Installed 680M on p150HMx, High Idle Temps!

Hello, I just recently installed a new 680M with new heat sinks included and diamond thermal paste. According toe HWMonitor:

Idle temps are in the 80s.
Gaming [ Left 4 Dead 2 reached 118...] [Not even a demanding game... Only played for 10 minutes.]

Had the fans on max with manuel fn + 1

Yet the fan behind the GPU did not feel warm, nor was the air coming out warm. Culprits?

1) Bad thermal paste job? [ I redid it 3 times all with the same result.]
2) Bad Heat sink?
3) Bad thermal paste? [ Tried arctic silver and had even worse results.]

I mean there was no heat felt, I'm assuming that the heat is not sinking properly from the GPU core. Is this correct?

I appreciate any comments or help.

Thank You
-SpheX

 

Either the heat isn't escaping the GPU to the Heatsink or the thermal reporting software/hardware is broken.

 

This right here raises some severe red flags for me.

The GPU generally has an auto-shutdown at (someone correct me if I'm wrong?) about 105c.

The fact you hit nearly 120 at all means there are some things that are problematic.

1. The firmware is borked on the 680M and the autoshutdown, heat sensors, the whole shebang are wrong.
2. The bios on the machine itself isn't capable of understanding the 680M
3. Something in the airway is clogged where it's not blowing the air correctly over the heatsinks.

#3 is unlikely as you feel air flow.
#2 is doubtful unless it thinks the card is different than what it is.

My bet is on #1, but I'll defer to someone who has more experience with this than I do.

 

What you are saying makes sense to me. I believe the problem might be a faulty sensor on the gpu. I did the following today:

Thermal repaste 6 times.

I used 2 different tubes of ICCC Diamond to [ twice with each one ]
Did it twice with Arctic Silver

During these attempts, I swapped out the heatsink for the one with my 6990M once and did not see any improvement at all.

I used the line method, the X method, spread method and pea method. The pea method was the best but there was no real significant improvements over the displayed temperatures.

I looked at the temp with 2 different programs in windows to rule out faulty software... They both read the same.



So, my laptop does not automatically shut off even the GPU is reaching over 100F according to the sensor. Is this because the bios does not recognize the GPU?

I believe that there is 1 of 2 things causing this.

1) Faulty Sensor, GPU is fine.
2) Faulty GPU that was repaired to working condition and resold as new, however, is unable to transfer heat correctly from the central core to the heat sink.


Advice, What should I do in this situation?

 

Ok... wait wait wait...

Are you comparing your temperatures in fahrenheit or celsius?

Because the shutoff at 105 is in celsius.

If your 118 degrees was in fahrenheit, that's not even 50c, that is amazing cooling.

Hot for a GPU is considered 90c or higher, dangerously hot is considered 95c or more.

You tend to want to keep a GPU under 88c or so. A range of 70-84c is best for under load. The CPU's you generally want to keep under 85c. 70-80c is generally a good area for 'high load' temperatures for a CPU.

And remember,, 90c ~= 194f

 

Lol. Damn imperial measures. When he said 118, I was thinking what xPU doesn't shutdown whole system at the max 100* C.

Sent from my SM-N9005

 

Lol you guys and your ancient measurements

PS kelvin or go home!

 

if I was going ancient measurements, I'd use Newtons, Reaumurs, or Romers.

Wouldn't that be fun.

 

Seriously that is good cooling? Wow maybe I was mistaken. Well when I had my 6990M...

The temps never went beyond 80F under full load lol.


I will take a quick look to see what is acceptable range of temps again. On boot up 30C.

Full load reaches upto 120F or nearly 50C.


Wow I just looked it up and everything is in C not F lol... I didn't realize how silly my mistake was.


No more late night Kim Jong Un Rap sessions for me X_X

 

Buddy 50C is nothing to worry about at all. My 780M can go up to 85C when playing Crysis 2 and I'm not the least bit concerned.

And yeah, Farenheit can go suck the big one. Metric for life

 

forget Fahrenheit and only post temps in centigrade.

 

So you are getting 119C?

 

Wait what you actually are seeing 118C?

 

I think they meant that they read the values as limits in C, but when they used HWmonitor it was displaying in F and they ... confused the numbers

 

Yep this post is exactly correct. Funny thing is I was concerned because my 6990M never reached beyond 82F under full load. I didn't realize actually how amazing the cooling was. I am now very relieved and am going to try some more demanding games to see what the temps reach.

As always, this community has great members with very informative posts. Thank you everyone for your input and help.

 

As a certain character would say:

"Play more games!"

 

I hear you there.

I know Witcher 2 on uber cooks my 880M a bit. She's hit about 86c with uber and everything maxed out. But I also get close to 40 fps