3570k Isn't Running as Hot as Expected

EarthDog said:

So you are saying that the IHS and TIM have 100% thermal conductivity (when we know for a fact the thermal conductivity of said items already and its not remotely close to that efficient). It doesnt. The temps should be less regardless as you are taking out the IHS and thinning out the TIM with direct contact.

Click to expand...

then again it could be the heatsink like you said before which = higher temps than should have been had, in which case the temperature similarities are purely coincidence, nothing more. Either way that article will have skewed results.

Its not often i agree with ED but on this occasion what he says makes sense, if the alloy heat disperser hasn't got efficient, full and tight contact with the heat source its not going to do a very good job of dissipating that heat.

Regardless, take out a lump of copper that is otherwise in the way the temp should be lower (even if the fit from the CPU cooler is similar on the DIE to that of the HIS)

Frakk said:

Its not often i agree with ED but on this occasion what he says makes sense, if the alloy heat disperser hasn't got efficient full and tight contact with the heat source its not going to do a very good job of dissipating that heat.

Regardless, take out a lump of copper that is otherwise in the way the temp should be lower (even if the fit from the CPU cooler is similar on the DIE to that of the HIS)

Click to expand...

either way the results seem very unlikely+ there is no proof it was actually carried out.

xander89 said:

either way the results seem very unlikely+ there is no proof it was actually carried out.

Click to expand...

yeah..

hokiealumnus said:

They are definitely a part of Intel and Mr. Wood comes complete with an @intel.com email address. I'm not sure their full mission but at least part of it is similar to Yellowbeard & Corsair - going around trying to help with the occasional tech support thread.

Click to expand...

The Intel Enthusiast Team is a group put together to address the enthusiast space (i.e. DIY'er, Overclocker, Gamers etc.). My job is to address the space in a social media role, forums and to pass feedback both ways up and down. Since I don't post unless I have something to add I don't post much on this forum since I am not the best overclocker. I am not an engineer or with technical support but I do have resource to check with them if I need but it can be slow to get an answer that way.


Christian Wood
Intel Enthusiast Team

IntelEnthusiast said:

The Intel Enthusiast Team is a group put together to address the enthusiast space (i.e. DIY'er, Overclocker, Gamers etc.). My job is to address the space in a social media role, forums and to pass feedback both ways up and down. Since I don't post unless I have something to add I don't post much on this forum since I am not the best overclocker. I am not an engineer or with technical support but I do have resource to check with them if I need but it can be slow to get an answer that way.


Christian Wood
Intel Enthusiast Team

Click to expand...

There is also the added onus of being INTEL...thus if you did respond on a whim and were wrong, there are annoying implications there.

Thank you for clarifying, now we know. As an aside, you have come to the right place if you want to become the best overclocker. Just head over to our benchmarking team's section and let the peer pressure do you some good. You might need to look into a partner board if you want to go cold though.

(Just messing with you.)

IntelEnthusiast said:

The Intel Enthusiast Team is a group put together to address the enthusiast space (i.e. DIY'er, Overclocker, Gamers etc.). My job is to address the space in a social media role, forums and to pass feedback both ways up and down. Since I don't post unless I have something to add I don't post much on this forum since I am not the best overclocker. I am not an engineer or with technical support but I do have resource to check with them if I need but it can be slow to get an answer that way.


Christian Wood
Intel Enthusiast Team

Click to expand...

ur not the best overclocker?! but you are part of the intel enthusiast team, poor show, you better get reading

stickies are your friend hehe

IntelEnthusiast said:

The Intel Enthusiast Team is a group put together to address the enthusiast space (i.e. DIY'er, Overclocker, Gamers etc.). My job is to address the space in a social media role, forums and to pass feedback both ways up and down. Since I don't post unless I have something to add I don't post much on this forum since I am not the best overclocker. I am not an engineer or with technical support but I do have resource to check with them if I need but it can be slow to get an answer that way.


Christian Wood
Intel Enthusiast Team

Click to expand...

haha, settle down xander... I'd rather have some1 around who knows nothing about OC but has sympathy/empathy for our community; at this point the OC community is let down due to the poor performance of IB in comparison to SB. Bring this up in an internal email, maybe your weekly meeting, something? Good to have you around christian.

I think what we'll do is collect all the OC.c IB i7/i5, mail them to you, and then you can have the IHS properly mounted. Coolio?

dtrunk said:

haha, settle down xander... I'd rather have some1 around who knows nothing about OC but has sympathy/empathy for our community; at this point the OC community is let down due to the poor performance of IB in comparison to SB. Bring this up in an internal email, maybe your weekly meeting, something? Good to have you around christian.

I think what we'll do is collect all the OC.c IB i7/i5, mail them to you, and then you can have the IHS properly mounted. Coolio?

Click to expand...

i was only teasing

dtrunk said:

haha, settle down xander... I'd rather have some1 around who knows nothing about OC but has sympathy/empathy for our community; at this point the OC community is let down due to the poor performance of IB in comparison to SB. Bring this up in an internal email, maybe your weekly meeting, something? Good to have you around christian.

I think what we'll do is collect all the OC.c IB i7/i5, mail them to you, and then you can have the IHS properly mounted. Coolio?

Click to expand...

IB actually OCs extremely well. The issue is less with its performance and it's OC-ability and more of what is capable with standard cooling.

GeneO said:

The fact that the results are virtually identical is compelling enough for me.

Click to expand...

Actually to me that test they did is proof that solder tim does make a difference.

When removing a soldered IHS, and placing waterblock directly on die with user tim, cpu temps go up by at least 6-8C, or up 10-15C if doing actual direct die with water, threads on xtreme previously linked with results of testing. However when removing the non-soldered, temps did not go up, clearly showing that non-soldered die attaches are less effective than soldered ones.

Problem is your beliefs that any time you remove a layer, you should get better temps. Try removing your heatsink and just blow air on IHS, since after all the heatsink and user tim is just another layer, or does surface area matter. If surface area doesnt matter, no point in reading any further.

The most important interface is tim1, since the die has very small surface area especially since relative hot spots (non-uniform temps).

Take a soldered IHS like E8400. Thermal conductance through die 125-150 W/mK >> solder tim 87 w/mk >> IHS 400 w/mk, now to a much larger more uniform area of heat before trying to cool relatively small areas with tim 4-6 w/mk.

Now remove solder IHS, and place user tim directly on die, temps go UP. Because now you are trying to cool small relative hot spots on die at very low thermal conductance of 4-6 w/mk with very small surface area. Or do so directly with water, again temps go up, on xtreme fallwind showed temps up 15C at stock TDP alone, any significant overclock and got throttling. Cant cool low surface area with water 0.58 w/mk it is futile.

Best temps in order would be
1) direct die >> solder 87 w/mk >> waterblock (now heatspreader)
2) die >> solder 87w/mk IHS 400 w/mk (to spread heat very quickly) before user tim 4-6 w/mk
3)and depending on non-solder die attach used in terms of conductance, bondline thickness, contact resistance, you might do better with removing IHS or leaving IHS alone. In that experiment you linked, nearly same temps, NOT SURPRISING at all.

If you want to prove soldered IHS doesnt matter (you cant it would defy physics), but if you want to show the difference, you need to use solder, not replace intel nonsolder die attach with even worse user tim1. I wouldnt have been surprised to see worse temps.

When one realizes that an extra layer of copper that increases surface area isnt always a bad thing, you can come to a different conclusion. Until then you can equally argue to cut the cooling fins off the heatsink, as they are just an extra layer of copper in way of air getting to copper base, or better yet get rid of that extra copper layer and user tim altogether and just put a fan on IHS.

Well, my testing of my 2500k as compared to my new 3570k has been dealt a setback. The Biostar board has a really flaky (and early) bios version that just barely will get a 3570k to boot at it's stock speed, much less overclock with it! And I do have the latest bios for it, as confirmed by an answer from tech support. I just fired off another question to their tech support, telling them that the bios is buggy as hell and not acceptable. It will not boot if you try to set any manual multi in bios and it also won't boot if you try to manually set the ram speed and timings! Plus, even if it would boot, you can only select up to a 38 multi in bios too.

So I guess I will have to tear down my other Asrock P67 system to do the testing with. And I wasted a day getting solid data for the 2500k on the Biostar system too.

That sucks... Did you try any of the BIOS versions in the hwbot thread?

Mudd, I have found some older biostar boards need a tiny boost in vcore for stability. even at stock settings. not sure if it was chip based or board based, but worth a shot.

@Matt, no, none of those bioses will work in my P67 chipset board. Except to bork the board of course. Which the way I feel about it right now would give me an excuse to shoot it!

@dejo, it's not a matter of vcore Jon, because I tried setting 1.25 v vcore in bios and no boot. It just won't boot if you manually try setting the multi or ram timings. The bios is real old too, dated like 1/4/12 or so. So you know that if Intel made any changes between then and now in the final silicon, the bios doesn't know how to play nice with it. My Asrock P67 Extreme6 boards on the other hand, have a bios dated 4/20/12 and I had zero problems with the 3770K with it (besides heat, which isn't Asrock's fault).

EDIT: BTW, how well do Gigabyte's Z68 boards play with IB? I seem to have read that they have a multi problem too, but know for a fact about this. I have a couple of Gigabyte Z68 mATX boards sitting here if they don't have the multi problem.

Oops, thought you were on Z77 and were going to tear down and go back to P67

muddocktor said:

Well, my testing of my 2500k as compared to my new 3570k has been dealt a setback. The Biostar board has a really flaky (and early) bios version that just barely will get a 3570k to boot at it's stock speed, much less overclock with it! And I do have the latest bios for it, as confirmed by an answer from tech support. I just fired off another question to their tech support, telling them that the bios is buggy as hell and not acceptable. It will not boot if you try to set any manual multi in bios and it also won't boot if you try to manually set the ram speed and timings! Plus, even if it would boot, you can only select up to a 38 multi in bios too.

So I guess I will have to tear down my other Asrock P67 system to do the testing with. And I wasted a day getting solid data for the 2500k on the Biostar system too.

Click to expand...

XE4 motherboard?