Thermal Paste Longevity

[WhitehawkEQ]

What is the longest anyone has gone before replacing thermal paste? I have a PC I built in 11/2010 that I have not changed the thermal paste, only thing I replaced was the fans on the CPU HS. That's almost 13 years of 24/7 running other than power outages.


Post what is the longest you've gone before you replaced your thermal paste

 

[EarthDog]

Not that long, wow!

A couple of years on my kid's machines at most.

 

[QuantumCookie]

I was about to buy some fancy thermal paste earlier, but then I think I read that it was expected that one would have to change the thermal paste once a year, or more frequently. :O
So I stuck with some basic thermal paste for my cpu.

I vaguely recall trying out some liquid metal years ago, and I had that one for, who knows how long. On removal, I think it looked like it had dried up or something.

There is now, also some kind of pads you can buy, super thin, and can be used multiple times and will last for many years I think I've read. Pads made of some kind of graphene substance. Apparently rips easily when if you touch the fabric.

 

[WhitehawkEQ]

This is what I use.

 

[Kenrou]

My NT-H1 (CPU and GPU) is going on 14 months and counting, still no sign of temp changes

 

[repo man11]

I was about to buy some fancy thermal paste earlier, but then I think I read that it was expected that one would have to change the thermal paste once a year, or more frequently. :O
So I stuck with some basic thermal paste for my cpu.

I vaguely recall trying out some liquid metal years ago, and I had that one for, who knows how long. On removal, I think it looked like it had dried up or something.

There is now, also some kind of pads you can buy, super thin, and can be used multiple times and will last for many years I think I've read. Pads made of some kind of graphene substance. Apparently rips easily when if you touch the fabric.

Phil's Computer Lab has been using those very successfully for some time now.

 

[freakdiablo]

Replace it every year...? *Quickly opens HWMonitor*

I think even cheap thermal paste lasts longer than that, lol. Keep in mind, most PCs out there are basic OEM office machines that may sit in someone's den for close to a decade without as much as cleaning a fan filter.

I've had OC'ed CPUs and GPUs go for 3, 4 years under load on a single paste application and been fine. Longest application may have been ~7 years, a laptop, that was still going fine when I sold it.

 

[ihrsetrdr]

Thermal paste shouldn't need replacement unless the heat sink has been removed / disturbed.

 

[Woomack]

Most modern thermal pastes don't need replacement. Some older needed, like Arctic Silver 5, used to dry out, oil was on the sides of the CPU die/IHS, and the structure of the paste was changing. The higher temps, the faster it needed a replacement. There were other TIMs that eventually disappeared from the market. Modern stuff is designed to last for at least 10 years (so like 3 times the average/expected PC life). On the other hand, laptop users are barely ever changing paste or pads, and they work fine for 5, 10, or more years. Only enthusiasts who love a specific laptop replace and refresh everything every couple of years.

I don't think I have any PC that has been working for more than 8 months without rebuilding so also changing paste. I guess the longest is one of the RX6800XT cards or the ASUS laptop that I gave my mother recently. All should have between 2-3 years. I'm using mainly Noctua NT-H1 or H2, and there are no problems. Sometimes other brands. Recently I used Alphacool Apex just because it was added to CPU blocks (and it is supposed to be the best non-liquid metal TIM).
I have plans to set something to run 24/7 for at least 2 years, but I'm not sure yet if I set up more regular ITX PC or the recently bought NUC.

 

[Mpegger]

The correct answer is, it depends on the thermal paste.

Nactua NT-H2 has about a 2-3 year usage life, which I myself can attest to being correct in that time frame before a repaste is required because of poor temps.

One of the best TIMs available in the entire market that's not a liquid metal, under the Thermal Grizzly brand, has about a 6 month to 1 year life in use before a repaste is required, which plenty of reviewers can attest to as well. Plenty of modern, more recent TIM reviews give the average usage life in their review. The good TIM producers also give shelf life as well as usage life before repaste estimates too, though many seem to be more on the optimistic side of usage life.

 

[Woomack]

The correct answer is, it depends on the thermal paste.

Nactua NH-2 has about a 2-3 year usage life, which I myself can attest to being correct in that time frame before a repaste is required because of poor temps.

Noctua says 3 years of storage, 5 years of usage. It was released about 3 years ago, so it's hard to have more precise test results on multiple home computers. I'm using it daily, but I'm changing PCs way too often to confirm long-term results.

One of the best TIMs available in the entire market that's not a liquid metal, under the Thermal Grizzly brand, has about a 6 month to 1 year life in use before a repaste is required, which plenty of reviewers can attest to as well. Plenty of modern, more recent TIM reviews give the average usage life in their review. The good TIM producers also give shelf life as well as usage life before repaste estimates too, though many seem to be more on the optimistic side of usage life.

Most reviews won't tell you anything about how long the TIM can be used on a CPU as tests last barely a few hours. There are only specifications provided by manufacturer and most of them, don't even provide that. Barely anyone reviews old products and it's hard to say if reality meets specifications regarding new products.

I can only say that I had more than one tube of Thermal Grizzly which was dried-out and supposed to be new. I've never had problems like that with Noctua products.
Arctic MX-4 is specified for at least 7 years and if I'm right, it's the only TIM which is long enough on the market and popular enough to confirm that. Even though some products perform a bit better, then MX-4 works fine for long years.

 

[mackerel]

On Intel delid CPUs I had problems with pumping between core and IHS usually needing paste update in the order of 6 months to a year. I forget what paste(s) I was using around then but maybe could have used thicker ones to prevent that. The problem disappeared when I switched to liquid metal and now I don't delid at all.

Not sure how long other pastes last, but "long enough" would be my answer. If there isn't a problem, I don't renew it. More likely than not, as a tinkerer I'd mess around with the system thus it gets a repaste long before the paste goes off.

At times I've repasted GPUs once they get several years old and not once did that result in any improvement over manufacturer's paste.

Most recent pasts I use are Cryonaut and MX-4.

 

[freeagent]

I mess with my stuff too much, but my sons rig is the long term mule. I have SYY-157 on my GTX 980 Classified, and GTX 580 Matrix Platinum.. both have a couple of years use, and so far so good. Before that I was using TF8 on them.. same deal. Right now my GPUS in our rigs are fairly new, so they are still on stock TIM and pads. I am using SYY-157 on our CPUS. The stuff is fantastic imo. Cheap, and performs really well. With that stuff there is such a thing as too much, unlike what Steve says. Best to go on a little thin.. you will know if its enough of not enough. I also use a spatula now..

People using the same hardware as me, but with AIO get higher temps across the board than I do, usually.. from what I have seen

 

[Mpegger]

Noctua says 3 years of storage, 5 years of usage. It was released about 3 years ago, so it's hard to have more precise test results on multiple home computers. I'm using it daily, but I'm changing PCs way too often to confirm long-term results.

My tube is about 3 years that I have owned in about 1 more month, and it still looks ok (I have no idea about manufacture date). I'll keep my eye on it when it comes time to repaste my CPU, but 5 years of usage? Maybe if you only use your PC a couple of hours a day, but my PC is 24/7 and the temps at almost 2 years were nearing 55-60C idle. By the time I replaced the TIM, it was near 65C idle, and that was near 2 1/2 years. At that point, I was curious about the TIM drying out and temps rising and when I searched around, I did come across some obscure "reviews" (more like blogs) and some Reddit posts on TIM life/replacement/high temps over time/etc, and saw that a number of others also had around 2-3 years of usage life with NT-H2, as well as mentions of other TIMs and the varying usage life as well.

Most reviews won't tell you anything about how long the TIM can be used on a CPU as tests last barely a few hours. There are only specifications provided by manufacturer and most of them, don't even provide that. Barely anyone reviews old products and it's hard to say if reality meets specifications regarding new products.

I can only say that I had more than one tube of Thermal Grizzly which was dried-out and supposed to be new. I've never had problems like that with Noctua products.
Arctic MX-4 is specified for at least 7 years and if I'm right, it's the only TIM which is long enough on the market and popular enough to confirm that. Even though some products perform a bit better, then MX-4 works fine for long years.

My tube of MX-4 is about 10 years old, it's what I used previously to the NT-H2. 7 years sound right as that's when I was making upgrades and noticed that the MX-4 wasn't flowing as well as I remember. It was most like a soft clay and that's what made me look for a new TIM and I decided on the NT-H2, even though the cooler I purchased came with a tube of NT-H1 (I still don't think I've ever opened that tube).

At the time I also had a very old tube of Artic Silver (20 years at least), that still looked good, but there wasnt enough in it to cover anything. I'll have to look around if I still have it, but I think I finally threw it out at that time.

During my search for new and "better" TIM, even 3 years ago there were various posts and comments about longevity in use before repaste is needed, which is why I said "it depends on the thermal paste". I purchsed the NT-H2 cause at the time it was said that repasting shouldn't be neccessary in a normal PC usage lifetime, though no hard numbers as to what that was. But after a few years out in the market, it looks like the 2-3 years is the norm for NT-H2.

You wouldn't be the first with receiving dried up Thermal Grizzly, as thats a common comment I see on various forums and even customer reviews on Amazon. Best rated performance in TIMs, but horrible longevity, with some seeing temps rise in as little as 1 month.

 

[Woomack]

You wouldn't be the first with receiving dried up Thermal Grizzly, as thats a common comment I see on various forums and even customer reviews on Amazon. Best rated performance in TIMs, but horrible longevity, with some seeing temps rise in as little as 1 month.

There are at least 5 other TIMs similar or better than TG. However, I have no idea what would be results after a couple of years. I mean less popular stuff that isn't liquid metal, like Alphacool Apex or Streacom. I used Alphacool on delidded Ryzen 7600, but I bet I remove the cooler in the upcoming weeks.

 

[Sologuy]

There are at least 5 other TIMs similar or better than TG. However, I have no idea what would be results after a couple of years. I mean less popular stuff that isn't liquid metal, like Alphacool Apex or Streacom. I used Alphacool on delidded Ryzen 7600, but I bet I remove the cooler in the upcoming weeks.

Would that not depend on what specific product of TG your comparing too, or is that for the whole range of them?
Personally, I use TG Kryonaut on both my AM5 rigs.

 

[Woomack]

Maybe I should be more specific that I was using mainly Kryonaut. It's simply not equal in every batch, even though it's supposed to be the same product. I doubt it's bad storing temperature as the product has a wide range of temp, and I don't think anyone keeps it in very high or very low temps as a typical warehouse does not have much different conditions than a house or office.
I see that some users say it's great, and some complain. For some reason, I've never had problems with small tubes that are with various coolers of other brands (like recently, I used one from Jonsbo cooler, I didn't know they use TG TIM in their products). Some water block brands use it too. EK used to have it, but now they have their own TIM. I only had problems with larger tubes sold separately, like 50/50 it was fine or too dry. It's just my experience, but I see that the same comments are from other users around the web.

 

[Mpegger]

Update: After having gone full water on both the CPU and GPU with NT-H2 being used on both, I've had to repaste again the other day, mainly because of the GPU. I was gaming the other day and noticed that my GPU temps were suddenly hitting and staying around 100C, with even 110C cropping up. Having already experienced NT-H2 drying out on the CPU previously where idle temps were in the 70's, I figured it was time to repaste the GPU, though it had only been a few months since initially setting up the loop (not too long before my initial post in this thread).

CPU paste was ok, though it was obvious I used too much. It was pretty thick and I even managed to pull the CPU out of the socket when I tried to remove the waterblock from it (pucker moment if anyone knows of how tiny those Zen2 pins are). The GPU was another story entirely...

When I went to pull the block off the GPU side of the card, it surprisingly came apart very easily. I had the block on the table (backside of the GPU card facing up to me), so I didn't see the GPU itself first, but the surface of the water block that would make contact with the GPU die, was strangely devoid of TIM. When I flipped the card over to take a look at the GPU die itself, it too was devoid of paste, and instead covered in what I assume it the suspension liquid used in NT-H2. It looked like a clear, thin oil was on the GPU die, and all the TIM particles was completely pushed out around the sides/edges of the GPU die. In just a couple of months, the NT-H2 mix had completely seperated itself, and nothing but the suspension liquid was left between the GPU and water block contact surface.

In all the years I have been building PCs, overclocking, and watercooling, I have never experienced the TIM seperating while in use in a system. Dry out? Yes. Seperate the particles from the suspension fluid? Only on really old, cheap TIM while it was still in the syringe sitting forgotten in a box, definetely not while in use! I don't know if it's because of the age of the NT-H2 I have (its definetely over 3 years since I purchased it), though it does appear to still look fine, or if the conditions it was under (high temperture fluctuations from the GPU and high pressure betwen the GPU & block), but I'll be looking for a new TIM as having to repaste every few months on a complete loop is not fun.

For now, I've repasted with the the NT-H2 untill I find something better. I'm even considering those carbon nano pads or whatever they're made of, as the difference of a few C between the top/best performers isn't critical in my setup. I'd rather get something I can set and forget, even if it means the components will run 3-5C hotter. I'll still be well below max T with my setup.

 

[Flamethrower1972]

What is the longest anyone has gone before replacing thermal paste? I have a PC I built in 11/2010 that I have not changed the thermal paste, only thing I replaced was the fans on the CPU HS. That's almost 13 years of 24/7 running other than power outages.
View attachment 362675

Post what is the longest you've gone before you replaced your thermal paste

I have gone 5 years without changing paste still have great temps.

 

[Mpegger]

Another Update: So barely a month later and I have to repaste the CPU, as the temps were already hovering around the high 50's, low 60's C range when barely using any CPU, which for my system, 2-3% (1 core) is normal idle/light load. High 80's to low 90's C was becoming the norm when under any load, not even all cores fully loaded, just a normal everyday heavy load of a few cores. GPU temps were still looking normal but had gone up a couple of C from first repaste. In anticipation of having to repaste, I had already purchased some thermal grizzly KyroSheets for both the CPU and GPU, so this time around I ditched the thermal paste, and I installed the KryoSheet on both.

So far it's been a week and temps seem normal as normal can be. The KryoSheets work surprisingly well, with CPU temp lows in the 35-40C range, idling under light load in the 40-45C range, and so far maxing out only a couple of times in the low 80's, staying most of the time under a normal heavy load in the 75'ish C range on the hottest CCD. The less used CCD stays under 50C unless I load it up. GPU temps follow a similar pattern. As hot as the GPU can easily get, I've yet to see it hit more then mid 80's C, while idling in the low to mid 30's.

So far I'm impressed by the KryoSheets, as they are performing the same as a fresh paste of the Nactua NT-H2 I was using. So long as it continues to perform as they say it does (doesn't dry out or break down so temps stay the same for years), it'll be a worth while purchase for me so that I don't have to drain and breakdown my water loop every single time to repaste. Hopefully it really is set it and forget it, but only time will tell.