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

[dejanh]

Proof? Accoding to Vapor's review here Indigo is 2C better than MX2, and Liquid Pro is defintely more than 2C better than MX2.

First it's slightly more than 3C, second...

http://www.vortez.co.uk/contenttelle..._review,6.html

http://www.overclock3d.net/reviews/c...e_comparison/5

Should I keep going? At the very worst these two will be equivalent but IX will not suffer from any of the problems associated with LMP.

Edit: This is not even taking into account the fact that you will almost never get a paste mount or LMP mount as good as Vapor but you will indeed get an IX mount as good as Vapor which will equate to more like 5C-6C difference at load

[Vapor]

Testbeds can't be compared....I was using a well bowed block (GTZ) that reduces the overall effect of TIM performance (or at least that's part of the intention of the bow). Heck, my IHS is slightly convex as well.

If I went with a flat-IHS and flat-block testbed, my results would be very different, even with the same heatloads

So in all, please don't compare tests amongst each other--if a reviewer has good methods, the data is good.

TIM review 2.0 will be with a CPU-360 and a Phenom II (though I am nervous about the quality of its temp sensors), so hopefully I'll be able to use Liquid Pro--no diffusion into copper when it's already nickel plated--and Liquid Ultra. Finishing up WB roundup #2 (CPU-360 will be done soonish--then I have a small cache of data to post from the review [Supreme Gold, Supreme + Copper top, HK 3.0 Cu, Supreme top + Supreme LT base, and a couple others, IIRC]) and TIM review 2.0 will be underway. After that, it's back to waterblocks for WB roundup 3.0

[Church]

Vapor: will wb roundup 3.0 feature Supreme HF full coper or full nickel too (or it's what you ment with Supreme+Copper top)? After wanting to see how well does 360, that's next thing i'm most curious about

[Nickel020]

First it's slightly more than 3C, second...

http://www.vortez.co.uk/contenttelle..._review,6.html

http://www.overclock3d.net/reviews/c...e_comparison/5

Should I keep going? At the very worst these two will be equivalent but IX will not suffer from any of the problems associated with LMP.

I don't exactly understand what you're trying to say here... You linked two reviews in which Liquid Pro performs 5+C better than AS5/MX2 on a "heavily" overclocked system. How does that prove that IX is better?

And as Vapor said, you can't compare between TIM reviews. There are too many variables affecting the relative difference between the TIMs. The Vortez review actually shows that, the relative differences are much bigger at 4GHz than at stock.

Edit: This is not even taking into account the fact that you will almost never get a paste mount or LMP mount as good as Vapor but you will indeed get an IX mount as good as Vapor which will equate to more like 5C-6C difference at load

Vapor puts a blob on the middle of the CPU and mounts the waterblock, why on earth should anybody be doing this worse than he does it?
I actually even believe that the variances you see between his mounts are due to him using differing amounts of paste since it's very hard to judge how much paste there is in a blob so the amount of paste you use will vary significantly.
The AS5 results on the hand are much closer together and here he used the line method, and it's pretty easy to apply a consisten amount of paste using the line method. Whether line or blob leads to better temps is a different question though, and not really relevant to TIM testing.

[dejanh]

Ok, agreed that comparing TIM reviews is not reliable and posting those therefore did not really illustrate much point. I will be the one to admit my own mistake. What I was trying to illustrate is that the performance is on par at worst which those reviews can somewhat illustrate.

I will now have to go out and get some LMP (well, Ultra) and compare it myself to IX if I want to actually have a reliable test. Considering that they are both based on the idea of pure metal serving as the interface between two surfaces and that IX is not suspended in graphite-copper matrix like LMP but at least based on best available information truly 100% metal interface I would be inclined to say that IX will perform as good as or possibly better than LMP.

As paste goes, most people do not do 5 mounts to average the best mount. They do one mount and however it turns out it turns out. In this case applying IX is more reliable for consistent performance. The same argument can easily be made about LMP as it requires one to "evenly spread" on the IHS.

Considering the extremely low thermal resistance (overall resistance) of IX found here http://indigo-xtreme.com/comparison.html it is extremely unlikely that LMP will have overall resistance lower than this figure. Unfortunately, I cannot back that up as I have no published figures on resistance of LMP.

[colex]

Previously I used the Artic Silver 5... Please I need help to choice the best thermal compound for my i7 920@3.36mhz and GTZ... From this shop: http://www.chilledpc.co.uk/shop/index.php?cPath=66

Thanks a lot!

[Church]

I wonder if it makes sense using heatgun to melt IX? Somehow switching off pump to melt IX from recent case of damaging pump in H50 makes me bit careful about choosing it as TIM in liquid cooling :/

[Nickel020]

Considering that they are both based on the idea of pure metal serving as the interface between two surfaces and that IX is not suspended in graphite-copper matrix like LMP but at least based on best available information truly 100% metal interface I would be inclined to say that IX will perform as good as or possibly better than LMP.

I got no clue what their graphite copper matrix is, but graphite is a great heat conductor, as is copper. Don't judge it based upon pure speculation.

As paste goes, most people do not do 5 mounts to average the best mount. They do one mount and however it turns out it turns out. In this case applying IX is more reliable for consistent performance. The same argument can easily be made about LMP as it requires one to "evenly spread" on the IHS.

Whether you get a good mount or not depends mostly on the block, and not the paste, so you can get a relatively bad mount with IX as well. I guess you could apply too much or too little TIM which can't happen with IX, but if you're considering using IX you should know how to apply TIM so the only variable is the block itself.
If you look at Vapor's data then you'll see that all but one of his AS5 mounts are very close to each other and they vary about the same as the IX mounts do.

Unfortunately, I cannot back that up as I have no published figures on resistance of LMP.

Comparing different manufacturers data is never good, you don't know if they used the same measurement setup, data is probably not comparable.

Previously I used the Artic Silver 5... Please I need help to choice the best thermal compound for my i7 920@3.36mhz and GTZ... From this shop: http://www.chilledpc.co.uk/shop/index.php?cPath=66

Thanks a lot!

They have some Shin Etsu there but I don't know if that stuff is similar to the one Vapor tested or not. A safe choice would be Arctic Cooling MX2 or MX3.

I wonder if it makes sense using heatgun to melt IX? Somehow switching off pump to melt IX from recent case of damaging pump in H50 makes me bit careful about choosing it as TIM in liquid cooling :/

Well using a heatgun on the H50 will damage it as much as turning off the pump. As for normal metal waterblocks, it doesn't matter whether you heat it up with a heat gun or let the CPU do it, the heat will not reach the pump in any case.

[dejanh]

Four tubes of Liquid Metal Ultra incoming. I'm curious to see the results once it arrives. Vapor, let me know if you want one for testing.

@Nickel020 - Thermal resistance of material is not "variable". It is what it is. Unless somebody is falsifying data it can be used as a pretty good indicator of performance. In regards to graphite-copper matrix I pointed it out to show that the claim of 100% metal interface is not quite accurate. There are non-metallic compounds in the interface, probably used to further lower the melting point/increase viscosity of the interface as gallium alone requires roughly 30C temperature to melt and would not be spreadable.

[Nickel020]

@Nickel020 - Thermal resistance of material is not "variable".

Wrong, thermal conductivity and resistance, just like electrical conductivity and resistance greatly depends on temperature.

With all the bull that fan manufacturers are putting out with their cfm and dB claims I just don't trust manufacturer data very much anymore. Neither should you

[dejanh]

Wrong, thermal conductivity and resistance, just like electrical conductivity and resistance greatly depends on temperature.

With all the bull that fan manufacturers are putting out with their cfm and dB claims I just don't trust manufacturer data very much anymore. Neither should you

I think you did not understand what I was saying. The same type of metal will have the same thermal conductivity and resistance properties at the same temperatures, and different types of metals each have different thermal conductivity and resistance properties. However, these are consistent and comparable. It's not like one day copper acts like an insulator and another day it acts like a conductor at the same temperature. Hence, they are perfectly comparable based on their thermal conductivity and resistance. You can easily (well relatively easily) determine which will perform better.

Point taken on trusting manufacturer data however. I am not one to trust it generally. They always skew results to make their products seem more appealing. As you said fans are a good example.

[Nickel020]

. You can easily (well relatively easily) determine which will perform better.

Not if the measurements are taken under different conditions, which was exactly my point.

I finally got some Liquid Ultra now, will slap it on my old Thermalright XP-120 and an Athlon 64 and have a look at how hard it is to get the stuff off again.

PS: The packaging says it's 100 % metal.

[Soulwind]

Well I'm really intrigued by the Indigo Extreme stuff, and I'm thinking about trying to use it with my new machine.

Even though the installation, no pump, overheat chip, re-flow bit still scares the bejezzus out of me.

One thing I haven't seen mentioned though.

That initial (pre-reflow) ring of material looks kind of large from all the screenshots I've seen.

Does it cause any contact issues or problems with getting the block level?

What I mean is, you tighten down the block and it ends up kind of like this --__ between the front and back portions of the block?

Then, after the reflow, do you need to give the back screws an extra twist to bring them back down to tight?

Does that make sense? (and does it actually happen?)

[rge]

I had same concerns, but it does not happen. Just 1) make sure the cpu socket is LEVEL. 2) tighten the block down evenly and completely like you normally would 3) Then when it reflows it will reflow evenly (both of mine did anyways, when I took them apart to change blocks, very thin, even layer). And you wont have to adjust anything after.

[dejanh]

I had same concerns, but it does not happen. Just 1) make sure the cpu socket is LEVEL. 2) tighten the block down evenly and completely like you normally would 3) Then when it reflows it will reflow evenly (both of mine did anyways, when I took them apart to change blocks, very thin, even layer). And you wont have to adjust anything after.

I actually found that to have an even mount with even pressure I had to account for the bump whe mounting the HeatKiller. I just basically had to make sure that all of the fasteners were equally screwed in by measuring differently than in the manual since one side of the block is slightly elevated compared to the other.

[rge]

yeah, I guess it matters how you get an even mount. I use 9 full turns on all screws, so bump or not, mine was even when I finished.

[LiquidReactor]

A heads up regarding Indigo Extreme.
If you lap both CPU and your heatsink / waterblock, make sure not to mount it too tightly before you re-flow.
I lapped both. Couldn't do the re-flow using the first set, so had to remove it and throw it away. Carefully applied the second one, same thing.
Good I was smart enough to loosen the screws of my Megashadow. It finally managed to re-flow properly, thank God.

Even after un tightening the screws I still couldnt get Indigo Xtreme to reflow.

Both chip and Megahalems lapped to 2500 grit, as flat as possible.

oH well, guess I am back to Liquid Pro. BTW has the updated version, "Liquid Pro Ultra" shown up in any USA shops yet?

[LarsMarkelson]

Even after un tightening the screws I still couldnt get Indigo Xtreme to reflow.

Both chip and Megahalems lapped to 2500 grit, as flat as possible.

oH well, guess I am back to Liquid Pro. BTW has the updated version, "Liquid Pro Ultra" shown up in any USA shops yet?

Mega is likely too good of a heatsink and is not getting hot enough to reflow.

Same thing happened to me with my NH-D14. I found a trick on youtube though, in the comments. He said you have to put a heatlamp directed at the NH-D14 to heat it up hot enough so you get reflow. I took a halogen worklight I have and pointed it at my D14 and after a few minutes I got reflow!! Try it out, I think you'll be very very happy with the results afterwards.

Hopefully you still have an application... if not, you can complain to Enerdyne on the Indigo Xtreme site and they'll probably send you a couple free.

[blazarcher]

The way I see it, SHIN-ETSU X23-7783D is the best "paste-based" TIM. IC Diamond is great but takes too long to cure and if you ever want to sell your CPU, you'd look elsewhere. The lil' diamonds have a tendency to polish away at the CPU's surface. So no one can tell if you "really" have a.. 980X for example.

That Indigo Xtreme is great too. Offering better cooling than shin-etsu by about 2C. But to me there is just too many steps to go through. Plus leaving your computer on without a CPU heatsink isn't something I'd do.

Shin-etsu is great on CPU's, but is it any good on GPU's?

[Captain_Harlock]

As a rule of thumb, anything that works well for CPU's will work well for GPU's. The problem used to be this, some of the stock coolers on GPU's need a thicker TIM like a pad becuase they didnt always have the best contact. But thats really all in the past. Today, GPU's run as hot or hotter than many CPU's and therefore require a good cooler with a good mating surface. Due to the large IHS on the higher end GPU's, you may find that a bead or line method for the TIM will work much better than the little grain of rice. I havent tried the current Shin-Etsu's (there are a couple of them), but from what I hear they are somehwhat thicker than many TIM's. You may want to heat it up a bit by letting the syringe soak for a few minutes in a hot cup of water. That should help in laying down a nice bead from one side to the next.

As far as IX goes, will you arent going without a heatsink. Wether you use a heatsink or waterblock, it is attached to the CPU. For a waterblock, it really isnt providing any cooling when your trying to get the reflow to work but the CPU is designed to handle that much heat without toruble. Heatsinks on the other hand seem to be a problem. They are cooling and tend to prevent the CPU from getting hot enough to allow the reflow. Thats why some people are using things like heat lamps, blow dryers, and even heat guns on low power to help get the reflow to start.

I stil have a tube of IC-7 left (couldnt find the IC-24 in stock when I was buying so I got a couple of IC-7's), and I can say that its a fine TIM, but it wasnt worth the asking price at the time. Now its down from the original $12-$15 to a more appropriate $6 - $7 in many places. This makes it a much better TIM for the money IMHO. I never saw that much of a "polishing" effect on anything I put it on, but I've read about quite a few people who seemed to say it was rather extreme. Then again, I dont twist my blocks or heatsinks when I put them on, and I let arctic clean soak in on TIM cleanup so I dont have to rub old TIM off a surface.

@LiquidReactor, they have Liquid Ultra in stock over at performance-pcs and I remember seeing it at one or two other U.S. sites. MemoryC and avadirect are not based in the U.S. but the do sales to all of N. America and they have it in stock. I've used MemoryC a few times now and they are good for getting stuff in the U.S. when your in a pinch and dont want to wait a month or two for U.S. retailers to get stock in. But they can take a few days to get the order into the mail. I really like Liquid Ultra so far, but I've yet to remove it to see if its leeching into the metal. Figured I'd give it a half year before I try thatm mainly as there is no need to remove it from the workstation I'm testing it on. Only been using it for about 3-4 weeks at this point. So still have some time to go.

[Nickel020]

Liquid Ultra does definitely stain the CPU and a nickel plated heatsink a little bit. I tried it on my old A64 3200+ (unlapped) and a nickel-plated Thermalright XP-120 for about 24h with Prime. Using just isopropyl alcohol and a cloth is definitely not enough to get it all the stains off. It only stained the surfaces though, similar to how untreated copper stains when you touch it, there was no bonding noticeable, and no hardening of the paste. I will have to try and make some fotos where you can see the effect and then try Nevr Dull to see if that gets the stains off without getting rid of the model markings on the CPU, but I haven't gotten around to it yet.

[Captain_Harlock]

Well, who cares if it stains. From my limited use it seems to be one of the better TIM's on the market and as long as Ultra is used on it again then it should have no effect in performance. I'd bet that it wouldnt even hurt performance if a normal grease was used afterwards. Or am I missing the point and its more about just the cosmetics of the IHS and cooler that people are worried about?

[LiquidReactor]

Liquid Ultra does definitely stain the CPU and a nickel plated heatsink a little bit. I tried it on my old A64 3200+ (unlapped) and a nickel-plated Thermalright XP-120 for about 24h with Prime. Using just isopropyl alcohol and a cloth is definitely not enough to get it all the stains off. It only stained the surfaces though, similar to how untreated copper stains when you touch it, there was no bonding noticeable, and no hardening of the paste. I will have to try and make some fotos where you can see the effect and then try Nevr Dull to see if that gets the stains off without getting rid of the model markings on the CPU, but I haven't gotten around to it yet.

Well, who cares if it stains. From my limited use it seems to be one of the better TIM's on the market and as long as Ultra is used on it again then it should have no effect in performance. I'd bet that it wouldnt even hurt performance if a normal grease was used afterwards. Or am I missing the point and its more about just the cosmetics of the IHS and cooler that people are worried about?

Where did you guys get the Liquid Ultra? I want to move away from liquid pro because I have to relap the chip and heatsink with high grit everytime just to get it off.

EDIT: Found that Performance PCs already sells some.

http://www.performance-pcs.com/catal...ducts_id=27832

[Captain_Harlock]

Ah... yeah, thats why I said "@LiquidReactor, they have Liquid Ultra in stock over at performance-pcs" in the previous post, lol. But as I said in the next post, who cares if it leeches if it doesnt have an effect on performance? If your using liquid pro, removing the cooler for whatever reason, and putting back on with liquid pro, then why bother relapping? The leeching would seem to only fill in the microscopic areas on the metal and have no effect if using liquid pro again on the leeched surface. It may have an adverse effect with a grease based TIM or another metal TIM like IX or even the coollaboratory metal pads (as they have a different chemical makeup). I also wouldnt mix liquid pro with liquid ultra. But if you still ahve plenty of pro and your surfaces are lapped, is there a erason that you lapp them again to remove the leeching?

[LiquidReactor]

Ah... yeah, thats why I said "@LiquidReactor, they have Liquid Ultra in stock over at performance-pcs" in the previous post, lol. But as I said in the next post, who cares if it leeches if it doesnt have an effect on performance? If your using liquid pro, removing the cooler for whatever reason, and putting back on with liquid pro, then why bother relapping? The leeching would seem to only fill in the microscopic areas on the metal and have no effect if using liquid pro again on the leeched surface. It may have an adverse effect with a grease based TIM or another metal TIM like IX or even the coollaboratory metal pads (as they have a different chemical makeup). I also wouldnt mix liquid pro with liquid ultra. But if you still ahve plenty of pro and your surfaces are lapped, is there a erason that you lapp them again to remove the leeching?

Argh...was a bit tired when reading your post.

Reason I relapped both cpu and hs before reapplication was that I think that the left over liquid pro that seeped into the "micro cracks" might be interfering with performance of fresh application.

I don't know if thats correct or not but I didnt want to take a chance. Also my last lap I only went up to 2000grit and now I got 2500 grit

Anyways, I ordered Liquid Ultra. Will be here by Friday hopefully so I can experiment over the weekend over the amount necessary for optimum temps.

Since both my chip and hs are lapped I probably need a tiny amount.