Originally Posted by TLM-610
So the main problem is a blade that digs into the PCB and not so the method itself right?
Yes. The problem isn't with the blade itself (unless it's not perfectly flat for some reason) the problem is in the direction of force. As long as you are really careful to always keep the blade flat or slightly pushed away from the PCB (toward the IHS) while cutting the glue holding the two together - you shouldn't gouge the PCB and should have a successful delid.
Originally Posted by Aparition
You need something thin enough and strong enough to cut through the seal and fit between the pcb and ihs. A razor is the easiest item to find and use.
This. Some people have succeeded using a thin filament wire or fishing line as well. In my case the IHS was so well pressed and glued that I couldn't even get a scalpel between the IHS and the PCB... so the vise method was about my only option. Much depends on the very slight variances in manufacturing - unfortunately my IHS was very well applied... the TIM in between it and the die was NOT however... so I had more than 18C difference between core temps under full load. Now I have 3C difference and the average core temp is 13C lower than before.
Originally Posted by Spritanium
Why are there two different threads if it's literally the exact same process?
Same reason there are different thread for the delidding process itself? Not sure but since there's a dedicated thread for every case, GPU, etc... on OCN - it kind of makes sense there would be a separate one for Haswell delidding.