Anyone ever get a bad batch of TIM?

[DrMrLordX]

Weird thing is last summer when I was handed a tube to try I had great results. 30G tubes?

At this rate give me a 11oz tube that fits in a caulk gun!

30G tube of MX-2:

http://bestbyte.net/merchant/mercha...e=COTIACMX23&Category_Code=COTI&Store_Code=BB

Fearsome to behold.

Also, that Coollaboratory Liquid Metal Pro has apparently been updated and is now Coollaboratory Liquid Metal Ultra. It's more-or-less the same formula, but less runny; you don't have to spread it out on the ihs with a syringe needle any longer. I have not tested it myself, but I hear it is the best out there.

There's also Indigo Xtreme which is . . . very interesting. Expensive, but interesting.

 

[Rubycon]

CLM is good but VERY subject to criticism due to application methodology. It also alters both surfaces after a remount. Kiss your CPU warranty goodbye if you use it. It can deliver real gains (or losses hehe) in core temps, however.

Care must be taken not to disturb the mount, however. Thermalright heatsinks, including the VX should not be used with this TIM unless you can be sure the heatsink will not be twisted at all. This is where the Megahalems mount shines.

Glad to see that I'm not the only one having issues with MX3. The first tube I was provided (the stuff wasn't even in main distribution channels yet!) provided really good results. Perhaps something changed for retail batches. I would not recommend it based on my experience with retail samples. I sent mine back.

 

[DrMrLordX]

Care must be taken not to disturb the mount, however. Thermalright heatsinks, including the VX should not be used with this TIM unless you can be sure the heatsink will not be twisted at all. This is where the Megahalems mount shines.

You know, I should do some research to see if this holds true for the new Ultra version. I have found few actual reviews of the product, such as this one (im Deutsch!):

http://translate.google.de/translat...erste-ergebnisse.htm&sl=de&tl=en&hl=&ie=UTF-8

 

[Gillbot]

Ceramique FTW! My tube is years old and still works great!

 

[Rubycon]

You know, I should do some research to see if this holds true for the new Ultra version. I have found few actual reviews of the product, such as this one (im Deutsch!):

I'm sure the rule holds with any compound based on liquid metal/extremely low melting point solder, etc.
It just won't give the same transfer once disturbed. I was advised about it and believe what I was told.

My tube is years old and still works great!

With or without the blue pill?!

:biggrin:

 

[Gillbot]

With or without the blue pill?!

:biggrin:

I prefer herbal... shhh.....

 

[Tiamat]

Ceramique FTW! My tube is years old and still works great!

Ceramique is all I use for TIM on my CPUs as well. I used to use a credit card to spread it evenly on the CPU surface, then they had to come out with IHS and I got way too lazy. Now I take a cloth and "buff" a thin film of ceramique onto the clean IHS and heatsink surface, then use a ball of the stuff on the center of the IHS...

 

[Gillbot]

i put a grain size drop in the center and let the pressure spread it out. I've tried all the various methods and I can't get more than 1C spread max so the effort isn't worth it.

 

[Zap]

Ceramique FTW! My tube is years old and still works great!

I've had some old Ceramique at work that kind of dried out, so it is tough to squeeze out and is pretty thick. I lent it to someone and he squeezed it so hard trying to get it out faster that the plastic syringe cracked.

I use Arctic Alumina. Basically it seems like cheap white stuff, but I use it because it is indeed cheap, and easy to apply with the huge syringe.

In the past I've gone through the whole expensive TIM phase, using cards and even razor blades to spread the perfect thin layer.

I also used to OC to the last MHz.

These days I just put a line, X or blob big enough so that it spreads nice and even but small enough so that it isn't too much. Then, I OC to a reasonable amount and call it a day.

I guess I have less time these days so that fiddling and tweaking past the point of diminishing returns is wasting time that I could be using killing zombies, keeping the enemy away from the HQ or taking dog tags away from snipers.

 

[Modelworks]

I'm still using the old silicon grease I've used for a decade. It is very hard to apply wrong and it works. Might be a bit higher in temps, but not OC so I don't care.

 

[Gillbot]

In the past I've gone through the whole expensive TIM phase, using cards and even razor blades to spread the perfect thin layer.

I also used to OC to the last MHz.

These days I just put a line, X or blob big enough so that it spreads nice and even but small enough so that it isn't too much. Then, I OC to a reasonable amount and call it a day.

I guess I have less time these days so that fiddling and tweaking past the point of diminishing returns is wasting time that I could be using killing zombies, keeping the enemy away from the HQ or taking dog tags away from snipers.

That's me exactly. And usually, the difference between top OC and my "lazy" oc is only a few hundred MHz anyway. Hardly worth the effort IMHO.

 

[Rubycon]

I do like to see what the absolute limit is (can take a good week or more of hard testing to find it nowadays!) so I know how much safety cushion I'm working with.

Overclocking becomes silly where if your room is 2 degrees hotter the PC crashes. Backing off a bit is certainly not going to make that much of a difference outside of some dumb number Vantage produces.

And besides if you look at it in a real world scenario - a large video project is saving you a lot of time with a 980 o/c to 4.5GHz BUT if it crashes and you have to start all over - even stock is faster.

 

[DrMrLordX]

If you have rock-solid cooling that can take your chip to its limits, you may find yourself at a higher clockspeed when you lower voltages and back clockspeeds off for a sane 24/7 OC. Besides, flogging chips like that for days on end can be fun.

 

[MagickMan]

Ceramique FTW! My tube is years old and still works great!

Yeah, it's awesome stuff for GPUs. On CPUs I prefer SE X23-7783D, it can't be beat (especially not for $1 /syringe).

 

[nipplefish]

I ended up keeping the tube of MX3 for my water build... seems to be working ok, 30C idle and ~65C LinX load at 3.8GHz. Above 4GHz/1.4V temps got up to about 75 max.

 

[rge2]

I searched for bad tim mx3 on google and came up here...between my HK3, swiftech xt, supreme HF, I have mounted my blocks at least 50X and never had a really horrible mount. then had 4 in a row with a new tube of mx3 (~bought 2 months ago). All my blocks/mounts were between 68 and 72C avg core temps...now getting 74 to 77 avg core temps with 2 new tubes of mx3.

Since I had just taken my loop down to change reservoir, I spent 2 hrs thinking it was air, etc, till I wised up and checked water temp...then looked at mx3 mount. First mount had 3 clumps of tim covering about 20% IHS, and watery crap running out sides...never seen anything like it...should have taken pics. mounted it 3 more times, same screwy tim pattern each time along with terrible temps, and different core temp pattern.

Found a very old 95% gone tube of mx3, and had just enough for 1 mount, and temps went back down to normal 68.5C.

Saw this thread about an hour ago, read post about possible separation not a bad thought,...so I took about 90% rest of that tube and squirted it out, and mixed it up, but same high temps and watery/screwy non-coverage of IHS, even trying second mount with 9mm pea size...still end up with almost no tim on block...just watery/clumpy mess.

Well, 5 hours to apply tim, must be a new record in stubbornness or stupidity...I was just starting to like mx3 after consistently getting about .5C better temps then mx2....but done with it now.