TIM ...

[2Tired2Tango]

Now that is some scarey stuff.... boiling water? Not on any of my computers you don't!

Ever heard of Loctite Thermstrate? (package photo below). This stuff is available from Loctite in a number of different thermal ratings, the TC 175 -125 pictured will reflow, under pressure, at about 55c and puddles to a liquid at about 90c.

It applies to the CPU (GPU, Chipset etc) direct from the "pen" in a semi-smooth film that is very soft so it begins reforming itself the minute pressure is applied from the heatsink. To "cure" it you simply assemble your system and run it like normal. It will run hot for a couple of hours and then begin to cool off as the compound flows and conforms to the gap between CPU and heatsink.

In prolonged use the material (Camphor, Zinc Oxide and Silicon) will set when cool and reflow each time it is thermally challenged. The hotter it gets the wetter it gets and the more thermally conductive it becomes... up to it's maximum of 90c. So this stuff continuously reflows itself throughout it's life cycle.

In use the CPU temps are about the same as for Arctic Silver, perhaps a degree or so higher, sometimes lower depending on the state of the compound.

Interestingly this works better than Arctic Silver on epoxy top chips (like some chipsets) because it has better gap filling ability.

Now here's the good part... That applicator tube is about 1 1/2 inches wide, 1/2 inch thick and 4 inches long... enough for literally hundreds of chips.... for $15.45us.

I got one as a sample from Loctite several years ago when they were developing the stuff for production use and haven't even used half of it yet.

 

[Psi*]

I am intrigued. I have another machine to build in the next few weeks.

 

[2Tired2Tango]

I am intrigued. I have another machine to build in the next few weeks.

You can order it from Stanley at the link provided.

The only caution I would give you is that the application depends a lot on how steady your hand is... If you can get a nice thin film in one pass, it will settle in and work perfectly almost immediately... Your chip temps will go up to about 60c for 10 or so minutes and then start to come back down.

The problems come in if you a) use too much or b) get lumps on the surface. It will eventually figure itself out but it could take quite a while during which your chip is not getting the best cooling.

Ideally you want a thin film, barely opaque on the roughest surface before you mate the chip and sink. (On anodized sinks I usually put it on the sink. On epoxy topped chips I'll put it on the chip. For polished metal surfaces I apply it on both chip and sink because the smooth surface produces a thinner film.)

I've never done any specific "curing" tricks. I just let it figure itself out and it's always worked pretty good.

This stuff works best on large surfaces like the X2 and X4 CPUs and most bridge chips.

The ultra tiny "exposed core" cpus (like the old Athlons) don't like this stuff at all and will eventually squeeze it entirely out of the junction but my good old zinc oxide works just fine in that case.

I'm sure you'll write back and tell us how it went....

 

[b1lk1]

I know I am coming late to the show, but Arctic Silver has not been the "in" paste for some time. After testing nearly every compound on the planet sold in the retail channel, I can state that it is absolutely not near the best. Most will be within 1-2C of each other and they are, as has been said, all basically the same. Right now I use ArcticCooling MX-2 for all my Techgage needs as well as personal use. Is this because it was given to me? No. It is because it only cost me $25CAD for a 30g tube which will probably last through 100+ builds.

Stay away from the overpriced and over-hyped stuff.

 

[2Tired2Tango]

I know I am coming late to the show, but Arctic Silver has not been the "in" paste for some time. After testing nearly every compound on the planet sold in the retail channel, I can state that it is absolutely not near the best. Most will be within 1-2C of each other and they are, as has been said, all basically the same. Right now I use ArcticCooling MX-2 for all my Techgage needs as well as personal use. Is this because it was given to me? No. It is because it only cost me $25CAD for a 30g tube which will probably last through 100+ builds.

Stay away from the overpriced and over-hyped stuff.

FWIW ... I agree. Most pastes are pretty much the same thing. As you say they all run within a couple of degrees of each other and the only real comparison is cost.

I think you'll find things are a bit different when you get into phase change (reflow) compounds. Here, there are very real differences.

Mostly it's the curing phase that worries me. Having to get your CPU (GPU, etc.) up to some ridiculously high temperature before the stuff will reflow is risky. It puts expensive components at a high risk of heat induced failure. And, as I'm sure you know, no manufacturer is obligated to honor warranties for heat damaged components.

This Indigo extreme is a metalic phase change compound that needs at least 90C --very close to the failure point of many chips-- to flow and it will not reflow until it gets that hot again. Beyond the horror of using a CPU chip as a soldering iron, this stuff just gives me the willies because it's brittle. One good bump and the thermal junction is broken and the only way to fix it is to again expose your equipment to the dangerous heats of the curing phase with a fresh $20 patch.

The reason I brought up Thermstrate is that it works. This is a waxy substance with about the consistency of a bar of soap. It is soft enough to move and reform itself under moderate clamping pressure such as that applied by the heat sink clip used on a chipset cooler. But unlike pastes it's cohesive, it's not going to ooze itself out of the thermal junction. Finally, as I mentioned earlier this stuff continuously reflows throughout it's lifetime... bump a heatsink, break the junction, and the chip runs warm for half an hour or so while it heals itself... try that with thermal solders.

If that applicator from Stanley is the same size as the one I have, it's going to do several hundred of chips... So it's a real bargain. Of course that's an industrial supply price, not a retail one where you're paying for packaging, middle-men, store overheads and sales staff salaries... so right off you're saving money. There are places where you can get a half pint tin of Zinc Oxide paste for under $50.00... it just won't say "Heat Sink Compound" on the can.

It's not about getting a free sample, Bill... it's about stuff that actually works.

You can see a review of Thermstrate, with test results at http://www.dansdata.com/powerdev.htm

 

[b1lk1]

Most CPU's will throttle well before they get damaged. I have routinely run the QX9650 I have for testing purposes over 100C time and time again with no loss of functionality at all. I certainly would not recommend this and I am not saying it is OK to do so on a continuous basis, but for short durations there should be no ill effects. I can remember vividly seeing a reading of 112C during a heatsink test when the heatsink itself had become partially dislodged from the board preventing a proper seating.

I will definitely take a look at that Thermstrate link. I love finding out about new compounds and such that can demonstrate actual gains. The only other "phase change" material I have heard of that was was some sort of "liquid metal" and the name of the product escapes me at the moment. It has a really major downside of being unfriendly to remove and it was also conductive.

 

[2Tired2Tango]

Most CPU's will throttle well before they get damaged. I have routinely run the QX9650 I have for testing purposes over 100C time and time again with no loss of functionality at all.

Yep... but how many years of usefull life did it cost you?

The sad part is that under testing conditions the chip will take that several, even dozens of times... but in daily use one such heat cycle can take years off the lifetime of the chip.

Most of the CPU failures I've seen have been heat related.

I found a data sheet for this same stuff branded as "UltraStick", attached

 

[Psi*]

Respectively ... slow down a bit ... I don't think you should jump to any conclusions based on the pic anymore than you should jump on using a TIM because you get a lot for few $$. I was surprised with the thermal response in the pic & I have been meaning to send the information Enerdyne Solutions for a response. Responsibly, that is what I should have done 1st, but still have not. But, I will ... ... ... sometime soon

Also, Intel does not publish nor is there a failure temperature. Although even finding a maximum recommended operating is not exactly forth coming. It appears to be 71 C or is it 69.7 C ... two values I have seen @ Intel.

One good bump to break the metal reflow would probably happen when the bump is sufficient to knock the heatsink off the board ... hmmmm, I wonder if that wouldresult in any board damage?

The price is definitely good for Thermstrate, but I have not been able to find any independent analysis of the stuff. Usually there is a manufacturer's web site if you drill into the net long & hard enough to at least find their their own reports. Nadda for that also. And yu can't really go by a PC reviewer's testing as some of those guys just throw things together not to mention sprinkles of pseudo-science.. There are many professional labs with free published reports for several of the TIMs that we use. The lab tests are repeatable & keep track of TIM thickness & also track the TIM's effectiveness over temperature and temperature cycles.

It is also a little frustrating to try to compare these materials for thermal resistivity or conductivity as the units are often presented a little differently. It is *not* that hard to convert, but for when I feel a bit lazy I'll hold the page open for a day or 2 until I either follow thru or forget why there is so much clutter. But, Thermstrate appears to have about 10X the thermal resistance to Indigo Extreme. And that is if you apply Thermstrate evenly ... although I have never had any problem applying any of the goo from the tubes although I do have a few credit cards with gray edges.

I will not say that you get what you pay for. I did get what I wanted for my term of applicaton. More expensive is not necessarily better, but the extra $5 in the scheme of things is nothing. I have misplaced more & just given away more left over tubes of this stuff than I can remember. When the hexa-cores are available I'll be back with what it is like to remove the stuff.

 

[b1lk1]

Honestly, I doubt it did any real damage at all. That CPU is still pulling off the same 4GHz overclock with all the same settings. Again, running it 24/7 would do some damage, but even a solid few hours of super-heat should cause next to no damage. Over all the years of punishing and abusing parts, I have yet to kill a CPU due to heat.

 

[Psi*]

Honestly, I doubt it did any real damage at all.
.
.
Over all the years of punishing and abusing parts, I have yet to kill a CPU due to heat.

I agree & same here. The system that this was built from ran Athlon MPs grossly overclocked and hot for several years.

 

[2Tired2Tango]

Respectively ... slow down a bit ...

My friend... I'm thinking you are over-thinking this by a few parsecs.

Here's what you got so far....

1) A suggestion from someone who's used the stuff in a couple of hundred projects (mostly computers, but some audio as well) over an 8 year period and has never had a problem with it.

2) A rebuttal from someone alluding to me going ga-ga over something I got for free.

Thermstate is an OEM product, mostly sold out of view of retail or even wholesale customers, the specs are there if you contact the parent company, as are all their application notes and research. But I seriously doubt that a 1-off customer is going to get much access to them. Any idea how many products in your own home might have this stuff already in them?

Consider this... If Intel is specing their chips at 70C, you've already taken your $300 CPU chip 30 degrees over that.... in the name of protecting it from overheating.


So, you're left with my word... make of that what you will.

 

[b1lk1]

Tango, I didn't mean to imply you got it for free and that was why you recommended it. I should have been far more specific when I said that so that there was no misinterpretation. In fact, that was pointed towards myself since I tend to get lots of stuff for free and that leads people to believe that I (not you) tend to recommend them.

Also, since I do get alot of free hardware, I willingly and intentionally punish it far worse than anyone that has to buy it. I purposely push my QX9650 machine to the upper limits of heat and then some. I don't even flinch when I see a 90C+ temperature reading. I let the tests finish and I record the results.

Honestly, anyone would not be wise to overlook the product you are talking about if they are indeed looking for an extreme solution. These phase-change type compounds are the best that can be used. I just use stuff that sets quick and can be wiped off fast since I never have a heatsink on any of my machines for more than a few weeks if that.

 

[2Tired2Tango]

The price is definitely good for Thermstrate, but I have not been able to find any independent analysis of the stuff. Usually there is a manufacturer's web site if you drill into the net long & hard enough to at least find their their own reports. Nadda for that also.

You sure didn't look very far.... Took me less than 5 minutes...

See attached...

 

[2Tired2Tango]

Tango, I didn't mean to imply...

I give up... Merry Christmas gang... I need a vacation.

 

[Psi*]

My friend... I'm thinking you are over-thinking this by a few parsecs.

Here's what you got so far....

1) A suggestion from someone who's used the stuff in a couple of hundred projects (mostly computers, but some audio as well) over an 8 year period and has never had a problem with it.

2) A rebuttal from someone alluding to me going ga-ga over something I got for free.

Thermstate is an OEM product, mostly sold out of view of retail or even wholesale customers, the specs are there if you contact the parent company, as are all their application notes and research. But I seriously doubt that a 1-off customer is going to get much access to them. Any idea how many products in your own home might have this stuff already in them?

Consider this... If Intel is specing their chips at 70C, you've already taken your $300 CPU chip 30 degrees over that.... in the name of protecting it from overheating.


So, you're left with my word... make of that what you will.

If your "word" is a suggestion or opinion, then it is only what it is ... and I was taking it as an experienced profession based on your other responses on this forum.

If you consider your *word* to be absolute unequivocally undeniable fact ... good luck with that.

This quoted response and other recent responses in this thread are suggesting that you feel very strongly that your opinion ought to have far greater respect. A hint of anything less is intolerable and you are intentionally being insulting, sarcastic, and unnecessarily defensive.

There are better places in this forum to take a rant than in a technical thread. It is normal internet forum courtesy to stay somewhere close to the topic.

 

[Psi*]

erst while a response from Enerdyne Solutions

[FONT=&quot]From:[/FONT][FONT=&quot] Chris Macris
Sent: Wednesday, December 23, 2009 5:47 PM
To: Jeff
Cc: 'John McCullough'
Subject: RE: Indigo Extreme reflow<o></o>[/FONT]

[FONT=&quot]Hi Jeff,<o></o>[/FONT]
[FONT=&quot]<o></o>[/FONT]
[FONT=&quot]Thank you for your purchase and glad to hear that you installed IX successfully.<o></o>[/FONT]
[FONT=&quot]<o></o>[/FONT]
[FONT=&quot]Re: Your questions:<o></o>[/FONT]
[FONT=&quot]<o></o>[/FONT][FONT=&quot]
An excerpt from Intel:[/FONT] [FONT=&quot]“Intel® processors have built-in thermal protection. If the processor gets too hot this built-in protection will actually shut down the computer. If your computer is not over-clocked and is running under the design specification, this should prevent damage” (http://www.intel.com/support/processors/sb/CS-029426.htm).[/FONT]<o></o>

• [FONT=&quot]The Intel throttling feature is designed to keep core temps below max junction temp; although Tj max is proprietary, internal tests and public domain data suggests that Tj max is ~100C; therefore, if you did not plug in the pump, you would see the cores peak at ~100C (typically bouncing between ~95-100C);[/FONT]<o></o>
• [FONT=&quot]Yes, looking at the screen shots, it would appear as if you installed IX properly; of course, if clamping force was non-uniform, you may see a similar reflow temperature profile and performance may be <optimal; what="" performance="" gain="" do="" you="" see="" (relative="" to="" your="" previous="" interface)?=""></optimal;>[/FONT]<o></o>

[FONT=&quot]<o></o>[/FONT]
[FONT=&quot]Sincerely,<o></o>[/FONT][FONT=&quot]<o></o>[/FONT]
[FONT=&quot]Chris M. [/FONT]<o></o>

[FONT=&quot] <hr align="center" size="2" width="100%"> [/FONT]​

[FONT=&quot]From:[/FONT][FONT=&quot] Jeff
Sent: Wednesday, December 23, 2009 1:04 PM
To: info@enerdynesolutions.com
Subject: Indigo Extreme reflow
Importance: High[/FONT][FONT=&quot]<o></o>[/FONT]

Hi,
I have just installed a water cooled system on an Intel i7 920 processor & used your material, Indigo Extreme for the 1366 socket, as the thermal interface material. I followed the installation instructions to the letter and have no problems.
<o></o>
Attached are screen photos of the reflow process. You will note that in the “thermal run away” photo the temperature begins to rise toward 100 deg C at which point I turned on the water pump for the WC system … plainly evident. The scale across the width of SpeedFan in 13 minutes.
<o></o>
The water block is a Swiftech Apogee XT and I used ½” ID tubing. To me that equals an ample volume of water. When the temperature started to rise I did notice a volume of small bubbles (sort of a “plume” for lack of a better word) forming in the tubing immediately connected to the water block.
<o></o>
So a few questions.
<!--[if !supportLists]-->1)<!--[endif]-->What basis do you say that bringing the CPU to 90 deg C is “safe”?
<!--[if !supportLists]-->2)<!--[endif]-->You indicate that there are thermal controls to prevent damage to the CPU. Referencing the 2^nd picture, when would one expect that to occur? Or, did I just linger too long before I turned the water pump on. That would have been about 3 or 4 minutes & I was waiting for the temperature to settle near 90 deg C again … somewhat suggested by the reflow instructions.
<!--[if !supportLists]-->3)<!--[endif]-->Can it be assumed from these photos that Indigo Extreme is properly installed? I have been running the system continuously and overclocked for the few days since the installation … 12/18/09 I believe.
<o></o>
Thanks in advance,
Jeff

 

[2Tired2Tango]

Given that Intel does not specify your processor (i7-920) for temperatures above 68C (intel data sheet, attached) what I see in the letter you got from the Indigo Extreme guys is somebody trying real hard to justify a very expensive BAD product that has put your CPU at risk of perminent damage or failure. Moreover, such a cavailer attitude toward manufacturers specified safety limits probably also carries exaggerated claims of product performance.

Thermal, over-voltage and over-current protections are akin to a climber grabbing at the edge of the cliff as he's about to fall over... they are put there in the hope of preventing damage but relying on them to save you is as stupid as climbing without safety ropes. In the case of electronic products it is widely considered to be unwise in the extreme to rely upon built in safetys to protect you money and emotional investment in your projects... only the most inept designers would even consider it.

There is a good and valid REASON that so much effort is made to keep these things cool...

Moreover, I'm betting you didn't even consider the ancilliary effects of producing so much heat on a PLASTIC socket so close to hot-air solder and fly poop sized components that surround it. It might not even be the CPU that gets damaged...

I've been in electronics a very long time. I've done tens of thousands of repairs on circuitry that is supposedly protected in all kinds of ways, but failed when those protections were challenged by inept and abusive operation of the eqipment. One of the most difficult things I have to do is, from time to time, try to explain these kinds of issues to someone who has, quite frankly just enough knowledge to be dangerous.

Bluntly stated... You appear to have been sucked in by a schister company hawking a bad product under the guise of advanced technology. I recommended Thermstrate precisely because it does respect the thermal limits of these chips and provides an adequate thermal transfer ability to keep them operating safely.

That said... I'm outa here. I don't see much point in continuing. As they say "You can't teach a pig to sing... All you will do is frustrate yourself and annoy the pig."

 

[Psi*]

... and annoy the pig."

You do seem quite annoyed

 

[2Tired2Tango]

You do seem quite annoyed

Well, lets see, now...

In the last few hours I've been accused of promoting a product because I got a free sample, told that I don't know what I'm talking about and had to listen to people who's biggest qualification is that they haven't burned down their house yet tell others to ignore manufacturer's specifications for the safe operation of their products... All in rebuttal to a suggestion that would likely have saved you a lot of money, fuss and trouble in the future...

So yeah.. I'm a tad annoyed.... and thanks to your childish rebuttal, now I feel insulted.

Enjoy yourselfs guys... but if you continue as you are, be sure to catch that sale on fire extinguishers.

 

[b1lk1]

Tango, once again, I did not imply you are promoting a product for free samples. I said what I did because people think I promote products because I get lots of free samples.

As for Intel heat specs, those are for 24/7 operation. They do not even heat throttle till well after 90c. They will not fail running over 68c from heat for short durations. I understand it is not recommended they run over these temps, but there is ALOT of thermal headroom.