Ronchan411

KingEngineRevUp

Ronchan411
So I have finished the assembly of my 3080 Ti today but I am not so happy with the temperatures. The delta is jumping immediately to 15C after the 1st second of full load and constantly climbs close to 20C. i have seen reviews showing 8-10C delta for this block, do I have a contact issue? This jump start from 25C to 40C in the 1st second is really unusual, I never had this on my past cards....
 
Ron
 
 
(I'm using an external radiator in case of confusion ;) ) 

I am 18-20C delta over my water and that appears to be the norm. My memory is 55-60C max. 
 
What is your temperature over water?

After 1 hour of gaming GPU was on 50C, memory at 51C and PWR at 51-56C. Delta over water as said 15-20. I will try this trick to tighten the screws again after the card has been warmed up. maybe this will help, otherwise I will repaste it with Kryonaut.

I'm already using Kryonaut and getting the same temperatures as you are. I don't think you can improve that much unless you use liquid metal. My census have shown me that multiple people have the same results. 
 
I think that EK has the GPU mounting pad a few micro meters too short so it's not making the best mounting pressure. That being said, I have spoke to a number of Hydro Copper people on reddit and they are getting similar results. I believe Optimus is the only block that gets around 10C delta over water. Is it worth the extra $200+? Up to user to decide. 

Ronchan411
Understand. I was just wondering as I had a 3080 MSI Ventus before, the Bykski block managed 13C delta and the Corsair around 15-16. expected EKWB to be the best… i will tighten the screws after warming it up and let it be as it is…

Ah, well the Ventus is pulling only 320W https://www.techpowerup.c...tx3080-10240-210308-1, the 3080 Ti FTW3 pulls 450W. Linearly scaling this, is around 18-20C. 
 
When I'm gaming in APEX, Warzone or Assassin creed Valhalla and drawing 350W, I am 13C over water, which is close to your Ventus. 

Maybe EK's pads are too hard. Has anyone tried replacing them?

KingEngineRevUp

talon951
Maybe EK's pads are too hard. Has anyone tried replacing them?

I've thought about using putty, but I feel the EK pads when compressed right do a good job getting memory to 55-60C.

What if softer pads get worse memo temperatures, what is worth it then? Lower core or lower memo temperatures?

I might do liquid metal on 9-12 months. But I want to test on the winter and see if there even is a boost bon worth chasing.

I still think it's the waterblock boss being a few micro meters off. The EKWB Strix block performs pretty well from what I hear with the same pads.

Softer pads tend to get better temps due to them conforming better around the chip. You would definitely get better mem temps using Gelid Extreme's. At least I did on my Corsair block. About 12C. And they are very soft. I doubt putty would perform any better.

Same thickness.

talon951

KingEngineRevUp

talon951
Maybe EK's pads are too hard. Has anyone tried replacing them?

I've thought about using putty, but I feel the EK pads when compressed right do a good job getting memory to 55-60C.

What if softer pads get worse memo temperatures, what is worth it then? Lower core or lower memo temperatures?

I might do liquid metal on 9-12 months. But I want to test on the winter and see if there even is a boost bon worth chasing.

I still think it's the waterblock boss being a few micro meters off. The EKWB Strix block performs pretty well from what I hear with the same pads.

Softer pads tend to get better temps due to them conforming better around the chip. You would definitely get better mem temps using Gelid Extreme's. At least I did on my Corsair block. About 12C. And they are very soft. I doubt putty would perform any better.

Maybe someone can test it, but my screws are so bottomed out that the standoffs get unscrewed when I unscrew then. The pads don't seem that hard either, I have them compressed possibly more than 50%

https://i.imgur.com/fursFsO.jpeg

talon951
That suggests they are a little thick I think.

I think I've shared this with you before, but thermal pads should be compressed to at least 40% to reach a point of diminishing returns and for the best thermal performance.
 

 
Even GELID test data shows they compress them to 40% for their "conductivity rating"
 

Ronchan411
So in case I want to repaste with Kryonaut and tighten the screws more, shall I warm up the whole block with a hair dryer or something to make the original pads softer? Cannot let it run hot in the system itself as I have hardtubes and it would take too long to remove the card.....

I'm using Kryonaut already and my results are 18-20C over water while drawing 450W. I have tightened my screws to the point where they are stuck in the standoff. Unscrewing them actually undoes the stand-off n the WB itself. 
 
Your time may not be worth it unless you have worse results than mine. But if you must, just run a benchmark for awhile like Timespy Extreme or Port Royal, do the 20 loops, and the card should be warmed up. Nothing beats having the card itself actually warm up and heat the pads directly. Then remove the card, remove the backplate and see if you can tighten all the screws any further. 

Ronchan411

KingEngineRevUp

Ronchan411
So in case I want to repaste with Kryonaut and tighten the screws more, shall I warm up the whole block with a hair dryer or something to make the original pads softer? Cannot let it run hot in the system itself as I have hardtubes and it would take too long to remove the card.....

I'm using Kryonaut already and my results are 18-20C over water while drawing 450W. I have tightened my screws to the point where they are stuck in the standoff. Unscrewing them actually undoes the stand-off n the WB itself. 
 
Your time may not be worth it unless you have worse results than mine. But if you must, just run a benchmark for awhile like Timespy Extreme or Port Royal, do the 20 loops, and the card should be warmed up. Nothing beats having the card itself actually warm up and heat the pads directly. Then remove the card, remove the backplate and see if you can tighten all the screws any further. 

Got it, I have 18-20 with just 400W, so worth a try for me to tighten them and a nice project for the coming weekend :). Thanks again. 

I also want to let you know that a user over at Overclock.net measured the actual reading of the card and it can draw more than 450W. 
 
https://www.overclock.net/threads/official-nvidia-rtx-3080-ti-owners-club.1791823/post-28835729
 
EVGA has come out and said that their cards are not read correctly by 3rd party software. 
 
So keep in mind, you're more than likely drawing a lot more heat than you think you are. That may also be a reason why the deltas are higher. 
 
490W vs your 320W 3080 Ventus is a 50% increase in heat, so imagine you were getting 13-15C before, that is 20-22.5C now. 
 
But yes, definitely try to tighten the screws a bit more if you can. 

TahoeDust

KingEngineRevUp
I also want to let you know that a user over at Overclock.net measured the actual reading of the card and it can draw more than 450W. 
 
https://www.overclock.net/threads/official-nvidia-rtx-3080-ti-owners-club.1791823/post-28835729
 
EVGA has come out and said that their cards are not read correctly by 3rd party software. 
 
So keep in mind, you're more than likely drawing a lot more heat than you think you are. That may also be a reason why the deltas are higher. 
 
490W vs your 320W 3080 Ventus is a 50% increase in heat, so imagine you were getting 13-15C before, that is 20-22.5C now. 
 
But yes, definitely try to tighten the screws a bit more if you can. 

Do you not worry at all about cracking the die when tightening down?  How tight are we really wanting to go here.  Is there a suggested number out there?  I snugged them pretty well, but I do a bit of gunsmithing and have some nice fitting driver bits and a in/lb torque wrench.  I could REALLY tighten them down if I wanted.

The standoffs, screws and boss (pad) height were already all engineered to a certain mounting pressure. That's what I did when I used to do electromechanical avionics in one of my past jobs.

I bottomed the screws out and my die is not cracked. I actually double washered the 4 screws too with extra plastic washers that came stock to make sure I bottomed it out.

The only reason I did this though is because I already know the EKWB puts "acceptable" mounting pressure. Not enough to crack the die.

This took retorque the card 3 times to double check to not crack the die. I didn't just go crank it down the first time.

Lastly, ymmv. I'm not responsible for anyone else here double washering and they end up damaging their card. I'm just reporting what I did to get the best deltas I can with this WB.

KingEngineRevUp

TahoeDust

KingEngineRevUp
I also want to let you know that a user over at Overclock.net measured the actual reading of the card and it can draw more than 450W. 
 
https://www.overclock.net/threads/official-nvidia-rtx-3080-ti-owners-club.1791823/post-28835729
 
EVGA has come out and said that their cards are not read correctly by 3rd party software. 
 
So keep in mind, you're more than likely drawing a lot more heat than you think you are. That may also be a reason why the deltas are higher. 
 
490W vs your 320W 3080 Ventus is a 50% increase in heat, so imagine you were getting 13-15C before, that is 20-22.5C now. 
 
But yes, definitely try to tighten the screws a bit more if you can. 

Do you not worry at all about cracking the die when tightening down?  How tight are we really wanting to go here.  Is there a suggested number out there?  I snugged them pretty well, but I do a bit of gunsmithing and have some nice fitting driver bits and a in/lb torque wrench.  I could REALLY tighten them down if I wanted.

The standoffs, screws and boss (pad) height were already all engineered to a certain mounting pressure. That's what I did when I used to do electromechanical avionics in one of my past jobs.

I bottomed the screws out and my die is not cracked. I actually double washered the 4 screws too with extra plastic washers that came stock to make sure I bottomed it out.

The only reason I did this though is because I already know the EKWB puts "acceptable" mounting pressure. Not enough to crack the die.

This took retorque the card 3 times to double check to not crack the die. I didn't just go crank it down the first time.

Lastly, ymmv. I'm not responsible for anyone else here double washering and they end up damaging their card. I'm just reporting what I did to get the best deltas I can with this WB.

10-4.  I guess the giant warning in the EKWB instruction about over tightening and cracking the dye gave me pause.  Maybe I will pull the card a re-tighten this weekend.
 
 
...or just order the Optimus block Friday.  If I am going to pull the card, may as well try something new.

KingEngineRevUp

TahoeDust

KingEngineRevUp

TahoeDust

KingEngineRevUp
I also want to let you know that a user over at Overclock.net measured the actual reading of the card and it can draw more than 450W. 

https://www.overclock.net/threads/official-nvidia-rtx-3080-ti-owners-club.1791823/post-28835729

EVGA has come out and said that their cards are not read correctly by 3rd party software. 

So keep in mind, you're more than likely drawing a lot more heat than you think you are. That may also be a reason why the deltas are higher. 

490W vs your 320W 3080 Ventus is a 50% increase in heat, so imagine you were getting 13-15C before, that is 20-22.5C now. 

But yes, definitely try to tighten the screws a bit more if you can. 

Do you not worry at all about cracking the die when tightening down?  How tight are we really wanting to go here.  Is there a suggested number out there?  I snugged them pretty well, but I do a bit of gunsmithing and have some nice fitting driver bits and a in/lb torque wrench.  I could REALLY tighten them down if I wanted.

The standoffs, screws and boss (pad) height were already all engineered to a certain mounting pressure. That's what I did when I used to do electromechanical avionics in one of my past jobs.

I bottomed the screws out and my die is not cracked. I actually double washered the 4 screws too with extra plastic washers that came stock to make sure I bottomed it out.

The only reason I did this though is because I already know the EKWB puts "acceptable" mounting pressure. Not enough to crack the die.

This took retorque the card 3 times to double check to not crack the die. I didn't just go crank it down the first time.

Lastly, ymmv. I'm not responsible for anyone else here double washering and they end up damaging their card. I'm just reporting what I did to get the best deltas I can with this WB.

10-4.  I guess the giant warning in the EKWB instruction about over tightening and cracking the dye gave me pause.  Maybe I will pull the card a re-tighten this weekend.
 
 
...or just order the Optimus block Friday.  If I am going to pull the card, may as well try something new.

I can't justify a Optimus block when that can go towards a 4080 Ti in like 12 months lol

Are they putting those up every Friday and can be easily snagged or is this another F5 battle?

KingEngineRevUp
I can't justify a Optimus block when that can go towards a 4080 Ti in like 12 months lol

Meh.  Optimus block will still be worth something in 12 months.  Even at half value, cost of ownership would only be ~$150.  I can live with that...not much in this hobby.

the FTW is easely overcklokable though

Here is my build. I'll take better pics once I get a better camera/phone.

Corsair 7000D
MSI MEG Z590 Ace
Intel i9-11900k w/Corsair Waterblock
EVGA 3090 FTW3 w/EK Waterblock
Corsair Vengeance 32gb RAM @ 3600mhz
WD_Black SN850 1tb SSD
WD_Black SN750 1tb SSD
WD_Blue SN550 2tb SSD
EVGA Supernova 1000 G2 80+ Gold
Corsair/Bitspower fittings
Corsair hardline 14mm tube
Corsair XD5 Pump/Res Combo
Corsair QL120 fans
480mm radiator
360mm radiator
120mm radiator
Mayhem Pastel Dark Grey Coolant




http://imgur.com/gallery/yrrEqwl