They look pretty beefy to me.
I wonder though... Does the processor die being so off center on the copper make any difference? NCK's pics clearly show the CPU side off center where it makes contact with the copper...
-
Sent from my Moto Z (2) using Tapatalk -
Sent from my XT1650 using Tapatalk -
If it's the same design, not sure how it can fix the problem. They have to adjust something even if it's padding.
-
"100c is the new 80c" wow just wow not ok
Sent from my XT1650 using Tapatalk -
Last edited: Dec 10, 2018Ashtrix, Darkhan, c69k and 1 other person like this. -
propeldragon Notebook Evangelist
-
Guy I got the laptop for mainly work with Linux but if you are interested in testing just point out what can I do. I am not a gamer and I've never using bechmarking software.
About the Linux side of the thing: does anyone know how to make the hybrid gpu to works? I spent the last three days reformatting the machine while trying to run on the intel GPU. Only the Nvidia card seems to work correctly but drains the battery in 2h (like a gaming session). -
-
Could that line/rib/bump that runs along the left side and bottom of the CPU side of the heatsink cause an issue?
Maybe it’s raised enough to compromise contact. The top of the die to the surrounding PCB is minimal as is.
Perhaps it’s nothing or this has already been discussed.billycuth likes this. -
Mech G2
(Tongfang GK5CN5Z / GK5CN6Z)
As for previous the Mech 15 ('G1' / rev1) you;re asking about:
Well all the signs show that older heatsink was made by the same engineer at Tongfang who also made the newer G2. And therefore it sports a very similar design in regards to the cpu side, the 2 parallel heatpipes.
Well here are some more annotated screencaps. For the other 2 machines we are discussing here. Again sourced from a previous teardown video, and other public source(s) online. And then drawn all over
Mech 15 - 'G1' / rev1
(Tongfang ??GK5CN67??)
So that also has the Triangle 3 mounting holes design. Although at least the 3rd (leftmost srew hole in the image above) is approximately positioned in-line with where the CPU die is actually located.
Alienware m15
(OEM unknown)
I think you need ask yourself this:
What other factors at play here? Where are the 6 IA cores precisely located underneath the cpu heatsink? The 2 rows of 3 cores. How much lateral distance (the first few mm) does the heat have to travel *sideways* from an IA core, to get itself into the nearest heatpipe(s) ?
You ask about the z height of the copper transfer block. Because z-height is something that permits lateral (sideways) spreading of heat, into a larger area. However all of these thin and light notebooks, the z height is always going to be a pretty constrained resource. Because we have to first subtract the thickness of the outer case, the thickness of the mainboard PCB, then the carrier PCB (of the bga chip) which the silicon die sits on. Then the thickniess of the die itself. Plus the thickness of the flattened out heatpipe(s) placed ontop. And that's already some significant minimum thickness. And we have omitted entirely the consideration of copper contact plate (not all have one). Some do direct contact heatpipes. Plus keyboard most times. So we really have to innovate in other ways. And that is what Tongfang seems to have done. Every other notebook. Has only 1 heatpipe directly over all 6 of those tiny IA cores. Every one of them? I have not comprehensively checked. But certainly in this specific segment, the 'thin and lights'.
So how does the heat trave? Well you have to imagine the heat itself as a special kind of an invisible (and somewhat visous) liquid. That is impeded by different materials at different rates.
At the very beginning of it's journey, most of the heat is generated within smaller portions of the IA core. Which is itself very small already. The 'cpu temp' is also measured in 1 point location somewhere on that specific core. But the cpu temp itself is just a kind of an indicator. It does not communicate to use the essential information about how much heat is transferred (which we might think of as the volume of that heat liquid). And it does not tell us how fast that liquid is flowing away from the source either. So instead we try to augment our understanding by using other terms like heat soaking, etc. But that does not paint a dynamical picture either. (because a soaked heatsink represents 1 steady state condition, after the entire thermal transfer system has already reached equilibrium, including fan speeds and everything else). So that obfustates our understanding significantly, what is going on earlier up the chain, in the first few mm around the core. That is way before the heat gets to the fans.
So: We have these tiny little squares. 6 of them. There a 2 rows of 3, along a central 'spine'. So the internal spine side of the cores will be the hottest. And the extremities (as with the human body), will be the coolest edges.
Now,
The thermal conductivity of silicon = TG150
The thermal conductivity of pure copper = TG350
The thermal conductivity of heatpipe = variable, greater than copper
Which isn't bad. Remember that silicon is in the middle of the periodic table. So it can conduct heat somewhat like it's neighbours, on one side the metal. So you can think of the die as semi-metal. Not as thermally conductive as a true metal, but you know, we will take whatever we can get at this point.
So a LOT of heat energy will be trying to move through the silicon die itself. From the central spine to the outward edge. Moving through the IA core itself. On a kind of a directional vector, something of a temperature gradient. And whilst silicone is only half as conductive as copper (= twice the resistance, against the heat flow), there is a lot more 'heat force' within the core itself. And the rate of conduction is also proportional to the temperature gradient / the concentration of heat energy. Which is high. And it is inversely proportional to the distances / scale here, which is very small. So you know, its probably still pretty good relatively speaking. Compared to certain other elements in the cooling system. For example like the copper-air interface, which requires much bigger surface areas and volumes of air to be effective at heat transfer.
Now this is the bit where it comes together:
Tongfang laptops (and only those laptops), they have the 2x heatpipes above the cores. And as we have just discussed: Despite the very hottest area, the central spine is not directly covered by a heatpipe (it is the crack / line between the 2 heatpipes), the silicone die itself is conductive enough to laterally move the heat away from the spine, to the better transfer areas. Nearer the edges, where the heat has the shortest lateral (x-y) distances to travel to get into the heatpipe. As we discussed previously, because the lack of z-height to enable a more natural sideways (lateral) spreading of the heat. Then about half the CPU's total heat 'oozes' out into 'heatpipe A'. From the top 3 cores. And the other half the heat, from the bottom 3 line of cores ends up in 'heatpipe B'. Mostly thanks to: the thermal conductivity of silicone. A contact plate (copper slug) might not even be necessary at all. (Except for better mechanical stiffness / clamping force reasons). For example the newest 2018 macbook pro. It does not have any contact plate. Just has 2 very flattened out large diameter heatpipes. Running parallel in a similar way to these Tongfang. However the MBP has other kinds deficiencies in its overall solution (2 fin stacks instead of 4, greater air restriction). Speaking of which, the rest comes down to:
The other Primary cooling stages
Total volume of vapour inside the heatpipes (2x). Total amount of contact surface area inside the heatpipes (2x). Thermal interface material, LM or paste.
The Secondary cooling stages
Total amount of contact surface area outside the heatpipes. The fin stacks. Fan efficiency. Fan speed. Air flow restrictions, (meaning air vents size, air blockages, dust bunnies).
Later stages in the chain don't have much to work with unless they are given something by the earlier stages. Which is why I spend all this time talking only about the earlier stages. It is more important that earlier stages are improved first of all.
Of course, later stages do also need to provide a sufficient level of performance and heat dissipation capacity, to cope with whatever is provided from earlier upstream. But point discussing that on designs which already fail at the 1st hurdle. But later stages consideration would be to measure the total fin stack area. And consider the fans performance. That is actually a bit of a let-down on the current Tongfang notebooks.
Other designs:
* Massive Vapor chamber in new razer 15 --> dont know / haven't looked at it
* However new nvidia 2080 gpus also have large vapour chamber. And they also get pretty hot --> ?perhaps? might be the way vapour / liquid moves internally. And circulates (re-cycles) within a single larger unified chamberLast edited: Mar 28, 2019 -
CSHawkeye81 likes this.
-
Agreed, based on my experience with the 15 R3/R4 issues with the cooling the m15 that I have is definitely an improvement on cooling design. It's nowhere near perfect but it is definitely better designed then the other units.
-
Core 0 and Core 2 in my unit are the ones that hit temps 9 degrees hotter than the other 4 cores, pretty consistently. I wonder if they are the internal 2 cores OR if they specifically have poor contact with the heat sink.
I've not opened my unit up mostly because I am trying to send it back, but Dell is negotiating with me to take more money off so I will keep it. I've requested a repair/replacement of the heatsink and an extension of the warranty before I will consider keeping it. In the meantime, I don't want to go pulling it apart or give them any reason not to fully refund me if I end up do sending it back. -
I think I'll hold off the repaste until I hear how the replacement heat sink works.
Sent from my Moto Z (2) using Tapatalk -
Not yet, but I will definitely keep you all updated.
-
-
Sent from my Moto Z (2) using Tapatalk -
-
Sent from my Moto Z (2) using Tapatalk -
They won't extend the warranty and they won't address the heatsink issue as part of the offer for me to keep the machine. I will be getting a call within 48 hours from a rep who has been assigned to my case. At this point, I am leaning towards just sending it back and being done with it.
Drinky likes this. -
Curious, did anyone in this forum experienced good temperatures at all? And regarding that, what do you guys consider to be "good enough" thermals?
Skickat från min SM-G960F via Tapatalk -
Staying under 90C without having to manually throttle the thing would be a good start.
Drinky likes this. -
Skickat från min SM-G960F via Tapatalk -
I just got of the phone with a dell tech. He said, that this is the first case he heared off with temperature problems. After showing him the temperatures, he agreed that I send it back and get my money. I think this is the best solution until the problems with the m15 are fixed.
-
swordofsilence Notebook Consultant
Propping up the back and using the performance profile in AWCC helped a lot too (assuming you don't mind the noise). Undervolting also helps.M-Fire likes this. -
Skickat från min SM-G960F via Tapatalk -
Vistar Shook, Drinky and billycuth like this.
-
Cap the clock speed (cap of the bells) to keep it (accept the flaws) will only favorite/help DELL.
Or better say encouraging Dell to continue in the same way with coming new refreshed models. See... Not much changed after release of the Skylake models.Last edited: Dec 10, 2018Dennismungai, Vistar Shook and alberty like this. -
swordofsilence Notebook Consultant
I definitely think it is crappy of Dell to make it sound like this thing has all of the performance of it's bigger siblings and then it throttles and overheats out of the box, which I think is what so many are bitter about and it's totally justified.Last edited: Dec 11, 2018 -
Sent from my Moto Z (2) using Tapatalk -
It's super disappointing but I really need to know I can trust this thing to last, and I just don't. -
Stock Temperatures after disabling One Drive (On Idle):
After disabling One Drive and turning on Speed Shift to 200, Undervolt 0.125:
Did not restrict clock speeds, Battery mode on Battery Saver.
AWCC:
I can safely say after around 2 hours of just surfing the web on my bed that it's been cool as a cucumber. Fan is only audible if your put your ear next to them!
I have also played on battery I used different thermal settings and had it plugged in (Middle setting battery)
I will add temperatures when I get home and start playing it again, but I have not gotten anything above 80c in medium settings (Aiming to hit around 120-144fps) on PoE. This 13 hours of play time was just yesterday in one sitting.
To those who are wondering what the temperatures really are for normal use/not benchmarks.
Won't be planning on repasting until I see my temperatures go above around the high 80c~Last edited: Dec 10, 2018M-Fire likes this. -
swordofsilence Notebook Consultant
I'll have to listen to my fans a little more closely next time I test it and see if I notice anything. I don't remember if it sounded funny while spooling up and down, but I know at full speed, the fans on mine sounded fine. -
-
http://forum.notebookreview.com/thr...15-owners-lounge.825553/page-92#post-10831912
Clearly the only heatpipe over the CPU cores is the middle one. Which is also the main heatpipe shared with the GPU. And the other 2 heatpipes are both smaller, and off to either side.
Now you pair that with --> the fact that these laptops have a 1080p screen. Well 1080p is the resolution that stresses out the CPU most, for a lot of games.
In UK i originally thought that the m15 was priced 'just about' within my budget. With a spec that came out to about £1700. However aparrently the price I was looking at in the configurator was excluding tax. Because when I went back there more recently and actually went through to the order page the tax made it something like £2100 ?
Well that is double the price of the PCS recoil 2 in a barebones configuration (excluding any storage or ram). I could literally buy 2 of them for the same price. And what would I be giving up?
* Full fat gtx1060 instead of 1070mq
* Thunderbolt 3
* AGA port (pci-e x4)
* Max after market battery size is 72wh, vs 90wh
I really did want all of those features in my notebook. To me it might have been 'just enough' to justify the £1700 price point with 'somewhat inferior CPU thermals and a lot of tweaking required'. Because I really dont mind doing any of these tweaks. But at £2100 its 'way way too much' money. Rather 'quite a lot of money'. -
The reason I want to stick with the m15 is because I was planning on handing my son my 13r3 and I already own a graphics amplifier so I can have portable and desktop level gaming performance without the bulk. Sucks that we have to deal with these overheating issues.
-
I had of wrong 4.1 is usually used doe 4 core's but some boost 4.1 all core instead of 4 -
swordofsilence likes this.
-
Skickat från min SM-G960F via Tapatalk -
That was me - I was offered $100 more off.
I don't feel ripped off by any means. This is a solid laptop with a poorly designed or implemented thermal solution (in my case) that I just don't feel good about even at the price drop. I am REALLY having a hard time giving up on it because I like the design and all the features a lot. But sometimes you just have to go with your gut. My gut is telling me to move on. It's telling me to spend less and go a little bit more conservative for something that will last.
It's a done deal for me. Dell sent me the return label. I'll clean/pack it up tonight and send it back.propeldragon and c69k like this. -
Already waited 3 weeks for it, might have something to do with Dell having to setup a Nordic layout since I live in Sweden. Which laptop are you aiming for as a replacement?
Skickat från min SM-G960F via Tapatalk -
Honestly if you want a laptop with no thermal issues that's thin and quiet but don't care about battery life, zephyrus hands down
Sent from my Moto Z (2) using Tapatalk -
Skickat från min SM-G960F via Tapatalk -
Core i7-8750H - Intel - WikiChipDarkhan likes this. -
Its difficult to get an accurate picture from reading these posts here. Because as useful as these detailed responses are, its hard to keep track which users are reporting new issues. And which ones have come back to make another post. And that goes both ways for good and bad results. What would probably make it easier to tell, the true extent of these problem(s) is:
A shared google spreadsheet. Or wiki page. Where each individual can maintain a single row per laptop. And whilst that would still be a somewhat biased data set, it would at least be easier to check up on over the months.
When Linus had their company fleet of Razer Blade 14 laptops. They did just exactly that.
Columns would be stuff like:
* Initial purchase date
* Benchmark scores, before repaste.
* Links to benchmark scores (before).
* Fan profile.
* Repaste TIM used
* Undervolt
* Link to XTU or Throttlestop tweaks if not applied by Alienware Control Centre
* Benchmark scores, after repaste.
* Links to benchmark scores (after).
* Fan profile.
* If unit was RMA'd in any way.
* What RMA was for / what was like afterwards.
* If unit was returned.
Other relevant information such as...
* Power brick wattage
* Firmware & EC version
* Control Centre software version
* Windows version + build number
* Link to a search terms string, that pulls out all of the posts [made by that forum user, for that specific unit]
By also being clearer about test conditions, it then becomes easier to compare across different units. Rather than having the ambiguity of not knowing whether there was also a lack of tuning at play, or some other sort of misconfiguration. By including a recommended guide for throttlestop, recommended BIOS settings, Control Centre profile, etc.
[edit] Oh, and length of time since the last re-paste. Since that is also a pretty significant factor. Combined with the specific choice of TIM paste.
[edit 2] And a checkbox to say whether the laptop was either propped up (air gap at the back). Or was flat on the desk.Last edited: Dec 10, 2018 -
I totally agree with @dreamcat4. I received mine 5 days ago and I love it. It's much better than my Clevo 2230st I had for 5 years, I know that isn't saying much. I noticed some serious heat from my m15 as well but nothing like you guys were talking about. I found that propping it up on a thick coaster lowered my temps by 8-10C when gaming. The fans really sounded like they were trying hard without the machine propped up. Playing Assassins Creed Odyssey for over an hour I saw GPU temps around mid 70's and CPU around 80's to mid 80s. I am running a 1060 in mine but I also have a later build date so maybe something was updated? I know the CPU can withstand 100c. I am no comp guru but that doesn't seem horrible to me to run at 80's under full load. You guys want me to open it up and take some pics? I was going to apply Kryonaught anyways.
I will do a better review tonight and show my stats. Maybe even open a shared doc for everyone to contribute their info like dreamcat said. I assume we would need two spreadsheets for 1060 and 1070mq's. -
-
B0B likes this.
-
I could go on all day about this topic alone. It’s partly why I started BOAT.Alias and NeWCuLTKiNG like this.
*OFFICIAL* Alienware m15 Owner's Lounge
Discussion in '2015+ Alienware 13 / 15 / 17' started by ssj92, Oct 25, 2018.