Relid Your Intel CPU on the Cheap with no Delid Tools @ [H]

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
55,602
Relid Your Intel CPU on the Cheap with no Delid Tools - So we have covered a lot about taking your shiny new Intel CPUs apart lately, affectionately known as "delidding." What we have found in our journey is that "relidding" the processor might be an important part of the process as well. But what if you do not have a fancy tool that will help you put Humpty back together again?
 
  • Like
Reactions: Dew
like this
Great walk through. I just have one question. Why did you leave the one spot without sealant? Is that so heat could escape from under the spreader?
 
You want there to be a way for pressure to escape. The die gets very hot and will make the air that's trapped expand, so you need to have an escape for that.
As soon as I read the word "pressure" in your post I was like :banghead: For a minute there I forgot about the relationship between heat and pressure.
 
As soon as I read the word "pressure" in your post I was like :banghead: For a minute there I forgot about the relationship between heat and pressure.
Intel used to put a hole in top of the IHS. Guess they finally got tired of those getting plugged up with crap.
 
As soon as I read the word "pressure" in your post I was like :banghead: For a minute there I forgot about the relationship between heat and pressure.

That camera is so good, and he did such a great job with the angles and lighting I think I felt the pressure on Kyle's thumb! I think I'll try to get my new build finished up first, but I'm giving some serious thought about a de/relid. Thanks for taking us along for the ride. Even if I never attempt it, the videos are excellently done such that anyone should be able to do it.
 
Kyle. Can you delid using a frozen block of bacon?

...CPUs and Bacon. All it needs is an explosion, an american eagle and some alcohol.
 
Kyle. Can you delid using a frozen block of bacon?

...CPUs and Bacon. All it needs is an explosion, an american eagle and some alcohol.
I would but I used all the bacon cooking this weekend...well, full slab of pork belly.
20170219_175236.jpg


That camera is so good, and he did such a great job with the angles and lighting I think I felt the pressure on Kyle's thumb! I think I'll try to get my new build finished up first, but I'm giving some serious thought about a de/relid. Thanks for taking us along for the ride. Even if I never attempt it, the videos are excellently done such that anyone should be able to do it.
That new camera continues to impress me. Sony AX53. Need to bump my 5820K up though to the 6950X sitting on my test bench when I get done with it soon. Grinding out 4K video is resource intensive...
 
It would be a better idea to mark the edge of the IHS BEFORE delidding.
Getting the substrate clean and not cleaning off the edge markings is a bit difficult though. I mention this because that is exactly what I did, at least on the first one.
 
As usual damn fine work sir.

Any chance you got the Scythe FUMA back from Marc?

I'm still itching to know if there is a benefit between a good AIO water cooler and top end air when subjected to a 5.0/5.1GHz Kaby Lake beast.
 
Two words.......but once you do it, you ain't going back in.......JB Weld.

Or maybe Arctic Cooling ceramic epoxy.
 
You've got a point there. Not sure if there's a marker that would hold up tp alcohol. Another way would be to clean all but 2 (preferably opposite) corners, and use them to remark the sides, and then clean those parts. A bit more work, but would make for better alignment. Unless the RTV is provides enough flex for everything to align anyway.
The IHS alignment is not exactly rocket science. If you are covering with it being perfect, a pair of calipers can make that happen easily.
 
I just don't see how the RTV silicone makes any difference in the amount of flex that's going to be in that substrate, RTV doesn't really harden up with it dries - it's fairly pliable and actually quite flexible which is why it's useful for making gaskets and whatnot in the automotive field. I'd definitely be interested in hearing about alternate adhesives that would do the job but still be easier to undo than whatever Intel uses.

I'm also a bit concerned about 'permanently' re-lidding when using certain TIMs like the Coollabs Liquid Ultra. I've heard multiple stories about it drying out within a year and thermal conductivity changing for the worse.

But man I sure LOVE these videos, as a Youtuber myself it's my favorite medium and I have to say you are doing AWESOME at it, Kyle! :)
 
  • Like
Reactions: Jspr
like this
I love that you trigger me so hard when you call a capacitor a transistor. Just what I need to get my blood pressure up in the morning. ;)
 
I'm also a bit concerned about 'permanently' re-lidding when using certain TIMs like the Coollabs Liquid Ultra. I've heard multiple stories about it drying out within a year and thermal conductivity changing for the worse.
This is why I decided not to go liquid metal.
 
As a delidder of several cpus now I suggest sanding the underside of the IHS down about 2mm to compensate for the RTV. From what I've read the real performance gain from delidding comes from the decreased core to IHS gap.
 
I did a delid with some waxed dental floss, no risk of chipping or scratching stuff. Just saw back and forth under the corners and work in to the flats.
 
As a delidder of several cpus now I suggest sanding the underside of the IHS down about 2mm to compensate for the RTV. From what I've read the real performance gain from delidding comes from the decreased core to IHS gap.

you mean 2 mils?
 
As a delidder of several cpus now I suggest sanding the underside of the IHS down about 2mm to compensate for the RTV. From what I've read the real performance gain from delidding comes from the decreased core to IHS gap.
With the Kaby Lake, I would suggest that if you need to compensate for the new sealant, you are not relidding properly.
 
How about silicone sealant?

That's what RTV sealant is. It's a silicone sealant. It's meant to have some kind of elasticity to it. Normally, you're just using it to fill gaps inbetween two objects that are bolted together. Like exhaust pipes, oil pans, etc. Using it to "glue" to items together isn't what it's meant to do.

--------------

There are many adhesive that can be used. Gorilla Glue, Loctite, J-B weld, DAP, etc, but I'd worry about being able to delid the CPU again in the future. It'd be tough to get it open again with those types of adhesives.

I do wonder if a thermal adhesive would be easier to take off.
 
FYI...0.2 mm != 0.2 mils (a mil is 1/1000 of an inch)


Thank you but I do mean .2 mm. If you want to be 100% safe and conservative you probably only want to do 0.1mm worth of sanding. But like Kyle said if you are able to put a really thin layer of RTV and it hold you can still get good results with absolutely no sanding.
 
I managed to almost mess up my 6700k relid attempt. 1st try core 1 was 20 deg higher. Somehow I seated the spreader on cockeye'd so one corner wasn't well seated. Broke it open and used some black ATV sealer I had from nitro RC helicopter work and that seemed to work way better. Re-purposed a re-loading press to help hold the hold thing together during seal. Previously only stable to 4.6ghz and hit 70-80's loaded with an Artic Freezer 240 AIO cooler with only 2 fans(case limitation). Now trying 4.7ghz and temps are mid 50's. :)
 
I simply used some super glue to reattach my heat spreader. Just a little drop in each corner. When glued the super glue takes basically no space so the heat spreader sits much closer to the die then with all that black rubber adhesive it comes with. If I need to remove the lid again I can easily dissolve the super glue with some acetone.
 
I tried a lot of glues from E6000, a few different superglues from loctite and krazy. I found G02 to have best elasticity while give an ok hold. If you want near permanent you have to head to epoxy. I just use JB quick. You need to redo the original delid to get that back off.
 
May I ask a potentially dumb question? What happens if you do not relid and place your cooler directly on to the core?
 
Question--why do you guys also do a Noctua application? Is that just to show how its done with liquid metal and a traditional TIM? Are there disadvantages to using a typical paste? I've read on some places where people say traditional pastes will flow out over time or whatever...
 
Hi, about to do my first delid. Great video, i went and ordered some 18ga needle tips to try. Only using Noctua NT-H1, not overclocking, just want better temps.

Confused on one part, why cant you just use your motherboards CPU socket to relid? Will vapors from the permatex sealant damage something? Am i about to do some unintentional damage?
 
The clamp mechanism slides forward moving the lid as it is clamped.
It needs to be already held fast to the CPU before it is clamped in the socket to get an exact position and no slide marks.
 
Just concerned about proper force and the delicate bits underneath. I mean the socket is perfectly made to hold the CPU. Google-fu wasnt coming up with it as common solution so that was a little red alert.

Gonna try it if only issue is location.
 
Just concerned about proper force and the delicate bits underneath. I mean the socket is perfectly made to hold the CPU. Google-fu wasnt coming up with it as common solution so that was a little red alert.

Gonna try it if only issue is location.
Just use clamp. Unless you physically crush the silicon, which is highly unlikely. The socket mechanism is going to move the IHS when you lock it down.
 
When I did my 3300K's and 4790K's, the easiest way to "re" lid was apply the RTV and stick the CPU back into the motherboard socket and lock the Socket clamps. It would automatically center the lid properly onto the substrate and provide a LOT of even pressure on the lid from the socket hold downs. Those hundreds of pins pushing up against the bottom of the CPU apply a LOT of clamping force and since the socket clamp keys to the lid, I didn't have to much with centering the lid on perfectly.
 
When I did my 3300K's and 4790K's, the easiest way to "re" lid was apply the RTV and stick the CPU back into the motherboard socket and lock the Socket clamps. It would automatically center the lid properly onto the substrate and provide a LOT of even pressure on the lid from the socket hold downs. Those hundreds of pins pushing up against the bottom of the CPU apply a LOT of clamping force and since the socket clamp keys to the lid, I didn't have to much with centering the lid on perfectly.

This partially defies the point of it having a lid.
Without the lid the CPU will bow in the socket causing more pressure on the pins in the centre and less at the edges and will develop a permanent flex of the CPU substrate and possibly squash some pins in the socket.
With the lid already solidly applied the load is spread much further out from the centre so it is much more even over all the pins.

When a CPU is delidded it will float on top of the cpu core and leave a gap all round. the lid does not touch the substrate.
If you clamp the CPU in place to seal the lid on, the CPU is already bent out of shape before the sealant sets.
The lid does not provide relief as intended.
 
I'm wondering how that extremely tacky urethane adhesive they use for automotive window placement will work. Having worked with it quite a bit, I feel like it's pretty similar to the stuff that is already used on the chip when it is sealed from the factory. I may try some on my 6600k, I've been running it without a relid sealant for a few months, and have wanted to find a good solution anyways.
 
Back
Top