[somethingbrite]

MrTOOSHORT
Thanks for the info, good to know.

I'll want to cover up the nicks anyway so I'll find some silicon glue similar to that used to fix the IHS in place. (I read the post about nail varnish not working.)

Also, There's a 3 week wait time on the direct die adaptor so I think I'll lap the IHS further with some 2000 and 3000 paper and see what happens. I figure that way I can put some Liquid Ultra in between and get playful with the clocks.

 

[somethingbrite]

Ok, short update.
Despite my fears of damage my CPU appears to be working.

Stress testing with either Prime95 large or Intel Burn Test Maximum and I am peaking at around 56c
No overclock yet. Turbo was around 3.6Ghz in last stress test. Cooler is Corsair H60 with 1 fan running about 1400 and another fan on the other side (push pull) not even plugged in.

TIM used was Noctua....but the Liquid Ultra arrived today so.....

 

[MoGTy]

Quote:

Originally Posted by somethingbrite

Ok, short update.
Despite my fears of damage my CPU appears to be working.

Stress testing with either Prime95 large or Intel Burn Test Maximum and I am peaking at around 56c
No overclock yet. Turbo was around 3.6Ghz in last stress test. Cooler is Corsair H60 with 1 fan running about 1400 and another fan on the other side (push pull) not even plugged in.

TIM used was Noctua....but the Liquid Ultra arrived today so.....

You'd be more satisfied with a H80i slapped on it

 

[somethingbrite]

I've got a custom cooling loop on the way in.
Putting the H60 back on was for comparative testing and checking the CPU even worked after my rough handling.

That said. I will transfer the H60 to my smaller build (media + file server box) I'm more than happy with its performance within the limits of its design and I think it should have another year of life at least. I seriously prefer a smaller footprint water cooler to a large and heavy air cool heatsink stressing the motherboard.

If I do have criticisms of the H60 it is that having cleaned the stock TIM it is delivered with, the finish of the heat block is a little rough and the design means it can not easily be lapped or polished.

 

[MoGTy]

Amen to that, closed loop water cooling is a blessing. Especially to those people who can't be bothered with custom loops, like me

As for the polishing/lapping. You can get some specialized tools to do that but it really isn't worth it. I did and it made 0 to 1 degree difference in my case. However lapping aircoolers still pays-off, IIRC my chip dropped 3 degrees C, although delidding it did the most with a ~5°C drop.

 

[somethingbrite]

Quote:

Originally Posted by MoGTy

As for the polishing/lapping. You can get some specialized tools to do that but it really isn't worth it. I did and it made 0 to 1 degree difference in my case. However lapping aircoolers still pays-off, IIRC my chip dropped 3 degrees C, although delidding it did the most with a ~5°C drop.

Lapping - every little bit helps ;-)

Progress Update.
Closed Loop cooling stripped out.
Custom cooling implemented.

Ambient Temp Room 24°C approx
Stock Clock (3.4Ghz Turbo @ 3600) Stress Load. Prime95 small = Temps hovering around 44°C
Overclock to 4.5Ghz Stress Load. Prime95 small = Temps hovering around 49°C (max logged temp have seen is 52°C)

I have only run the OC stress test for a couple of hours so far but the CPU definitely seems to have survived my attempts to kill it and these are the best temps I have ever seen on this system by quite a margin. I think I'll vote that a first time success after all.

The system
Intel i5 3570K, Asus P8Z77-V Pro
Delidded, IHS Lapped to 1200 Grit on top and a tiny bit on the lower edge. (Internally was only polished with a dremel - not lapped.)
Liquid Ultra between Die and IHS
Noctua NT H1 between IHS and Waterblock
EK-Supremacy Waterblock
XSPC RX 480 V3
Noctua NF F12's in pull.

 

[BossMac]

Just sharing my delid experience a couple of hours ago. Went with the razor method and popping off two IHS didn't take more than 10 minutes.

Delidded both my 3770K and 4770K. At first I was going for a TIM replacement but since I'm running a custom loop, I decided to mount the Raystorm blocks directly on both CPUs without the IHS(running two systems). Got the idea from when I mounted an ALC to my GPU and thought, what the heck I'll try it with a CPU. So far the 3770K is showing great temp improvements at 4.8Ghz @1.44v. Great improvements vs. the 100*C~ I get at that voltage.



Here's the AIDA stress test results:


Here's the one for my Haswell rig with the 4770K 4.8Ghz @1.35v. Both systems running direct-to-die with the waterblock.


Parsed through through the thread and was wondering if anyone has tried mounting direct-to-die with an ALC cooler like the Antec H20 920. I'm thinking it would involve adding some washers on the screws to compensate for the lack of a socket retention brace. Ideas anyone?

Thanks!

 

[deepor]

Quote:

Originally Posted by BossMac

Parsed through through the thread and was wondering if anyone has tried mounting direct-to-die with an ALC cooler like the Antec H20 920. I'm thinking it would involve adding some washers on the screws to compensate for the lack of a socket retention brace. Ideas anyone?

I've seen people mention they managed to do it. You might want to try searching by using the name of some other cooler, H100i for example. You might also want to use other acronyms like "CLC" and "AIO".

The posts I remember, they said pressure created by the four corners of the mounting system of the cooler had to be adjusted carefully or it won't boot. There seems to be something going on with the CPU's contacts pads touching the pins in the socket. If you keep the IHS and keep the bracket thingy that locks the CPU into the socket, there will be guaranteed uniform pressure on all pins and that's what you have to manually recreate yourself with the screws of the mounting system of your H2O 920.

 

[BossMac]

Quote:

Originally Posted by deepor

I've seen people mention they managed to do it. You might want to try searching by using the name of some other cooler, H100i for example. You might also want to use other acronyms like "CLC" and "AIO".

The posts I remember, they said pressure created by the four corners of the mounting system of the cooler had to be adjusted carefully or it won't boot. There seems to be something going on with the CPU's contacts pads touching the pins in the socket. If you keep the IHS and keep the bracket thingy that locks the CPU into the socket, there will be guaranteed uniform pressure on all pins and that's what you have to manually recreate yourself with the screws of the mounting system of your H2O 920.

Thanks for the tip, man. I'll look around. I'll eventually be moving the delidded 3770K on an ITX rig and wanted to stick with an H100i or so without putting the IHS back on. I may need a sacrificial chip lol.

 

[MoGTy]

Quote:

Originally Posted by BossMac

Parsed through through the thread and was wondering if anyone has tried mounting direct-to-die with an ALC cooler like the Antec H20 920. I'm thinking it would involve adding some washers on the screws to compensate for the lack of a socket retention brace. Ideas anyone?

Thanks!

I actually did try that with an H80i and some washers, I needed about 4 x 2 of those millimeter thick washers. However, this mounting process was rather fragile in my case, one day my pc just wouldn't boot anymore, I put the IHS back on with the normal mounting bracket and things were back to normal. I guess I could've tried more, maybe another washer on every screw, but I haven't had the courage (read : motivation) to try it again

 

[BossMac]

Quote:

Originally Posted by MoGTy

I actually did try that with an H80i and some washers, I needed about 4 x 2 of those millimeter thick washers. However, this mounting process was rather fragile in my case, one day my pc just wouldn't boot anymore, I put the IHS back on with the normal mounting bracket and things were back to normal. I guess I could've tried more, maybe another washer on every screw, but I haven't had the courage (read : motivation) to try it again

During the times that it did work, were the temps better overall compared to the mounting it with an IHS?

Its been a couple of days since I've mounted my blocks on the naked chip and I've noticed that a mild bump would crash the system. Might be that the block retention isn't tight enough and I'm afraid this could be much more of a problem on the ALCs.

 

[MoGTy]

Quote:

Originally Posted by BossMac

During the times that it did work, were the temps better overall compared to the mounting it with an IHS?

Its been a couple of days since I've mounted my blocks on the naked chip and I've noticed that a mild bump would crash the system. Might be that the block retention isn't tight enough and I'm afraid this could be much more of a problem on the ALCs.

Sadly it didn't, 0 to 1°C drop. I think the surface of the die is too small and the heat too concentrated on one spot, that it doesn't really make a big difference. Maybe it was my mounting method or maybe these chips don't benefit as much from direct contact. I have no idea. IIRC I have seen people who had great succes with it, so it could very well be my method.

Yes, the system is very fragile in this state, you'll need some method of fixating the chip in the right position or any bump can cause a crash. Come to think of it, given the pressure a normal mounting bracket puts on a CPU. Maybe my chip needed more pressure and thus only one washer to keep it from crashing with every nudge the case got.

 

[BossMac]

Quote:

Originally Posted by MoGTy

Sadly it didn't, 0 to 1°C drop. I think the surface of the die is too small and the heat too concentrated on one spot, that it doesn't really make a big difference. Maybe it was my mounting method or maybe these chips don't benefit as much from direct contact. I have no idea. IIRC I have seen people who had great succes with it, so it could very well be my method.

Yes, the system is very fragile in this state, you'll need some method of fixating the chip in the right position or any bump can cause a crash. Come to think of it, given the pressure a normal mounting bracket puts on a CPU. Maybe my chip needed more pressure and thus only one washer to keep it from crashing with every nudge the case got.

Well in my case I'm getting better temps with direct die mounting so I'm really optimistic at what can be done with ALCs. I guess I should try it myself and report back with my findings. Might get a shim fabricated so I can mount the socket retention bracket again and keep the chip from moving.

 

[MoGTy]

Quote:

Originally Posted by BossMac

Well in my case I'm getting better temps with direct die mounting so I'm really optimistic at what can be done with ALCs. I guess I should try it myself and report back with my findings. Might get a shim fabricated so I can mount the socket retention bracket again and keep the chip from moving.

Good luck, report back to us.

I might give it another try if you're succesful.

 

[BossMac]

Quote:

Originally Posted by MoGTy

Good luck, report back to us.

I might give it another try if you're succesful.

Will do. Wish me luck that I don't crush the die also LOL

On another topic though, I am very comfortable cooling my 3770K via direct die. Running 4.8Ghz @1.38v (my 4770K, also delidded runs 4.8Ghz @1.33v fine also) and so far no errors in games and benchmarks (did a lot of HWBOT subs last few days with it) but I was just wondering if there is any negative concern I should be worried about aside from the occasional bump that crashes the system.

 

[MoGTy]

Apart from crushing the pins on the socket, the die or normal degredation with high voltage. I don't think so. Just make sure you don't short out anything while your mind is occupied with tinkering with a custom mounting system.

 

[Moparman]

Just de lid my 6.737ghz ln2 3570k. Now daily it was running [email protected] 78c max on this mugen 4000 with p/p 2100rpm fans. Now same cooler used Z5 on die and IHS now I'm maxing at 61c in IBT. So great temp drop.

 

[dilster97]

First time success.