Warning: Some i7-6820HKs and i7-6700HQ have Uneven Core Temps due to Uneven Heatsink

Have you repasted a laptop before? If you're comfortable with it then it won't be much of a stretch to go liquid metal, however, if you are wanting to go with traditional paste that's completely fine as well.

Use Grizzly Kyronaut and/or Gelid Extreme.

 

Nice thanks for the update. Can you post some hwinfo64 screen shots? Thanks!

 

http://imgur.com/a/w71sg this is after a match and half of infinite war, max settings 1080p vsync on.

 

Very nice....great temps.

 

In regards to the thermal pads...what thickness is recommended? 0.5mm or 1mm?

 

How recently did you guys receive your laptops? Frank said on twitter that the manufacturing process was fixed 2 weeks ago..

 

I received mine on 31st October (build date 25th). Though I've read people have build dates of 6th November and have the same issues we're all experiencing.

 

I received my replacement on the 11th of November (built in early November) and it's showing core differentials of ~20°C :-(
Therefore, there is at least some reason to doubt the correction in the assembly process.

This time around, I would have actually wanted to keep it due to the CPU hitting 4,0 GHz (OC level 3) and showing somewhat decent temperatures, but the dedicated graphics isn't being detected (neither in BIOS nor in Windows)...
So I guess, I have to go through Dell's support process again.

Or maybe there is someone that is able to give me some hints on what to do when the BIOS is only detecting the Intel graphics card.
Strangely enough, the onboard diagnostics does not show any faults and passes all tests.

 

I received mine on the 9th ordered it on the 4th.

 

If people want to keep their machines even with 20c degrees difference between cores... I understand very well why Dell said only 20+- reviewers have complained about thermal problems. And 5000+ is happy with their purchase. Dell will never ever fix the design flaws, if people don't return back their new faulty machines and complaining.

 

It baffles me that some people think (are content) that a 20C differential in core temps is normal.

Most of those who purchase the gaming laptop's probably never check the temps so they don't know what is even happening anyway and Dell is getting away with this ignorance.

::iunlock::

 

I am afraid... he s right :/

I did the repaste with lm and got a drop of about 20-25c with my aw15. Sadly its kind of a lemon cpu... i cant undervolt it Even 20mv. Freezes as soon i Start anything demanding.

I thought about something else: as far as i understood the Problem with r3 heatsinks its Kind of a pressure issues in the upper half of the cpu. If we could lower the upper mounting Point in which the screw goes, then it could be enough pressure to evening the heatsink? Maybe its possible to "cut" of a bit of the mounting by using a dremel or something similar? What do u think about this?

 

Frank Azor‏@AzorFrank
news coming today or over the weekend working on it now. 1080 on Nov 21st

are we still waiting on this?

 

All this info is in the OP. Yes, modding the tension arm on the upper part of the heat sink on the CPU side does help. So far so good...it is holding up well.

 

You are applying a *negative* undervolt, right?

It will read -20mV *not* 20mV, the negative sign is important

 

It will be interesting to see how much faster the 1080 model performs than the 1070 model, or will it be like the Aorus that only performs about 2% faster Graphics Scores between the 1070 model and the 1080 model.

Aorus X7 DT V6 Desktop GTX 1080
Laptop Review / Benchmarks
http://forum.notebookreview.com/thr...scussion-thread.794965/page-323#post-10386438

Hopefully the Alienware 17 GTX 1080 model will come with a 330w PSU, and have enough power to outperform the 1070 version enough to make it worth buying.

 

There are still people reporting the problems even now.

http://forum.notebookreview.com/thr...-uneven-heatsink.797477/page-36#post-10386501

If you don't want to chance getting a bad one, and don't feel up to repairing it yourself, you should expect to interact with AW Support for an RMA in house or sending back to AW for repair.

You could wait and look for positive owners reports before ordering.

 

You can call Dell/Alienware Support and ask if they can send a brand new heatsink which has "upgraded" attachment mechanism or a so-called brand new "flawless "one. Need to fix a brand new heatsink yourself with a bandaids half fix (modding the tension arm), isn't the correct way to do.
You could call AW support.., Tell them that Alienware CEO Mr. Azor said they now have fixed the heatsink flaws on the new Aw laptop models and you can replace the heatsink by yourself. They only need to send it to you. Or send the laptop back for a replacement or refund!! You have paid a high price for a working laptop out of the box. Not a flawed one who needs modding so the new laptop can work as expected/intended.

 

It shouldn't freeze if u apply more voltage afaik ;-) so yes I meant -20mV

I think he s referring to the post from umar, where he states that 95c are within the specification

 

So what exactly do you guys suggest to do now? Just return it since it seems like they dont want to acknowledge their mistake or did they actually fix the heatsink so i can just get them to replace the heatsink?

 

Return it. Then buy the 17 inch.

 

Unfortunately it isn't as simple as that for those of us wanting either portability or the 120Hz panel (or both). The 17" suffers in both respects.

For me I'm returning my unit and waiting until Kaby Lake. I'll just have to see what the heatsink is like by then...

 

It's not that unusual for someone to do undervolting the first time, and end up increasing voltage rather than reducing it, that's why I brought it up

Yeah, I had a G750JH that would only do -15mV the first time. It took me a week or more to get -20mV, months turned into more than a year, and finally it was running stable at -45mV.

You can push the CPU / GPU tuning from wherever your laptop first settles, but you need to be patient, consistent using small increments, and relentless

Sometimes it gets worse too.

 

Call AW support and ask when a new heatsink with enhanced / improved tension arm on the upper part of the heatsink will come so the laptop cooling can function as intended. Or ask Frank Azor‏@AzorFrank - Umar on tvitter or other social media

 

Why doesnt everyone on here just get the Eurocom Tornado F5, not need to mod anything and comes with a fully working processor.

 

Maybe because if taken care of the problems the Dellienwares perform admirably like @iunlock clearly demonstrated!

Regardless of the brands: at least a repaste is needed and, in the worst case, some minor hard mods like bending something, etc.

 

LOL bending something to make it work properly in a brand new machine is not really a fix for most people...

 

Jupp, I know that! But the (sad) truth is that there is NONE which runs perfectly stable, cool, awesome kicks a$$ as intended out of da box.

Never owned that one, but after taking care of the issues they performed good - admirably.

That's what I meant to say.

 

True, but when have you ever had a machine where it ran 100% out of the box like it should? The whole point here is to address Dell's screw up and to provide a solution for those who are wanting to try...

I've clearly said...RETURN IT and SHOVE IT IN DELL's FACE....and the ONLY exception to keep it is if one has a real nice chip, because those are rare to come across.

The fix so far has been holding up well....only time will tell, but I am also experimenting with something that can provide a more permanent solution to ensure that the tension arm stays put. (Not that I doubt that it won't....)

Also, just to point out, that tension arm isn't like it's easy to bend. It's pretty meaty with a good amount of force so I'm skeptical that the heat alone is enough to bend that arm severely like it is being assumed by those who haven't even experienced this first hand.

It's easy to speculate, we're all guilty of that, but the reason why I along with others report on this issue is to put the theories to rest.

So far so good.

 

The screen isn't as good as the Alienware 120Hz one, the battery isn't as large, and most importantly for me, it's not available locally in Australia which means I can't salary sacrifice it for tax savings.

 

Maybe because the Alienware name sounds so good in your ears? Remember the famous older AW models that was shipped with socket hardware...? Perhaps people associate the new AW models with these older ones and expected same quality and best performance a laptop can give? Nothing last forever. You don't needed to mod any heatsink in those AW models. Only a re-paste and no other brands laptop models showed same good performance. Just ask @Mr. Fox

 

The good old days of AW are gone and the only thing keeping them alive hanging by a string is the 6820HK, Build Quality (less heat sink mounting design) and Pascal.

If, AW ever loses the unlocked CPU, it may be the end for good for a lot of us.

So AW if you're listening, don't be a FOOL. MAKE SURE TO KEEP AN UNLOCKED CPU IN YOUR LINE UP AND DON'T FLUFF UP THE BIOS. NO FUNNY BUSINESS...

BTW the new BIOS update is a joke. DO NOT DOWNLOAD. There's some funny business going on there...it looks like they've messed with the power limits and other strange limitations deep within the BIOS that prevents old OC profiles to be applied without being limited.

I just flashed back to 1.0.2 and so far back to normal.

 

And wind up like me with a 17 inch /6700HQ that has uneven temps....

 

Almost all the hardcore overclockers/performance enthusiasts have jumped from the AW ships after the introduction of the BGA models!! This can tell you a lot. I don't need to say more!!

And the bandaid fix for the "delicate" heat problems, is most likely the Bios Upgrade you tested


Last time Dell's engineers created a similar bandaid fix... Was the Hybrid bios for the early 2015 AW models!! Yeah, Times has changed!!

 

They are not worthy to be called engineers. More like fools that are experts at crippling software.

I challenge any dells so called engineers to to justify these ridiculous temps in front of my face. Yea...exactly. Facts are facts....whoever is behind this madness of poor QC and ridiculous circus stunts behind the software, you need to be fired.

 

That's too bad. Most of what we've been focusing on are the 6820HK's, not the 6700HQ.

The fact is, these gaming laptops still fall short of being able to run them straight out of the box. If you stay stock...maybe....but if you don't plan on repasting these gaming laptops, then from the facts that are being presented it would NOT be wise to buy one.

Yes it is a bit of an oxymoron, but if one can't or plan to repaste with some help, then don't buy these things.

 

It's extremely unlikely to be the fault of the engineers themselves. For one, from the sounds of it the heatsink and cooling design it's actually extremely good, it's just the implementation that's failed. As for the software engineers doing the BIOS hack, it's most likely a product manager saying 'we need to get these temps down; can anything be done in the BIOS?' and the engineers were honest about what could be done...

 

I can be the first to tell you as I've been preaching it from day one upon receiving the unit that the heat sink is great and not a gimmick. Tested and proven.

What I am implying here is how ridiculous anyone at the department in charge of the heat sink mount design can justify such rubbish going with a 3 screw mounting system on a square heat sink that's not even rectangular.

I'm just feeling for all those who have fell victim to this madness.

 

Stock on a K series chip is a waist... If there's not the room to OC it thermally why bother. Sadly most people who buy them will not repaste or overclock, little own bend brackets and lap there heat sinks.

 

If Dell's software engineers continue to use the same new crippled power setting changes in newest Bios/Ec upgrade that @iunlock discovered... Forget upgrading to newer bios later. You are screwed!!

 

As I mentioned before, I don't think the fact that it's only 3 screws is the issue; the Razer heatsink is also only 3 screws, but the screws go through a solid plate, so the pressure would be easy to keep even. The issue here is the difference between the bottom mounts which are quite well attached vs the top one which has a lot of flex. If the single mount was fixed and the two bottom ones had more flex it might not be so bad, but as it is it just doesn't work well.

 

So, you're saying that "normal" 90°C can't melt steel beams?

On a serious note, how bad is the design on the AW15r3 vs. AW17r4? Seems everyone talks about how terrible the AW15r3 is, but not much talk about how the AW17r4 compares when it comes to how bad the tension arm is in that model.

And why is the issue seemingly more prevalent in the 15" model, than the 17" model? Or is it just simply because the AW15r3 is much more popular? (i.e., larger sample size)

I hope my AW17r4 from HIDevolution skips the madness, as it's getting repasted, and tension arm adjusted before I even get it. Regardless I'll be undervolting my i7-6820HK

 

Isn't the razer hear sink plate rectangular? In that case a 3 mounting point design would work since it is proportional.

What I'm referring to is the idiotic design flaw of a 3 mounting point design on a square plate.

::iunlock::

 

Hello again!

I finally sourced some Fujipoly 17W/mk pad (that ships to Canada) and should be here in the next couple of days. Very excited to get this swapped out. Instead of going with liquid metal (for now), I picked up some Gelid GC-3 paste which should still be a million times better than the stock paste. I will do some before and after stuff to show the difference. I am not going to mess around with the heatsink for now just to see what it does as-is.

 

Just a wild guess here, but is this somehow related to the issues here?

See:
http://www.tomshardware.com/news/evga-addresses-geforce-1080-temperature-problems,32967.html

 

no evga had another problem. their heatsink isnt even touching ram and pwm's. The one here is touching but it isnt tight

 

Remember? I still want them back. I still remember the m18x r2 I had with 6.5tb of storage. It seems every laptop that I get after that heavy weight, I have less storage. The current one, Eurocom F5 only has 3tb, before that, p370sm3 had 5tb.

 

Reading through this thread has me wondering a few things.

1) Is it worth buying replacement thermal pads? How much of a difference would that make?
2) I repasted a few weeks ago, if I buy thermal pads, should I repaste again to make sure there are no bubbles or gaps? I'm fine with doing this, just curious.
3) When I monitor my temps on my processor they're relatively close to each other (within 5 degrees) however once in a while, for one reading (not even a full second) one of the cores will spike to possibly a 10-12 degree difference with the lowest temperature core. How is that possible? I know what's going on with the heatsink, but I don't get how it could go from like 45 to almost 60, and then down to 45 again within 1 second. Is that possible that the temperatures are being reported wrong? Or is that the issue that everyone is having?

Thanks everyone!

 

Sounds great! Keep us posted...

::iunlock::

 

1. YES.
2. YES.

3. It's due to the uneven contact on the die-> Heat Sink and the thermal paste is thinner or thicker on that side to where after it reaches the thermal pastes threshold of thermal conductivity, then it'll spike.



::iunlock::