possibly anywhere around 17-19C above ambient.

You have not mentioned whether or not you have increased the power limit on your graphics card. Stock is 270 watts, and maxed out PL is 351 watts.

"best case" scenario (1300 RPM fans in open air)
Stock = 270 watts / 254.6 = 10.6C delta
Max PL = 351 / 254.6 = 13.8C delta

"more realistic" scenario (radiator suffocating inside case)
Stock = 270 watts / 200 = 13.5C delta
Max PL = 351 / 200 = 17.5C delta

So depending on your actual ambient temperature, how restrictive your case is, and what kind of power limit you're running (along with actual average power consumption of the card while gaming), your 44C water temperature might not be too unreasonable.
 

·
Overclocker in training
Joined
·
13,062 Posts
Hi,
What a rambling boob

You'd have to show how much timm wasn't enough and how much was a little better
If you take it apart use a thermal pad 0.5 and assemble and disassemble and see if the gpu die leaves a impression on the thermal pad keep adding until it does leave an impression
Not sure about evga reputation on hydro copper blocks is, gpu's are fine though but contact on the die is a must.
Water temp matters too.

So take it apart again and take some images this time.
 

·
Registered
Joined
·
1,623 Posts
I think that pewpewlazer has a point, the coolant is way too hot at 44*C, 55-65*C load temp is nothing surprising then.
 

·
Overclocker in training
Joined
·
13,062 Posts
I think that pewpewlazer has a point, the coolant is way too hot at 44*C, 55-65*C load temp is nothing surprising then.
Hi,
Nothing new added though except his obnoxious remarks towards others otherwise water temp was already mentioned on post 6.
 

·
Village Idiot
Joined
·
2,368 Posts
I think that pewpewlazer has a point, the coolant is way too hot at 44*C, 55-65*C load temp is nothing surprising then.
That wasn't my point at all.

1. 44c coolant temp is most definitely hot, but it's within the realm of what he should expect for his setup (relatively high ambient temperature, single 280 rad, 1200 RPM fans, not so good case, etc). There's nothing glaringly wrong with the thermal performance of his actual water loop, and he certainly should not be expecting a 3c water delta.

2. Regardless of his water temperature, the difference between water temp and load temp won't change.

44c water -> 55c load temp = good, this is the kind of performance he should expect.
44c water -> 65c load temp = bad, this is not what he should expect and it is indicative of a block mounting issue like he seems to suspect.
 

·
Invalid Media
Joined
·
2,162 Posts
In waiting for the rest of my wc supplies I decided to put a loop on my GPU to play with it for a bit. Initially my 2080 super ftw3 was getting super hot right away with the Hydrocopper wc block. I reseated it a few times with no luck until evga support advised to put extra thermal paste and so I did and temps improved. Now since I'm new at this wc stuff I'm not sure what temps to expect. I have an alphacool nexxos XT45 280mm radiator at the front with two arctic pst 140mm fans in push config. The flow meter shows around 5-6L/min at 40% pump, fans around 1200 rpm. Ambient is around 25C and with my pc at idle the fluid temp is 33-34 with the GPU at 36. During a two hr gaming session fluid will get as high as 44C and gpu anywhere between 55-65C. Is this ok? The rest of the system is a 3700X inside a BeQuiet Dark base 700 mid tower. I will be adding the cpu when the stuff comes in as well as two more rads, one 140 at the back and one 280 at the top, both 30mm thick. Airflow will be in at the front, out thru the top and rear.

...temps seem high to me (I run 2x 2080 Ti w-c, with GPUs temp in the low to mid 30C range , though I also use more pumps and rads). One thing I noticed in your first pic is that there seems to be air in the loop - was that fixed after you took the pic but before the observed and reported temps ?
 

·
Registered
Joined
·
263 Posts
Discussion Starter #20 (Edited)
Water cooled Turing cards typically run around 10-15*C above water temperature, but I don't understand your comment about GPU temps being anywhere between 55-65C. Are you saying the card runs at 65C when the water temperature is 44C? What is the water temperature when the GPU temps are 55C? What happens to GPU temps if you run the pump at 100% speed?
//55-65 is just the difference from the GPU being highly used to when other times the usage is not as high, for example during in game menu etc. Water temperature is steady throughout. So I suspect a contact issue as I've read about it in evga forums.(I'm waiting for new screws to see if it helps, if not then shorter standoffs or an RMA)

I'm assuming you mean more thermal paste on the die
//correct...
how did you spread the paste? //I don't have a spatula thingy so i sliced a q-tip and tried using the flat part of the plastic tube. I know, ghetto, I'll cut a credit card or something straight and try that or just order new TIM that will come with a spreader.
How much did you use?
//a lot, maybe three grains of rice worth, large grains of rice.
Did it look like the scene out of the infamous Verge PC build video?
//hahaha never seen those



You have not mentioned whether or not you have increased the power limit on your graphics card. Stock is 270 watts, and maxed out PL is 351 watts.
//I have not, it is running stock.

"best case" scenario (1300 RPM fans in open air)
Stock = 270 watts / 254.6 = 10.6C delta
Max PL = 351 / 254.6 = 13.8C delta

"more realistic" scenario (radiator suffocating inside case)
Stock = 270 watts / 200 = 13.5C delta
Max PL = 351 / 200 = 17.5C delta

So depending on your actual ambient temperature, how restrictive your case is, and what kind of power limit you're running (along with actual average power consumption of the card while gaming), your 44C water temperature might not be too unreasonable.

//I just got home at I managed to fin a digital thermometer, ambient was 30C and loop was 36.6, call it 37. So I'm seeing a 7C delta T from ambient. AC has been running for a bit, temperature went down to 28C and loop is sitting at 34.7C

thank you
Hi,
What a rambling boob

You'd have to show how much timm wasn't enough //rice/pea size like you would put on a CPU IHS
and how much was a little better
//read above

If you take it apart use a thermal pad 0.5 and assemble and disassemble and see if the gpu die leaves a impression on the thermal pad keep adding until it does leave an impression
Not sure about evga reputation on hydro copper blocks is, gpu's are fine though but contact on the die is a must.
//I will try this tomorrow if I don't work or next time I have a lot of free time
That wasn't my point at all.

1. 44c coolant temp is most definitely hot, but it's within the realm of what he should expect for his setup (relatively high ambient temperature, single 280 rad, 1200 RPM fans, not so good case, etc). There's nothing glaringly wrong with the thermal performance of his actual water loop, and he certainly should not be expecting a 3c water delta.

2. Regardless of his water temperature, the difference between water temp and load temp won't change.

44c water -> 55c load temp = good, this is the kind of performance he should expect.
44c water -> 65c load temp = bad, this is not what he should expect and it is indicative of a block mounting issue like he seems to suspect.
I'm just going to wait guys for the rest of the fittings/tubing, then reassemble everything and try getting an even thick layer of TIM on the GPU die. I'm also waiting for a new set of screws from EVGA as they did not ship with my waterblock so I used the ones that came with the card. Instructions read to use M3x7mm screws and I measured the ones I had on hand with a caliper and they were 6.9mm so I presume that is not the issue but who knows.
If it doesn't improve I will try to get an RMA, if it keeps happening then I will grind down the standoffs with a dremel or a file, 1/16th of an inch as another person did and see if that provides a better contact. I would rather file down standoffs than get shims or slam all kinds of thermal paste on the die.

I am not aiming for insane good performance but I want to make sure it is reasonable since this hobby is expensive and I don't want to feel like I've spent wads of cash just to get the same level of performance as air cooling. I also understand that my case is not the best for airflow but I picked it because my current desk setup needs an inverted motherboard layout so I can look at the eye candy.

The two pictures attached are 1: when I applied TIM in an X pattern and 2: current status has a little more than this picture, extra TIM

Quick question, how many applications would I be able to get out of 1gram of thermal paste? Just deciding on whether to buy 1g tube or 5.5g tube of Thermal Grizzly Kryonaut
Thanks
@J7SC there is no air in the loop right now.
 

Attachments

1 - 20 of 21 Posts
Top