Asus G73JH Temps Question

I have an Asus G73JH with the I7 720qm and the Mobility 5870. As you can see in the pic I have the graphics core OCed to 800 MHz. I ran furmark for 20 minutes and those are the max temps that it read on GPU-Z. My question is how accurate is that 120 degree reading? I usually just base my gpu temp off of HWmonitor. Furmark's temp matched HWmonitor's temp btw. Is that 120 too high for this 5870?

Thanks



Edit: I somehow messed up the image trying to trade it out for a smaller version. Here is a link to it.

 

Your temps are through the roof, time to get the air cans out and clean the heatsinks. If that doesn't reduce the temps, time for a repaste, either by you or Asus.

Also, check this here on how to resize your image: http://forum.notebookreview.com/asus/486972-how-properly-insert-huge-image-into-post.html.

Your picture is way too large according to the forum rules. Resize so that at least someone viewing it on a 1366x768 lcd won't have to scroll sideways.

 

Brady , welcome to the forum. 120 Celcius is very high, i am surprized it did not shutdown at that temp. http://forum.notebookreview.com/asu...m/516113-asus-g73-temperature-question-4.html

In game to do ever exceed 100 Celcius? I would recommend cleaning the vents with canned air, and check temps again. If the temps are still high you may need to repaste the GPU heatsink.

 

My bad on the pic. I had a smaller version but it made it hard to read the numbers. I guess full size was too big. haha.

That was doing an xtreme burn in full screen on furmark for 20 minutes. My games don't get that high. With skyrim on Ultra detail my gpu temp reading on HWmonitor doesn't break 70 C after hours of continuous play. Any idea where gpu temps 1-3 are reading? I assume gpu temp 1 is the gpu core since that is the same one that furmark and HWmonitor read.

Maybe I will have to repaste this thing. I am a bit nervous to tear it apart though. I have read/watched walkthrough on it though. I air dusted it out right after I got back from Iraq, not that my laptop saw anymore dust there than I do here. It spent no time out doors, unlike me.

On a side note. At 800/1000 it ran fine with no artifacts (that I noticed anyway) for 20 minutes on furmark. At 825/1000 furmark sort of crashed... the screen went black. I did not see an artifact leading up to this and it happened within a few minutes. I just had to hit alt+f4 or alt+tab to move around so windows was fine. I turned it down to 815/1000 and got the same result only it last a few minutes longer before crashing, again I saw no artifacts. It also wasn't a temp issue because none of the gpu-z temps were even close to what I have in my pic up top yet. Is that a good indicator that 815/1000 is unstable? Also since there were no artifacts and I don't think that my games push my gpu as hard as furmark, is it still possible to run at those clocks in game?

After further testing, I had furmark black out on me with the core at 805, but I had previously ran 800 for 20 minutes fine. So I tried 800 again and again it ran just fine and due to my house's furnace being on room temperature was higher and the card achieved a couple degrees hotter at 800 the second time than 805 did just before it so temp is probably not the cause of the black crashing in furmark... My OC so far is 803/1072. I haven't tried 804 but 805 and anything higher blacks out and 1073 and anything higher cause artifacts.

 

Well, even after a few months, you'd be surprised at the amount of dust that gets in there, give it a good dusting just in case. If your laptop is still under warranty (it is if it is not a BB model and you bought in the US or Canada) you can ask Asus for a RMA and have them repaste it or bring it to the nearest ASP and have it repasted there. I recommend the ASP if it's still under warranty.

For the image, a link is fine, imageshack also has a nice thumbnail code you can embed in forums that will link to the image. It's usually what i use for very large images in my posts.

 

So I repasted the gpu and the cpu and cleaned out the fins on them. It helped GPU temp 1 which I believe is the core but gpu temp 2 is worse now... I am guessing it is the gpu memory temp. I apparently need to replace the sticky thermal pads on them. Any recommendations on which ones to buy or which size they are?

 

The memory temp is nothing to do with the VRAM the {black modules covered by the thermal pads}. The MEM (IO) temp {Memory controllers} on the 5870M is located at the top edge of the die itself for instance:



^^Memory controllers lead to the VRAM Modules^^

It is quite common that the heatsink clearance or thermal compound coverage does not conduct properly with the memory controllers because of their location and the fact the heatsink for the JH is poorly constructed, I found myself make a kind of rectangle blob of compound on the die when I repasted instead of just a pea shape and this helped but if you are not already I highly recommend a thick TIM like IC Diamond to cure your clearance problems it was the only thing that worked for me.

You can attempt thinner thermal pads over the Vram to lessen the clearance, I believe my stock ones measured at 1.5mm which is too thick if you replace them with 1mm it will mean the heatsink will sit flusher over the die and conduct better which may sort out your temperature issues. You are not alone with this problem.

 

My gpu temp 1 seemed to be much lower after repasting but temp 2 hit 128 when I shut it off (was still climbing) where I had previously not had it break 122 even after 20+ minutes of furmark. So you think I should just apply more paste and try again. I used my artic silver 5 but rather than using the dot method you used, I smear it out evenly onto the die. So I guess it is possible that I did not get a good enough contact with the core due to those heat pads.

 

A word of warning anything above 110oC you are in the instant death zone and no GPU should be hitting above 120oC and you are the first person I have heard of who has. Congratulations

The scraping/credit card/smeering etc way of applying compound is not one I enjoy using, this is seconded by the fact it is a total no go on the G73JH because as I just mentioned the DIE clearance to the heatsink is a joke a not very funny one created by Asus so you must apply a thick blob of compound at the centre of the DIE to allow the heatsink to smooth it out evenly and make sure it is conducting. You can also go down the route of the 1mm pads in your case it sounds like a good plan but by just smeering it out you are causing the same problem you had before.

Go back in and remove the heatsink and I bet you will not have a rectangle shape compound mark on the copper and the section that is not being covered your {GPU 2 temp} will be the section that is not conducting.

To prevent damage do not put the GPU through another thermal shutdown.

 

So I tore it apart and repasted again. There was a full square on my heatsink but it was a little light on one side I think, so it may not have been making good contact with the die. After reassembling and rerunning furmark for 15 minutes I have much better temps. Link to temps.

On a side note I smashed a wire on reassembly and I believe it was the one to my webcam... Luckily I no longer have a need for my webcam now that I am home from Iraq.

Another side note, so furmark ran just fine for 15 minutes OCed to 803/1074, I never saw an artifact but when I was playing skyrim it threw several artifacts right off that bat and froze in less than a minute. Is Skyrim just more prone to artifacts than furmark or something?

 

I kept to 800/1050 for gaming and my temps were much lower than yours so I wouldn't worry just drop them down a bit on the Memory as your MEM(IO) if your problem temperature.

Glad your repaste was a success even with the busted wire

 

That's about where my temps are. There are a few champs with a JH at crazy low 77C temps, but there's also a thread I had going with a list of 20+ people with 85-95C temps... they're just not as vocal.

 

So I took it apart again (the power button wasn't pushing correctly, turned out to be one of the wireless antenna cables was interfering) and that cable that I damaged is labeled as going to the "G73 LCD Light bar" Which I hate anyway so I might just cut it and forget about it. That doesn't explain why my webcam doesn't work though... Does it share the same cable as the mic? Otherwise I dont know what cable it uses... There are 5 cables going into the screen. 2 are for the wireless antenna, 1 for the light bar, 1 is labeled "G73 CMOS Mic cable" and the other one, which I assume is for the screen itself, isn't labeled. It is the one that has the black pull tab that you pull straight up on and is just to the left of the GPU fan.

And after reassembly the light bar actually still works.
When I pull up the webcam driver in device manager it just says that this device is currently unplugged. I'm pretty sure I didnt miss a cable anywhere though...