i7-7700K @4.9GHz w/ NH-D15S (Delid + Liquid Metal)

[volkgren]

I figured the NH-D15S (w/ second fan) would be enough to keep it cool at 4.9GHz and 1.35V. The voltage is enough to run 4.9GHz, but temps go up to 87C within a few minutes of Prime95 Small FFTs.

Would an AiO do any better? It seems at this/that point, ~$450 would have been better spent on a better CPU and a budget cooler. I have not seen any significant advantage of 4.9GHz over 4.5GHz other than the personal satisfaction of doing it.

I just upgraded from my Cryorig H7. The NH-D15S gives the same temps at 4.7GHz as the H7 gives at 4.5GHz. All of this is seeming like a huge waste of money. Intel is taking away the overclocking experience by pushing their K series processors to their economical limit with their junk TIM.


**UPDATE**

Successful delid and Thermal Grizzly Conductonaut reduced temps by ~20C.

Achieved 5.0GHz 1.400V w/ max of 73C Prime95 (20 mins).

Screenshot: https://imgur.com/TUGAQ2Q

 

[SumTingW0ng]

http://www.coollaboratory.com/pdf/manual_liquid_pro_englisch.pdf


LAST PARAGRAPH.

But it still left marks and residues behind even after you wiped it with isopropyl alcohol. This is why you should only use liquid metal for delid not for apply on top of the CPU HS like typical thermal compound.

Are we not talking about delid, or can you just not admit that you are wrong?

https://www.youtube.com/watch?v=7H4ptevACfw skip to 8:18

https://www.youtube.com/watch?v=IIikmLf_H48 skip to 1:47

Not, I'm not wrong but rather you are the one without doing your homework.

Liquid metal only worth it if you are delid, if you are not just use a good non conductive thermal compound like MX-4 will do fine. Remember, your method apply thermal compound, cooler, and ambient temperature will affect your CPU temperature.

Jesus the manufacturer states that it can be removed.

Your WRONG PERIOD.

Any brand recommendations on liquid metal?

coollaboratory liquid pro

liquid metal thermal compound is very conductive, and once apply must never take it off.

Jesus the manufacturer states that it can be removed.

Your WRONG PERIOD.

Take you bruised ego and parrot your misinformation somewhere else

Have you look at the second video? It can be remove but not 100% completely remove. Look at the second video and see it as yourself before spouting nonsense and false accusation.

You stated that liquid metal is electrically conductive and so shouldn't ever be removed.

Proven wrong, Your now trying to walk back your misinformed opinion, by talking about nickel plated cpu cooler bases and discoloration/corrosion. This doesn't explain away why you think electrical conductivity means you can never remove Liquid metal after using it when deliding a cpu.

Once again we are talking about deliding a cpu. The liquid metal goes in between the IHS and the DIE.

You have a looked at what the Manufacturer says?

Do you know what a manufacturer is?

So how many delids do you have under your belt?

I have done 7. I've taken them back apart and put them back together and mercy me nothing bad happened.

I didn't say couldn't remove, I said " liquid metal thermal compound is very conductive, and once apply must never take it off." which is for delid procedure because of what happened to the video #2 showed, it will affect the temperature because it left stubborn residues behind. Once again, you're reading wrong and therefore deploy a false accusation on me. Re-read what I wrote because there's a different meaning between couldn't and must never take it off.

I met a lot of people that used liquid metal on delid CPU, and they say the same thing once you apply liquid metal onto the CPU die, it should not take it off.

Never, couldn't. Your still trying to side step.

It can be taken off, it can be reapplied. I have done it. The manufacturer says it can be done.

You are wrong.

Then explain on the video why there's residues left behind after wiping it off using isopropyl alcohol. There's a different meaning between must never take it off and couldn't, and yet you are the one try to side step not me. The manufacturer can say whatever it want, but the fact the video proven that liquid metal can left stubborn residues behind make it more difficult to get rid of it, and by re-apply it with previous thermal or liquid metal residues left behind will affects the CPU overall temperature. This is why a lot of people suggested using liquid metal on the die because once you apply it you shouldn't take it off and re-apply it again.

As matter of fact, this thread will say it was well: http://www.tomshardware.com/answers/id-3495322/thermal-paste-heat-sink-liquid-metal-delidding-thing.html

Best selected answer from that thread said : Liquid metal is really for situations where you don't intend to ever take it back apart, which is great for delidding. Between the IHS and heatsink, however, I'd use a high quality thermal paste of some sort. Be careful with liquid metal, it's really conductive and can ruin electronics if it gets into crevices it shouldn't be in

I'm done arguing with you because I already provided my facts and statements and I'll stand with it.

I see what the problem is, you believe everything you see on you-tube.

You should consider a career in politics.

You flat out called me a liar right in this very thread and when presented with proof(as good of proof as anyone on these forums can provide), you couldn't admit you were wrong.

You stated that because of electrical conductivity that LM must not be taken off(your opinion) Not what the manufacturer states and Not what I personally have experienced.

Once again you cannot admit you wrong. This along with your moniker( I think your subconscious was hard at work when you decided to use it) is telling to say the least.

But the fact is you should only used liquid metal on the die not on the HS. Liquid metal was meant to be not taken apart once apply which is useful for delid. I provided with the fact about this, and this suggestion have be around for a decade i believe, and a lot people will recommend liquid metal on the die for better temperature improve than apply standard thermal compound onto the die.

I have provided you with my multiple facts and yet you refused to accept them that liquid metal was meant to apply once not multiple times like thermal compound and the fact is I said "MUST NEVER TAKE IT OFF".and not " COULDN'T". Yet you failed to comprehend the difference between must never take it off and couldn't. As for the part of you said you can achieve 4.9Ghz on air, I believe that statement have to be a lie first because I knew that i7 7700k runs hot as i7 4790k, therefore achieve 4.9Ghz on air seems to be exaggerated, but not until you provided with a picture of it.

The manufacturer say it can be remove, which of course if you can apply it than you can remove it. But the biggest problem is it will left a big junk residue or stain behind make it more difficult to remove it than standard thermal compound and will affect the CPU overall temperature.

Other than that, I'm out. Peace

 

[volkgren]

Okay so if I apply liquid metal to the die, what if I need to remove the CPU from the motherboard later on? Will the IHS not remove from the die? Will this affect the liquid metal application?

*Also, I right now I have a 10C difference in 2 of the cores.

 

[biglizard]

http://www.coollaboratory.com/pdf/manual_liquid_pro_englisch.pdf


LAST PARAGRAPH.

But it still left marks and residues behind even after you wiped it with isopropyl alcohol. This is why you should only use liquid metal for delid not for apply on top of the CPU HS like typical thermal compound.

Are we not talking about delid, or can you just not admit that you are wrong?

https://www.youtube.com/watch?v=7H4ptevACfw skip to 8:18

https://www.youtube.com/watch?v=IIikmLf_H48 skip to 1:47

Not, I'm not wrong but rather you are the one without doing your homework.

Liquid metal only worth it if you are delid, if you are not just use a good non conductive thermal compound like MX-4 will do fine. Remember, your method apply thermal compound, cooler, and ambient temperature will affect your CPU temperature.

Jesus the manufacturer states that it can be removed.

Your WRONG PERIOD.

Any brand recommendations on liquid metal?

coollaboratory liquid pro

liquid metal thermal compound is very conductive, and once apply must never take it off.

Jesus the manufacturer states that it can be removed.

Your WRONG PERIOD.

Take you bruised ego and parrot your misinformation somewhere else

Have you look at the second video? It can be remove but not 100% completely remove. Look at the second video and see it as yourself before spouting nonsense and false accusation.

You stated that liquid metal is electrically conductive and so shouldn't ever be removed.

Proven wrong, Your now trying to walk back your misinformed opinion, by talking about nickel plated cpu cooler bases and discoloration/corrosion. This doesn't explain away why you think electrical conductivity means you can never remove Liquid metal after using it when deliding a cpu.

Once again we are talking about deliding a cpu. The liquid metal goes in between the IHS and the DIE.

You have a looked at what the Manufacturer says?

Do you know what a manufacturer is?

So how many delids do you have under your belt?

I have done 7. I've taken them back apart and put them back together and mercy me nothing bad happened.

I didn't say couldn't remove, I said " liquid metal thermal compound is very conductive, and once apply must never take it off." which is for delid procedure because of what happened to the video #2 showed, it will affect the temperature because it left stubborn residues behind. Once again, you're reading wrong and therefore deploy a false accusation on me. Re-read what I wrote because there's a different meaning between couldn't and must never take it off.

I met a lot of people that used liquid metal on delid CPU, and they say the same thing once you apply liquid metal onto the CPU die, it should not take it off.

Never, couldn't. Your still trying to side step.

It can be taken off, it can be reapplied. I have done it. The manufacturer says it can be done.

You are wrong.

Then explain on the video why there's residues left behind after wiping it off using isopropyl alcohol. There's a different meaning between must never take it off and couldn't, and yet you are the one try to side step not me. The manufacturer can say whatever it want, but the fact the video proven that liquid metal can left stubborn residues behind make it more difficult to get rid of it, and by re-apply it with previous thermal or liquid metal residues left behind will affects the CPU overall temperature. This is why a lot of people suggested using liquid metal on the die because once you apply it you shouldn't take it off and re-apply it again.

As matter of fact, this thread will say it was well: http://www.tomshardware.com/answers/id-3495322/thermal-paste-heat-sink-liquid-metal-delidding-thing.html

Best selected answer from that thread said : Liquid metal is really for situations where you don't intend to ever take it back apart, which is great for delidding. Between the IHS and heatsink, however, I'd use a high quality thermal paste of some sort. Be careful with liquid metal, it's really conductive and can ruin electronics if it gets into crevices it shouldn't be in

I'm done arguing with you because I already provided my facts and statements and I'll stand with it.

I see what the problem is, you believe everything you see on you-tube.

You should consider a career in politics.

You flat out called me a liar right in this very thread and when presented with proof(as good of proof as anyone on these forums can provide), you couldn't admit you were wrong.

You stated that because of electrical conductivity that LM must not be taken off(your opinion) Not what the manufacturer states and Not what I personally have experienced.

Once again you cannot admit you wrong. This along with your moniker( I think your subconscious was hard at work when you decided to use it) is telling to say the least.

But the fact is you should only used liquid metal on the die not on the HS. Liquid metal was meant to be not taken apart once apply which is useful for delid. I provided with the fact about this, and this suggestion have be around for a decade i believe, and a lot people will recommend liquid metal on the die for better temperature improve than apply standard thermal compound onto the die.

I have provided you with my multiple facts and yet you refused to accept them that liquid metal was meant to apply once not multiple times like thermal compound and the fact is I said "MUST NEVER TAKE IT OFF".and not " COULDN'T". Yet you failed to comprehend the difference between must never take it off and couldn't. As for the part of you said you can achieve 4.9Ghz on air, I believe that statement have to be a lie first because I knew that i7 7700k runs hot as i7 4790k, therefore achieve 4.9Ghz on air seems to be exaggerated, but not until you provided with a picture of it.

The manufacturer say it can be remove, which of course if you can apply it than you can remove it. But the biggest problem is it will left a big junk residue or stain behind make it more difficult to remove it than standard thermal compound and will affect the CPU overall temperature.

Other than that, I'm out. Peace

Okay must NEVER is still WRONG.

 

[biglizard]

Okay so if I apply liquid metal to the die, what if I need to remove the CPU from the motherboard later on? Will the IHS not remove from the die? Will this affect the liquid metal application?

*Also, I right now I have a 10C difference in 2 of the cores.

No, You will have to be a little more careful then normal. But its no big deal if it moves, you can always clean it up and reapply LM.

 

[volkgren]

I reapplied the thermal paste to the die once and of course the IHS a couple times but I'm still getting 10C difference between cores, which makes no difference in max temp from the delid/non-delid on the hottest core. Upon removing the heatsink the thermal paste appears evenly spread. I'm not sure what is wrong here.

The lowest temp core is 10C difference from non-delid and the hottest core is zero difference from non-delid. There is the tiniest amount of rubber silicon left on the IHS, but I scraped it pretty good and it seemed to sit flush.

This is at 4.9GHz 1.35V after 20mins of Prime95 26.6 with memory at 3200MHz 1.35V and Uncore at 4.5GHz

Current | Minimum | Maximum | Average

 

[biglizard]

How are you holding IHS when clamping cpu?

Did you put IHS on with cpu installed in board or out side of board?

I cover die with TIM then place cpu into board socket, I then place IHS. Then I place clamp into place and while holding IHS in place with my finger to prevent it moving I clamp it down.

Are you sure the IHS is oriented with CPU correctly?

 

[volkgren]

I placed IHS on after inserting CPU board, but I moved it up a little so when it slid down it would be centered. I'm sure it's correctly oriented, same way as it was before the delid.

I've ordered some Thermal Grizzly Conductonaut, I'll try your method when I get it in. Maybe do some more practice installs before then with some of my regular thermal paste.

*Do you know where on the die the cores are located? Is core 0 at the top with core 3 at the bottom?

*I found an image showing where the cores are located https://en.wikichip.org/wiki/intel/core_i7/i7-7700k#Die_Shot

 

[volkgren]

After just a couple hours of use the temps went back up to the same as before delid, and then went higher under stress to 95C. After removing the lid to see, I had too little thermal paste and it looked like it mostly evaporated. Anyway, I'm still waiting on my Thermal Grizzly Conductonaut so I'm still using regular thermal grease.

I watched this video and they show AS5 doesn't really help much on a delid.

[video="https://www.youtube.com/watch?v=TCu4IsUEyCA"][/video]

 

[biglizard]

After just a couple hours of use the temps went back up to the same as before delid, and then went higher under stress to 95C. After removing the lid to see, I had too little thermal paste and it looked like it mostly evaporated. Anyway, I'm still waiting on my Thermal Grizzly Conductonaut so I'm still using regular thermal grease.

I watched this video and they show AS5 doesn't really help much on a delid.

[video="https://www.youtube.com/watch?v=TCu4IsUEyCA"][/video]

AS5 is an old school TIM, that has a break in period of something like 200 hrs.

I doubt that this youtuber put the time in needed to show what AS5 may or may not be truly capable of. Clicks not facts.

In my opinion it doesn't really matter, if your going to delid then LM is the way to go.

 

[envy14tpe]

Don't know if I said this before, but delidding and using liquid metal made my temps drop by 10-20C. So worth it. But I may replace the LM in a year or 2. I have heard that that may be necessary depending....

I used Coolaboratory Liquid Ultra.

 

[CompuTronix]

Guys,

Silicon Lottery - https://siliconlottery.com/collections/all/products/delid - sells only delidded, overclocked "K" CPU's, and give the following figures on how much Core temperatures at 100% workload are improved by delidding:

8th Generation ... Coffee Lake - 12° to 25°C
7th Generation ... Kaby Lake X - 12° to 25°C
7th Generation ... Kaby Lake - 12° to 25°C
6th Generation ... Skylake X - 7°C to 20°C
6th Generation ... Skylake - 8°C to 18°C
5th Generation ... Broadwell - 8°C to 18°C
4th Generation ... Devil's Canyon - 7°C to 15°C
4th Generation ... Haswell - 10°C to 25°C
3rd Generation ... Ivy Bridge - 10°C to 25°C

The following graph will provide you with a perspective on the severity of Intel's TIM problem. Shown below are the thermal characteristics compared among different Generations of soldered and TIM’d processors:
Core temperatures on 2nd Generation Sandy Bridge and earlier processors using Indium solder between the Die and IHS have consistently tight gradients within 5°C above IHS temperature, which indicates good thermal conductivity. However, Core temperatures on later Generations using TIM between the Die and IHS have inconsistent gradients up to 25°C above IHS temperature, which indicates relatively poor thermal conductivity and uniformity.

Note: Intel uses engineering samples with soldered Integrated Heat Spreaders for testing and developing specifications.

Delidding requires that you use only liquid metal TIM between the Die and IHS. Conventional silicon TIM will fail in a relatively brief period of time. Here's why:

A process known as “pump-out” will cause silicon TIM to ooze out from between the Die and IHS due to thermal cycling. Intel's TIM, which has relatively poor thermal conductivity, is, however, very resistant to pump-out, as are liquid metal TIM's. As has already been mentioned in this Thread, a highly recommended liquid metal TIM is Thermal Grizzly Conductonaut, which Silicon Lottery currently uses - http://www.thermal-grizzly.com/en/products/26-conductonaut-en

Thermal Grizzly Conductonaut is a relatively new product. Prior to Conductonaut, Silicon Lottery used CoolLaboratory Liquid Ultra.

Here’s a short list in order of thermal conductivity:

Indium - 81.8 W/mk (Indium solder was used on 2nd Generation and earlier mainstream processors)

Liquid Metal TIM (IHS to Die)

Thermal Grizzly Conductonaut - 73.0 W/mk
CoolLaboratory Liquid Ultra - 38.4 W/mk
CoolLaboratory Liquid Pro - 32.6 W/mk

Conventional Silicon TIM (IHS to Cooler)

Thermal Grizzly Kryonaut - 12.5 W/mk
Arctic Silver 5 - 9.0 W/mk
Gelid GC Extreme: 8.5 W/mk
Arctic Cooling MX4 - 8.5 W/mk

Here's a good read: The Truth about CPU Soldering - http://overclocking.guide/the-truth-about-cpu-soldering/

All this information and much more is available here: Intel Temperature Guide - http://www.tomshardware.com/forum/id-1800828/intel-temperature-guide.html

CT

 

[volkgren]

Still waiting on the Liquid Metal to arrive. But I've replaced the thermal paste under the IHS again, this time with AS5 because I don't want to keep wasting my NH-T1. I was a little more generous with it this time. Lowered my clocks to 4.5GHz at 1.200V and Prime95 during the first 30 seconds stays 57-60C during the first 30 seconds. (*Ambient is 20C) I dont' want to stress it any more until I get the LM. But I see how the TIM matters.

Thanks for all of your input. I'm sure I'll post back the results of the Conductonaut.

 

[volkgren]

Okay I've got the LM applied, but I'm afraid it might be too much. How do I know if it is enough?

 

[volkgren]

I cleaned them both up some. Now there is barely some puddle. Should there be any puddle?

 

[volkgren]

Well, there you have it folks. 4.9GHz 1.35V Prime95 (v26.6) 67C max over 2 and a half minutes.

(Full screenshot of HWInfo-> https://imgur.com/PwVSQx0)

 

[biglizard]

5GHZ awaits.

IO and SA are a little high.

You might have applied a little to much LM. Barely any is enough.

I don't use any on IHS.

 

[volkgren]

5GHZ awaits.

IO and SA are a little high.

You might have applied a little to much LM. Barely any is enough.

Well I was taking as much off as I thought wasn't needed. But yeah, I might have left a little more on. Would that cause slightly higher temps?

What can I set IO and SA to in BIOS?

 

[volkgren]

Here's 5.0GHz. It kept freezing and BSOD during P95 until I got to 1.380V. It ran for over 2:30 until I stopped it. I know I see reviewers and overclockers going to 1.40V, but is this voltage range safe for long term if temps are good like this?

Full screenshot: https://imgur.com/HBKFHch

 

[volkgren]

And will the too much LM possibly run and short circuit something? Would you remove the extra?

 

[biglizard]

And will the too much LM possibly run and short circuit something? Would you remove the extra?

Better safe than sorry.

If higher voltages make you feel uncomfortable, then is it really worth it?

I believe that the 1.4 number is for longevity and an added margin for safety. I think Intel says that they are good up to 1.5V.

I was running my 7700K at 1.44, BUT I knew that I was upgrading to 8700K.

At 72c, I would run stress longer and see if is truly stable and if temps stay in line.

Temps gaming?

Edit: You could dab a small amount of clear nail polish or liquid electric tape to exposed portions on PCB for added safety.

 

[envy14tpe]

Here is a chart that shows results of 7700k owners and how they did it.
http://www.overclock.net/t/1621347/kaby-lake-overclocking-guide-with-statistics

https://docs.google.com/spreadsheets/d/1NoxceLMU9dnVev8QmYmBT16fnjwrGkwdIRjTQzzKaVk/edit#gid=0

 

[volkgren]

I adjusted I/O and SA to 1.15V.

Also found it wasn't stable at 1.38V or 1.39V. So I ran 1.40V at 5.0GHz for 20 minutes. Max 73C. I was even browsing the forums while running Prime95, which I like to do to test stability. It will usually crash if not stable. I'm not saying it's rock solid, but I'm impressed with the temps.

https://imgur.com/TUGAQ2Q

 

[biglizard]

I adjusted I/O and SA to 1.15V.

Also found it wasn't stable at 1.38V or 1.39V. So I ran 1.40V at 5.0GHz for 20 minutes. Max 73C. I was even browsing the forums while running Prime95, which I like to do to test stability. It will usually crash if not stable. I'm not saying it's rock solid, but I'm impressed with the temps.

https://imgur.com/TUGAQ2Q

Yep, its a pretty cool process.

Gonna delid my 8700k this weekend and then put it under water, not the best chip but I think I can get 5ghz.

4.9 now @ 1.375 and I wont even attempt small fft.

SA usually needs to be a bit higher than IO, so if you need to make an adjustment to get stable at 5ghz I would start there, Say 1.25.