Overclock.net banner

[SOLVED]12900K delid issue

2268 21
###SOLVED###

Folks with delid experience,

I recently popped the top on a 12900K for liquid metal treatment. Heated the IHS with hot air station before using Rockit Cool tool for 12 gen. Followed their process via youtube video. No liquid metal is present on the PCB except for the die and IHS. All seemed to go well until I went to boot. The PC failed POST. I found the system runs normally with a single stick of RAM. RAM runs fine at XMP in this system --just not dual channel. RAM works fine in 2nd system and worked with this CPU previously. Different RAM kit has the same result. CPU handles gaming and bench normally for a missing channel. Any thoughts?

MSI Z690-A Pro DDR4
G.SKILL 2x8GB kit
DeepCool AK620
EVGA 850W GA
WD SN850 1TB
EVGA 3080ti

~Pistol

Attachments

1 - 20 of 22 Posts

·
Facepalm
Joined
·
11,386 Posts
Folks with delid experience,

I recently popped the top on a 12900K for liquid metal treatment. Heated the IHS with hot air station before using Rockit Cool tool for 12 gen. Followed their process via youtube video. No liquid metal is present on the PCB except for the die and IHS. All seemed to go well until I went to boot. The PC failed POST. I found the system runs normally with a single stick of RAM. RAM runs fine at XMP in this system --just not dual channel. RAM works fine in 2nd system and worked with this CPU previously. Different RAM kit has the same result. CPU handles gaming and bench normally for a missing channel. Any thoughts?

MSI Z690-A Pro DDR4
G.SKILL 2x8GB kit
DeepCool AK620
EVGA 850W GA
WD SN850 1TB
EVGA 3080ti

~Pistol
Are you using a rockitcool or bitspower IHS replacement or the original?

Please remove the CPU and check the socket. I know you don't want to do that, but please don't complain about it and just do it.

You either bent pins (good luck unbending them, you need special fine tools and some sort of magnifying glass to fix lightly bent pins, otherwise you are buying a new motherboard), or you have insufficient socket pressure downwards in the socket. Did you re-lid the chip with the original IHS? Relidding the chip should not cause problems like this.

On some previous chips (I know specifically the i9 9900K), you had to sand down the underside edges of the original IHS about 0.2mm if you intended to relid the chip (sanding it down until you saw clear copper along the entire edge was good enough for a 0.2mm measurement if you didn't have tools to get this precise), because the removal of the indium solder would prevent middle section of the IHS from making proper contact with the CPU die, thus causing bad temps/bad core deltas).

Check the back of the CPU. Did you rip off any caps or smd's from either the front or back side of the CPU? If you aren't sure there are many photos available of stock CPU's and delidded front ones so you can compare them.

If your socket pins are pristine with no bent pins, PLEASE highly consider buying a thermalright anti-bend bracket. This will work with delidded chips that have been resealed again and give you proper contact pressure as well. AFAIK, neither the TR bracket nor the der8auer grizzly brackets will work with the rockitcool copper IHS.
 

·
Registered
Joined
·
26 Posts
Discussion Starter · #3 · (Edited)
I inspected the board for bent pins under microscope. I could not see any damaged pins. Caps look fine. I used small amount of conformal coating to protect the smallest on top. Bottom side is factory condition.

I am currently using Intel IHS sitting ontop of chip (not glued in place) without a contact frame.

It makes sense that perhaps there isn't enough pressure due to adhesive between IHS and cpu PCB is no longer present. I removed the black adhesive from IHS and PCB.
 

·
Registered
Joined
·
648 Posts
The same thing happened to me what cooler you are using cooling because I was using the heatkiller iv for socket 1200 and it seems that it was making a lot of force, I changed to an lga1700 waterblock and problem solved
 

·
Registered
Joined
·
814 Posts
You have got a scratch in the PCB or substrate ? picture of surface of the CPU after cleaning and application of LM ?

right lower edge on the surface of your cpu looks a bit as if there is a scratch.
 

·
Registered
Joined
·
26 Posts
Discussion Starter · #7 ·
It is possible to be a scratch on the top. It is difficult to discern if it's adhesive material residue or scratches even under 180x magnification. I put down a bit of UV mask in case. I can not tell if there are damaged traces. I assume the test points outside edge of cpu may verify this?
 

·
Registered
Joined
·
5,233 Posts
This is usually mounting pressure related. Either it isn't pressing straight down or is too strong. I had this several times with a 10th gen direct die with a EK block. The EK springs for example are quite strong and without a IHS the pressure on the die is a LOT if you crank it all the way down. I always had to leave it 2 full turns loose for all the memory and PCI-E slots to work.
 
  • Rep+
Reactions: pistol3d

·
Registered
Joined
·
650 Posts
Funny sidenote I have the same MOBO new would not detect gpu 1st slot or my other channel ram. I im going to pull mine out and check something for you with the ILM
Edit: IMO might be worth removing the black spacer from the ILM and remount everything might help? make sure not to over tighten
 

Attachments

·
Registered
Joined
·
26 Posts
Discussion Starter · #20 · (Edited)
Check this out. Found contamination on cpu pads. This is a hardened substance. I'm assuming it's cyanoacrylate glue from the re-lid effort.

Update: that was the issue! It's working with both sticks now. Never use super glue to lid an IHS. I don't care if the de-lid companies recommend it.

I used the smallest bit of nail polish remover on a cotton swab and very gently scraped it off with super fine precision tweezers under the microscope. Works like a charm!!!
 

Attachments

1 - 20 of 22 Posts
Top