Question Conductonaut on a 2080 with the copper Vapour chamber?

[mildewman]

I have recently aquired a 2080, and it has a large copper vapor chamber with a bunch of aluminium fins attached.
I conductonauted my 980 and my 1070TI, and would like to apply it to this card as well. ( Im aware that its a bit iffy with copper, but i dont mind redoing it yearly when the temps are so good )

Has anyone done this process, is there anything in particular i need to watch out for?
Things that worry me is if there are a number of heat pads on the ram chips that i have to preserve the integrity of, and the problem where the much high viscousity of the liquid metal makes a way thinner layer thats not actually thick enough to join the chip and the heatsink unless you sand the standoff legs of the heatsink down slightly - but not so much you crack the gpu....

Im currently getting about 2Ghz @ 60C when gaming, is it worth me risking frying / damaging my expensive card to go LM, and if so what should i watch out for in particular?

 

[BonzaiDuck]

I have recently aquired a 2080, and it has a large copper vapor chamber with a bunch of aluminium fins attached.
I conductonauted my 980 and my 1070TI, and would like to apply it to this card as well. ( Im aware that its a bit iffy with copper, but i dont mind redoing it yearly when the temps are so good )

Has anyone done this process, is there anything in particular i need to watch out for?
Things that worry me is if there are a number of heat pads on the ram chips that i have to preserve the integrity of, and the problem where the much high viscousity of the liquid metal makes a way thinner layer thats not actually thick enough to join the chip and the heatsink unless you sand the standoff legs of the heatsink down slightly - but not so much you crack the gpu....

I'm currently getting about 2Ghz @ 60C when gaming, is it worth me risking frying / damaging my expensive card to go LM, and if so what should i watch out for in particular?

I've had the same idea for some time. I've concocted schemes in my mind to assure that the Thermal Grizzly doesn't "pump out" as some say it might over time. For instance, you could make a "dam" of something thick like IC diamond along the edge of the chip where it mates with the heatsink base.

But then, you'd want to make sure that the VRAM chips were also given a similar treatment. Too much complexity -- too much to go wrong.

On a more positive note, the guy at Silicon Lottery doesn't believe TG Conductonaut "pumps out". He says he would've "heard something" from among over 10,000 customers over the last five or six years. That, of course, is the application of TG to CPU dies. I never noticed any changes on the Skylake I had done over four years of use, though.

But if the heatsink base is bare copper as opposed to copper with nickel-plate, I've heard stories. Some stories say it forms an amalgam with the copper. Others suggest that it erodes or damages the copper.

 

[mildewman]

Yes, in my experience if the LM is only touching Nickel and Sillicon and there are no air gaps it lasts forever without temps degrading. If there is copper then the bits of amalgam will eventually cause things to decay till you clean up the heatsink and reapply the goop - but you get a few years before this happens.

Today i attempt to put a Noctua double tower and conductonaut on my R5 3600. Putting LM on the stock Prisim cooler made very little improvement, but this total overkill cooler should


**edit - Installed the Noctua D14 with conductonaut, has made perhaps 2 degress difference .......

 

[BonzaiDuck]

Yes, in my experience if the LM is only touching Nickel and Sillicon and there are no air gaps it lasts forever without temps degrading. If there is copper then the bits of amalgam will eventually cause things to decay till you clean up the heatsink and reapply the goop - but you get a few years before this happens.

Today i attempt to put a Noctua double tower and conductonaut on my R5 3600. Putting LM on the stock Prisim cooler made very little improvement, but this total overkill cooler should


**edit - Installed the Noctua D14 with conductonaut, has made perhaps 2 degress difference .......

I take it this was a conventional installation without removing the processor cap. That's what you get, then. It's all out there in comparison reviews. For me, it just seems more reasonable to use nano-diamond paste than to fiddle with LM, for which I could anticipate a 2C improvement over a normal range of thermal power. But you apparently aren't wise to use nano-diamond paste on a bare-die GPU -- or on any bare die.

I've got a D14 I used in a SB 2600K system. The D15 was of course a small improvement over that. The ThermalRight Le Grand Macho RT beats the D15 by another 1 or 2C. For the CPU, I won't hesitate to pay Silicon Lottery to delid and re-lid with LM. With something like the Der8auer De-lid Die-mate, you can do it yourself, but I'd still just as soon pay Silly Lots the 40 clams over my "doing it right the first time".

Also, it took me a while to understand what the young German engineer meant with the "8". I think one is supposed to think (or say) "Der-Acht-auer".

This is the 21st century; I'm a 20th-century rocka-roller, and I think I'm just visiting . . .

 

[Leeea]

Im currently getting about 2Ghz @ 60C when gaming, is it worth me risking frying / damaging my expensive card to go LM, and if so what should i watch out for in particular?

no

The question is, is it possible to replace the GPU if anything goes wrong?