Core 2: remove IHS or not?

[L&MFatty]

well done i must say

[0DIGy]

good job

what about the temps?

[menlatin]

well my ds3 is awaiting RMA return, but hopefully in a couple hrs i'm going to try it on a co-worker's machine. I'm going to put the IHS back on and use As5 on both sides. And when i get mine i'll try naked core to the Apogee. I really dont want to grind down my apogee so its going to hold the chip in place, instead of the latch.

[dinos22]

LOL nice one mate

too bad about the intel mounting mechanism which you'll have to butcher next or your cooler.......anyways eagerly awaiting some results

[nooh]

Amazed , does it still work? also the temps difference it has made?

[menlatin]

OK the IHS did not work too well. Since it was originally concave, before lapping, the opposite side is still convex... soooo... there was only contact in the very CENTER of the die. Temps were horrible. 1.43v, 333fsb, 61c bios. I didn't go any further. After i pulled out my cpu, my co-worker was cleaning up the artic ceramique that got around the socket.. and.. well.. bent like 4 pins... So we're not going to mess with it any more. Just going to put his e6600 back in there and let it be. . I may have to break down and buy another Ds3 from Fry's tomorrow....


here's some pics of the inside of IHS. I already touched the cpu die so you can't see it as well.





Notice how almost 1/4 had near no contact.

[ak_47_boy]

i wouldent try to put that ihs back on its probly warped or bent from takeing it off

[mds47]

posting in legendary thread. c2d IHS removed!

please give us more detail, pictures so far are great though.

[menlatin]

Ok here's a little results. I'm currently at 3.52ghz @ 1.45 bios volts. Before i could only get 3.44ghz, so 80mhz gain, and also with warmer ambient by about 10C. Once again though, the reason i can't clock too high is these damn mosfets and coil what ever boxes are getting hot, evne with a fan on them, and my apogee is coveriong half of them up.

My temps seems to have not changed much in coretemp, but in speed fan an easy tuner (when it opens) temps have dropped about 12C under load, and idle down to about 35c. Much better than the 42c of previous.

I'll post a screen later when it proves prime stable. Gotta go!

[DltdApS]

really awesome ! what is your stepping/package date ? your ihs is not solder ?

i want to remove my ihs ! what is your tactic ?

[menlatin]

There we go.

[mds47]

nice! looks like the cpu is still in good shape

So those results are from 1) removing the IHS, 2) cleaning out the stock sodder, and 3) puting the IHS back on but filled with AS5?

What HSF are you using?

Anxious to see what you get with no IHS at all. good luck!

[vintage_guitar]

after all these people said impossible good work. i love proving people wrong!

[oohms]

Does the core stick up above the retention mechanism? Or did you have to take the top metal part off?
I was thinking of doing this but i only have air cooling, which probably wouldnt sit properly on the core since it is heavy

[menlatin]

This is with NO IHS . As noted eariler, the ihs was no flat in the inside either, so only the center of the die was getting contact.
I removed the top part of th metal bracket and my apogee is the only thing holding the cpu in place (plus a little bit of scotch tape)
Im trying 1.5v @ 3.6ghz right now. idle temps is 34c and load is up to 62c.

Also, did i mention water cooled NB? I couldn't get higher than 420 with the stock HS.

I'm really starting to get an itch for my 320watt peltier....

Also i think a storm would be a much much much better cooler with this naked die. It seems like the heat is too concentrated for the apogee.

[vintage_guitar]

can you take pics of the socket now with the processor in it please? this is interesting

[CrazyXP1700]

CONGRATS MENLATIN!!!!!
did you use my method from the other threads?

BTW... i did this a few weeks ago... (i ended up having a bent pin in my socket says ASUS... i checked for that before RMA... but ok ASUS!)







i joined in the thread after a buddy (ZX) rememberd i removed the IHS off a Prescott Celly D a year ago... and the thread was going on in the LN2, and Dry Ice, Intel Heatspreader Removal thread

my method outlined above and in the other thread is for SMD's cpu's mostly... can be applied to all cpu's without SMD's ofcourse since you dont have to worry about how far the razors stick under the core (just dont go too far or apply pressure)

and to everybody wondering in the other threads... it's solder, or some kind of alloy! ive removed IHS's a few times (2 successful, 3 unsuccessful) eventually i cracked the cpus mounting large heatsinks... so be careful

i am trying to sell an old PC so i can finally get my core2duo and take off the IHS

[menlatin]

CrazyXp : Yes i used your method. I was amazed how fast and easy it was! The way i've avoided crushed cores wth all my IHS removals (3 A64's and this one) is using 4 layers of electric tape for "feet" around the edges of the cpu.

I'll try and remember to take pics when i move the setup into my case. I decided on 3.5ghz @ 1.45v 24/7. I can keep my fans on 5v now instead of having to crank them to 12v.

I'm thinking that a little bit better results may be had with a a good air cooler over my setup since they have all the pipes pulling heat from the center of the cpu.

I still want to try the peltier, but thats going to have to wait til next weekend.

[ewsforos]

Btw, as mentioned in a post before... Noone said that it is impossible... We said that it is impractical.

1) you just have to forget about the default mounting system, and tape etc to rebuild an edge to mount is useless as the core will be below the level of the socket cap mechanism.

2) You have an extremely high risk to bend/break pins in the socket itself while mounting a heavy heatsink/block -> bye bye mobo

3) You can't properly mount anything heavy (ie. large heatsinks), or waterblocks because of the risk of cracking the core either due to tension from hoses, weight of heatsink, force from mounting etc. In short, 1/4 of a revolution more than it should on one of the screws and you'll hear a very expensive *crack*

4) CPU warranty goes out of the window

and 5) All that for what? 2-3C REAL temps (it's the temp of the core that counts... and as menlatin said, he didn't gain anything worth mentioning from coretemp).

So... is 80Mhz worth all that risk and work, given of course that he obviously ain't after any world records? I tend to think "no".

Again, noone said it was impossible... Just very impractical.

[Entsafter]

1) You have to remove the old socket cap
2) Why? With an IHS and the old mechanism you can also mount very large heatsinks and they push the cpu very hard in the socket
3) Why? There were 10000 of AMD CPUs without IHS
4) CPU warranty isn't important
5) I think it will be more than 2-3C real temps...

6) I've also removed the IHS of my E6600 a few days ago and the cpu still works great.. tomorrow I'll test it with a waterblock and without IHS... I've temporarly remounted it with coolaboratry liquid pro and the temps went down 10-15°C ! without the IHS they will be a few degrees better than atm...

btw: this is extremesystems!

[menlatin]

^^ The 4 layer feet really help to prevent any cracking. You could probably go about 2-3 turns uneven and you would just see bad temps. I've sat there and cranked down on one side until i see temps rise, then let off and try the other ones. Remeber, the feet are right about the same height as the core, and they only squish a tiny bit.

But yes this is a very impractical mod, especially if you're using an Apogee or similar pin style block, as there is way too much heat in the center for the block to dissipate efficiently.

[CrazyXP1700]

Yep... this is Xtreme Systems...
If your worried about your waranty... you wont be pushing your hardware (doesn't overclocking itself void warantee's?)

not necessarly trying to get the best temp or anything...
I would see it for the Xtreme... in benchmarkers systems, guys pushing lots of voltage...i wouldn't want to wait for heat to go thru the TIM, into the IHS, then thru more TIM and then be released to whatever cooling method provided... id wont the most responcive solution possible...

really i though of this being used for those LN2 and phase guys, also some waterblock designs would benifit from IHS removal too (i think anything that dumps water directly over the cpu, with a decent amount of surface area for heat transfer)... i believe Cathar said best performance is with a naked cpu... with the storm blocks... might be wrong...

i mean i relayed my guide to Shimano at VR Zone... im sure he'll be breaking WR's none the least... maybe with my method... he pushes a few more Mhz now...
If someone, anyone drops my name for gettin a few more mhz, for lettin them know how it's done... i'll be a happy man!

my method works... my cpu didn't work... but the method works... and thats all right with me
its simple, it's easy... its successful...

sure as hell isn't practical...
but my name is Crazy... and this is XS!

Edit:
when i get another cpu, im going to write a detailed guide on intel IHS removal... and get it stickied...
that way we can get it all in one post... for all those Xtreme members lookin to have fun!
i think i fit in great (dont'ya think?)

[menlatin]

Oh yes Ln2!! I just remembered i need to pick up my Alu tube from the welders.... uh oh.....

[ewsforos]

1) You have to remove the old socket cap
2) Why? With an IHS and the old mechanism you can also mount very large heatsinks and they push the cpu very hard in the socket
3) Why? There were 10000 of AMD CPUs without IHS
4) CPU warranty isn't important
5) I think it will be more than 2-3C real temps...

6) I've also removed the IHS of my E6600 a few days ago and the cpu still works great.. tomorrow I'll test it with a waterblock and without IHS... I've temporarly remounted it with coolaboratry liquid pro and the temps went down 10-15°C ! without the IHS they will be a few degrees better than atm...

btw: this is extremesystems!

Sure

Though:

1) Then you have no secure way to mount the cooler/block, if you move the cooler/block around while you are mounting the cooler/block that will hold the cpu down for whatever reason you are bound to bend/break some pins...

2) Pushing the CPU down is NOT a problem, the PCB can take it no problem, problem is though that the force is applied on the IHS (large surface area) and from there to the PCB itself... On the other hand the core itself is not so forgiving and it is pretty easy to crack. The IHS _is not_ putting the force of the mounting system on the core, it is putting it on the PCB through its base.

3)They have those feet to take the load off the core, and they are engineered to be used in that way, so there are some measures that help to prevent such problems. On a naked C2D though, you have nothing there.

4) Until you need it

5) No, not really. It can make a difference with LN2 or exteme temps, but with air you will see minimal difference, and with watercooling, only a few degrees. Whether it is worth it, is up for debate... I am not forcing anyone to not do what he thinks, by all means, go ahead... I'm just saying it is impractical and risky. How many people will break their core, especially first timers, that read here that "it is so easy, and simple" and figure out there is no risk involved. For seasoned people, or people who know and understand the risks that's fine. But sooner or later you WILL have someone screaming "I cracked my core and have no warranty and no money to get a new CPU". That's what I'm trying to point out.

If you are going for a WR, by all means, there even 80Mhz will count. But for the average Joe, there is really no point

and 6) The difference will NOT be as much as you think. Sure you have less material in between, but the heatspreader does exactly that, spreads the heat so you have a larger surface area to dissipate it from. Coppers thermal conductivity is only that much, and trying to dissipate 150W+ from an area smaller than your fingernail using only 10-15% of the base of your heatsink, can be a PITA...

Finally, someone said that he used the coll. liquid pro. Well by having his block to keep the processor down with force, I am REALLY curious to see what will happen when the "paste" solidifies with time and pretty much is stuck there, like in the other chap who was trying to get his G4 off his CPU for a week. Since the core is not like its riveted on the PCB, chances are that he will be selling on ebay "Storm block. Comes with a free, bonus C2D core"

In any case, by all means do whatever you like with your hardware, hell, its your money and I have no saying in what you do with it. Just stating some facts so that inexperienced people who read this and think "Ah, cool, that seems so easy, I'll do it too" will at least know the risks involved.

Enjoy

[ewsforos]

Yep... this is Xtreme Systems...
If your worried about your waranty... you wont be pushing your hardware (doesn't overclocking itself void warantee's?)

There are several reasons for a cpu to fail, if the reason is something related to overclocking, sure, it would be unethical to go and RMA your CPU. But while you are overclocking something _not_ related might happen and even though you overclocked your cpu, if you are CERTAIN that it wasn't the culprit, I don't see a problem RMAing it. Good luck sending a cpu without an IHS for RMA though...

Also I am fully aware that this is Xtreme Systems, but are you aware that 90%+ of the userbase of this forum are not seasoned overclockers, and do what they see the big names do? There is a reason there is 1 Fugger, 1 Coolaler, 1 Shamino, etc... Most of the userbase is just looking to what they are doing and mimic them.

Watching Ferrero Pipin free-diving to -138m sure looks easy... If you try though chances are you'll be fishfood before you even go halfway that. He has the experience, the knowledge and he is practicing his whole life for this.

I am not saying that one who knows shoudn't attempt this, but one should make CLEAR to the people who lack the knowledge and experience and tend to mimic the "masters" the dangers and the risks involved in doing so instead of going around screaming "look, look, I free-dived at -138! If I can do it, you can do it too! Just learn advanced techniques to equalize the pressure and lower your heart-rate, and hold your breath for 6 and a half minutes! That's all there is to it, it's THAT easy!"