Thermalgrizzly conductonaut LM install on 1080ti: what did I do wrong?

[magellan]

I have a MSI 1080Ti Duke that I tried using thermalgrizzly conductonaut liquid metal on. I carefully cleaned the GPU die and the cold plate several times with carb cleaner until the cotton tipped swabs came back clean. I then applied two coats of clear nail polish to the SMD's surrounding the GPU die and made sure none of it got on the die or the surrounding shim.

I applied the conductonaut to the GPU surface until it was coated with a fine coat then did the same thing with where I thought the die contacted the cold plate. After all this I re-assembled the heatsink w/the PCB and slowly tightened down the spring-loaded screws on back in a criss-cross pattern.

When I fired up the card it runs much hotter than it did with Phanteks' ph-ndc TIM that I had on before. Did I not put on enough conductonaut?

 

[EarthDog]

Clearly something is up and a reapply necessary. Exactly what the reason is, no idea. A picture could help, but regardless you need to try again.

That said, I dont see much of a point to use that as the couple/few degrees C difference (from review sites) isn't going to make a big difference in clocks/performance so I wouldn't sweat it too hard.

 

[dfonda]

Carb cleaner? Do you mean Brake cleaner (Mostly alcohol). Depending on the brand Carb cleaner can leave a residue....I would not choose that for PC parts...Especially if the area is not well ventilated. Substitute Alcohol or at least Brake cleaner....

That probably isn't the issue here but just check on what is in your particular brand.

and how many Grizzlies had to die to make a bottle of Thermal Grizzly...

 

[EarthDog]

Great point. I thought that leaves a petroleum based reidue...but really haven't used it in over a decade.

We tell people to use isopropyl rubbing alcohol (highest % you can find) to clean off the IHS/die/heatsink base.

Maybe try that..

Edit: PS - only metallic grizzly bears were harmed in the making...

 

[Janus67]

Agreed with the above.

I only apply new paste if the stock paste isn't doing well or I had to disassemble for some reason. I wouldn't replace known good/third party paste unless it was actually causing an issue.

As was stated there's enough variables with application of paste that there's guaranteed variance every time you do it.

 

[magellan]

Carb cleaner? Do you mean Brake cleaner (Mostly alcohol). Depending on the brand Carb cleaner can leave a residue....I would not choose that for PC parts...Especially if the area is not well ventilated. Substitute Alcohol or at least Brake cleaner....

That probably isn't the issue here but just check on what is in your particular brand.

and how many Grizzlies had to die to make a bottle of Thermal Grizzly...

The chemical components of carb cleaner are all gases at STP. I'll bet some of them are carcinogens too. Maybe that's a good reason not to use it at all.

 

[magellan]

Well it looks like the problem was how I had strapped one of the 127x38mm fan and shroud assys. to the heatsink. The pressure was uneven preventing the cold plate from applying even pressure to the GPU die (i.e. it was canted to one side).

 

[dfonda]

Well it looks like the problem was how I had strapped one of the 127x38mm fan and shroud assys. to the heatsink. The pressure was uneven preventing the cold plate from applying even pressure to the GPU die (i.e. it was canted to one side).

Great Glad you got her fixed Mag's

 

[magellan]

The card still doesn't run as cool as it did with the Phantek's ph-ndc TIM so I'm going to have to yank it and disassemble it again. I'll take pics and see if anyone can figure it out.

Does liquid metal require the cold plate and the GPU fit together better than thermal paste to work well?

 

[EarthDog]

Does liquid metal require the cold plate and the GPU fit together better than thermal paste to work well?

Not entirely sure what you mean, but still going with no.

 

[dfonda]

Not entirely sure what you mean, but still going with no.

Meaning does the paste fill in better than the LM...so the two parts fit tighter...but yes....no...

Just wrench those suckers together.

 

[bassnut]

Great point. I thought that leaves a petroleum based reidue...but really haven't used it in over a decade.

We tell people to use isopropyl rubbing alcohol (highest % you can find) to clean off the IHS/die/heatsink base.

Maybe try that..

Edit: PS - only metallic grizzly bears were harmed in the making...

I used carb cleaner to flush crap out of my shotgun triggers for years, works great that being said some do leave a residue. I used to spray a bit in my hand the good stuff just evaporated the crapy stuff left an oily film.

I would avoid it completely for electronics cleaning it's quite harsh.

 

[magellan]

I used carb cleaner to flush crap out of my shotgun triggers for years, works great that being said some do leave a residue. I used to spray a bit in my hand the good stuff just evaporated the crapy stuff left an oily film.

I would avoid it completely for electronics cleaning it's quite harsh.

The carb cleaner I use consists entirely of VOC's, which are gases at STP. It works excellently to clean off TIM, much better than isopropyl-alcohol. Brake cleaner might work well too.
AFAICT there's no evidence of any residue and I don't like the idea of getting VOC's on my skin as I believe some of them are carcinogenic.

 

[EarthDog]

AFAICT there's no evidence of any residue and I don't like the idea of getting VOC's on my skin as I believe some of them are carcinogenic.

Weird sentence... feels like there should be a 'but' instead of 'and', lol.

Knowing that, why use it when there's a perfectly viable alternative that doesn't have that issue? We get that it works, but....



So..... any pics? Updated results?

 

[magellan]

I just yanked the 1080ti, I haven't had time to take it apart yet and take pics.

 

[magellan]

It looks like I didn't paint the proper part of the cold plate with the conductonaut, which resulted in a thin spot on the GPU:

 

[EarthDog]

It looks like I didn't paint the proper part of the cold plate with the conductonaut, which resulted in a thin spot on the GPU:
View attachment 214428

Glad you found the problem!

......sooooooooooo, are the results better???

 

[magellan]

I need to do more testing, particularly WRT to the hotspot temps which were really bad previously.

 

[magellan]

The temps are not looking significantly better than Phanteks PH-NDC TIM, it's looking like only a couple of celsius degrees (or is that centigrade). Does liquid metal between a bare GPU die and a cold plate work better with a higher clamping pressure than with paste TIM's?

 

[EarthDog]

The temps are not looking significantly better than Phanteks PH-NDC TIM, it's looking like only a couple of celsius degrees

See post 2...

Makes sense considering most empirical testing/reviews on the web only puts it a couple/few C better than most pastes anyway.

(or is that centigrade)

Celsius is the modern and more commonly used term. They are one in the same.

Does liquid metal between a bare GPU die and a cold plate work better with a higher clamping pressure than with paste TIM's?

I haven't seen any tests. Typically testing is done with 'stock' pressure anyway. That said, no way I'd mess with more pressure on a bare die......especially a GPU who's replacement cost is sky high these days.