Aspire E5-573 heatsink construction and overheating

jimveta

I just sent my Aspire model: E5-573-5653 with Core i5 5257U in for RMA due what I think may be a thermal conductivity issue.  The details that I included in a letter for the technician is below along with pics.

 

I would like to know fi other people are experiencing the same thing and if you have opened up your laptop and see the same type of construction with a portion of the bare cpu + gpu chip exposed, not covered by the solder thermal interface. If others see the same then it could be an engineering design flaw I believe.  On other laptops like some recent Asus, they cover the whole chip with the solder compound to the heatpipe.

 

==========================================================

 

Description of problem:
-----------------------
When running CPU and integrated GPU stress tests and Passmark's Burn-in test:
http://www.passmark.com/products/bit.htm
The fan spins up to max speed as expected, but the temperature reaches
100 degress Celcius for the duration of the testing.

Even though the absolute max temp for the i5-5257u is listed as 105C in
its specificiations, I think the cpu may be throttling a bit since the
temp continuously fluctuates slightly between the peak near 100 C and
upper 80's C even thought the workload is entirely continous without idle.

As a side note, I don't think you can detect throttling by observing cpu
clockspeed since I believe it is done by injecting idle (halt state) cycles
now.

As a result, I removed the back cover and found the cpu chip only about 2/3
connected to the heatpipe that connects to the fan.  See the attached pics.
The permenent solder-type thermal compound that provides the thermal conduit
to the heatpipe does not cover the entire chip.  About 1/3 of the bare chip
is exposed and unconnected to the heat pipe.

I have removed the password for the local account.  You can log in and
run BurnInTest.  It should have CPU, RAM, Network, Audio and GPU tests
all selected at 100% duty cycle already under Test selection.  You can

then press the Play, the green [ > ] button and monitor cpu temps under
the temp tab.

 

==========================================================

 

top view of board:

http://i.imgur.com/rF7PFdE.jpg

 

side view, close up of cpu:

http://i.imgur.com/X78ClNQ.jpg

 

philetus

Either the CPU and graphics are in the side covered by the heatsink and the chipset is in the uncovered side, which is probably the case, or the CPU or graphics would have a cooling issue.

 

• A new multichip package integrates the chipset on one die with the CPU and graphics on one die connected via a lowpower, on-package interface

https://cyberraiden.wordpress.com/2013/07/05/4th-gen-intel-core-i3i5i7-mobile-processor-haswellcrystal-well-specification/

jimveta

Thank you for that information. I found Broadwell's raw package:

 

and I assume the chipset die is the smaller one on the right.

You are probably right that the chipset portion is probably the one uncovered as I have seen that practice on desktop mobos when it's low powred enough.

 

I actually called Acer support and described the high temps I was seeing and asked if I could talk to a technician to ask about this issue, to see if it is an intentional design. But the only solution offered was RMA.

 

If it is the seperate chipset die then it's probably better to leave it unconnected to dedicated more of the heatpipe to the cpu+gpu. Still I hope something can be done to address the high heat. When loaded, it's constantly bouncing off near Tjunction in the spec.

philetus

I pulled the fan out of my V3-551 to replace the fan and when I untaped the fan from the heatsink, I found it blocked in between with dirt and junk.I couldn't tell it from looking in the exaust port.