Viscous thermal paste for thermal pad replacement

Hi everybody. I found a thermal paste desinged for thermal pad replacement and I would like to share it with you because it really solved my problem of replacing thermal pads after cleaning a laptop's or desktop's heatsink. The product is called K5-PRO and it is designed for Apple iMac computers but it should work just fine on all computers. It is very viscus (like a gum) and can replace practically all thermal pads. Also it is not electrically conductive and if you place a biger amount of it on the GPU you can get more heat from the GPU to the heatsink not only from it's core but from the entire chip. This results to very low temperatures....
Anyway you can see a video about it here https://www.youtube.com/watch?v=e9v5MDivZfc
and you can find it on ebay here Thermal Paste Set for Apple iMac 27" AMD MXM III Video Board A1312 K5 Pro K4 Pro | eBay
it is also avaliable on Amazon...
Has anyone else tryed it?

 

Apple computers are well known for their thermal issues and throttling. I wouldn't trust anything cooling-related made for or designed by Apple. Cooling efficiency in laptops is rather important, I guess most users here would stick to more well known and tested brands of TIMs, like IC, Arctic, Gelid, Coolaboratory etc.

No offense, but "bigger amount on the GPU" part of your post is complete rubbish.

 

Hello,
Thank's for your reply. How do you know that it is rubish? Have you ever tryed it? This product was propably designed for Apple computers exactly because they give overheating issues. It was not made by Apple but by a Greek company....The brands that you mention are great indeed but they didn't spent any time to produce a thermal paste for thermal pad replacement as far as I know...

 

It's just standard skepticism expressed towards something that promises great improvement over existing things, yet offers no independent testing or hard numbers but just a vague user description.

Also factually incorrect statements like "biger amount of it on the GPU you can get more heat from the GPU to the heatsink not only from it's core but from the entire chip".

 

I was also skeptical when I found it but it costs around 7$ including postage so it was worth trying...However I can't understand why my statement is wrong....The entire chip gets warm...not as warm as the core part but still this paste gives a direct connection from the entire chip to the heatsink..Can you please explain? If I am wrong indeed I would like to know it so that I don't waste extra paste in the future....

 

The effectiveness of a TIM is directly related to the thickness of the layer of paste - thinner is better. If you glob more on there, even if it's touching other areas, the heat transfer from those areas will be negligible.

 

The thermal transfer of an interface is indirectly proportional (almost linear) to the thickness. The thermal transfer is asymptotic with respect to conductivity (ie its irrelevant at a certain thickness)

I'll link you to the equations when I get home but basically, this product has promise only as a thermal pad replacement provided its thermal conductivity is sufficient. Its semi liquid nature makes it superior to thermal pads due to higher wetting and contact but inferior to a true TIM in tight contact scenarios because it makes it impossible to achieve an intimate contact.

 

You got my point in a wrong way. This product is very flexible. If originally there was a thermal pad of let's say 1mm thickness, no matter how much K5 you will place, when the heatsink is attached the layer of K5 will be also 1mm because of the heatsink presure. The extra amount of this paste will go on the chip and will allow contact of other parts of the chip to the heatsink. I see serious changes in GPU temperatures (around 2-4 degrees lower) so it is not negligible. Also many serious brands place a thermal pad on some CPUs around the core for better contact of the CPU to the heatsink so I am not claiming that I descovered America here... The thickness of the layer between the component's core and the heatsink will be the same. Contact is definitelly better than thermal pads because this product is quite sticky.
Finally I don't use it when there is direct contact between the heatsink and component. In those cases I use a low viscocity thermal paste called K4-PRO (yes it is made by the same company). K5-PRO is usefull when you need to fill a gap that was originally filled by a thermal pad and it is imposible to achive direct contact between the heatsink and component and if you would install normal paste it would split away.
I hope that I could get an opinion of someone who has also tryed it. Also I can't understand why there so much rejection without even trying a product that sugest a solution to a problem that all of us who deal with computers and gamestations have....To find the proper thermal pad every time I open a laptop for cleaning...It's price is lower than a good thermal paste including postage and from what I understand on their ebay offer you can try it and if you don't like it you can return it and get a refund...

 

I might try it on my vRAM to replace those thermal pads, but for CPU and GPU dies I'll stick with a proven TIM paste.

I don't understand the negativity either. It's inexpensive, and worth a shot if you need/want something better than a thermal pad that already exists. At 4W/m-K it's higher thermal conductivity than most quality pads, and should get a better contact with the surface you're trying to cool. Although the comment " practically zero electrical conductivity" has me worried. Practically doesn't sound like none, and even a little bit of conductivity is enough to cause a short.

Edit: Just watched the video. Issue I see is that thermal pads for vRAM are usually 2-3mm thick and I don't see that paste being that thick to take up that gap appropriately. I definitely wouldn't use it on the GPU die though still. And also realized the M5-Pro is the 4W/m-K and the M4-Pro on the RAM sinks are 3W/m-K. Even less impressive. Just buy some decent TIM paste and some thermal pads instead...

 

I am sorry if I sounded negative - that was not my intention. I am a bit skeptical towards anything Apple+thermals, besides if the K4-PRO or K5-PRO were any good, there would have been some reviews available online. What is more, there are thermal pads available with much better parameters than the above products and they're made by reputable companies, like 3M, Akasa, Alphacool, Phobya, EK, Fujipoly (even 17W/m-K) etc.

 

Thank you for your sugetion. I think you got it a bit wrong. On the RAM chips you should instal K5-PRO which is very viscus and has 4W/m-K . It can stick perfectly in gaps up to 3mm thick, I try it almost every day and I know it does.Once you try it you will see how wrong all your assumptions are. K4-PRO is for the GPU or CPU when there is no thermal pad.

About the electrical conductivity and since I am electronics engineer I know that all matterials have some electrical conductivity. If they would say it is zero then this would be an obvious lie...Maybe it would sound nice to users with limited knowledge but I think that this product refers to proffessionals mostly. In our language practically zero means that on sugested aplications there is no electrical conductivity. Don't forget that it is made in EU and not in China....

I don't understand your final sugetion...Do you sugest to place thermal paste of usual viscocity on the RAM chips? But it will split around especially when it gets warm and it will leave a gap...Or do you sugest to place thermal pad and thermal paste the same time? I don't like this idea but even if it is correct I still have the problem of having a load of thermal pads in order to find the correct one every time...I say that once you try it instead of a thermal pad you will forget about thermal pads. Honestly it is so simple. It comes in a matchbox size container and it has enough amount to replace around 30-40 thermal pads. If you work with computer repairs just consider how many thermal pads you have around in your workbench and how much time you spend each time to find the correct one....

They have a lot of reviews on their feedback on ebay where I found it. However this is a new product....Yes there are oviously great thermal pads. However with K5 the equivalant of 30 pads costs less than 7$...how much one Fujipoly pad costs? Can you aford to use it if you are computer technitian like me and you clean a student's laptop for 15$-20$ (yes this is how much it costs where I live...)

 

I understand your point, but not everyone here cleans students' laptops for living. I do all maintenance of my hardware myself (and I am a student too ). I have 2 laptops and I'd rather use a bit more expensive TIMs which provide better performance in them. I don't need a drawer full of different pads, I know exactly which ones I need. Many people here on NBR forum are hardware enthusiasts and their approach is similar to mine. They want top performing products, even at a higher cost, rather than cheaper ones, which are "good enough", especially that many of them (like me) overclock their hardware. In such usage scenarios having best performing Thermal Interface Materials is crucial.

Thermal pads were designed for a reason, they come in various thickness and sizes, you can buy them in bulk when needed and they are easier to apply then a paste.

Don't get me wrong, I am not against the products you mentioned, it's just that for my usage they are no good, that's all.

 

That's correct. I guess I see the situation from my point of view and I missed your's. Obviously when it comes to overclocking all details are important and speniding a few extra $ for the best is well worth it. Still for my work this was a life saving product and that's why I wanted to share my experience with the forum. I now hope that someone else who has used it will give his review...

 

Thanks for sharing. That's what this forum is all about.

 

Any more actual reports on it?

I really hate thermal pads, they are so uncomfortable to use, and you can't usually use typical thermal paste instead unless you put something like polished copper sheet in the gap which may end up damaging chip if your measurements ain't correct. Overall, that's too much work if it's not CPU or VGA but northbridge or memory you're trying to cool.

 

Actually, this K5-Pro stuff reminds me a lot of IC diamond except even more viscous and cheaper. Would be fantastic for applications where intimate contact is not desired or possible.
It'll be brilliant for RAM and GPU VRM assemblies which have been the traditional users of Thermal Pads. Some Laptop heatsinks which would normally require copper shims may also benefit.

 

In my opinion every situation has its needs. In case of a high performing GPU/CPU its best to use a paste & pads TIM combo to ensure the fastest and most efficient heat transfer. And that usually applies to us enthusiasts here in the NBR forums. For common folks (90% of the population) maybe it could be a solution. I say maybe because even in that case easy does not mean trouble free. It can cool down an i3 or a lowend GPU but I wouldn't use it to cool down my i7 and 780m. I'm not being skeptical here, I'm just curious.

 

So your sugetion is to place a layer of TIM on the GPU then thermal pad, again a layer of TIM and then the heatsink? But this will create a 3 layer sandwitch between the heatsink and GPU with different thermal conductivity....I am no expert here but from what I learned in thermodynamics at the university when you have layers of means with different thermal conductivity you seriously reduce the total thermal conductivity of the set... This is actually a good way to create a thermal insulator as far as I know. Also I have never seen any computer or other electronics brand to place both thermal pads and thermal paste at the same time...

So I guess the only way to find out is to do some testing. I have a load of thermal pads and copper shims and some different thermal pastes. However I am not sure what kind of testing I should do. I guess I will need a software to count the temperatures and one more to make the computer warm. Any sugetions?

 

Run Prime95+Kombustor's Furnark test for an hour, while measuring temps with Speedfan. And don't forget to check your CPU&GPU datasheets for max allowed temp before testing, so that you could stop right before they meet the threshold - that means a fail for current theermal solution anyway.

 

I know this slightly irrelevant but will the following get the old thermal paste of my CPU and heatsink?
20 IPA NHS GRADE 70% ISOPROPYL ALCOHOL SWABS WIPES CPU HEATSINK TATTOO | eBay

Thanks

 

Medical or optical spirit = best choice. Cheapest vodka will do, though.

 

But what about this? this is the best I can get... Can't get the pure one...

 

I wouldn't have messed with this stuff, rather go to a shop and buy the cheapest vodka. You can always use the rest for cocktails anyway.

 

Screw that, like I said in another thread, HPLC grade or bust!
Don't though, that stuff is expensive!

 

Well I don't drink.. Has anyone used these swabs before? I was told that they work..

 

I haven't used them (sorry), but I too didn't drink...

before I joined this forum.



Seriously, if you were told they work... do you doubt the veracity of the one who told you?

 

I have used similar swabs before and they did work. Q-Tips + 99% IPA remains my favorite though.

 

I always use optical spirit. Never had any problems and old paste is cleaned very easily. Just make sure that CPU is dry after the cleaning.

I will try the testing sugetion on the weekend and I will let you know.

 

The thing with the swabs is they're way cheaper then getting 99% alcohol for me.. I can get the 99% but if the swabs do it, I rather save the money and get them.. The only issue is have you used it with Dell's stock thermal paste? Does it come off easily with the swabs?

 

70% isopropyl alcohol (basically rubbing alcohol) does the trick for me, even on dried out paste that haven't seen the light in 2 years.

 

Have used the swabs on stock Asus paste, shouldn't be too different than Dell's. I used 99% IPA to clean Dell paste since I had some on hand when I repasted my M6700.

 

Gone ahead and bought the swabs.. Let see how it goes... Buying Arctic Cooling MX-4 to repaste since ICD seems to be a bit lethal...

 

Lethal? LOL. Not lethal, been using it for years on my own and many family, friends, and client laptops.

 

As promised I did some testing with this thermal paste today and here is the results:

Testing system nVidia GeForce 9800GT
Room temperature 28C

I used a software called rthdribl in order to push the GPU to it's limits and measured the temperatures with Speedfan

This is the video board from the computer I am mostly using. I had it cleaned and applied Arctic Silver MX-2 thermal paste about 4 months ago. I removed the board, cleaned some dust that was in the heatsink and istalled it again. Run rthdribl and 5 minutes later the GPU temperature at speedfan was stabilized at 76C

I removed the board again. Cleaned old thermal paste and installed new Arctic Silver MX-2. Run rthdribl and after five minutes GPu temperature on speedfan was 76C again. So for me this proved that MX-2 didn't change it's cunductivity after 4 months of use (more than 10 hours a day)

I removed the board again, cleand thermal paste and istalled it without any paste at all (just to see how important paste is....). I run rthdribl and after 2 minutes the GPU temperature was 89C and going higher.....I turned off computer imediatelly....If anyone tells you that thermal paste is not important then he is completelly wrong...even with clean heatsink I could fry the GPU if I kept using it...

Removed the board again and installed a small drop of K4-PRO thermal paste on the center of the GPU. I also replaced the thermal pads on the v-memories with K5-PRO thermal paste. I run rthdribl and 5 minutes later the GPU temperature on speedfan was 75C. I let it running for 5 more minutes to be sure but there was no change (it would drop to 74C some times for a second but never got higher).

K4-PRO claims thermal conductivity K>3 while MX-2 claims K>3.5....So why did K4-PRO gave lower temperatures? I am not an expert here but when I removed the heat sink after both testings the difference was that MX-2 is more viscus that K4-PRO and it would almost create a film on the GPU while K4-PRO would leave only small pins on the GPU ( I guess where the gaps where). So I believe that K4-PRO alows biger area of direct contact between the heatsink and GPU while the gaps are filled. Anyway the difference is not very serious here.

Final testing was the most interesting. I removed the board and placed a small drop of K4-PRO on the GPU and filled the area around the GPU core with K5-PRO thermal paste so that compete GPU would be in direct contact with the heatsink through K5-PRO. I am not sure if I am making it clear so I am attaching a photo. I also placed K5-PRO on the V-Memories. I run rthdribl and the GPU temperature on speedfan was 71-72C. I left it running for half an hour and it never got higher than 72C . So by applying K5-PRO thermal paste around the GPU core where there is usually nothing but air decreased the (maximum?) temperature by 3C. I think that this is quite a success.

Unfortunatelly speedfan didn't show the v-memory temperature so I didn't actually test the temperature difference between thermal pad and K5-PRO thermal paste...So I will try to get a laptop with thermal pad on the GPU and do the same testing.

I hope that this will be usefull to someone....It took me a couple of hours to complete the testing but it was quite interested so I enjoyed it

 

This looks like a VERY, VERY, VERY BAD paste job.

 

Did you read the my testing description or are you refering to the photo only? Yes it looks very bad (the photo didn't get good also) but the result was 3C lower...So it looks bad but it works great....Do you want your video board to look nice or to operate cooler? I am not a game user so 3C lower is not important for me but I know people who installed very exensive heat sinks and cooling systems to get the same result that I got with just this thermal paste....

I hope that it is clear that there are two types of paste on the GPU...Low viscosity K4-PRO paste on the core and viscous / gammy K5-PRO paste around the core and on memories. This paste is very very viscous and this is why it is very hard to make the result "look nice"

 

I was referring to the photo only.

My video board costs ~600USD (new). It's not a cheap, expendable, obsolete desktop 9800GT I can experiment on. If I want it to run cooler, I will buy a more advanced thermal paste (or upgrade the heatsink, like I have done a couple of weeks ago, but it's a rare possibility in laptops).
I will not experiment with something I have never heard of. Don't get me wrong, my first overclocked chip was Intel 386SX 33MHz (OCed to 40MHz) back in 1992. I know what I'm doing and how to do it, and what to use and what not...

Have a look at this thread:
http://forum.notebookreview.com/har...et-upgrades/754618-best-thermal-compound.html

In laptops, there are factors influencing their cooling systems' efficiency, such as die/heatsink contact. In every laptop it's different, so results might be very different. Your paste might have worked on your desktop GPU, but in laptops it can be a completely different story.

 

I didn't ask you to experiment...I did the testing and gave the results....This is a new product so I wouldn't risk a 600USD board also . I tryed it on a board that is cheap now but cosidered high end on it's time.... I now have a toshiba laptop with thermal pad on the GPU. I will do some testing with copper pads and thermal pads and this paste and I will let you know.

 

I did one more testing on a laptop Toshiba Satelite L450D with Ati HD 3200 GPU. This laptop has a thermal pad on the GPU.

Room temperature 27C
Testing software Furmark and speedfan

This laptop has a thermal pad installed on the GPU so this is exactly what I needed.


First I used the air blower to clean the dust from the heatsink externally. I run Furmark and after 5 minutes the GPU temperature on speedfan was 61C

I removed the heatsink carefully, removed the thermal pad, cleaned some dust that was still around and installed again the old thermal pad and heatsink. Again did the same testing and after 5 minutes the temperature was 66C and going higher....So reusing the same thermal pad is definitelly a mistake.

Removed the heatsink again and installed a new thermal pad. This laptop needs a 0.8mm thermal pad but I still have a lot of them so I found an exact match. Didi the testing again and now GPU temperature was 59C. I used to buy the best pads that I could find so the pad I used was propably better than the one originally installed.

Removed the heatsink again and placed a new thermal pad with thermal paste on both sides. After the testing temperature was 60C. So using both thermal pads and thermal paste is not a good idea.....

Removed the heatsink and placed a copper sim 0.8mm with thermal K4-PRO paste on both sides. After 5 minutes of testing temperature was 50C!!!

Removed the heatsink and placed a thiner copper sim 0.5mm with MX-2 thermal paste on both sides (I didn't use K4-PRO because it is very liquid and wouldn't fill the gap). After 5 minutes temperature was 60C and would go higher...propably the thermal paste was getting more and more liquid increasing the gap

I tryed to place a copper pad of 1mm thick but the heatsink wouldn't stand properly so I didn't push it.

Removed the heatsink and placed only MX-2 thermal paste. Did same testing and after 5 minutes temperature was 68C and rising.....Again I didn't use K4-PRO because it is very liquid and wouldn't fill the gap. So replacing a thermal pad of this thickness with normal thermal paste is a mistake.

Removed the heatsink again, cleaned thermal paste and placed a quite big amount of K5-PRO thermal paste on the GPU. Did the same testing again and after 5 minutes temperature was 54C.

My last testing was to add even more K5-PRO thermal paste so that entire GPU was covered with it. This gave direct contact of the GPU to the heatsink. I did the same testing again and 5 minutes latter temperature was 53C.

So finally I used the copper pad....

From this testing it was clear that if you can place a copper pad of exact same thickness as original thermal pad then you will get the best result. However if your copper pad is not thick enouth then it will not do any good. K5-PRO gave much better result than the new thermal pad and also it was much easier to use because I didn't have to measure the thickness of the thermal pad exactly. I just placed a ball of K5-PRO on the GPU and the heat sink did the rest of the work.

So at this point I will forget my box of thermal pads permenently and use K5-PRO on normal cases and copper pads when I need if for a gaming computer.

 

This all comes back to what a lot of us were saying, it's the wetting of the surface that matters the most when it comes to paste. When you get at thicknesses like 0.8 mm, then thermal conductivity can make a decent difference, but the benefits will always end up being lower with thinner TIM layers. Thermal pads are convenient because of their texture and ease of application in some cases, if you feel safe going with copper shims, that will definitely offer better performance.

 

That's exactly my point. Obvioulsy I would never place a copper shim where there was no thermal pad originally but since I mostly deal with laptops my problem is the replacement of thermal pads. I still don't understand why laptop companies are using thermal pads on sensitive components like the GPU.... Anyway for thermal pad replacement at them moment I have a solution with this viscous thermal paste. However if I understand your point correctly (and I also feel the same) it is better to use thermal paste with low viscosity when you have direct contact with the heat sink. Unfortunatelly most high thermal conductivity pastes on the market are quite viscous. Except this K4-PRO that I am using and is a new product so it needs some time to prove it's quality, can you point another paste with good thermal conductivity and low viscosity?

 

Hi there. On my previous post I didn't mean placing both thermal pads and thermal paste on the same GPU die. I ment using both as a combo: pads on memory and mosfets and paste on GPU die