[Review] Indigo Xtreme vs. AS5, MX-2, IC Diamond, Shin-Etsu X23-7783D

[Vampiyer]

I would like to draw attention to TIM Consultants and their new Matrix 2 TIM:
http://www.tim-consultants.com/products.html

As the successor to TC grease 0098 (which is competitive with MX-2), I'm wondering how much better (if at all) it could be.

If MX-3 is MORE of a PITA than IC7, then no way. I like my MX-2 for now...

Vapor: thanks for the tests. You too "rge."

[~CS~]

That indigo is something very cool concept,cant believe someone put so much thought on a tim , but it certainly paid off . Thanks for the review , really well done. Am sticking with my MX-2 for now and seems like most people are .

[Vapor]

Okay...few updates! One, MX-3 tests are done but they'll be qualified with some very important words (you'll see when I post up the data).

And I also tried something that I think a few people were wondering about, but didn't want to ask: is Indigo Xtreme reusable?

Short answer: yes.

Long answer: yes, I reattached and redid the reflow process on a used application of Indigo Xtreme. Everything behaved normally but the end temperatures were 1-2C worse than they should be. When I removed it, the metal panel looked kind of like swiss cheese (the 'holes' were actually pieces of the PCMA that were so thin they had more adhesive strength than structural strength and were lightly attached to the block/IHS....they came off with a wipe of a paper towel )

Anyway, Ceramique mount 1 is underway and I'll be posting MX-3 data shortly

[rge]

That Matrix 2 tim does look interesting.

Actually, since I was btchn about the mx-3, I should clarify it is not that bad if your used to bead in the middle mounts instead of spreading like me and you use enough (more than I was initially using) such that it will cover IHS most mounts. I had always spread mx-2 and that is not happening with mx-3. But now after 20 mounts, I am getting used to bead in middle, though I still measure with ruler. I tried 8mm bead, just one mount and still only got .6C better than mx2, but that amount easily covered IHS, have not tried 1cm yet like Vapor posted he was using.

Other reason I was btchn, I was expecting to see 2-3C better than mx2 like some other published review, but I am not seeing that. Also will be interesting to see if mx3 cures better in vapors tests...curing just exceeds my testing patience.

[Vapor]

Alright, here are the MX-3 results

I'm going to say this twice in this post and I can't emphasize it enough: due to time constraints in my schedule, I cannot do more mounts than this. Yes, I realize each mount is better than the last and that is because I simply learned how to use the TIM better as time went on. MX-3 is unlike any other paste I've used before--it performs better with more paste being present. I'm sure there's an upper limit, but I didn't reach it. First mount I used a ~6mm bead, second mount I used ~7.5mm, third mount was ~8mm, fourth mount was ~9mm and fifth mount was ~1cm in diameter. Yes, 1cm in diameter....

So with that, I would expect performance to actually end up better than my numbers portray as long as it is used properly. I estimate expected performance at somewhere between what it currently shows and X23-7783D's performance....maybe .2-.5C worse than X23-7783D on my testbed...I've added a fabricated curve to signify roughly where I think it will perform, based on what I saw with the last two mounts

The formula I used to create the expected curve is the average of mount 4 twice and mount 5 once.

This could have all been avoided if Arctic Cooling just provided instructions that say "use a lot" but whatever

Without further adieu, results....
New 'big' chart:



The individual mount chart:



Again, due to time constraints in my schedule, I cannot do more mounts than this. Yes, I realize each mount is better than the last and that is because I simply learned how to use the TIM better as time went on. MX-3 is unlike any other paste I've used before--it performs better with more paste being present. I'm sure there's an upper limit, but I didn't reach it. First mount I used a ~6mm bead, second mount I used ~7.5mm, third mount was ~8mm, fourth mount was ~9mm and fifth mount was ~1cm in diameter. Yes, 1cm in diameter....

So with that, I would expect performance to actually end up better than my numbers portray as long as it is used properly. I estimate expected performance at somewhere between what it currently shows and X23-7783D's performance....maybe .2-.5C worse than X23-7783D on my testbed...I've added a fabricated curve to signify roughly where I think it will perform, based on what I saw with the last two mounts

The formula I used to create the expected curve is the average of mount 4 twice and mount 5 once.

This could have all been avoided if Arctic Cooling just provided instructions that say "use a lot" but whatever

(yes, I told you I would say the same thing twice in one post)

[skinnee]

how many mounts can you get out of a 4 or 5g tube considering a 1cm bead per mount?

[Vapor]

Hmmm, not sure. I've only done <5 mounts (5 mounts, but 3 of them weren't nearly full-sized) and the plunger is already getting pretty low.

I'm pretty sure less than 10 from the 4g tube...

[Nickel020]

Thanks a lot for adding the MX3 results so quickly!

I've got one more question though:
Seeing as the AS5 mounts are more consistent wouldn't you conclude that a line method (or double line) is superior to a bead method? From my experience it's much easier to get a line to be consistent than a bead and your results seem to match that theory.

[zalbard]

Is it just me who is disappointed by MX3 results? I thought it would be more than 1C better than MX2...
Thanks for your update! Nice job!

[Darth_Penguin]

Is it just me who is disappointed by MX3 results? I thought it would be more than 1C better than MX2...
Thanks for your update! Nice job!

Actually I cant believe that MX3 performed that well. But, Im still in shock about the Indigo stuff. People switch water blocks and dont get the temp decrease that indigo offers over its competitors. Really of all the components of a computer this is the area that I thought was "as good as it gets". This is probably why I have still been using AS5.

[Vapor]

Thanks a lot for adding the MX3 results so quickly!

I've got one more question though:
Seeing as the AS5 mounts are more consistent wouldn't you conclude that a line method (or double line) is superior to a bead method? From my experience it's much easier to get a line to be consistent than a bead and your results seem to match that theory.

Maybe?

I don't think there's nearly enough data here to begin to make that judgement.

[slim142]

From now on, I will apply paste in "rice" form.
And will definitely get rid of any AS5 that I have.

[HESmelaugh]

Wow, the MX-3 thing is pretty weird, eh?
Well done for even catching that. I think I would have probably just stuck to my usual amout of TIM and never seen the correlation between larger amounts and better results...

[Martinm210]

FYI, I notice the TIM consultants TC Grease was also thicker and worked better with more of a thin layer method over thin line. There's definately something to the thickness of the material and the method that's best.

[tiro_uspsss]

Here's what it looks like when you remove it:

so thats the Indigo Xtreme? never heard of it.... admittedly I only skim-read this thread - could u put in bold link to store & price etc please?

[Darth_Penguin]

so thats the Indigo Xtreme? never heard of it.... admittedly I only skim-read this thread - could u put in bold link to store & price etc please?

a quick google search later
http://indigo-xtreme.com/

[tiro_uspsss]

a quick google search later
http://indigo-xtreme.com/

sorry just got out of bed! thats my excuse, & I'm sticking to it!

[tiro_uspsss]

IX comes with gloves?!? this stuff is potent!

[Vapor]

IX comes with gloves?!? this stuff is potent!

It's more so that the end user doesn't get their grubby paws directly on it

The three Ceramique mounts are almost done (2 done, 1 will be done when I wake up tomorrow)....then the testbed gets dismantled and upgraded slightly and then I go visit family for a week...and when I return, it's back to CPU blocks!

(On a side note, I really can't wait to go back to CPU blocks...I think some of the data this new testbed will be able to provide will be really fun. It also allows me to test blocks 2.5 times as fast so hopefully I can test more blocks on more CPUs )

[woffen]

It's more so that the end user doesn't get their grubby paws directly on it

The three Ceramique mounts are almost done (2 done, 1 will be done when I wake up tomorrow)....then the testbed gets dismantled and upgraded slightly and then I go visit family for a week...and when I return, it's back to CPU blocks!

(On a side note, I really can't wait to go back to CPU blocks...I think some of the data this new testbed will be able to provide will be really fun. It also allows me to test blocks 2.5 times as fast so hopefully I can test more blocks on more CPUs )

Sounds really promising, waiting eagerly for more of your well done tests

[Vapor]

Alright, Ceramique data

Well, I'm going to write-up a short supplemental review for the http://vapor.skinneelabs.com posting (sometime tomorrow, maybe on the plane), but here's the data from it:

Main Chart:



Ceramique Data:



MX-3 Data:




Too tired for many words right now...I've already gone over the flaws with the MX-3 data and explained why I expect the end performance to be better than presented so long as the end user uses a ton of it.

Ceramique...not really much to say I guess. I wish I could do 5 mounts, but time constraints are preventing that...regardless, I do think it's basically done breaking in after this kind of load and that it's bringing up the rear in performance, even if only by a little. Just a small half-grain of rice was used in the middle for each mount and the "post-squeeze flow" (those who have used the 22g tubes know what I'm talking about) was minimized by pulling the plunger back pretty soon after squeezing it down. At $7 for 100+ mounts it's obviously a great deal...though I don't think it's adequate as a testing TIM due to the steep initial break-in and progressive curing that goes on for the next 11+ hours. In almost all cases, I'd use it over AS5...no capacitance, cheaper, similar performance, slightly easier to clean up (especially from clothing), and a little easier to apply.

[Machinus]

The ETI installation procedure says to "burn-in" the layer by letting the cores reach 100 degrees...is this safe???

[dreamaxx]

Wow won't catch me putting the cores at 100 deg C (at least not on purpose!).

[Boogerlad]

use a torch? rofl.

[Nickel020]

Great stuff, this came right in time for me and convinced me to use MX2 for my air cooling heatsink tests (just testing them, the PC I'm writing on is water cooled of course).

But now what am I gonna do with my 22g syringe of Ceramique and the 12g of AS5