280x vapor-x crazy temps

[Shelster]

Since the debacle of waterblock on my Sapphire 280x Vapor-X, have re-fitted the stock cooler this morning, yet am now seeing some very crazy temperature readings.

When I was running this last week, was getting approx 50-55 degrees with fan on 100%.

Now when I start anything up that puts the card under load, temps jump straight up to 100 degrees

Have had the cooler off and checked TIM layer and that all looks good. Even touching the back of the card it does not feel like it should be hot enough to boil water.

Sat at idle / desktop work have temp of 38 degrees with fan at 21%.

Any ideas as to what is going on here? Borked temp sensor? Have some more / new thermal pads coming at some point next week (when OcUK get some in...) so will be re-seating then too

 

[Gregster]

What are you using to monitor temps? With the WB, you had very high temps and with the air cooler, you have very high temps. I wonder if it is a faulty sensor perhaps?

 

[Shelster]

Temps being measured from MSIAB, GPU-Z and also CGMiner. All three read the same.

Never got round to fitting the WB onto the card as it is a non ref PEC, which I did not realise until I had removed the stock cooler.

Stock cooler was giving the lower temps prior to removal.

Temps rise and fall iaw load, so am kind of thinking the sensor is ok, it is just something in the cooler mounting that is being funny.

Will try it with different TIM next week (have left the stuff I wanted to use in work....doh) and see if that makes any difference.

 

[Gregster]

Temps being measured from MSIAB, GPU-Z and also CGMiner. All three read the same.

Never got round to fitting the WB onto the card as it is a non ref PEC, which I did not realise until I had removed the stock cooler.

Stock cooler was giving the lower temps prior to removal.

Temps rise and fall iaw load, so am kind of thinking the sensor is ok, it is just something in the cooler mounting that is being funny.

Will try it with different TIM next week (have left the stuff I wanted to use in work....doh) and see if that makes any difference.

Ahhh ok

It sounds like poor contact and with idle temps at 38c with 21% fan, that does seem high. Apart from reseating the card and making sure you screw the cooler GPU screws first and do the diagonals, no idea sorry.

 

[thefogo]

A
It sounds like poor contact and with idle temps at 38c with 21% fan, that does seem high. Apart from reseating the card and making sure you screw the cooler GPU screws first and do the diagonals, no idea sorry.

Yep this sounds like the problem to me.

 

[wazza300]

don't those coolers have a vapour chamber? idk if removing the cooler affects it?

 

[GefGz]

don't those coolers have a vapour chamber? idk if removing the cooler affects it?

I assume it's built into the heatsink assembly so doesn't really matter, unless you plan on cracking it when taking it off that is..

 

[wazza300]

yh can only be poor fitting or paste application

 

[Shelster]

Think I may have sussed the issue.

When I was loosely putting the VRM heatsink (long one on the right of the chip as looking down on it with I/O panel to left) on so that it was in place for tomorrows refit in work, I noticed that there is rather a large air gap between the thermal pad I put on there and the heatsink itself.

This I feel is the issue.

Think I have some thicker pads in work left over from when I over from when I fitted the waterblock, so will give this a go and see what occurs.

 

[Shelster]

Temps now lower than they were...phew.

Was peaking at 56 - 60 the other night when I ran a brief mining run to see what would happen....now have another issue (typical)

If i try to run anything that stresses the GPU to 100% (mining or Furmark) it will run briefly then crash the whole system.

Am using latest beta drivers so will put them back to official release later and see what happens then. Have a feeling it will be the same as is now though