G1S Wierd heating issue

Hey, I've had my G1s-A1 since August and I just reinstalled CoreTemp/Speedfan today just to check my temps. Right now, it's idling at 57C which I find quite high as I used to get high 30s low 40s consistently. What I find wierd though is that the temperatures actually drop when in battery saving mode and not plugged in, as soon as it goes to AC mode, the temps go up to 57-58 from 50. I cleaned out the vent by the video card/processor earlier. any other suggestions?

 

what's your ambient temperature? and of course get a notebook cooler. but, i just realized you said you cleaned out the vent. do you have access to the cover of the cpu? clean that area out. hold the fan and blow with compressed air, and don't forget to clean the intake, too. and see what your thermal pad/thermal compound look like.

 

I have a cooler and its making no noticeable difference like it used to. I've throughly cleaned all the fans on both the GPU and the cooler. And ive cleaned out the processor area aswell. I don't actually think there is an intake on the G1S

 

did you check the condition of what's in between your processor/heatsink/copper?

 

no but wouldnt that void warrantee?

 

not sure. i don't think they could tell you checked it unless you screw it up. call tech supp and see if it does.

 

So, what and why is weird?
May be it suppose to be that way that computer consumes less when runs on battery?

 

the power gear i think comes into play when you use with battery. maybe by default the power gear goes to the maximum setting thus squeezing the processor for more juice

 

Whats wierd is that my idle temps are almost 20 degrees hotter than they were back in the fall... my ambient temperature is about the same as well...

 

this temps are normal...

 

How did you manage to get to your G1S fans? I can't see an easy way ...

 

Does the G1S have more than one fan??

 

I don't think it has more than one fan although I can't say for sure.

To the OP: I suggest installing something like RMClock for explicit and clear control of the CPU frequencies. Get rid of Power4Gear. I think the difference in temps is due to Power4Gear always forcing the CPU into a higher-performance mode on AC power, which is not really necessary. Check the Guides forum or my signature (Tips and Tricks) for guides on using RMClock.

 

G1S has only (1) fan. It is coupled with the GPU cooling solution. If you were able to "look through" your laptop, the actual fan placement is pretty much directly under the [Tab] and [Q] keys on the keyboard. Air from the fan flows through the heatsink stemming from the CPU heat transfer pipe.

Furthermore, the actual speed of the fan seems to be directed from the temperature of the CPU and NOT the GPU. The fan speed ramps up @ 50C, 55, 62 (65?), 70, 75, 80 repectively. I'm not sure if it goes higher.

I've got pics of a complete teardown of my G1S if anyone's inderested. It was during a CPU/GPU repasting. Now that I think about how the fan works, it might be beneficial to the GPU if I had NOT repasted the CPU. A higher average CPU temp may result in the fan kicking on sooner. The result might be a noisier laptop, but a cooler GPU. Conversly, maybe temperature from the CPU radiates out quicker, resulting in a lower total system temp, and thus keeping the GPU cooler as a result.

All in all I am satisfied with the repasting of both as it seemed to have lowered my average temps all around. I used Arctic MX-2.

A side note:

I've actually been thinking of a mod to connect the heatpipes coming from the GPU, to the CPU to kick the fan on sooner. Playing games while overclocking results in the temperature to spike up as high as 95C+ (which is a stuttering point for some people). Meanwhile, the CPU can be at a happy temperature of 74 or below, keeping the fan from working full-blast. If some of the heat from the GPU was tranferred (within safe limits of the CPU of course), faster to the CPU, the fan would kick on sooner, hypothetically resulting in a cooler GPU, and a more stable overclock. Again, it would probably mean a noisier laptop, but maybe a more stable one.

 

Ive got 70 - 109!!

 

post here that pictures....
i have de same idea to change the paste...

 

I made a thread specifically for my teardown. Ask me any questions about it here:

http://forum.notebookreview.com/showthread.php?t=251619

On there you'll find the links to the pictures as well.

http://www.flickr.com/photos/2673196...7605149783571/

It was relatively simple, just a TON of tiny screws.

 

Damn thats hot! I'm not sure what these chips are rated for, but that's definitely the highest temp I've see reported. I'd be VERY careful with anything over 100. I tend to start worrying at about 95C. I've forced some temps via extreme overclocking and ATItool. Once I start to reach over 97C I have driver stability issues, and hangups usually. So, getting to 95C is like a warning to me. Mind you, every chip is different, and there is differing yeilds. What may be completely safe for some, can be devastating to others and vice-versa. I suggest trying to "feel out" your chips capabilities before going gun-ho on overclocking. Or ideally, buy a better vid card/laptop. :wink:

 

The 109°C is nothing unusual, my GPU reaches the same temps in long gaming runs or when I stress test it. It's still not ok and definitely needs to be fixed, though. So whatever 3Dlink said besides that: 100% agree.

 

Wow, 109C for you too? I can't push my chip past 97C for long or it starts to get angry at me. Crashing, freezing up, etc...

 

Really? I guess I should be happy that mine doesn't. If it did I couldn't play at all. If I push mine up to 110°C (only possible after the sleep fix that disables the first throttling mechanism) using FurMark 1.3 in stability testing mode you can observe in RivaTuner how the clocks are lowered in order to stay at least at 110°C. If it was not for the periodical downclocking it would quite possibly melt.
So for everyone to know: That's just HOW crappy the cooling on the G1s can be. Again, I'd recommend anyone with that problem to either turn his in or reapply the heatsinks as 3Dlink showed in his thread.

@the 3dlink:
Since you are the only one in reach who has applied arctic silver that I know of, would you mind getting the FurMark ( http://www.ozone3d.net/benchmarks/fur/) and running it on 400x400, not fullscreen in stress testing mode and observe your temps, clocks and fpss? Make sure not to use any external cooling.
I'd be prepaired to praise your greatness if you took a screenshot with the following things on it:

1)FurMark with the amount of time it has been running (is displayed on screen anyway) >400 seconds
2)A program showing that your current GPU temperature is <95°C

That would be _teh_ awesome for then I do not only know how to exactly reproduce the issue but I'd also have proof that the G1s can perfectly handle it if the cooling is not all factory f***ed up

 

I'm confused, the temps also depend on the type of driver you use, i use 175.70 right now

 

I've heard that- still it's not very well covered by investigations and I'd also go as far as to neglect that point when talking to an Asus technician for just another person telling us that We're using the wrong drivers is the last thing I think we need.

 

@gr33nf4c3: here ya go chap... I was surprised at how well it handled this test. It seemed to stabilize @ around 86-88C. Notice my CPU temp in the corner too(RMClock). I put markers in RivaTuner so you can see how it progressed. It was a real slow roast. Took about 10 mins to get up to a stable temp. I had my laptop resting on a binder, on my lap during the entire test. I also took a shot of it cooling down. Also, there's the laptops information in RivaTuner/Everest too (for all the Area-51 folks). You can see I'm running 175.16 for drivers.

Oh, and I used Arctic MX-2, which has no silver. It's non-conducting, and it's the best rated stuff Arctic sells.

http://www.arctic-cooling.com/further_prod2.php?idx=140

P.S. - Tomorrow, err... later today I'm gonna try and stress my CPU at the same time. (After some much needed ZZzzzz's...) I would't be too surprised if I saw cooler/slowed GPU temps due to the fan kicking on sooner, then again maybe the ambient temp inside the laptop will increase it.. .who knows? I will... later today....

 

Thanks bud, you are my hero!

For everyone who owns a G1s and hasn't grasped this so far:
Replicate the benchmarking setup I listed somewhere up there and compare your findings to those of 3dlink and you will have unquestionable proof on whether or not your G1s behaves the way it is supposed to.

Again, thanks very much, dude. Someone +Rep 3dlink please- I can't because I've already done so for the G1s disassembly guide.

 

Ok, so I conducted my test on the relationship between the CPU triggered system fan. What I found was interesting, but just as I suspected. Take a look at the graph from RivaTuner. How I ran the test was I let furmark run for about ten minutes to get the GPU up to a stabilized temp (It's somewhat cooler today so I got lower temps than my last test with furmark). Then, during the test I stress the CPU with Everest as well. This was to induce the fan to kick on. I then stop the CPU test to see it's effects on the GPU temp.

As the picture shows, in the orange box: the CPU starts to heat up quickly, the fan is ramped up twice, and the GPU temp falls a notable amount.

In the yellow box: I believe the CPU temp starts to affect the GPU and it's temperature rises.

Finally, in the blue box: is where I cut the CPU stress test, and the CPU cools ALOT quicker than the GPU. This results in the fan slowing down quickly, and the GPU temps soaring.

The last marker is roughly when I stop the GPU FurMark stress as well.

I believe most G1S owners would benefit from being able to change the parameters of when/how the fan kicks on. It may result in a louder laptop, but a cooler/more stable one, and potentially one that does not stutter(for folks having that issue).

I DON'T think the fan that's in there could have handled the GPU/CPU stressing lots longer. The ambient temperature was getting too high, and the fan is just not powerful enough to expel all the heat being generated by both processors being stressed. In reality, most people would not run into conditions such as that.

Gaming however, can keep the GPU quite busy, while the processor is not very stressed. This keeps the fan from efficiently removing the heat from the GPU.

 

Yeah I experienced the same, therefore I was trying out pretty much any fan-control software available, unfortunately none was able to detect the fan.
The only software that was at least able to read out fan rpm was Asus' NBProbe 3.0...there's also absolutely nothing about fan settings in the BIOS, actually you can pretty much forget this BIOS at all, since you can't access anything useful with it...

I think with access to fan-control you could get all max. temps to reasonable values like 75°C GPU and 55°C CPU in combination with a notebook-cooler.

 

Yeah, once that fan gets going at it's max speed, it does a decent job with intense gaming...albiet - it's loud! Nothing some headphones can't handle. Only problem, it's not always going @ max during gaming. I'm thinking since the fan speeds are set by the bios, prolly have to flash a custom/edited version to change them(know anyone?).

For a laptop aimed at gamers, I'm surprised the bios is so stripped. I don't understand why they don't let people have more control over certain things. I can understand if there are set minimums for fan speeds, but what about people who want to bump it up a notch? Or people who live in hot climates? Or actually want to use it on their lap without getting 3rd degree burns?

It's too bad they couldn't have set up the fan to work at max under Performace 3D mode. Or be triggered by the GPU temp as well.

 

Yeah, that crappy BIOS is really a pain. But don't mind trying to edit it by any other means. I've had a friend read it out with some app and it turned out there weren't even hidden options in there. So unless you know how to decompile and rewrite a BIOS any thought spent on this is one too much.