So I initially had some issues with my hybrid temps that I figured out (with a repaste) now I max out at 59C core and 86C mem junction after hours of gaming. I don’t mine but I will game for 6 or more hours at a time. Currently I max the power, temp, and voltage sliders. I know the temps are fine but what about voltage? I frequently see 1.1 volts and 430+ watts on vanguard, control, cyberpunk etc. am I significantly reducing the lifespan of my card with the increased voltage? I do get higher clocks with the increased Voltage.

I've got daily driver OC profiles that I'll use for games that demand it. Some games I play are so old that the GPU speed doesn't stay boosted to max with or without my OC so it has no impact at all. My 3080ti is water cooled all over, so I'm not worried about temperature-induced lifetime reduction. My daily driver OC also isn't flat-out right on the bleeding edge either. Given my monitor has a 120hz refresh rate my own rule is that for any game where I already see over 120fps without an OC I just won't bother OCing it. The OC Club isn't going to cut a corner off your virtual member's card if they find out you don't OC 24/7.
 
I'd say test out the games that you play and see what framerates you get with and without the OC, or with a milder OC. If you wouldn't notice the difference then don't run it OC in those games. If you would then go for it and don't sweat it. I don't have a hybrid but have followed some of the threads that discuss potential issues with how the non-water-cooled parts on that card are cooled. I'd say use your best judgment on whether your usage patterns represent some sort of temperature abuse that makes you worried. 

Those are all valid points. I game at 4k 120hz with a cx, and yeah the frame difference I’m seeing is only 3-7 at max. I think I’ll just keep the oc profile for games that I’m right around 60fps to make sure I don’t dip below. Thanks for the responses!

kev012in
So I initially had some issues with my hybrid temps that I figured out (with a repaste) now I max out at 59C core and 86C mem junction after hours of gaming. I don’t mine but I will game for 6 or more hours at a time. Currently I max the power, temp, and voltage sliders. I know the temps are fine but what about voltage? I frequently see 1.1 volts and 430+ watts on vanguard, control, cyberpunk etc. am I significantly reducing the lifespan of my card with the increased voltage? I do get higher clocks with the increased Voltage.

I'm not fond of the EVGA memory plate to hybrid interface.  Thankfully this solves it https://www.coolmygpu.com/product/hybrid/4.
Stock I was seeing temps in the high 90s for memory which is shocking because it's ~16c+ higher than the air cards.  The core temp runs good.  Hybrids offer a lot for what they are, the memory temps on the 80 and up cards with 6x leaves a lot to be desired. 
Asus's AIO design with its large copper plate is a better solution, not sure why EVGA didn't opt for it considering asetek makes pretty much every hybrid GPU on the market including Dell's OEM stuff. 

nosomo

kev012in
So I initially had some issues with my hybrid temps that I figured out (with a repaste) now I max out at 59C core and 86C mem junction after hours of gaming. I don’t mine but I will game for 6 or more hours at a time. Currently I max the power, temp, and voltage sliders. I know the temps are fine but what about voltage? I frequently see 1.1 volts and 430+ watts on vanguard, control, cyberpunk etc. am I significantly reducing the lifespan of my card with the increased voltage? I do get higher clocks with the increased Voltage.

I'm not fond of the EVGA memory plate to hybrid interface.  Thankfully this solves it https://www.coolmygpu.com/product/hybrid/4.
Stock I was seeing temps in the high 90s for memory which is shocking because it's ~16c+ higher than the air cards.  The core temp runs good.  Hybrids offer a lot for what they are, the memory temps on the 80 and up cards with 6x leaves a lot to be desired. 
Asus's AIO design with its large copper plate is a better solution, not sure why EVGA didn't opt for it considering asetek makes pretty much every hybrid GPU on the market including Dell's OEM stuff. 

So that plate does not use pads at all?!?!  How can that possibly not screw up core contact pressure?  I could see it working for some cards and not others depending on tolerances.  Feels like a really bad idea.

what does that plate do differently? looks the same to me.

talon951
 
 
So that plate does not use pads at all?!?!  How can that possibly not screw up core contact pressure?  I could see it working for some cards and not others depending on tolerances.  Feels like a really bad idea.

No pads on memory, inductor / VRM pads are still used.  There's a little play / flex in system as a whole.  These are machined parts, not a simple one size fits all approach.  I have 4 in operation.  My kingpin results are awesome.

talon951
According to the site it is thicker and takes up the gap the thermal pads normally fill. So then I guess you just use paste, same as the core.

I've seen plenty of instances where people just used pads that were too hard and/or thick that hurt core temps significantly. How can this possibly work with the mem making hard contact with the copper plate?

I could see this working if it left a small gap and putty was used instead of pads on the mem.

oh, that makes sense. quick question though, are you supposed to have paste between the two plates?

Kingofroam

oh, that makes sense. quick question though, are you supposed to have paste between the two plates?

Yes, on the hybrid there's 2 plates and a little paste is applied between them.   Pump block - paste as usual - flat plate - paste above mem location - machined mem plate - paste.  There's imgur link that shows it.   I did several meticulous fitment tests with everything cleaned.  I put a real small drop on the core and then put the pump assembly / memory plate on the card w/ its 4 screws (didn't bother with the whole shroud for test fitments), then unscrewed the pump from the card and with the plates in place the little bit on the core was completely spread so there's clear pressure on the core which is unquestionably important. 

talon951
The paste getting spread out on the core doesn't guarantee the core has sufficient pressure on it.  I'd have to see a before and after on core temp to be convinced it works.  And with tolerance build up, some could work well and not others.  Even if the new plate is held to very tight tolerance, the water block itself may not be.  

The water block / pump assemblies are machined.  I've got 6 different hybrid kits and they're all machined to within .02 mm of each other.  Asetek is incredibly precise with their assemblies, I'm impressed.  Just wish EVGA opted for the larger 3mm thick plate like asus AIO cards have.

Here's one that's been in for 9 months now untouched running the most difficult rendering there is...
Just noticed the SRAM power -- 80 watts on the mem!!!!  Furmark is nasty


https://imgur.com/a/wqQMRWM
visit the link for larger size w/ date (it's running right now)

Kingofroam
 
HAHAHAHA. Hey everyone I just found out why my memory junc temps are 85-95. oops, looks like i need to order new pads and some paste.
 
Thanks for the clarification.
 
In my own defense I bought the hybrid kit used, it was cleaned and the instructions online didn't mention any paste there (as far as I can remember). I remember thinking paste should go there, but I didn't design it so whoever did probably knows better.

Well to be fair your temps aren't out of the norm for the hybrids.  There's a lot of pressure in that little area.  I'd wager that not using paste there is only worth 4c, 6c at worst.  I've done a stupid amount of testing.  For instance using huge heatsink on the back of the 3090 only picked up 4c reduction in mem temps.

Kingofroam
 
Would you say its worth buying more pads and paste and most importantly running the risk of a novice (me) breaking my card? 
Also, I have since put a better fan curve on which has lowered the temps but all I can tell you is that it doesn't hit 90 now.

That's up to you.  If I  understand you installed the hybrid kit onto your air card yourself?  It's a pretty straight forward process. Hybrid cards use 2mm thermal pads on the VRM (thin) and 1mm on the inductors.  That's what I use at least, I've found the stock hybrid kit pads to be a little thinner --- on the display output side of the card I found that stock wasn't even touching the inductors.  I put the card together, took it apart and there was no witness marks on the thermal pad at all --- as in it touched nothing.

It's a good habit to clean all of the components and then without using any pads / paste, go through the motions of putting the card back together (basically just seating the assembly onto the card, don't plug/unplug items.  Just get a feel for the process itself so it's not a cumbersome motion once you have pads and paste on.  Use ample light, I use a little head lamp myself so I can see alignment.  Once back together with pads / paste on before fully screwing it together, give the fan a little spin and make sure neither the pump power wire nor RGB wires are protruding into the fan area.  It can get a little tight in there and one of them can end up hitting the fan.   Best to catch it before it's all compressed with screws. 

nosomo

talon951
The paste getting spread out on the core doesn't guarantee the core has sufficient pressure on it.  I'd have to see a before and after on core temp to be convinced it works.  And with tolerance build up, some could work well and not others.  Even if the new plate is held to very tight tolerance, the water block itself may not be.  

The water block / pump assemblies are machined.  I've got 6 different hybrid kits and they're all machined to within .02 mm of each other.  Asetek is incredibly precise with their assemblies, I'm impressed.  Just wish EVGA opted for the larger 3mm thick plate like asus AIO cards have.

Here's one that's been in for 9 months now untouched running the most difficult rendering there is...
Just noticed the SRAM power -- 80 watts on the mem!!!!  Furmark is nasty


https://imgur.com/a/wqQMRWM
visit the link for larger size w/ date (it's running right now)

Fair enough. I tried running that same config on my 3080ti hybrid. I set the power limit so it would be close to what you show in the screenshot. I saw almost identical core temp.

But I also saw very close to the same mem temp too. So I'm not completely sold. I did remount my hybrid with Gelid Ultimates along with pads on the backplate and heatsinks. 24C ambient at the time.

nosomo

talon951
Fair enough. I tried running that same config on my 3080ti hybrid. I set the power limit so it would be close to what you show in the screenshot. I saw almost identical core temp.

But I also saw very close to the same mem temp too. So I'm not completely sold. I did remount my hybrid with Gelid Ultimates along with pads on the backplate and heatsinks. 24C ambient at the time.

You set +100 core and +500 mem and got the same temp?  This card stock was doing 96c with those settings.

Yes. That was about what mine did before the remount and other changes I mentioned.

I decided to retest it anyway.  I did have my rad fans at 100% (ML120's).  They are on a mobo header, so it shows up as zero in GPUz.  Also don't let the gpu temp in the Kombustor window confuse things.  It mistakenly picks up my 2080 super that is also in that machine.  Here you go,