Delidded my i7-3770K, loaded temperatures drop by 20°C at 4.7GHz

[Idontcare]

my current result:
3570k Vcore 1.212 avg core temp 68.5 room temp 28 Delta:40.5 test used:IBT
i think that is right in the ballpark for h80i stock fans max non-lapped.
what do you say?

Yeah that is reasonable, good ballpark result.

 

[sanurocks]

i will soon get my liquid pro and ultra..
i was reading through some threads on this site about CLP/CLU..
everybody mentioned that they solidify over time..
my question is that if it solidifies how to remove it from the bare die?
as it is silicon we cant use sandpaper right?
So what is the method to clean the die after CLP/CLU solidifies..
removing form cpu cooler is as easy as lapping.

 

[Idontcare]

i will soon get my liquid pro and ultra..
i was reading through some threads on this site about CLP/CLU..
everybody mentioned that they solidify over time..
my question is that if it solidifies how to remove it from the bare die?
as it is silicon we cant use sandpaper right?
So what is the method to clean the die after CLP/CLU solidifies..
removing form cpu cooler is as easy as lapping.

Yes that is a problem...so don't use it if you want to remove the IHS at a time in the future.

Basically no one cares to remove it once they've delidded and used the liquid metal TIM.

But if you want to or need to, then don't use the liquid metal TIM. It will eventually seal the IHS to the silicon die just like a regular solder would.

 

[Mean MrMustard]

I have a Haswell i5-4670 that will be air-cooled with a Zalman CNPS10x. I lapped both the IHS and the heatsink.

If I delid the processor, I'll want to use CLU between the IHS and the CPU. Should I bother using CLU between the IHS and heatsink? Will it make a difference if I use NT-H1? I don't want it to solidify on the heatsink over time...

 

[Idontcare]

I have a Haswell i5-4670 that will be air-cooled with a Zalman CNPS10x. I lapped both the IHS and the heatsink.

If I delid the processor, I'll want to use CLU between the IHS and the CPU. Should I bother using CLU between the IHS and heatsink? Will it make a difference if I use NT-H1? I don't want it to solidify on the heatsink over time...

Just use a regular top-end TIM for the IHS to HSF interface. Yes CLU/CLP will give you superior temperatures but it becomes a complete PITA to remove your HSF from the IHS and to clean up the HSF again for later use.

 

[Mean MrMustard]

Just use a regular top-end TIM for the IHS to HSF interface. Yes CLU/CLP will give you superior temperatures but it becomes a complete PITA to remove your HSF from the IHS and to clean up the HSF again for later use.

Thanks. That's exactly what my concern was. I didn't want the HSF fused to the IHS. I only plan on moderately overclocking anyway.

 

[sanurocks]

Basically no one cares to remove it once they've delidded and used the liquid metal TIM.

But if you want to or need to, then don't use the liquid metal TIM. It will eventually seal the IHS to the silicon die just like a regular solder would.

My bad i forgot to add the most important detail.. I am on bare die now with my cooler..
My ihs is in bad shape it dropped and someone stepped on it, slightly bent i suppose but i will lap it..
So golden words are:
CLP/CLU between die and ihs good idea..:thumbsup:
Between ihs and cooler,not so much..:thumbsdown:

 

[Idontcare]

My bad i forgot to add the most important detail.. I am on bare die now with my cooler..
My ihs is in bad shape it dropped and someone stepped on it, slightly bent i suppose but i will lap it..
So golden words are:
CLP/CLU between die and ihs good idea..:thumbsup:
Between ihs and cooler,not so much..:thumbsdown:

Your H80i cost barely anything compared to your CPU+mobo+ram (all of which won't be useful for upgrades come Skylake and beyond), do you really care if you fuse it to your bare die?

If you do then yes, use NT-H1 (or equivalent) between the die and the cooler. But I personally wouldn't care so much about an H80i (or my H100 in my case) in the bigger picture. Just get a new one when you are ready to replace your CPU.

 

[sanurocks]

Good point..

 

[iNs_]

I may not contribute anything at this moment into the subject, but I spent like 6 hours reading the (almost) entire topic and I am simply shocked and astounded how useful and meticulous information in this topic is (especially the scientific side of the subject). IDC (and others as well) - tremendous work and thumbs up for more. Personally, I am in the waiting process for my 4670K based system - perhaps I will come back with my own delid story, in which case I'd be delighted to share my results, if they turn out to be any of use for you guys (honestly I hope I won't be forced to do so, nevertheless I wouldn't be surprised if I delid it anyway, just for the fun of it and because I like to fiddle with stuff and hardware )

Cheers, looking forward to reading some more awesome stuff

 

[Idontcare]

Welcome to the forums iNs_ :thumbsup:

Thanks for the kind words :$ but don't be surprised if you find yourself really wanting to delid that CPU once you get it!

You don't need to be trying to get 5GHz OC's to benefit from delidding, even a mild 4.3GHz OC significantly benefits from delidding because of the lowered temperatures and lowered Vcore needed for stability at those lower temperatures.

Your OC'ed CPU will last longer and consume less power...what's not to like?

 

[iNs_]

Welcome to the forums iNs_ :thumbsup:

Thanks for the kind words :$ but don't be surprised if you find yourself really wanting to delid that CPU once you get it!

You don't need to be trying to get 5GHz OC's to benefit from delidding, even a mild 4.3GHz OC significantly benefits from delidding because of the lowered temperatures and lowered Vcore needed for stability at those lower temperatures.

Your OC'ed CPU will last longer and consume less power...what's not to like?

I won't be surprised, not at all. Im no newbie to the overall hardware world, just hasn't been in a probable need of delidding cpu (switching from a clocked e8400 right now, i know, thats old ). So yes, it is probably just the matter of time and pre-delid benches before I gently take his hat off

 

[JDG1980]

Are there any pre-made shims for Haswell you can buy, sort of like the ones they used to have for bare-die Athlons? It seems that with a big heatsink it would be a lot safer to mount if you had a shim to ensure flat contact and no rocking back and forth that could chip the edges.

 

[Idontcare]

Are there any pre-made shims for Haswell you can buy, sort of like the ones they used to have for bare-die Athlons? It seems that with a big heatsink it would be a lot safer to mount if you had a shim to ensure flat contact and no rocking back and forth that could chip the edges.

Given that delidding has been in vogue for nearly 2yrs now, and there is a notable dearth of reports of chips being killed from chipped or cracked die, I would say the concerns are over-rated when it comes to delidding IB and HW.

My own theory on how/why this is the case is that it comes down to two things - the die itself is silly thin compared to the thick silicon dies from 8 yrs ago, and the socket being LGA means the CPU itself is basically cushioned on a big huge spring mattress that ensures downforce from the cooler is not concentrated on any specific location across the CPU.

This combination results in cracked die and chipped corners being more of an academic concern now, not much of a practical concern these days.

 

[PliotronX]

Given that delidding has been in vogue for nearly 2yrs now, and there is a notable dearth of reports of chips being killed from chipped or cracked die, I would say the concerns are over-rated when it comes to delidding IB and HW.

My own theory on how/why this is the case is that it comes down to two things - the die itself is silly thin compared to the thick silicon dies from 8 yrs ago, and the socket being LGA means the CPU itself is basically cushioned on a big huge spring mattress that ensures downforce from the cooler is not concentrated on any specific location across the CPU.

This combination results in cracked die and chipped corners being more of an academic concern now, not much of a practical concern these days.

Methinks you are correct. I have removed every IHS dating back to the Opteron and have never chipped anything nor regretted removing them. Following a few simple rules of thumb, it's idiot proof. Those double sided paper thin razors are what I use, gentle twisting, even pressure when mounting. It's so easy I can't believe more don't do it.

 

[C.C.]

Sorry, no experience with Antec formula 7, but if it has abrasives in it like diamond paste then yes I would be worried about putting that in direct contact with the silicon die.

I know it has been awhile since I popped back over from [H]ard, but I wanted to point out that I used IC Diamond on my 3770K's die for quite some time and when I removed it I saw no visible damage to the die or the block..I did hit the block with a few passes of 1500 grit wet-sand since it was a pure copper block, and had slightly oxidized..After 10-15 passes looked good as new (before I used the IC Diamond)...

 

[PliotronX]

I know it has been awhile since I popped back over from [H]ard, but I wanted to point out that I used IC Diamond on my 3770K's die for quite some time and when I removed it I saw no visible damage to the die or the block..I did hit the block with a few passes of 1500 grit wet-sand since it was a pure copper block, and had slightly oxidized..After 10-15 passes looked good as new (before I used the IC Diamond)...

Yeah silicon won't be etched its just soft metals like copper tat are affected. Also if it were to etch the die there is plenty of it before it becomes an issue. Sometimes the die itself is concaved and I have lapped them with no I'll effects.

 

[alkemyst]

Any info on a de-lidded chip without the IHS put back on it or does this not work anymore with the smaller die size.

Also no matter how light, always wear eyewear and gloves when working with hammers, blades and dremels.

I had a dremel wheel fragment just testing it out in the air. One piece went about 1/4" into my cheek about 1/2" away from my eye dead center on it. A little more up my cheekbone wouldn't have saved me.

Since wearing ear-wear I have had a few near tragic accidents from broken things.

I don't buy cheap tools.

 

[ehume]

Seventy-two thousand posts.

 

[Idontcare]

Any info on a de-lidded chip without the IHS put back on it or does this not work anymore with the smaller die size.

Bare-die testing: A delidded 3770k, an H100, and 9 different TIMs

Also no matter how light, always wear eyewear and gloves when working with hammers, blades and dremels.

No worries for the OP, wore both when using all of the above

 

[felang]

Hopefully you guys can help. So I got hold of a delidded 3770K, I used NT-H1 between the Die and IHS and also between IHS and my CPU waterblock. I´m using only socket retention mechanism pressure to hold everything together, I didn´t use any adhesive to join the IHS to die. According to IDC´s meticulous testing NT-H1 seemed like the best compromise so as to not have to mess with liquid TIMs.

My problem is, while after I first set everything up I was getting great temps (max 65 deg at 4.5ghz and 1.25V), after 2 weeks my temps are skyrocketing (80 deg running OCCT´s standard torture test.

I´m guessing the IHS must have moved so now I´m getting bad contact. Any tips as to what I should look for when I tear everything apart and reapply TIM to the die? Should I use an adhesive to reattach the IHS to DIE?

 

[PliotronX]

Hopefully you guys can help. So I got hold of a delidded 3770K, I used NT-H1 between the Die and IHS and also between IHS and my CPU waterblock. I´m using only socket retention mechanism pressure to hold everything together, I didn´t use any adhesive to join the IHS to die. According to IDC´s meticulous testing NT-H1 seemed like the best compromise so as to not have to mess with liquid TIMs.

My problem is, while after I first set everything up I was getting great temps (max 65 deg at 4.5ghz and 1.25V), after 2 weeks my temps are skyrocketing (80 deg running OCCT´s standard torture test.

I´m guessing the IHS must have moved so now I´m getting bad contact. Any tips as to what I should look for when I tear everything apart and reapply TIM to the die? Should I use an adhesive to reattach the IHS to DIE?

Can't hurt to reapply the TIM and you could use arctic alumina epoxy mixed with a bit of your NT-H1 (1:1:1 ratio) so it isn't completely permanent and enhances thermal conductivity.

edit- oh and did you scrape any of the black sealant off? If not, from my understanding that stuff can hurt mating with the die.

 

[felang]

Can't hurt to reapply the TIM and you could use arctic alumina epoxy mixed with a bit of your NT-H1 (1:1:1 ratio) so it isn't completely permanent and enhances thermal conductivity.

edit- oh and did you scrape any of the black sealant off? If not, from my understanding that stuff can hurt mating with the die.

Yeah the black goo was completely scraped off, I guess I´ll try again and really tighten the waterblock retentión mechanism as much as possible, maybe use more TIM than the first time in order to make sure I get decent contact, hopefully I won´t have to reapply every couple of weeks!

 

[kalamar182]

Wow i have an asrock z77 exfreme 4 m an a i7 3770k oc to 4.8 rising 90° with his off and a watercooling refrigeration. I saw your post its awesome what you do with your i7. I have a cap a questions, cos i really wanna do dat asap. Tellme your system its workin 7/24. At 5 ghz??? Did you have any problem doing that??? Do you have any guide to dothat. Thanks dude congrats and sorry my english.

 

[Idontcare]

Welcome to the forums kalamar182!

Wow i have an asrock z77 exfreme 4 m an a i7 3770k oc to 4.8 rising 90° with his off and a watercooling refrigeration. I saw your post its awesome what you do with your i7. I have a cap a questions, cos i really wanna do dat asap. Tellme your system its workin 7/24. At 5 ghz??? Did you have any problem doing that??? Do you have any guide to dothat. Thanks dude congrats and sorry my english.

My 3770K's (three of them) could get to 5GHz stable for 24/7 operation, but I never left them there because the voltages were way too high for my comfort level.

They run comfortably at 4.7-4.8GHz.

Delidding doesn't really seem to bring much in terms of increased overclocking potential, but it does dramatically lower the operating temperatures which will substantially increase the lifespan of your already over-stressed CPU (because it is overclocked/over-volted).