How to FIX AW17R4/15R3/13R3 CPU Core Temperature Differential Issue

That they kick on late is indeed an AW problem. But I mean if you just look at the max temp section in HWinfo a person might be fooled if he/she takes a look at the temeprature history in XTU and only see that peak for a second during full load.

 

3 laptops, both 6870HK and 7820HK, and yes, i went through all 15 heatsinks from earlier revision to latest revision, all show more than 10-15C some even 20C temperature difference once you unlock and overclock. The only way to reduce the temperature difference is to reshape the heatsink metal. Or use liquid metal.

If you don't unlock and overclock, yes some are tolerable, under 10C difference.

 

Like I said I saw all latest models go under 5c difference even on HK systems and I never used liquid metal myself (which is just not tolerable but excellent in general). Always ICD7 or Kryonaut. I also never bent the CPU arm, always started to tighten screw 1 first while keeping pressure on the heatsink to keep it level, after that tighten screw 2 and 3 (according to the CCI heatsink). You are pretty much the only user that I see with complaints regarding cooling AFTER repaste.

 

to be honest, you are pretty much the only person claiming cpu temperature difference is well under control with an unlocked and overclocked HK CPU using traditional paste. you need to read more from other users in this forum. if i remembered correctly, you previously had a quarrel with another guy who strongly disagree with your claim, so i'm definately not the only one.

you have very limited experience with a HK CPU (if any) and your HQ CPU is a walk in the park for alienware cooling. an unlocked CPU at 100W PL1,2 is at a totally different level playing field.

 

Now you make assumptions about my experiences. My personal laptop is not the only system i touch by a long shot. Nedoo didnt even believe my screenshots and started trolling. Thats not even disagreeing. From my experiences the people on reddit are far more positive about recent systems. Also do you really believe your cpu pulls 100watt? XD the vrms cant even push that much clean power.

Also check your own poll results. Not all people have issues even on a forum which primarily exists for troubleshooting. Statistics 101.

edit: also do you know why I never opted for a HK cpu? Because even in the Asus ROG watercooled laptops they tend to run TOO hot. It takes a lot of effort to keep those cool and unfortunately, the majority of the manufacturers cant keep the HK under control. They become enormously inefficient after clocking them higher, CPU lottery is also part of the problem in this case.

 

And none have the number on how many owners have flawed products in their hands without knowing. Not all out there have proper tech knowledge. A hell lot people think Intel's processors can handle up to 100C and thats it. Aka work as intended.!! And not all post on forums. The HS design is flawed. No doubt about this.

 

THIS.. LOL..
My cousins laptop was a 15r3 with 7700HQ 1060 and 10C core difference and was bought last year Dec 2017.. I fixed it with repaste and repad with loads of gelid extreme paste.. Thermals from 95 on core 0 and 3, 85 core 1 and core 2. Now with repad and repaste it went down to 70+ with 5C to 7C core difference.. I am sure it is still uneven but got compensated due to loads of thermal paste.. LOL.. also -0.150mV undervolt.. Another friend bought(Jan 2018) a 15 r3 with 7700HQ 1070 and still the same core difference of 10C.. Fixed it through repad and loads of thermal paste.. LOL.. so from my experience still have core difference problem.. Also the main key of iunlock's guide was that all pads on his guide was just a reference and cannot be followed religiously..If I remember correctly this hint pop up when alexnvidia was moaning about the problem which helped alot.. Balancing the heatsink with the thermal pads is extremely difficuly for me.. maybe because I am new to this.. BUT just want to emphasize again that the main key to help me with the problem is that you must check the thermal pad size yourself and not follow iunlocks guide religiously.. For me Both iunlock and alex has been really helpful.. Also both side critizing each other is also helpful from my point of you.. keep up the good work guyz.. and please dont take things so personally.. your arguments are helpful.. THANK YOU THANK YOU THANK YOU..

 

@rinneh is only known living person to have 63 C max teps on stress testing 7700HQ...legend is there was another one but he miseriously disappeared

 

You might hit up Papusan's BGA collection, even it would be below 50C during stress test?
@Papusan Am I correct?
Under strict temp. controlled condition, even OCCT maxes out at 60-65C at stock fan, with aggressive fans or full speed it will drop more.
Earlier I wanted a super silent gaming PC but now if I don't hear any fans spinning, my intuition says there's a problem.
I was able to undervolt my cousins 7700HQ and even it barely broke past 65C. Without UV, it hit 97C instantly on HP Omen 15.

 

Well the poll results already show it is clearly not every machine. Reddit users are also more positive. But i doubt that issues would fly under the radar for every user out there.

 

Pst my friend peace.. not to offend you or anything.. I think last week I was lurking around reddit to look for the bios 1.4.4 problem and most people who was in the forum was having a problem. I think it was you who mentioned that people that dont have a problem or not aware of the problem dont usually go to forums.. So don't know where you found that people are happy with there alienware laptop.. OH i did remember I only like 2(or more) out of 100 from people's post that were saying that they dont have any issues.. But generally people are having a massive problem with this tripod mess and weak tension arm of the heatsink(was really surprised as the 17 r3 was different, felt like the new heatsink model are cheap).. Also just to add up.. I have a friend who lives in Australia and we play PUBG.. he always complain about spike lag or suttering.. When we did a screen share and asked him to install HWinfo his laptop was a mess.. heat problem and uneven core temp.. but this was not a alienware.. it was a dell gaming laptop..but my point is he is one of the people who buys a laptop and is not aware of the problem the manufacture giving to there customers.. Hopefully though with all this dell non sense.. I hope I hope there sales will go down abit.. even the slightest so they will realized that they need to fix issues mainly QC and change there toothpaste to a proper paste.. but it would be like a miracle tough.. hehehehe.. Anyways again iunlocks thread and alexnvidia thread are extremely helpful and also the people in the forum brainstorming to solve the problem and make newbies aware of this problems.. I will keep lurking and learning.. LOL..

 

There are about 100 people at most on this forum that complain that they have issues (I think less, just a large estimation from looking at the polls and general topics). There are a vocal minority complaining but its nothing compared to the actual sales numbers which run in the hundreds of thousands. If it was such a widespread large issue it would already have destroyed the brands.The ironic thing is that I had a larger core temp differential with my previous alienware than my current. I dont want to say that the people that do have errors are lying. They are obviously not and I hope they can resolve these issues. But to think that no other brand has QC issues would be plain wrong.

It shouldnt be needed to repaste unfortunately with 9 out of 10 gaming laptop it is the case however and we have to blame intel for that with their large variation and just out of this world operation specs which arent healthy. No CPU should run at 100c max. If they state that 80c shoul dbe the max, manufacturers should act accordingly.

But yes I am lucky that I did not receive a machine with large core temp differentials. But it would be naive to think that I am the only one with such a machine.

 

Hundreds of thousands? Nope, I don’t think so. Alienware is small in the Gamingbook segment. As I said in previous post... Up to 100C is allowed for Intel processors. This is the target... Stay below shutdown temp. People are so different. Not all have tech knowledge. A Hell lots computer users complain only when they get problems like BSOD or shutdown due overheating. Not before.

 

I am new to this gaming laptop thing.. hahaha.. Used to use a desktop.. Don't really know about sales and etc.. Anyways just wanted to know how long does your dell toothpaste thermal paste last before you need to do a repaste? I am just curious and envy you as your one of the lucky person who has less issues with Alienware in the forum..

 

OEM’s thermal paste is equal in design as the thermal paste Intel use between die and IHS. Meant to last long(If the cooling work as intended!!) But not designed to give best thermal results.

 

Well I am going to repaste next week anyway. The lower the temps the longer my laptop will live. But in the end yeah I have been lucky. Started out with an AW15R1, which had a battery problem, but the technician that wanted to replace it ruined my whole laptop. So I received an AW15R2 instead. Which ran pretty good but had the DC in cable problem after a while even though the cable was already replaced and the motherboard developed an AlienFX problem. In the end got a refund and bought the R3 which I use now. Very happy with it. The thing is I tried so many brands and AW is not perfect but it served me the best and they have good service in my country. Used Gigabyte, MSI, Samsung etc before and they all dont offer this support while those laptops also develop issues over time. The Gigabyte was replaced 3 times within half a year by Amazon after that they gave me a full refund. Only Dell, Alienware and back in the days Sony served me well. MSI was just poorly made in my opinion in general.

 

I remember having seen sales numbers somewhere. Will try to find them again. Alienware seems to be huge in China with even dedicated Alienware shops.

But even Razer managed to sell a huge amount.

 

Undervolt is cool

 

Update on my situation. Both shots below are at 4.1ghz with a -80mv undervolt. The first image is how I received the laptop and the second is after a Kryonaut repaste and curing time.

Before repaste/repad after < 1min Prime95 blend (Throttled once it hit 94c):


After repaste/repad > 10min Prime 95 blend (No throttle!):



Still not great, still have 10c difference (I had it down to 6c difference by loosening the back 2 CPU mounts across from the single arm but I wasn't comfortable making a bigger uniform gap).

It does not throttle now. Witcher 3 runs no problem with onboard 1080 running on maxed out settings (temps end up being similar to the Prime95 run).

I can run 4.1ghz at -110mv but I left it at -80mv for this comparison to be equal. I think I'm going to call it quits on working with the core differentials... I have liquid metal but I'm not comfortable using it with that temperature spread and the amount of movement my laptop gets. Considering my voltage requirements spike after 4.2ghz I'll settle at 4.1ghz with a nice undervolt.

Thanks for your help everyone. I really hope Alienware can fix this and produce even temps out of the factory. I will be watching and will not buy another Alienware unless I hear the next generation has fixed this issue. It kills me to see them say they fixed it in an AMA on Reddit just a few weeks ago. They obviously have not.

 

They had to. Less sales doesn't pay their bill. What's correct doesn't matter.

 

Beware of some forum members stating the same...cough...rinneh...cough...

 

they have not fixed the cpu temperature differential issue. period. the extremely rare few lucky ones got their system with little temperature difference, most if not all are due to weaker cpu and those who never unlock and overclock the cpu. otherwise for 99.9% of all the system that came out of their factory to this date are all plauged with this issue.

 

I ask again. But where is this statement based on! 99% of all system. Did you check all those hundreds of thousands?

 

Hundreds of thousands, where did this amount came from?

 

Sales numbers from almost 3 years ago are 100000 https://www.notebookcheck.net/Asus-and-MSI-account-for-50-of-gaming-laptops-sales.156832.0.html.

They have grown a lot during those years according to paid reports. The gaming laptop has a very strong growth past 2 years so they expect them selling over 200.000 laptops past year. Alienware themselves has not released any individual sales numbers but they are number 3 in the industry overall.

 

When I see you talk about numbers I go

 

After even a mod corrected you, you still dont learn. Also you have trouble reading whole posts. Check the expected market growth for 2016. And after that 2017. You can make a nice estimate based on actual market research.

 

Mod didn't corrected me, he corrected you.
He just warned me to be less harsh on you.
Alienware has big share in the market, that is true, and it obviously reeaches few hundred thousands units sold.
Still after we count numerous customers returned-exchanged three-four units it could drop

 

I think you live in a different universe. Also i saw your post after that and got deleted as well . Your like trump when it comes to bending truth. He said the same as me you can read it back if you need. Count from zero

Also as far as i know. Sales numbers do not incorporate exchanged units. Only units sold as in from invoices etc. Business reports often show sales figures next to income, profit and AS costs. Based on that you can create an actual and factual conclusion on how succesful they are. Only razer released a report on that because they released an IPO on the Hong kong stock market and in their case they made a very small profit because of high AS costs.

 

Works great for me

Sent from my SM-G935F using Tapatalk

 

Exactly, I've fixed dozens of these units and always with the same heatsink it came with. I've never seen a reason to get hold of another if it's just a little bit misaligned

Sent from my SM-G935F using Tapatalk

 

Ask MOD on that matter if you like

It's just amusing to tease you once in a while...sorry
I'm just waiting for release of next gen AW laptops anyway...I hope it will happen in less then 4 months...

 

It baffles me how much you can go on about that by the way. Since i jsut state the core difference that I experience myself. Maybe it would be beneficial for everyone to discuss that. Also because others want to know it as well. Frank Azor stated they will launch new models as soon as Intel releases the 8th gen HQ CPU's.

 

Well I double on that, it baffles me how stuborn some human beings can be.

Quoting @rinneh:

"My hottest core is core 1. But only 3c higher than the others.for example core 0= 61c core 1=62c and core 2 & 3=60c."

Now he tries to presue me I cant count:

Core 1=62c

Core 2 & 3=60c

So 62c is 3c higher then 60c???

60c+1c=61c 61c+1c=62c

Now he was posting comments like I'm stupid for saying 62c is 2c higher then 60c not 3c.

Note: he said 3c higher, not within 3c difference...

When I said it is false, his ego was so hurt he tried to ridicule me, and reported me numerous times "for questioning his math "skills""...
After that he said my English is terrible (translates to STFU!!!)...

Why an I spending time on such pointless argue?

Because there are some forum members so arrogant if you point to any false info they wrote they will try to " bully" you, and I just can't stand "bully" behaviour...

 

Again if 60 is 1, 61 is 2, 62 is 3. 3 degrees variation spread over 4 cores. Can i make it more clear than that! Can you sleep now? I am counting the difference while including the baseline and you do not. That's all there is to it and you throw a hissy fit about that?

Talking about arrogance and bullying while constantly trolling and trying to discredit my proven results while pointing at reviews that are a year old.

This whole discussion is truly pathetic.

Tomorrow I just block you, giving you time to read my reply. spend already enough time that I just had at that time. Wont waste it more on you. Feel sorry for the mods that needs to clean this crap up.

 

You are on the right path.
If 0=1 you are right, and if 1=1 you are wrong, you couldn't make it more clear...now go to sleep, but please, just please dont count sheeps

 

Repasted twice this morning. One time with spreading the paste myself and after just a drop and let the heatsink pressure do it's job.

Very weird behaviour regarding the temps.

First of all in my case Dell used actual thermal paste and not a stamp for the CPU, Check the attached photo for details. As can be seen in previously posted Prime95 stress test photos I had a small core temperature difference with this paste. But From the photo it seems that there was uneven pressure since the top has more paste on the heatsink compared to the bottom.

After repasting with Phobya paste I noticed a larger temperature gap in Prime (7c) but only in prime. while playing heavy load games such as Assassins creed it was the same as before. But the overall temperature are lower when both the GPU and CPU are stressed. Now I just wonder why the core temp difference is larger. Even though the maximum temperature is still very good. Maybe Phobya paste needs more time to settle.

Also I recommend anyone not to remove the mid frame with a pry tool such as a creditcard or dedicated plastic pry tool. Lost 2 of the clips on the bottom, the official Alienware diassembly video is showing another method of just pulling towards you while grabbing the top.. Doesnt seem to do anything since there are 2 screen holes right next to them but it does itch my OCD. So ordered a new palmrest to replace that later.

 

They still use a stamp like rectangular pattern. Same method is followed by all vendors.

 

I dont think the CPU side was a stamp, compared to the GPU side it feels like actual thermal paste but it also doesnt has those straight stamp like lines.

 

From your pict.... It seems Dell have removed the pads in the Orange box on TRIPOD as seen below. Or is it on components ?

Long time since they switched to paste on Cpu HS. The so called fix.

 

The other 2 pads are on the components.

But the paste was quite soft compared to the stamp which crumbles when you try to remove it. Have to try other screw patterns to remove the thermal gap I think for the heatsink even though it seems to make good contact.

But so annoyed about those 2 clips. Even though they dont seen to do anything, it just irritates me to know end knowing they are not there now. New palm rest is going to cost 40 euro's and have to fully dissemble it. So for the other users, DONT USE A PRY TO TO UNLOCK THOSE CLIPS! Just pull upwards. First time I ever saw those plastic tabs/clips break.

 

Well, I figured I'd throw my two cents in, for what its worth. I picked up a 15 R3 a few weeks ago after my Aorus X5V7 went belly up. The choice was based on Dell's reputation and my personal experience in getting my wife's XPS 13 fixed after it had some issues. Here's what I can add to the discussion:

My unit does not have the HK processor. It is bone stock except for an additional PCIe SSD I added (because the 128 GB primary it came with is too small). I have considered a repaste but haven't yet because I have no problems with thermals. No big difference between cores; the last time I checked using CPUID and HWINFO it reflected a difference of < 5 degrees between cores, which in my experience is within tolerance for longevity and performance needs. I haven't hit it with OCCT, Furmark, or Prime95, but I do a fair bit of rendering using Maya and Blender that does hit the CPU fairly hard. Temps level off around 80C after sustained workloads of 30 to 45 minutes. Gaming sees similar results, depending on the title. Ghost Recon Wildlands hits the CPU harder than Rise of the Tomb Raider, for example.

I don't think I'm a statistical anomaly either. Here's my logic for why I believe that to be the case:

Dell is the 3rd largest segment owner for laptops worldwide. The market for laptop sales is around 150 million globally in 2017. Their market share is roughly 16%, which is around 24 million units total. Obviously, that isn't 24 million Alienware laptops but we can conservatively assume it to be a single digit percentage, say 5%. That would be 1.2 million Alienware laptops sold in 2017 worldwide. That may seem like a lot but considering market sizes in the west and east, I don't believe that number to be too far off. The poll on this thread had 83 respondents with 49 stating they had a less than an ideal unit, or roughly 59%. Now, if that percentage was representative of the 1.2 million sold in 2017 and well over half having annoying to serious issues with thermal management, I would imagine a class action lawsuit being in the news, which we don't see.

Here's what we can learn from the data; the number of complaints must be viewed in the context of the number of units in the wild. Forums like this typically exist for owners to find solutions to problems they might be experiencing with their product, not to rejoice over how awesome and flawless it is. Meaning, more often than not, the posts found on Reddit, here, and elsewhere are not representative of the number of units that are working fine. It is a sample that is nearly entirely comprised of individuals that are having problems (or in this case 59%).

It isn't that @alexnvidia is right or wrong, or that @rinneh is right or wrong. The sample, in this case, is loaded because of the audience and the reason why those people have found this thread. Most folks are looking for solutions to the same problem and congregate; they are not a fair sample of the total product owner base. I found this thread because I've been toying with the notion of possibly fixing something that isn't broken and wanted to see if it was something I would benefit from.

I don't think the 2016 Alienware model line is critically flawed. There is, however, some sort of manufacturing issue that results in an X number of units being assembled incorrectly. I say this because nearly every person on this thread has stated that the thermals are manageable after repaste, even with units with overclockable CPUs. That would indicate a less than ideal assembly process, not necessarily a flawed thermal solution. Rather than looking at who manufactured the heat sink, I would look at where Alienware laptops are assembled. Something in my gut tells me that these are likely put together at two or three different locations (maybe more) and one or more of those locations has shoddy management or overworked labor that aren't taking the time to do this part of the assembly correctly.

FWIW.

 

So did anyone try out the aftermarket heatsink?

 

More paste at the top than the bottom = higher core temp on the core #1 and #3 and lower core temp on #2 and #4.....right??

 

With the stock paste on the photo I did not have a large core temperature difference. Only about 3c but not in the usual pattern. Core 1 and 2 where a tad warmer than 3 and 4. Now I have the usual pattern after repasting it. But the overall temperature is lower. But core 3 and core 4 about 6c lower but only in prime 95. In Assassins Creed Origins which I use as a real life situation (also shows about 100% CPU usage and 100% gpu) there are no core temperature differences. GPU is reaching 70c. So I leave it be for now since the overall result is better.

I would if they sold a 15inch 1070GTX version.

 

I can only agree with this. Also a lot seem to forget the time factor. Earlier units where definitely worse than current units that come from the assembly line.

 

I'm really curious about it but I haven't tried it yet. Would like to see some testing first.
Anyone willing to be a guinea pig?

 

If anyone finds good aftermarket heatsinks I'll give it a shot too.

 

Their on ebay in fact I'm working with a seller to make a heatsink for me and he also makes heatsinks for the R4. Might be worth trying, using 15 different sets of heatsinks from Dell seems like a complete waste of time. Id be looking for alternatives after 2 lol

https://www.ebay.com/itm/CPU-Vacuum...fff6c0eb|iid:1&_trksid=p2481888.c100675.m4236

 

___________________________________________________________________________________________________________________________________
http://forum.notebookreview.com/threads/official-alienware-13-r3-owners-lounge.797884/page-687

Just a quick follow up after a month with the above re-paste job.

Luckily, I still experience EXACTLY the same 0-1 C core temperature differences and the same good overall temperatures (maybe even better) after a month of heavy use.

I write this because a month ago this was my 4th liquid metal re-paste and this was the first time I took the processor thermal pads off. Before this last re-paste, after re-pasting with liquid metal the core differences/temps would slowly grow after few days.

Basically my theory is that if I had no 'perfect touch' between dies and heat-sink while heat-sink being unscrewed the LM would get pushed to the sides during reassembly and then dry during use (heat-up, cool down etc.) the gap between dies and heat-sink would be 'pulsating' up and down. I remember that LM would simply look DRY with 2-3 small liquid spots in a corner of the die after a week of use. Definitely the 2 thermal pads were causing all this.

During my previous re-pastes I used 2 stripes of thin paper, placed them on Proc and GPU, screwed the heat-sink ON to see 'proper' contact by not able to move them at all. Now I think that even the stripe of paper is a gap itself. Really, while unscrewed ... heat-sink VISUALLY not going down and no sound produced when tapping on top of Proc and GPU is so far the best way to get a good fit and stable LM job. It includes a lot of ALL AXIS heat-sink twisting but once good contact is achieved it pays.

Imagine. You have a perfect contact without tightened screws. Tightening them should take care of any gaps + LM >>> way to go

Peace from city of coffee_\|/_.

___________________________________________________________________________________________________________________________________
+1 to this.

I had huge core differentials (over 20c) and they completely went away once I removed BOTH the thermal pads from above the CPU. Going to replace them with this: https://www.amazon.com/gp/aw/d/B00K...95c-4720-be21-c25f9965f6be&pd_rd_i=B00K04D3UK (thanks to Papusan for the link). If anyone else has issues with core temp differentials, the first thing I would try is removing the thermal pads and briefly testing to see if differentials improve. If so, just use that gel, re-paste, and you will have major temp improvements compared to the stock setup .