The last thing you want is a heat shield!

You want something that can take the heat away from the CPU, and you're not likely to get better than is already there in the form of a copper sheet on the back. You just need something to do a better job of transferring the heat to the copper heatsink.

Yes, I am aware but I'll see whats what if and when I take it apart.

- Ordered 150 x 100 x 1.5mm - Thermal Pad to shape over the cpu/gpu area

I'll let you know how it goes....

One person on the Russian forum had a problem with overheating. However, once he uninstalled Swype keyboard (first thing he'd installed on the pad after unpacking), the heating problem disappeared (he didn't test the unit with setcpu\heavy games, but he used to have overheating when web-browsing, not he claims there's none).

Not sure, though, that everybody here is using swype keyboard )

Okay, So The copper back plate is MUCH larger (More surface area) than the heat spreaders on the GPU/CPU/whatever.
So the copper plate will take much more heat to reach a Thermal equilibrium with the heat sources.


I'm sure if I look it up I could give a super accurate answer of what is going on. but in stupid terms.

The copper will dissipate heat faster because it is larger.
Seeing as the mobile chips probably have a VERY LOW TDP a large cooling "fin" should be efficient enough to displace the heat produced.. given that the heat exchange is adequate

Just like with any cooler or radiator fins are implemented to increase the surface area..

These might be relevant?
http://en.wikipedia.org/wiki/Thermal_equilibrium
http://en.wikipedia.org/wiki/Heat_transfer
http://en.wikipedia.org/wiki/Thermal_design_power

I've been unable to source a suitable, better-performing replacement thermal pad capable of bridging the (relatively) wide gap between SoC and heatsink. I'm thinking that a 'sandwich' solution, similar to what I've used on GPU VRMs before, may help.

Basically you use a thermal adhesive such as Arctic Alumina to attach (permanently) a portion of metal such as Aluminium - or even better, Copper - to the SoC. You can then use your TIM of choice to bridge the air gap between the new 'heatspreader' and the heatsink inside the rear of the tablet. This eliminates the need for a 'liquid bridge' of TIM, which is largely ineffective.

I've had success using this method on graphics card VRMs in the past, albeit they're always assisted by active cooling.

I'm still looking at/working with replacement pads before I do anything too permanent. I'll report back in if I have any breakthroughs

EDIT: Fujipoly, more specifically their Sarcon range of products, looks like it may offer a viable solution, depending on the gap involved. Is anyone able to provide me with a hard measurement of the gap between SoC and copper heatspreader on the rear casing? Even just the thickness of the stock thermal pad would be a start.

Sarcon SPG-30A looks particularly interesting. This is a 'high conductivity', form-fitting conductive foam/elastomer, which should scale to fit the die/gap reasonably well. Conductivity isn't as high as I'd like to aim for, but we're talking a low-voltage SoC so should be okay. There's a higher performing (6W/mK) putty which may also be viable (cost permitting).

Now I just need to track some down

I'm actually using Fuji-poly thermal pads from FrozenCPU which have seemed to have reduced the maximum temperature at least 5°C in preliminary monitoring although I need to test some more. At first I put a strip over the CPU area but this enlarged the heat surface area and made the back hotter. So I reopened it and where the areas rise (Where the original heat pads were) cut about stamp size squares to use instead which seem to have worked better. I'm using 1.5mm thickness and it works, although I had to move the one nearest the hdmi port back some so it would close more flush without being slightly raised when the unit was closed.

I'm still looking into programs that are running that are making the device hotter then it should be. I was playing around with cpu/task programs last night and after killing some things (Not listed in the usual places - Settings>Applications) and it appeared to be running cooler, but I dunno need more time. One important thing I've learned is make sure you remove the Micro SD card before operating... In my idiocy I lost a 32gb card when I failed to remove it, then it bent and snapped. And make sure to use rubber or lite plastic to remove the trim on the tablet, I roughed mine up pretty good. Total noob action......

Let us know what you find or if you wanna brainstorm on some things...

Originally Posted by Itchiee

I'm actually using Fuji-poly thermal pads from FrozenCPU which have seemed to have reduced the maximum temperature at least 5°C in preliminary monitoring although I need to test some more. At first I put a strip over the CPU area but this enlarged the heat surface area and made the back hotter. So I reopened it and where the areas rise (Where the original heat pads were) cut about stamp size squares to use instead which seem to have worked better. I'm using 1.5mm thickness and it works, although I had to move the one nearest the hdmi port back some so it would close more flush without being slightly raised when the unit was closed.

I'm still looking into programs that are running that are making the device hotter then it should be. I was playing around with cpu/task programs last night and after killing some things (Not listed in the usual places - Settings>Applications) and it appeared to be running cooler, but I dunno need more time. One important thing I've learned is make sure you remove the Micro SD card before operating... In my idiocy I lost a 32gb card when I failed to remove it, then it bent and snapped. And make sure to use rubber or lite plastic to remove the trim on the tablet, I roughed mine up pretty good. Total noob action......

Let us know what you find or if you wanna brainstorm on some things...

My feelings, are that it is going to involve PWR management, and kernel/driver adjustments. There's no denying that. Just move your finger across the display, and you start seeing temperature increases. With NO INPUT. Just the action causes the temp to rise,

You guys are just putting a band-aid on what the core issue is. Kudos for trying the physical aspect of things though

A good thesis, but not the solution.

MD

Yes, that is a possibility. Also perhaps the CPU/GPU are over-volted could be an issue as well.
I do notice the cpu likes to ramp up and down on its own at times.

And my band aid lowered the temp by at least 5°C!!!!!

Originally Posted by Itchiee

Yes, that is a possibility. Also perhaps the CPU/GPU are over-volted could be an issue as well.
I do notice the cpu likes to ramp up and down on its own at times.

And my band aid lowered the temp by at least 5°C!!!!!

I can drill 9 holes in the back cover and 2 at the opposite end. Will probably decrease it by 10c. Just by thermal flow.

Doesn't fix the primary issue, which is, the CPU overheats till it re-boots.

It's a band-aid fix. The primary issue is not solved. Hell, the other Tegra-3 users don't complain.... And they OC to 1.8ghtz!!

Mine out of the box, was clocked to 1.4. Even though specs say 1.3. And please, don't anybody say things about "stepping". It doesn't apply to a stock device.

These 700/701 tabs had an issue, and Acer knew it. They were hoping us at XDA would find a solution. Typical behaviour.

MD

Yes, you are correct. Lets hope they can do something to fix it.
Mine has never rebooted no matter how hot it gets though for the record.

In Soviet Russia solution finds you!!!!


Sorry, had to, couldn't help it.