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

[woffen]

Can anyone tell me how OCZ Freeze fits in here, is it thick as MX-3 or as easy to apply as MX-2?

[dejanh]

Went ahead to place an order for a couple of Indigo Xtremes to do some testing myself and after the order was placed received an email saying that the order cannot be processed because they do not allow the shipping that I have selected and that there is a mistake in their store. O-k...

They refund me the money and simply say "re-order if you still want it with a different shipping option". Not impressed so far.

I emailed them now to ask for an affordable shipping option so I will see what they say. I cannot imagine that they would think people will buy their stuff if the cheapest shipping option you can get is more than the price of the product. I indicated to them that USPS parcel service should be an option since their products are cheap and light. Under 4lbs USPS is just fine for shipping.

Let's see what they say... or ...time to judge...I think this more than anything will determine the success of their product. Not going to be many resellers who carry this in stock @ $20 US a pop.

[century child]

There was a post in this thread a bit further back from someone who emailed them about this exact same thing. The poster received an answer saying this was a mistake and that they were trying to get other shipping options added to their site.

[dejanh]

There was a post in this thread a bit further back from someone who emailed them about this exact same thing. The poster received an answer saying this was a mistake and that they were trying to get other shipping options added to their site.

Well, after I indicated the USPS option to them they have quickly added it and now you ship 1 piece for $8.73 or 2 pieces for $11.93. Order placed. So far so good

Let's see what arrives in a couple of weeks.

[Snow Crash]

Based on my recent use of OCZFreeze I can say it is pretty easy to work with and it squished nicely under my Xiggy DK 1283 (fill in the gaps between pipes and base, razor excess off, then two thin lines, about 1/3 the length of and directly on the two center spans of the base) i7 at 3.6 GHz (1.224 v) WCG load 24/7 @ high 50s C. I have done three mounts and all about the same results (+- 1C).{edit} Yes, this is a C0 and I have HT on. {/edit}

[Vapor]

Quick and dirty supplemental write-up is now live at skinneelabs: http://vapor.skinneelabs.com/TIM/IXS...nt/IXSupp.html

Sorry for the recent AFK status...still visiting family in CT

EDIT: doesn't say anything too new....just had more time to reflect on the respective performances and I think both are winners. Ceramique is almost universally better than its sibling, except in actual thermal performance (and only then by a tiny amount).....while MX-3 really looks like it could be a great paste-based TIM once we all figure out the best way to use it

[astrodanco]

Despite the 1C resolution of the Intel DTS sensors, these tests can be considered statistically highly accurate due to the immense amount of data acquired from polling every sensor/probe/meter every second over the course of 12 hours.

So inaccurate data equals accurate data, so long as you have lots of it?

The above quote could use a little elaboration or rewording.
.

[Vapor]

Statistically speaking, you can increase the resolution of the data by increasing your population size. As for the accuracy of the Intel DTS sensors, it is undocumented, but can be assumed to be no better than +-.5C (resolution/2, although it's probably worse than that). [edit]Then the fact that you have four of them actually does further increase resolution/precision....since they are the same DTS sensors being used each time, bias is negated (i.e., accuracy of temp reports vs. actual temperatures, a.k.a., bias)[/edit]

I suppose I should change that sentence to revolve around the word "precision" rather than the word "accuracy."

Thanks for pointing that out I'm not actually out for accurate temperatures (that sounds so much worse than it is! ) so long as the precision, repeatability, and reproducibility remain very high

[aka_GK]

great review, thanks

[rge]

So inaccurate data equals accurate data, so long as you have lots of it?

The above quote could use a little elaboration or rewording.
.

I had the same issue with accuracy, used to think .5C difference was just noise in testing, in fact had pmed Unclewebb earlier about same issue. So I tested same mount 5 times on 1 day old mx2, so I could see the range of temps (btw, looking at Vapor's data over time, you can see same thing). Loaded prime large ffts (custom so stays on same fft) let water reach steady state temp, then monitored. Let it cool off, then repeat 4 more times. Got 39.67, 39.54, 39.65, 39.61, 39.73 delta water to cpu. I was amazed got that close, I was expecting something much different. A half a degree C would be detectable on different mounts. Still there is the variability of bad/good mount and need to do many mounts, but intel sensors are relatively accurate since 4 sensors are rounding at different times, and taking thousands of data points. I tested a few other mounts multiple times, again all very close results, within .2-.3C.

Many of national semiconductor DTS (which intel does use on some of theirs), have an accuracy of +/- 0.3C, some even better. Intel will still report accuracy of +/- 1C from DTS rounding to nearest 1 degree, and from an absolute accuracy after calibration, however once you filter out the rounding by taking thousands of data points at different times, the relative accuracy will be better.

[tastymannatees]

Another view on testing

Why does every test come to a different conclusion? The problem is that their sample size = 1.


Even collecting multiple readings the cluster size is = 1. An individual can collect all the data readings off one system, and will still have almost no statistical power (In statistics this is known as "Intra-Sample Cluster Correlation")because the test set up is dominated by methodology. This is a problem not only in paste reviews, but in other hardware reviews, heat sinks, etc. as well.

In the final analysis methodology ends up defining the final placement of all compounds- All test methodologies fail to take into account things that have a major impact on paste performance. For example the mounting system along with mechanical contact between IHS and sink as evidenced with our independent contact/pressure testing. Variability was very high on the contact results with perhaps 1 in 10 having any thing near what you might call full contact, even on those with lapped components.

In considering pressure related to mounting hardware some pastes perform relatively better under poor mounting, others perform relatively better under good mounting (viscous ones such as ICD). Considering ICD - people that had poor results with ICD had very poor mounting. Once they improved that mounting, ICD did considerably better. Of course, so did their old paste. But ICD improvement >> old paste improvement. Generally, this resulted in ICD>old.

In summary, sample size = 1 tells little. . Sample size matters!

80% see a positive result out of the gate ICD. 16 % that did not were determined based on our Pressure/contact test sampling were determined to have insufficient contact or Pressure. So from our data we can account for 96% of users results. Which is pretty good by most accounts


A single test is anecdotal, even my own with $100,000 in thermal test equipment at my disposal is still a sample of one.

For example IC Diamond comparisons vs AS5 were tested by 146 users in 11 forums. I can not say that ICD is better than AS5. I can say that out of 146 samples 18 resulted in higher temps, 8 reported no change in temperature and 120 users saw an improvement with an overall average improvement of 3.26C which is statistically significant





NOTE ON COMPARISONS

When comparing results more credence should be given to larger sample sizes, smaller less so.

1-5 samples 1 or two tests can flip results either way so usually get thrown onto the miscellaneous group.

6-10 samples may start to indicate trends but can be heavily influenced by outliers and so are lightly weighted other than a general trend indicator

11-20 samples - Starting to develop more of a confidence in the trend direction.

20-30 samples - Confidence level improves.

30+ Samples - High degree of confidence













Sample size on the ShinEtsu is about half of the 30+ samples I like to see where the averaged number only fluctuates a couple of a 10th's of a degree from there on out.

In any event if I take the latest Giglamesh review and the result posted here gives me a total of 14 results. 9 show improvement, one sample no change or "0.0" and 4 that have higher temps. Generally following our model of 80% improvement vs 20% that saw no improvement -sample size matters.







80 % of the results from independent users tests see a result out of the gate

From independent lab analysis of the remaining 20% of the total users, 80%(or 16% of all users) of those had inadequate contact and/or pressure.

This issue of contact and pressure is mostly a confirmation of what most experienced users already know and is the most recommended forum solution when trouble shooting heat sink thermal problems - If the temps are too high you tighten down the sink or try lapping the surfaces for better contact.

[rge]

I partially agree with you in using one waterblock and one IHS may not represent ideal contact, and may not be representative of all mounts, and there is certainly some error in that and will be variation in pressure.

But you grossly understate the error of using forum feedback as a legitimate means of testing tim. Are you suggesting 10C 14C 20C differences actually occur with different tims? I can compare IC7 to toothpaste and not see that much of an effect.

Many are eyeballing core temps after a fresh install where internal air has cooled which will cause temps to be a few C cooler, and some either had mounted with no tim before or are exaggerating or are people pleasing or just plain bsing, and some may have accurate testing equipment, but without sorting out what, I cant make any conclusions for it, other than I would rule out using forum feedback as reliable tim testing.

Now if you want to personally test 5-10 different cpus/blocks/setups, and show difference of one versus many using proper measurement equipment, I would like to see that.

[tastymannatees]

I partially agree with you in using one waterblock and one IHS may not represent ideal contact, and may not be representative of all mounts, and there is certainly some error in that and will be variation in pressure.

But you grossly understate the error of using forum feedback as a legitimate means of testing tim. Are you suggesting 10C 14C 20C differences actually occur with different tims? I can compare IC7 to toothpaste and not see that much of an effect.

Many are eyeballing core temps after a fresh install where internal air has cooled which will cause temps to be a few C cooler, and some either had mounted with no tim before or are exaggerating or are people pleasing or just plain bsing, and some may have accurate testing equipment, but without sorting out what, I cant make any conclusions for it, other than I would rule out using forum feedback as reliable tim testing.

Now if you want to personally test 5-10 different cpus/blocks/setups, and show difference of one versus many using proper measurement equipment, I would like to see that.

BAre you suggesting 10C 14C 20C differences actually occur with different tims? I can compare IC7 to toothpaste and not see that much of an effect.

Not at all if you look at the averages between tims -ocz -2.66, Shin Etsu -1, MX2 -3.3, ceramique -2.7, tx2 -2, AS5 -3.26 etc. Max of 5-6 C for the generics and generally a 3 C range for the competitive compounds



Most of the performance community is brighter than your average lot and experienced at mounting sinks, What makes you think they are doing anything different than anybody else?

you do have outliers but they usually cancel each other out and after 400 tests a -10 or +10C data point is only 1/400th of the calculation and affect the result maybe 1/100 of a degree.

Infinite testing on one machine is still only one data point, so goes the tooth paste comparison. I have had the paste tested at some of the best thermal labs in the country the result was a much narrower span of diff. in temps but generally they fell average wise within a few 10th's of a degree of the forum results and my own.

As a manufacturer my concern is what does the average user see at the desktop?

Benchmarking is a relative process, but relative to what? Most users do not go to the bank on a single test or review, instinctively they read multiple reviews, other peoples experiences in the forums and mentally average results to make a buy decision.

I could take my super duper lab results excruciatingly detailed to the nth degree over 3 years post them on the forums and it would mean poop. The response from most be "cute, what does it do in the real world?" Hence the forum testing, Real world tests

From the forum tests most see an improvement whether it be 1C 2C 3C 4C 5C etc. Variation in results are not really a mystery but are understood and can be narrowed with proper methodology (pressure/contact)

[genec57]

Hi Guys,
I just about two hours ago installed Indigo Extreme. First off, as Skinnee said, you absolutely MUST pay slavish attention to the installation instructions. On my first attempt I got the ETI a little off center and thought I could lift it off and reseat it - WRONG. Reflow did not go right and the resultant temps were in the 80's.

Second attempt and I really took my time to get it right.

This stuff kicks Butt. My heat killer was installed previously with MX2. From MX2 to Indigo Extreme I realized a good 4c drop.

Indigo Extreme takes TIM to a whole new level. Expensive (especially when you have to do it twice ) and a pain to install but it is majorly worth it. Any overclocker who is looking for that last few degrees needs to put Indigo Extreme at the top of their shopping list.

[masska]

Hi Guys,
I just about two hours ago installed Indigo Extreme. First off, as Skinnee said, you absolutely MUST pay slavish attention to the installation instructions. On my first attempt I got the ETI a little off center and thought I could lift it off and reseat it - WRONG. Reflow did not go right and the resultant temps were in the 80's.

Second attempt and I really took my time to get it right.

This stuff kicks Butt. My heat killer was installed previously with MX2. From MX2 to Indigo Extreme I realized a good 4c drop.

Indigo Extreme takes TIM to a whole new level. Expensive (especially when you have to do it twice ) and a pain to install but it is majorly worth it. Any overclocker who is looking for that last few degrees needs to put Indigo Extreme at the top of their shopping list.

you're right, you do have to install properly an follow procedure, however the extra effort is worth it! Happy to hear its all good now.
How long did install take you, and what are your idle/stress temps now?

[genec57]

Doing it twice took me a bit over two hours.

Temps across the board are down about 3-4c. It's hard to pin them down very precisely. Nothing over the mid sixties where before I was pushing 70. I have had an ambient variation of about 2c throughout the day and load temps tend to jump all over the place.

As I said, expensive but well worth it. I am looking forward to hearing from more courageous users .

[p3st]

sweet review

Just emailed their support and kept getting bounced back mails.
Do they only ship within the US?
I live in Asia and would love to test this out.

[masska]

sweet review

Just emailed their support and kept getting bounced back mails.
Do they only ship within the US?
I live in Asia and would love to test this out.

send me a pm if you like and I will try and help.
where in Asia are you situated? I may be able to supply you from here.

[Vapor]

send me a pm if you like and I will try and help.
where in Asia are you situated? I may be able to supply you from here.

I read on eVGA forums that you're the one who poked and prodded the Enerdyne guys to seed some reviewers, so I owe you some thanks

What other sites are due to post up a review? I'm still waiting for more results from people since I got so excited about IX's overall performance

@genec57--glad it's working out for you! Great drop in temps there, definitely a big drop considering you already have the best block available Sorry you goofed on the first install attempt, do you know what went wrong exactly?

[masska]

I read on eVGA forums that you're the one who poked and prodded the Enerdyne guys to seed some reviewers, so I owe you some thanks

What other sites are due to post up a review? I'm still waiting for more results from people since I got so excited about IX's overall performance

your very welcome I wanted to secure as many respected reviewers as I could. Allow the testing to show diverse methodology,application and above all accurate results. You're review and forum feedback shows that you're a true professional and your attention to format and detail should be congratulated.
I'm expecting another 2-3 reviews to be published soon. I know there has been a slight delay with receipt of TIM's to compare, so this has held things up a little. Will update as soon as I have more info.

[M.Fixit]

Doing it twice took me a bit over two hours.

Temps across the board are down about 3-4c. It's hard to pin them down very precisely. Nothing over the mid sixties where before I was pushing 70. I have had an ambient variation of about 2c throughout the day and load temps tend to jump all over the place.

As I said, expensive but well worth it. I am looking forward to hearing from more courageous users .

I know for the HK block that the manufacture states the spring height should to be set to 19.6 mm (and best measured with a vernier caliper) to provide the correct pressure on the mounting plate in order to achieve good contact with the IHS. Will this measurement be affected in any way by using Indigo Extreme as this appears (to me at least, unless I have got this wrong) to be thicker than normal TIM?

[genec57]

@genec57--glad it's working out for you! Great drop in temps there, definitely a big drop considering you already have the best block available Sorry you goofed on the first install attempt, do you know what went wrong exactly?[/QUOTE]

Yes,
I didn't like the positioning of the material at first and thought it would be ok to lift it off and reposition more accurately. WRONG. The reflow went up to ~100 ok but never really went back down. I turned the pump back on anyway and the temps went up to around 80.

When I removed the block the material came off in two pieces and had a crinkled look to it. Second time around I was ultra careful and everything went just as it should have.

[genec57]

I know for the HK block that the manufacture states the spring height should to be set to 19.6 mm (and best measured with a vernier caliper) to provide the correct pressure on the mounting plate in order to achieve good contact with the IHS. Will this measurement be affected in any way by using Indigo Extreme as this appears (to me at least, unless I have got this wrong) to be thicker than normal TIM?

Wow! I never saw that. I tightened the screws down the max and then backed off half a turn. I measured that and it was ~9mm.
I am at ~14 now and if I backed all the way out to 19 I think I would be almost all the way out.

edit: a detailed read of the instructions shows that the measurement if from the base of the spring to the top of the screw - that is a bit looser than I had it but reasonable. Any difference in temps remains to be seen.

[genec57]

I know for the HK block that the manufacture states the spring height should to be set to 19.6 mm (and best measured with a vernier caliper) to provide the correct pressure on the mounting plate in order to achieve good contact with the IHS. Will this measurement be affected in any way by using Indigo Extreme as this appears (to me at least, unless I have got this wrong) to be thicker than normal TIM?

To get back on topic The material is definitely thicker when applied but after reflow it is probably no different. This whole process bears little relationship to TIM as we have always known it. I doubt that I will ever go back unless I am forced to in an emergency.

[Utnorris]

Has anyone else tried it?