Thermal paste myth?

[Salmiakbal]

So I have laying here before me two kinds of thermal paste. One is a cheap no name paste they sell at a local computer shop, the other the well regarded Arctic MX-4.

A few months ago I redid the thermal paste on my desktop. I was in a hurry and went to the local computer shop to get the stuff. They didn't have any well known thermal pastes so I just got the paste of which they said ''this is what we always use here.''.

So I went back home and did my CPU and GPU.

Now a few months later I did a clean install of Windows 8 on the desktop and I figured it was time to overclock my GPU a bit and see how hot it got.

Using the giant doughnut of death (aka furmark) as a torturer, I stressed the GPU and it turned out it got to a fine 75C with the 1GHz overclock.

I recently got MX-4 to re-thermalpaste my laptop ()and of course I still had a lot left over, so I figured I should test out to see if it makes any difference with the no name paste.

I took the graphics card out of the case, undid the cooler etc etc and I put on the MX-4 compound, after which I redid the test.

...turns out it gets just as hot. 75C as a max temp and the fan speed the same (60%) the card is a Sapphire HD 6870.

I often see people saying you should get good paste, and I'n not saying you shouldn't. But seeing this, does it actually make a difference?

What are your experiences?

[fishymamba]

I did the same thing actually, except I use TIM Consultants Quantum instead of MX-4. Temps dropped ~10C on my CPU from the switch.

In my opinion temperatures are more dependent on how you mount the cooler than the TIM.A bad application(air bubbles, uneven pressure,too much/not enough tim), can have a huge impact on temperatures.

[DeamonFish]

I did the same thing actually, except I use TIM Consultants Quantum instead of MX-4. Temps dropped ~10C on my CPU from the switch.

In my opinion temperatures are more dependent on how you mount the cooler than the TIM.A bad application(air bubbles, uneven pressure,too much/not enough tim), can have a huge impact on temperatures.

This has been my experience with thermal paste as well.

[Wats]

Does it taste like MX-4?

This is just more (good) proof that the closer the heat to the copper, the better. We've all seen that chart where ketchup beat some TIMs while mustard failed :P

At least they were just trying to sell you something when they said "this is what we always use here.". Maybe they were o.O

Either way, thanks for sharing, next time I see some really cheap TIM I'm going to see how it compares.

[Salmiakbal]

Does it taste like MX-4?

This is just more (good) proof that the closer the heat to the copper, the better. We've all seen that chart where ketchup beat some TIMs while mustard failed :P

At least they were just trying to sell you something when they said "this is what we always use here.". Maybe they were o.O

Either way, thanks for sharing, next time I see some really cheap TIM I'm going to see how it compares.

I think it was what they always use there - it was their only paste :p

The cheap paste was pretty watery and if you got it on your finger the silver colour was hard to get off..

[Salmiakbal]

I did the same thing actually, except I use TIM Consultants Quantum instead of MX-4. Temps dropped ~10C on my CPU from the switch.

In my opinion temperatures are more dependent on how you mount the cooler than the TIM.A bad application(air bubbles, uneven pressure,too much/not enough tim), can have a huge impact on temperatures.

10C seems like a big drop in temps :p I think it depends from case to case

[IdeaStormer]

How many of us have run without any thermal paste (due to running out) and it is still fine? I still remember someone using toothpaste the chalky type (not gel) and it worked, sure not as good but still worked, as in the CPU did not get fried.

[Helltech]

The difference between thermal pastes is minimal at best, but I do know that putting GCExtreme on my CPU from what ever was on there (can't even remember), did drop my temps about 4 degrees.

[IdeaStormer]

The difference between thermal pastes is minimal at best, but I do know that putting GCExtreme on my CPU from what ever was on there (can't even remember), did drop my temps about 4 degrees.

I bet if you cleaned the old paste and re-applied the same paste you'd get better numbers as well. Has anyone tried that?

[Aphexx]

myself, I use cheapass thermal paste as well. I wonder if the "quality" paste is about how it performs over time and not the immediate effects of it. the long game over 5-10 years of heat cycles.

[fishymamba]

myself, I use cheapass thermal paste as well. I wonder if the "quality" paste is about how it performs over time and not the immediate effects of it. the long game over 5-10 years of heat cycles.

I actually bought a tube of cheap thermal paste at a local store to try it out. At first temps were better than MX-4, but after 2 weeks they started getting very high.

I removed the heatsink and the compound had pretty much turned into a fine dust.

[Whaler_99]

Typically a "quality" paste will have some better heat transfer properties... but also depends on the cooler and how quickly it can draw off the heat. Also, as seen by many techies, longevity. Will that past still produce quality temps 6 months later? A year? And one important thing, conductivity. The cheaper pastes can be somewhat conductive and 99% of the time this is fine. The one time you spill a bit though and it isn't? :)

I have used a lot of different kinds... I can usually spot a 2-3 degree difference between my "it came in the box" and my quality MX-4. Water cooling GPU's though can torture any TIM and a high quality one needs to be used or it will break down FAST. For the average system though, most any TIM is ok. As long as it is spread evenly and the heat sink is mounted properly...

Is the 20 dollar tube so much better then that 8 dollar tube which is so much better then the 2.50 one? I would say get the 8 dollar one - will be better then the cheapo stuff. 20 bucks? Not so much...

[gaza4]

i have in the past used automotive copper grease as a thermal compound.

my old laptop was having major heat issues a few years back and i was recommended to clean out the fan. being a tech noob at the time i didnt realise that the thermal paste would dry out and become useless.

being in the tight spot i was i improvised and used the copper grease and it worked out pretty nice. being copper gave great heat transfer although being a grease its consistency was far from desirable, the clean up when applying proper paste was messy to say the least!!!

[Kidbest100]

As Linus himself has said in many videos, GPU's do get hot, but they don't OUTPUT a whole lot of heat. They get hot in the actual core, but you cant cook an egg with one like you can with a hot CPU, because they don't put out the heat as well.

The actual cooler you have on the card is more important in most cases than the paste you use. This is why when you see things like Asus' DirectCUII cooler, with all the huge heatpipes and stuff, the card actually does have lower temperatures, because that cooler is designed to take the heat away from the GPU, and move it somewhere else (Heatpipes, fins, etc etc,)

The thermal paste is only the medium through which the heat transfers. No matter what paste you use, there is a maximum amount of heat that it can transfer from the things its cooling, to the cooler itself. Your MX4 isn't making that much of a difference because its transferring as much heat as it can, but being bottlenecked by the cards cooler.

A good analogy I heard once, may not be the best one but here it is. I won't be surprised if it makes no sense to you at all.

Cars = Temperature

Tunnel = Thermal paste

Cars destination = Cards cooler

"You can only fit as many cars through a tunnel as it has room for. You can make the cars go faster, and get them to their destination better and more effectively than if they were going slow, however putting more and more cars in is not going to make the tunnel wider by itself so that it can accommodate more cars. If you want to fit more cars in, get a bigger tunnel, and make sure their destination has a wide enough road to accommodate them.

Anyway, I'm not really a seasoned veteran when it comes to this stuff. I might be completely wrong.

Feel free to tell me I am if I'm wrong so that I can learn what is really going on.

[aaronu]

heres how i look at thermal compound. white goop stuff is cheaps stuff. not going to get the best cooling. grey silvery stuff is the good stuff.= good cooling

[aaronu]

How many of us have run without any thermal paste (due to running out) and it is still fine? I still remember someone using toothpaste the chalky type (not gel) and it worked' date=' sure not as good but still worked, as in the CPU did not get fried.[/quote']

why would ya risk frying ya computer lol. Not only that that would proably make ya cooler fan run at max speed. how annoying lol. I never run out of thermal compund. Cause i buy and build computers for family and friends. cause i always toss the stock coolers cause they blow balls. and I buy coolermaster heatsinks so they come with some. I personally use artic silver 5 on my gaming rig!

[IdeaStormer]

How many of us have run without any thermal paste (due to running out) and it is still fine? I still remember someone using toothpaste the chalky type (not gel) and it worked' date=' sure not as good but still worked, as in the CPU did not get fried.[/quote']

why would ya risk frying ya computer lol. Not only that that would proably make ya cooler fan run at max speed. how annoying lol. I never run out of thermal compund. Cause i buy and build computers for family and friends. cause i always toss the stock coolers cause they blow balls. and I buy coolermaster heatsinks so they come with some. I personally use artic silver 5 on my gaming rig!

Try it, then get back to me.

[Salmiakbal]

As Linus himself has said in many videos, GPU's do get hot, but they don't OUTPUT a whole lot of heat. They get hot in the actual core, but you cant cook an egg with one like you can with a hot CPU, because they don't put out the heat as well.

The actual cooler you have on the card is more important in most cases than the paste you use. This is why when you see things like Asus' DirectCUII cooler, with all the huge heatpipes and stuff, the card actually does have lower temperatures, because that cooler is designed to take the heat away from the GPU, and move it somewhere else (Heatpipes, fins, etc etc,)

The thermal paste is only the medium through which the heat transfers. No matter what paste you use, there is a maximum amount of heat that it can transfer from the things its cooling, to the cooler itself. Your MX4 isn't making that much of a difference because its transferring as much heat as it can, but being bottlenecked by the cards cooler.

A good analogy I heard once, may not be the best one but here it is. I won't be surprised if it makes no sense to you at all.

Cars = Temperature

Tunnel = Thermal paste

Cars destination = Cards cooler

"You can only fit as many cars through a tunnel as it has room for. You can make the cars go faster, and get them to their destination better and more effectively than if they were going slow, however putting more and more cars in is not going to make the tunnel wider by itself so that it can accommodate more cars. If you want to fit more cars in, get a bigger tunnel, and make sure their destination has a wide enough road to accommodate them.

Anyway, I'm not really a seasoned veteran when it comes to this stuff. I might be completely wrong.

Feel free to tell me I am if I'm wrong so that I can learn what is really going on.

''As Linus himself has said in many videos, GPU's do get hot, but they don't OUTPUT a whole lot of heat.''

That seems very weird. Graphics cards often consume up to 250 watt, while CPU's have a hard time even getting over 125 watt.

I believe graphic cards get hot because of the limited space they have in the PCIe slot, which limit the cooler size. This gets confirmed by the fact that graphic cards can be very cool when incorporated into a water cooling loop, where the cooling isn't constricted by the PCIe slots.

[Puppet]

How many of us have run without any thermal paste (due to running out) and it is still fine? I still remember someone using toothpaste the chalky type (not gel) and it worked' date=' sure not as good but still worked, as in the CPU did not get fried.[/quote']

why would ya risk frying ya computer lol. Not only that that would proably make ya cooler fan run at max speed. how annoying lol. I never run out of thermal compund. Cause i buy and build computers for family and friends. cause i always toss the stock coolers cause they blow balls. and I buy coolermaster heatsinks so they come with some. I personally use artic silver 5 on my gaming rig!

Are you seriously trying to get him to run his cooler without TIM?

[Salmiakbal]

How many of us have run without any thermal paste (due to running out) and it is still fine? I still remember someone using toothpaste the chalky type (not gel) and it worked' date=' sure not as good but still worked, as in the CPU did not get fried.[/quote']

why would ya risk frying ya computer lol. Not only that that would proably make ya cooler fan run at max speed. how annoying lol. I never run out of thermal compund. Cause i buy and build computers for family and friends. cause i always toss the stock coolers cause they blow balls. and I buy coolermaster heatsinks so they come with some. I personally use artic silver 5 on my gaming rig!

sssssssssssh!!!