Originally Posted by Justintoxicated
Yea I delidded and I am disappointed in the results as well, I only dropped 30c.. I was hoping delidding would keep my CPU at ambient or below.
BTW I believe CLP is better when applied to the core and CLU better for the top of the heatspreader which is less flat. People have mixed results with all these liquid metal TMs, but from what I gather CLU is better for uneven surfaces where more is used, and CLP better where you can get by with super flat surfaces and super thin layer is used. Some people say CLU some CLP and some even say Phobya Liquid metal is best. I used Phobya TIM because it was less than 1/2 as expensive $6.99 for a syringe, It was fun to spread and I was able to get a reflective spread super thin. My netted drop is over 30c. I haven't sceen anything higher than 58c (hottest peaked 63c but its dropped a few more degrees after setting up for some reason) where I was at 92c before with similar ambients and .01 less v core (give or take a few degrees). I'm sitting at 1.275c which I was afraid to stress test for 24 hours before delidding due to hitting 92c @ 1.265v.
I can't beleive the improvement but there is no disputing it. I verified my spread was about perfect when I pulled the waterblock the first time, so it's not like I had bad coverage either.
That's a hell lot of temperature drop, grats mate:) I am waitiing for my Indigo Xtreme before I delid it. Well Xmas time, I need wait bit more
Originally Posted by bilditup1
I'm having a bit of trouble with my 4770k delid/h100i solution myself.
I used the vise method to delid but unfortunately didn't use a big enough block of wood. As a result (I'm guessing???) it took many more smacks to remove the IHS, and it got pretty banged up. Went ahead and put CLU on the die and then on the IHS...with awful results. 50+ deg at idle, 100+ deg on load without an overclock and all fans on high. I tightened the screws of the H100i a bit, to no avail - still, awful temps.
Sure enough, when I removed the 'heatsink' portion of the H100i, the bottom looked basically clean:
(There was a smidgeon of metallic goop on there, but very, very little.)
Now, I applied the CLU to the top of the IHS the night before, and then put the motherboard with CPU installed back into the anti-static bag in which it came. I have no real experience with these liquid metal TIMs and don't really know how they are supposed to stick and spread, but that looked suspect.
Here's what the IHS looks like (cleaned off the CLU as best I could):
And note the scratches near the bottom of the heatsink:
So right now my assumption is that the raised bits at the bottom of the IHS is preventing all but the most cursory contact between the IHS and heatsink.
Ideas? Can the IHS be filed down easily? Should I simply eschew it and place the H100i upon the bare die? Buy a cheap CPU like a G1610, delid its IHS properly and use it? Or am I just screwed?
Hmmm, I think you need a pine wood and wide enough before hammer it
but well, you can always lap it to flat
Originally Posted by ladcrooks
Noctua paste has served me well
I have heard that Noctua paste has popup effects?
Originally Posted by defiler2k
I usually put tape on the vice grips and make sure the vice is making full contact with the IHS wall to prevent that.
Lapping is to sand down the top of the IHS to make it perfectly flat. Some people go as far as giving the IHS a mirror finish. That apparently improves temps. In your case it would solve your current problem.
Hi defiller2k, I am waiting for the IE before I delid it. Is there any suggestion for delid die protections? such as whether I need use elec tape to cover the outside the die area on PCB in case CLP pops or mx-4 will be a better idea.
Also am I really need a piece of pine for vise method or any kind of wood will be fine as long as wide enough to cover the edge of the IHS.
I should be able to delid it maybe next week.
I have received my CLP but waiting for IE now.
I was trying to stress the video card for max temperature, but well 49 was my highest I believe after about 1 hour of 3d mark tests @1220 mhz on core(1330 mv) 1800 on VRAM (1620mv)
I wasn't brave enough to keep add voltage as I am not sure how much I can add to them..... 1330mv on core was my comfort zone before I burn it.
Cheers mate, hopeing to learn some delid tricks before I do it with my shaking hands