Delidded 4790k temp drop then spike in temperature? Thermal Grizzly C-series Liquid Metal + MX2

[Bluescreendeath]

So I delidded my 4790k and applied a tiny amount of liquid metal (Thermal Grizzly C-series). I also reattached my lapped IHS with black silicone adhesive (with some gaps for air) and applied a layer of clear nail polish to the CPU circuits near the die so the liquid metal won't short it out on accidental contact. For the heatsink + cpu IHS, I used a small line + 4 dots in the corners of MX-2 thermal paste.

When I tested my cpu, initially everything was ok. 35'C idle, 65'C max Prime 95 Small FFT. Then an hour later, idle temperatures are around 45'C, and Prime 95 SFFT now maxes out at 88'C. This is not normal I presume?

Did I do something wrong? (eg. apply too little or too much liquid metal?, used too much pressure on the IHS and die?)

Here are pics of the liquid metal application:

http://imgur.com/a/1CakMlB

 

[PerfectWave]

Conductonaut thermal grease must not be used with aluminum heatsinks!

 

[Spaceman Spiff]

He said he used MX-2 between the ihs and heatsink...

I am not sure what would cause such a rise in temps after some time like that. I have heard this happen before with conductonaut but I have only ever used coolab liquid pro. Maybe reseat your cooler and reapply the MX-2 again and see if temps go back down and it happens again. Then you'd know that it was the MX-2 and not the conductonaut giving you issues. If no change, re-delid, check your conductonaut, clean it, reapply, and try again.

I would keep testing to see where the issue is. Once you find it, maybe switch thermal compounds and see it changes your results. Keep a good log of what temps you get with what compounds at what time interval after delid etc.

 

[Mussels]

the silicone you used probably expanded and made the IHS shift


Put it back into the socket and let the clamp hold it down, or use a small drop of superglue on a corner to hold it in

 

[Bluescreendeath]

the silicone you used probably expanded and made the IHS shift
Put it back into the socket and let the clamp hold it down, or use a small drop of superglue on a corner to hold it in

I put the ihs on the die and had the cpu clamped by the bracket on the motherboard as soon as I applied the liquid and silicone. I then mounted the heatsink - all within 30min. Is it possible for it to expand even with this pressure? How much does silicone expand?

Does my liquid metal application look incorrect/insufficient?

 

[Sasqui]

I put the ihs on the die and had the cpu clamped by the bracket on the motherboard as soon as I applied the liquid and silicone. I then mounted the heatsink - all within 30min. Is it possible for it to expand even with this pressure? How much does silicone expand?

Does my liquid metal application look incorrect/insufficient?

If the silicone hadn't cured, uneven pressure might have tilted the IHS slightly, that's my guess and probably what @Mussels is thinking too.

I did a 3570k both with IHS and naked. Never had a problem with the IHS (I never put silicone back on it, btw), but had similar problems to you with running it naked. Took some heatsink pressure tweaking.

Oh and I don't think silicon expands, but you'd have to look at the specs for it.

 

[Bluescreendeath]

If the silicone hadn't cured, uneven pressure might have tilted the IHS slightly, that's my guess and probably what @Mussels is thinking too.

I did a 3570k both with IHS and naked. Never had a problem with the IHS (I never put silicone back on it, btw), but had similar problems to you with running it naked. Took some heatsink pressure tweaking.

Oh and I don't think silicon expands, but you'd have to look at the specs for it.

What do you mean you had the 3570k naked? No IHS/heatsink directly on die?

Did you solve the problem with more pressure or alleviating pressure?

Since I have to take everything apart again, do I need to reapply the liquid metal or can I just leave on what I currently have?

 

[Sasqui]

What do you mean you had the 3570k naked? No IHS/heatsink directly on die?

Did you solve the problem with more pressure or alleviating pressure?

Since I have to take everything apart again, do I need to reapply the liquid metal or can I just leave on what I currently have?

Yes, I had to crank down pretty hard and try different pressures on the waterblock screws to get it directly on the die. Sometimes it wouldn't boot, other times it would be very unstable. When I got it set, it was stable. Ironically, the temps between naked and IHS were negligible. I ended up putting the IHS back on several months later when I was doing some maintenance.

I have never worked with liquid metal, so I don't know what the consistency is and I don't know how forgiving it is. I use AS5 paste, it flows into gaps nicely and stays there.

 

[Spaceman Spiff]

Ah yeah. I don't reseal my ihs either. I just use the cpu clamp to keep the ihs in place.

 

[jboydgolfer]

i had great results with my 4790. I used a spot of gasket sealer on each of the 4 corners, "painted" the die with grizzly conductonaut, and that was it. Im aware of people putting polish on , but never did myself when delidding. Just need to redo if its not working good for you. iirc, i had one hotter core for a day or two after i delidded, but its cold as ice now.

Since I have to take everything apart again, do I need to reapply the liquid metal or can I just leave on what I currently have?

Re do all.

 

[Bluescreendeath]

i had great results with my 4790. I used a spot of gasket sealer on each of the 4 corners, "painted" the die with grizzly conductonaut, and that was it. Im aware of people putting polish on , but never did myself when delidding. Just need to redo if its not working good for you. iirc, i had one hotter core for a day or two after i delidded, but its cold as ice now.
Re do all.

Gotcha. I think the mistake I did was I put the sealer on the bottom of the IHS corners. I'm not supposed to put the adhesive between the IHS and the CPU substrate right? I'm supposed to put on the IHS first without sealant, and then dab the edge where the IHS/CPU meets with the adhesive?

 

[Sasqui]

Gotcha. I think the mistake I did was I put the sealer on the bottom of the IHS corners. I'm not supposed to put the adhesive between the IHS and the CPU substrate right? I'm supposed to put on the IHS first without sealant, and then dab the edge where the IHS/CPU meets with the adhesive?

Either method should work, just use it sparingly. I never put sealer on mine, just put the chip in, IHS on top and let the socket clamp hold it in place before installing the heatsink. Did that several times and never had an issue. I did seal the chip when I sold it, I think I just dabbed the corners with white silicone sealer and tested it after curing a few days.

 

[jboydgolfer]

i will toss a couple pics i took when i did my 8600k, & my 4790. Ive used super glue after putting the IHS back on, and just put a dot on each corner. Ive also done spots of Gasket sealer in a few spots (under IHS like in pic below) as long as your not putting too much, it wont interfere with seal.


i like to put LM on both Die & IHS. (if you do this be careful your familiar with how much LM to use)

this is how i usually apply Gasket Sealant. Or i just put it under the four corners.

note the scotch tape, i use it to make LM application SUPER easy, as you cant get it anywhere but the Die.

i use a toothpick or similar small tool to apply.

you can see here how badly my 4790 needed delidding

 

[Bluescreendeath]

Thanks. Do you recommend putting a decent amount of liquid metal on both the IHS and die? I put it mostly on the die, and then rubbed smidgens of leftover stuff on the IHS. I rubbed nitrocellulose nail polish over the circuits next to the die to protect it, so I don't think short circuiting from residual LM is too much of an issue.

So both putting the silicone sealant underneath the IHS and putting it on the corners/edges of the IHS (not underneath it) both work? Did you notice the difference between the two?

I'm wondering if I messed up by using a bit too much underneath it, so maybe the edge/corners method would work better for me.

 

[jboydgolfer]

I'm wondering if I messed up by using a bit too much underneath it, so maybe the edge/corners method would work better for me.

Remember you're not trying to permanently fasten this IHS down ,or anything serious, don't try to do what the factory did ,just put enough to hold the IHS down and that's it. You could literally apply it with a paintbrush if you were able to spread it with one ,and then vice clamp it down. In the picture I posted earlier, what you see on the end of that toothpick ,is basically what you need for the entire job.

Also make sure your clamping it down with a vice for an hour or two while the sealant sets up , or longer. As long as you're applying enough pressure while the sealant cures, it won't matter what you put on ,because any excess is going to push out on the inside and on the outside, so you're going to want to put as little as possible while having enough toform a bond.

If you're not comfortable with this method, I would suggest if you don't have success using sealant, to just put the IHS back on the chip ,and put a couple drops of super glue on 2 of the corners.

remember all you need from whatever bonding agent you use ,is enough to hold the IHS on while you install it into the motherboard ,because you're going to clamp a heat sink or a cooler on top of it .

Do you recommend putting a decent amount of liquid metal on both the IHS and die?

i did it because i had enough out of the syringe to apply it to both, but you dont want too much, so be careful.

Edit:

I just noticed that you posted you lapped the ihs, I wouldn't of done that personally because there's no need (unless you have pitted surface, corrosion, etc). Did you lapp the top or the bottom of the IHS? Please post a clear & detailed pic. Unless you've gotten low temperatures after the lapping process, it could be the cause of higher/fluctuating temperatures.