Arctic MX-4 Lacks Longevity

[DefStar]

So I bought a couple ounces of Arctic's MX-4 since there are so many reviews that rave about how great it is...and decided not to buy Silver since it's, and I quote, "old". So here is my experience from the past year with MX-4. I bought a Powercolor 7950 and mounted an Arctic Accelero Plus II (By the way Arctic, nice 6-year warranty? WTF) up to it and used MX-4 as my thermal compound:

12 months ago: idle temp-34C, load temp-64C
8 months ago: idle temp-34C, load temp-90C

Removed Accelero and re-applied MX-4:

8 months ago: idle temp-34C, load temp-64C
4 months ago: idle temp-34C, load temp-90C

Removed Accelero and re-applied MX-4:
4 months ago: idle temp-34C, load temp-64C
Yesterday: idle temp-34C, load temp-100C

Removed Accelero and re-applied MX-4:
Today: idle temp-34C, load temp-64C

Notice a trend here???? It's not dried or brittle either, its still just as creamy as when I put it on...so its deteriorating/breaking down on me. I call shenanigans on Arctic's MX-4 claim that it is long lasting.
Or hey, I got the only tube of MX-4 in the entire world that does this.

Has anyone else noticed their temps are not as low as they use to be? Have you been just assuming its dust bunny build up when its really not?

[AMDforME]

It's difficult to say if the MX-4 or any other TIM deteriorates over time without a much larger sample size and controlled testing conditions.

While it's possible that the MX-4 is thermally changing over time it's also possible that other issues are causing the temp change that you are seeing such as the two surfaces not being extremely flat or the mount tension changing over time. These could allow the TIM to extrude some or lose good contact between any warped surfaces. There may be oxidation issues based on the thermal operating range, environmental issues, etc. that lead to the results that you are seeing. There is also the possibility that your results are very typical and a true indication of an issue long term with MX-4. We just don't know without a more comprehensive investigation. Maybe you should try AS5 or some other TIM on the same component and see what the results are 4 months down the road as a reference point.

[DefStar]

Yup, i'm calling out MX-4. I realize a lot of people wouldn't realize an issue that occurs over such a long period of time.

I'm asking if any other users of MX-4 that have supplemented their original GPU's compound with MX-4, if you have noticed a sharp rise in GPU temps after 3+ months of usage?

I was quite impressed with my Acellero keeping my 7950 at 64C load, so you can imagine it is quite noticeable when the compound has deteriorated and load temp goes up to +100C.

++The surface is smooth as I have lapped it twice now.
++The contact between the GPU and sink really cannot be any tighter, as I have already stripped (and replaced) the steel threads out of the base-plate of the sink once.
++I've tried getting away with just tightening the contact between the two when temps go up...but to no avail.

And the stripping of the steel threads is the problem here...I've had to reapply the compound so many times that the steel threads on my Arctic Acellero finally gave out. Arctic won't honor their 6-year warranty with me directly...so i'm currently having Newegg deal with Arctic for me (they say Arctic has to honor that warranty). How ironic, Arctic's MX-4 compound caused premature failure of my Arctic Acellero GPU-sink.

I am guessing MX-4 was only tested with CPUs by Arctic.

[Daveburt714]

I am guessing MX-4 was only tested with CPUs by Arctic.

FWIW, I've been using MX-4 for a few years now and have been quite happy with it compared to others.
There are more exotic (read expensive) TIM's that probably perform better, but I've done some testing on my own and decided to stick with MX-4....

I installed an aftermarket Zalman cooler on my 6870 back in 2010 (using MX-4) so it gave me a point of reference to look back on:
http://www.xtremesystems.org/forums/...ler&highlight=

I ran those same tests again tonight and temps did increase from 51c to 58c load (not sure what ambients were back then but it's 72f now so they should be close).

100c temps would concern me too, so this info probably doesn't help you, but it is another point of reference.

I'm sure all tims degrade a little over time, but 29 months is a pretty long stretch and a 7c increase doesn't seem all that drastic in my case.

[DefStar]

Yeah, that's a small increase but it doesn't depict the same scenario i'm experiencing. I wonder if my cooler is so big that maybe it's twisting and moving the compound around when I pull the card out of the case to swap something. That's something I've noticed to, I just happened to be doing an upgrade/swap of some component every couple months lately...requiring me to pull my vid-card out. Maybe disrupting the compounds initial position once it has been set and heated causes it to lose its thermal conductivity...or creates micro-air pockets.

[OldChap]

I'm not going to argue with your logic.... but instead of replacing the tim see what happens if you re-torque the mountings.

I have used this for short duration only and on spring mounted cpu so this is just an idea that comes to mind.

[DefStar]

Yeah, I retorqued the screws but it gave no improvement. I guess that rules out my micro-air bubble theory. I think MX-4 breaks down under the higher heat expelled from a modern GPU.

[AMDforME]

I have no personal experience with the GPU/cooler that you are using nor why the threads would continue to strip. Excessive tightening however can also cause the heatsink to warp over time with heat cycling.

It's possible MX-4 deteriorates over time. We just don't know for sure. The only means to know is by accurate, controlled testing. It would be nice to test your current set-up with another TIM to see how much the temps change in 4-8 months. That might be a useful data point.

[HKPolice]

Probably caused by the dust build up on the cooler's fins.

[Budwise]

Were you by chance removing dust every time you re-applied? Dust accumulation could definitely explain the rise in temps and the drop in temps every time you re-applied.

[apexracing]

Shine a light on the heatsink, you will probably see dust caked on the fins. Even if you clean between the fins, dust on the fins is still holding heat in. Remember, dust is a fantastic insulator.

[AMDforME]

If he isn't cleaning the heatsink every time then that isn't the issue.

[DefStar]

Got some "Prolimatech PRO-PK3" compound from newegg. It's pretty expensive but it's the only compound that advertises "Long-Term Stability"... . So I applied it and my load temp actually dropped 2C (from 64 down to 62), which is impressive. I've applied MX-4 to this thing 5 different times and load temp is always 64C in Heaven by the end of the first loop. So I'm now 6 loops into Heaven and the Prolimatech is keeping it at a steady 62C. Anyway, now for the test of longevity...I guess I'll update this thread in 3 months or so. Wow, 1.5 Grams cost $6.99 so it's pretty expensive.

[AMDforME]

The Prolimatech PRO-PK3 is expensive and more difficult to apply properly. It tend s to deliver a couple degrees reduction but that is insignificant in actual performance. It's lifespan is undocumented as most folks who use it tend to mod/mess with their systems a lot so it never is in operation for long periods of time before being replaced. Report back in 3+ months.

[EniGmA1987]

I put some MX-4 on when I installed my Vishera CPU. I dont remember what the max load temp was the day I applied it so I have no comparison here to give, but ill check this afternoon and see what the load temp is now days and let you know if it looks really bad or not.

[Kain665]

I've been running MX-4 on my 870 for over 2 years now without change. It's been running fine, load temps now are marginally higher than before simply due to dust buildup over the past couple of months.

[DefStar]

Just to be clear, for the people posting that their "CPU load temps" have not increased...I also use the MX-4 on my CPU and have also not had an issue with the load temps going up over time...only on my GPU have I noticed an issue, which I believe to be a result of the much higher temps experienced when using my overclocked GPU...just to be clear.

[Ket]

Been using MX4 on my HD7950 (OC 1100 / 1565... crappy overdrive, but I don't like needless apps launching with Windows.. nor do I like having to launch a app for OC to be applied) for 4-5 months now, temps are every bit as good as when it was first applied.

[apexracing]

All compound dries out over time, it's a case of watts over time. When you're pumping 200+ watts into the compound, it is exponentially worse than ~100w for the average CPU.

[AMDforME]

All compound dries out over time, it's a case of watts over time. When you're pumping 200+ watts into the compound, it is exponentially worse than ~100w for the average CPU.

Not necessarily... There are various formulations of TIM compounds and they do not all share the same physical characteristics including the perceived "drying out" characteristic which in some compounds may not be significantly altered by the thermal load or operating temp.

Without proper scientific testing, it's all speculation and not documented fact.

[DefStar]

Been using MX4 on my HD7950 (OC 1100 / 1565... crappy overdrive, but I don't like needless apps launching with Windows.. nor do I like having to launch a app for OC to be applied) for 4-5 months now, temps are every bit as good as when it was first applied.

What do overclocking applications launching with windows have to do with this thread? Since you seem to be awfully distraught over the situation, I will go ahead and remind you that you can set the app. not to launch when windows starts. Actually, I'm bored with compounds... lets start talking about overclocking needless crappy overdrive launching windows applications

[st0ned]

When doing direct Die cooling I noticed that MX-4 was good and then became really bad. Other more common TIMs such as MX-2 were average and also deteriorate fast although not so fast as MX-4. The "normal" TIM I had which lasted the longer was Prolimatech PK-3 although it only lasted for a couple of weeks ! All my problems were solve when using a Liquid Metal TIM.

My opinion on this matter is that conventional TIMs tend to deteriorate when you make with the watts they transport, in this case doing direct Die cooling it was clear that the small areas where the cores were located within the CPU were dryer then the other parts.

[AMDforME]

For those who don't know... TIM is only meant to fill the micro imperfections in the two surfaces. Ideally you want to apply a film of only a few thousands of an inch in thickness. In other words if you can't see thru the TIM on the CPU/GPU surface, you've used way too much. The goal is 100% metal-to-metal contact of the two surfaces. Since that is impossible you want as high of metal-to-metal contact and the TIM to only be in the micro imperfections where there isn't metal-to-metal contact.

As far as "drying out" some TIMs are meant to dry because they conduct heat better as a solid vs. a paste. That being said I have never personally seen any TIM "dry out" in a few weeks or even a few months. To determine the true thermal properties long term you'd need to conduct accurate controlled testing.