Overclock.net banner

I successfully delidded 7900X and killed its conductors with Thermal Grizzly Conductonaut. What else could've caused this, besides Conductonaut?

16K views 59 replies 16 participants last post by  TheReciever  
#1 ·
Hey there guys

I successfully delidded my 7900X and when I applied [too much, this one's on me] of TGC, it dripped away right into conductors. That wouldn't be that bad, CPU is really safe towards shortcircuit. What is a problem, is that Thermal Grizzly Conductonaut most likely reacted with solder tin, which led to conductor[s now] falloff.

I used only ArctiClean set for cleaning the PCB of the CPU + beznin alcohol [1% benzinum and 96% ethanol].

My question goes as follows. What could've caused this disaster?

a) Thermal Grizzly Condonaut dripped for 24 hrs on conductors

b) ArctiClean set [1+2]

c) benzin alcohol [1% benzinum and 96% ethanol]

Please advise

Thank You for Your eventual reply

best regards

uplink
 
#3 ·
Quote:
Originally Posted by GreedyMuffin View Post

Tip: Use nail-polish on the conductors.
Thank You for Your tip. Wanted to, but I'm afraid that nail polish will burst in flames eventually. I bought silical color, that's used for sealing of chimneys and fireplaces, but I'm still hesitant about it's conductivity. It should be non-conducting. Won't some random nail polish burst in flames eventually?

And also, any clue whether some of the above have caused this effect? Of desoldering the tin [and thus conductors?]
 
#8 ·
Quote:
Originally Posted by GreedyMuffin View Post

I used nail polish on my 7800X without any issues.
Hi, did you buy a delidding tool? - How difficult did you find it? - I've seen some videos of DerBauer doing this but I'd need to see a whole load more of each step of the process to minimise any major problems.

Did you practice on any cheaper chips or just dive straight in?
 
#9 ·
I bought the Rockit 99.

I've never delidded, relidded or used LM before. So everything was new to me.

I watcher gamersnexus's videos. So I knew everything I needed to know before I started. Tok 20-25 mintues from stsrt to finish. I let the loctite cure for 1 1/2 hours as I relidded before I installed it in my machine. Can post pictures tomorrow? Going to bed now (00:46 here..)

I was not even nervous! It was rather fun instead!

Was very easy tbh. Just you your time.
 
#11 ·
Quote:
Originally Posted by TheReciever View Post

This is why we (on laptops) use Super 33+ scotch tape to prevent spillover unto the CPU

Hopefully others will be able to help repair the CPU.
Hmm, I'll google that one out for sure, never even heard of it :(

Quote:
Originally Posted by fra z View Post

Hi, did you buy a delidding tool? - How difficult did you find it? - I've seen some videos of DerBauer doing this but I'd need to see a whole load more of each step of the process to minimise any major problems.

Did you practice on any cheaper chips or just dive straight in?
I used the Rockit 99. It's much better compared to unexisting Der8auer's tool, which I had preordered. iHS will fly off within Der8auers', whilst in Rockit it will stay put. Der8auer's is heavy duty [100+ units], Rockit is lower duty [up to 50 units I reckon, till it breaks].

Quote:
Originally Posted by GreedyMuffin View Post

I bought the Rockit 99.

I've never delidded, relidded or used LM before. So everything was new to me.

I watcher gamersnexus's videos. So I knew everything I needed to know before I started. Tok 20-25 mintues from stsrt to finish. I let the loctite cure for 1 1/2 hours as I relidded before I installed it in my machine. Can post pictures tomorrow? Going to bed now (00:46 here..)

I was not even nervous! It was rather fun instead!

Was very easy tbh. Just you your time.
Yup :), exactly.

Quote:
Originally Posted by fra z View Post

Hello Greedy Muffin,

Found the website for Rockit 88 - but couldn't see 99 but now I've looked again I can see.

This is the 2nd company I've seen doing this - That's cool if it's done safely
Quote:
Originally Posted by Snyderman34 View Post

What about liquid electrical tape? Maybe that would be an option?
I can't seem to find that in my country. I will try to order it from ebay/amazon. Thank You for Your input! :)

Quote:
Originally Posted by GreedyMuffin View Post

I used nail polish on my 7800X without any issues.
Thank You! I'll try it than as well! :)

Quote:
Originally Posted by fra z View Post

Arrmannn -
puke.gif
- Are you sure it's knackered? - I'm just looking into delidding and reading this makes me shudder ! -
When You avoid liquid metal, from any company [TG, or Coolaboratory], I'm 99% positive You'll succeed.
 
#12 ·
Quote:
Originally Posted by TheReciever View Post

This is why we (on laptops) use Super 33+ scotch tape to prevent spillover unto the CPU

Hopefully others will be able to help repair the CPU.
I cannot stress this enough.
Laptop users tend to use Super 33+ or Kapton tape to cover and insulate those conductors or whatever they're called.

https://www.amazon.com/3M-194NA-200-Inch-Super-Electrical/dp/B002YDZA1K/ref=sr_1_9?ie=UTF8&qid=1507487624&sr=8-9&keywords=Super+33%2B

https://www.amazon.com/BCP-Kapton-Polyimide-Silicone-Adhesive/dp/B00W3FAFP0/ref=sr_1_3?ie=UTF8&qid=1507487741&sr=8-3&keywords=Kapton+tape

These are commonly used to insulate the conductors around the silicon.

Some laptop users also use nail polish successfully, sometimes combined with a high compressible cutout foam dam (not inside delidded CPU's but OUTSIDE) to prevent LM from dropping over onto the mainboard if a drop ever escapes (only when LM is used for the copper or nickel plated heatsink itself (NOT aluminum--ever)). I used transparent nail polish myself on both my laptop GTX 1070, which I put Conductonaut on and TDP modded to 185W, and on the CPU. It will NOT burst in flames next to silicon. Some people also said that using nail polish gives lower temps than using Super33 or kapton tape, but I'm not going to get into that. It is a bit logical since you have less material in the way, and thus less things that can trap unnecessary heat.

However the most important safety measure you can do is to NOT apply too much LM. If you apply too much, all the protection in the world isn't going to stop a disaster. It's very, very important to apply a very tiny drop (just ONE drop will do! You won't believe how much a tiny pinhead sized drop actually spreads!), and then spread it with a lint free swab spreader like the black ones included in the package. How much a tiny little dot spreads is very mind blowing and disbelieving. If you had a similar sized dot of Kryonaut or MX4 or Nanogrease Extreme, it wouldn't hardly go anywhere whatsoever. But for a bare silicon die like bare CPU or Pascal cards, one tiny dot is enough to spread all around the silicon. You just start spreading it and it just...spreads and spread and spreads.

Your goal is to have as thin of a layer as possible while still having it cover the entire surface. You don't want it so thin that it's basically a trace layer (That's too thin), but you want it thinner than what regular thermal compound would be like. The poster iunlock over on notebookreview has excellent pictures of how much LM should be used.

From there, any little 'extra 'can then be used to "tin" the surface of either the IHS underside, or if used as TIM from chip to heatsink, to tin over the heatsink. This is because Gallium will react slightly with copper, and tinning it will cause that faint coat to react first, rather than what's on the chip, which can give you a few more C drop, and more assurance of longevity.
Nail polish on my GTX 1070N conductors worked just fine and protected them well.
 
#13 ·
If you check out this video Der8auer https://youtu.be/HUn5et2sqz0 it explains a little, and kind of gives some information about this. Also I think a lot of this depends on the material used in/with the solder that might cause the reaction and not just directly the use of it.

Also Gamernexus did a little explanation of using Nail Polish https://www.gamersnexus.net/guides/3068-how-to-delid-intel-i9-cpu-and-apply-liquid-metal
Quote:
"Nail polish typically has nitrocellulose or another functionalized cellulose in it which forms a thin, stable film without a gap over items such as fingernails, or the capacitors you used it over. An alkyl acetate helps move it into place, and then promptly evaporates leaving behind just the thin film.

"Some nail polish formulations have benzene derivates. Toluene is a prime candidate used by some companies, which does a similar job as nitrocellulose. Look at the composition before buying or using these, as toluene isn't nice to play with even at lower concentrations."
This sucks man you are being a champ about it, I hate to hear it buddy about all of this.
 
#15 ·
Quote:
Originally Posted by TheReciever View Post

Had not known of the nail polish trick

Thanks for sharing!
No problem, from what I have understand clear is preferred I am not sure why if the color adds something a person doesn't want in the nail polish but I have seen clear recommended mostly. Anyways welcome hope this can save others to please use something like this to help. For the cost of nail polish it is no more than $2-8USD the cost is def worth it trying to save a CPU. Anyways lol
 
#16 ·
It's nice because you can insulate while using less of it, and it can result in lower temps. However it's very highly recommended if doing this on a laptop, that you use some type of cutout (highly compressible) foam barrier around the CPU. The foam barrier is not to protect the conductors (that's what the nail polish is for). It's to stop conductive balls of doom from escaping the silicon area and shorting out the GPU or mainboard PCB. Free insurance is good insurance.

For LM, less is more. You don't want to use so little that there is insufficient contact, but you don't want to use so much that there is any runoff. With normal paste, using a bit too much is better than using not enough, since with sufficient pressure, the excess simply will get pushed out (but again using the right amount gives the best results, but on a laptop, with a low pressure, possibly poorly fitting heatsink, using a little too much is better). For LM, it's best not to use it on any badly fitting heatsink, but for a good fitting heatsink, a light, even coat is best. You see a lot of videos where people used a drop similar to the same size drop they use for traditional paste. That's too much. Always start with a VERY small drop and spread it. You'll know after awhile if you need a little bit more or not. Start small. Then if you find that you have a little too much excess, use that excess to tin over the heatsink or IHS.
 
#18 ·
Thank you for this video. Godly video on what is too much and proper/too little.
 
#19 ·
Hi,

Someone suggested not using liquid metal compound - Is this correct? -

Eventually I'll delid a CPU, hopefully successfully - Time for a semi-stupid question, when the CPU is delidded, after any *work has been perfomed on the CPU, how is it absolutely possible to know the lid (IHS) is going back on in the correct orientation? - EG - Not the wrong way round? - Is the IHS "universal?" meaning it can be put back on any way around?

After looking at a picture, if the arrow is in the bottom left corner it should be possible to read the writing on the back of the IHS

Is there an arrow on the IHS as well as the CPU? - This query came to mind thinking about mis-understandings or what could go wrong potentially even after successful *work on the CPU - If any of us are going to do this we need to be 100% on every step -
smile.gif
 
#21 ·
Also does anyone know if the Rockit 99 delid tool will work on X99 eg - Haswell-E / Broadwell-E? - Rockit 99 is advertised as LGA 2066

Someone else mentioned Haswell-E thermals are good so why would you want to de-lid a 5820K and they are soldered and any de-lidding would damage the CPU. Any comments?
 
#22 ·
Quote:
Originally Posted by TheReciever View Post

If you dont know what your doing and dont want to learn then I would suggest against it.

Otherwise full speed ahead.
It's a matter of getting everything organized so any of us who do wish to do this know which steps to take. As far as liquid metals are concerned it's in the realm of chemistry and the periodic table and knowing which elements are in play in any of the compounds used on the CPU to improve thermals.

One expert mentioned using liquid metal whilst someone else's opinion suggested not using liquid metal compound may have resulted in the 7900X in question not getting damaged. This could be user error / bad luck.

Personally I do wish to de-lid at some point but do need to know all the advantages and dis-advantages of all the options available -
smile.gif


A slow calculated approach to this is the best way beginning with obtaining a de-lidding tool -
smile.gif
then obtaining other items needed after working through the merits of the other options available.

- How about yourself - Have you de-lidded yet? - A little or a lot maybe?
 
#25 ·
Quote:
Originally Posted by fra z View Post

Also does anyone know if the Rockit 99 delid tool will work on X99 eg - Haswell-E / Broadwell-E? - Rockit 99 is advertised as LGA 2066

Someone else mentioned Haswell-E thermals are good so why would you want to de-lid a 5820K and they are soldered and any de-lidding would damage the CPU. Any comments?
The only reason to De-lid a cpu is to improve thermal conductivity between the CPU die and the IHS or direct-heatsink on a cpu which is not already soldered to the IHS (uses TIM instead)
AFAIK, all Intel high end EP/Server SKUs are still soldered, rather than using cheap TIM, so no reason to delid and most likely you would destroy the CPU if you tried.
 
#26 ·
Quote:
Originally Posted by Nemesis158 View Post

The only reason to De-lid a cpu is to improve thermal conductivity between the CPU die and the IHS or direct-heatsink on a cpu which is not already soldered to the IHS (uses TIM instead)
AFAIK, all Intel high end EP/Server SKUs are still soldered, rather than using cheap TIM, so no reason to delid and most likely you would destroy the CPU if you tried.
Hi Nemesis158,

I was watching a Derbauer video and unless I'm mistaken he did delid a Haswell-E processor, that is unless I'm mistaken. It's very easy to get mis-understandings here and I don't pretend to be an expert -
thumb.gif


I saw another video on Youtube where a guy de-lids his 7700K and shows all the bits and pieces he uses. It's a little quick flowing for the un-initiated like me but the pause botton and rewind are good -
smile.gif
- Anyway he showed minus 20 C on his 7700K so instead of gaming at around 60+ his temps were around 40+

And yes thanks for the input, regarding the contact of the IHS to the die and the efficiency of the thermal conductivity which is what we are all after

I suppose it's a question of watching this space - I'm gathering all the bits that are needed then when I do try this it should go as well as possible - I hope -
smile.gif