Ek 290x backplate Contact

J

Justintoxicated

[H]F Junkie
Joined
Apr 10, 2002
Messages
14,522
This seems odd.


290x ek backplate contact by Glamisduner, on Flickr

Ok another question, when I put on my thermal pads I use some thermal grease on the memory and VRMs then applied the pad. However when I attached the black I forgot to add any compound to the other side of the pad.

Should I bother re-doing it? If so can I simply just add compound or do I need to cut up more fujiploy? Stuff isn't exactly cheap...

For the backplate I just re-used the stock thermal pads and only added grease to the side that touches the card as well. I carefully removed the sticker that was on top of those capacitors or resistors or whatever they are, it was really sketchy. I didn't bother removing the entire sticker cause I was afraid of breaking one off with my finger nail. It was stuck on there good. Card is Sapphire brand.

Card is currently mining at 40c VRMS are at 37 and 33c
 
You don't put thermal paste on the pads. You just apply the pads.
 
Looking at the directions per their website, it looks like you were advised to apply non-conductive grease on top of the thermal pads on the front side on the GPU for the phase regulators and not the backplate section. The backplate + provided thermal pads should be more than sufficient to passively cool that side.

Front
Back

And from the looks of it, you totally missed them on the back. The pad supposed to align with those screws and from the picture you posted, it's a little off to the left.

Hope this helps!
 
nob0dey said:
Looking at the directions per their website, it looks like you were advised to apply non-conductive grease on top of the thermal pads on the front side on the GPU for the phase regulators and not the backplate section. The backplate + provided thermal pads should be more than sufficient to passively cool that side.

Front
Back

And from the looks of it, you totally missed them on the back. The pad supposed to align with those screws and from the picture you posted, it's a little off to the left.

Hope this helps!
Click to expand...

I don't think I missed them although I do agree that I thought it was supposed to hit the regulators between the screws.

The diagram you linked suggests contact is supposed to be made off to the side of the line between the 2 screws. However there is only one protruding place on the backplate the thermal pad can go, and my picture above shows where it makes contact, so it must be correct.

Heres a picture from Stren on extreme systems comparing the EK to the aquacomputer backplate.
0hJCSkjl.png
 
Last edited:
It looks like its trying to cool the Coils, not the VRMs with the back plate.

Also, I don't care what EK says, adding more material between the contact points will degrade cooling, period. Just use the pad.
 
Jorona said:
It looks like its trying to cool the Coils, not the VRMs with the back plate.

Also, I don't care what EK says, adding more material between the contact points will degrade cooling, period. Just use the pad.
Click to expand...

Has testing been done on this? Adding a TIM will ensure better contact into the pad, which his going to compress to a certain height anyways? Seems you would get better contact with a thin layer of tim spread out manually, just like researching online suggested using a tim between layers if you have to stack pads.

BTW the backplate is getting hot to the touch when mining.
 
Most pads either a oily or gummy surface, allowing them to make proper contact regardless of surface. It its a particularly crappy pad that wont mate properly with the surface to be cooled, then MAYBE tim might help, but Adding more material between the cooler and the surface to be cooled will only degrade performance. So it's a catch 22 with shitty pads, and a no no on good pads (Like Fujipoly)
 
Jorona said:
Most pads either a oily or gummy surface, allowing them to make proper contact regardless of surface. It its a particularly crappy pad that wont mate properly with the surface to be cooled, then MAYBE tim might help, but Adding more material between the cooler and the surface to be cooled will only degrade performance. So it's a catch 22 with shitty pads, and a no no on good pads (Like Fujipoly)
Click to expand...

Can you point me to some proof? Your stating that EK's directions are completely wrong? From what I read: Typicaly TIMS are significantly better at transferring heat than pads, and I still do not agree that you are adding any significant thickness to the pad. Can you link me to some tests because the links I find say that you should use a tim in addition to the pad for best results while others say it will run hotter if you do this. I could see how adding a glob of paste could significantly increase the thickness but most pads have some give and flatten out when they get compressed anyways, they will flatten to pretty much the same height regardless if a tim is used in addition...

Hell that's what took most of my time was spreading a thin layer of tim for each pad. I probably spent 2 hours doing this...Adding a more dense material between the cooler and the sink should increase performance while increasing the gap would decrease performance. So in sense there should be some balance when using pads. As for the backplate specifically, the pads on my block were already used and compressed, so I applied a thin layer of Thermal Grease on them to make sure I was getting proper contact. Lookup the cost of 1.5mm thick fujipoly and you will see why I went this route, the card requires .5mm, 1mm AND 1.5mm pads. I hardly think an additionl .1mm of thermal paste is going to make any difference when your using 1mm and 1.5mm thick pads to start with.

My thought is that the pads would compress to the same height regardless if you used a TIM between them or not, there's likely more variance from different brand cards than the thin TIM itself would make. I'm not after all putting a glob of tim down, I'm spreading it thin by hand. I don't see how this could be an issue but I don't want to burn out my $600 video card either. My temperatures are great so I probably won't be screwing with it. My VRMs get less hot than Stren's testing using the stock Pads.

So to clarify your telling me I need to buy another $70 worth of fujiply and start over? Here is the msallest 1.5mm piece I can find. http://www.frozencpu.com/products/1..._100_x_15_-_Thermal_Conductivity_110_WmK.html. I think I will wait for more opinions before I make such an investment. That would bring the cost of the pads to just under the cost of the waterblock... Also please note, I have years of experience overall, but this is my first time screwing with thermal pads and maybe what I did wasn't optimum but I believe I should be ok (as per EK's directions for the block, and my own discretion for the backplate. So far I hate them, they are crazy expensive and crappy compared to TIM.
 
Last edited:
From what I understand adding a small dab of tim to the thermal pads does little for cooling but can help to keep the pads in place when attaching the waterblock/backplate.
 
Justintoxicated said:
Can you point me to some proof? Your stating that EK's directions are completely wrong? From what I read: Typicaly TIMS are significantly better at transferring heat than pads, and I still do not agree that you are adding any significant thickness to the pad. Can you link me to some tests because the links I find say that you should use a tim in addition to the pad for best results while others say it will run hotter if you do this. I could see how adding a glob of paste could significantly increase the thickness but most pads have some give and flatten out when they get compressed anyways, they will flatten to pretty much the same height regardless if a tim is used in addition...

Hell that's what took most of my time was spreading a thin layer of tim for each pad. I probably spent 2 hours doing this...Adding a more dense material between the cooler and the sink should increase performance while increasing the gap would decrease performance. So in sense there should be some balance when using pads. As for the backplate specifically, the pads on my block were already used and compressed, so I applied a thin layer of Thermal Grease on them to make sure I was getting proper contact. Lookup the cost of 1.5mm thick fujipoly and you will see why I went this route, the card requires .5mm, 1mm AND 1.5mm pads. I hardly think an additionl .1mm of thermal paste is going to make any difference when your using 1mm and 1.5mm thick pads to start with.

My thought is that the pads would compress to the same height regardless if you used a TIM between them or not, there's likely more variance from different brand cards than the thin TIM itself would make. I'm not after all putting a glob of tim down, I'm spreading it thin by hand. I don't see how this could be an issue but I don't want to burn out my $600 video card either. My temperatures are great so I probably won't be screwing with it. My VRMs get less hot than Stren's testing using the stock Pads.

So to clarify your telling me I need to buy another $70 worth of fujiply and start over? Here is the msallest 1.5mm piece I can find. http://www.frozencpu.com/products/1..._100_x_15_-_Thermal_Conductivity_110_WmK.html. I think I will wait for more opinions before I make such an investment. That would bring the cost of the pads to just under the cost of the waterblock... Also please note, I have years of experience overall, but this is my first time screwing with thermal pads and maybe what I did wasn't optimum but I believe I should be ok (as per EK's directions for the block, and my own discretion for the backplate. So far I hate them, they are crazy expensive and crappy compared to TIM.
Click to expand...

Look, I have no proof, but my reasoning is sound. If you don't agree, that's fine. I've given you my opinion, and the reason why I've come to said opinion.

Its just like TIM, you spread it thin to minimize the distance from your block to the die. By adding TIM, you're adding distance. While I don't believe the difference will be more then a couple degrees, I do believe there will be a difference.

I know how much Fujipoly costs. I am part owner of a shop that does custom builds and repairs everyday. I know that FrozenCPU gets about 40% markup on those pads. They are some of the best pads on the market though, and are what I use in laptops and video cards when they come in for cleaning and remounting the thermal solution.

When it comes down to it, I doubt you're hurting anything, I'm just saying that you should just use the pad in the future, and save your TIM for direct contact areas.
 
You must log in or register to reply here.
Back
Top