Arctic Silver 5 vs. IC Diamond

I find that Arctic Silver is performing better than IC Diamond.

I have a P4 2.8 GHz chip which of course runs way hot, and it never was supposed to be placed into a laptop.

IC Diamond was running at 59-62C and running the fan constantly. Silver is showing significant difference right in the 50-52C range. The fan is only running on slow now. Both were using pea shape method of application.

Anyway what do you all think between the two thermal pastes?
Do they differ for different processors?

 

10 degree celsius difference? Either you look at different scenarios of both or you did a not so good paste job with the IC Diamond.

Both are equally good

 

Been running an hour after Silver application and has only got into 55C range (even dropped 1C - something never seen with IC). Applied both pastes equally same.

 

From what I have seen online, they should perform roughly the same when correctly applied. The ICD is a little thicker, so if applied exactly the same it might not spread as well, leaving a thicker layer and higher temps.

 

ICD is better by a good bit. depending on temps you will see different numbers. If your at 30-40C with AS5 ICD will not due much. if your 70-90C with AS5 ICD has the potential to drop you 10C or more due to the high temps. With ICD you have to apply it right otherwise you will see virtually no gain and might get worse over time. Also the tighter you get it with ICD the better results you will get. Also TX-4 is better than ICD. ICD is ~4Wmk and TX-4 is ~6.x?Wmk. ICD is very easy to apply if you understand how to apply it. If your below 50C on max load ICD is roughly pointless for the improved performance for the price. But if your laptop/desktop is up in the 70-100C ICD is higherly recommend or TX-4.

EDIT: you definitely pasted ICD wrong.

 

From what I've heard, ICD is better by a little in terms of temperatures but AS5 is easier to apply. The difference shouldn't be nearly as big as what you're getting though.

 

Difference shouldn't be that big, there was a problem with the application most likely. Personally, i vote ICD, but only because it has no cure time and it's been going strong for a year in my G73, thermal performance has absolutely nothing to do with my vote really.

 

Very little difference between the two in terms of tangable performance, though if I was forced to I would give the edge to properly applied ICD. The trick is to actually apply ICD properly since it is not your traditional paste.

 

What are the differences between cure times?

 

IC Diamond claims less than two hours with at least 50PSI: Innovation Cooling

" Curing Time: IC Diamond requires minimal time to attain peak performance; in most cases, IC Diamond will reach peak performance after two hours of use. *
Note* Cure Time Assumes an optimized pressure of 50 PSI Lighter load pressures will increase cure times to a week or more."


AS5 claims 200 hours: http://www.arcticsilver.com/as5.htm

" During the CPU's initial use, the compound thins out to enhance the filling of the microscopic valleys and ensure the best physical contact between the heatsink and the CPU core. Then the compound thickens slightly over the next 50 to 200 hours of use to its final consistency designed for long-term stability. "


In either case you won't see much temperature drop after it's cured, maybe a deg or two. I personally like IC Diamond because it has performed better for me and it also is not electrically conductive. So if you accidentally goop some on your motherboard or don't manage to clean everything up it won't cause issues.

 

You definately either had a bad ICD install or little no no preasure for the HS to die or just a bad contact patch. Whatever the reason some thing is way wrong with the ICD install.........

 

When I pulled the chip to apply the AS5, the ICD was mottled on top. So it more than likely was just a bad contact patch.

 

I was skeptical about the claimed differences between AS5 and ICD7. I have been applying thermal paste for a long time and know how its done. I have noticed temps of only a few degrees between factory and AS5, even with a perfect application, but 10C between factory and ICD7. I have swapped CPU's at least 3 times in my Acer and repasted once just becasue I had the unit open installing a backlit KB. I see the same results from ICD7 every time. At least on my Acer, ICD7 is far superior to AS5. I am a believer. Pass the koolaid.

 

1- curing time is a myth
2- ICD is better then AS5
3- AS5 is an ancient TIM and nowadays there's many better

these are facts.

 

agreed and bronksy...he got worse with ICD that is a bad pasting job. Also as i said many times the difference varies on how high your temps are. if your below 60C the difference isn't going to be huge..in the 80C plus it will be a big difference

 

How long AS5 has been around doesn't matter. It was the first of the good pastes that we have today. And it isn't a bad paste at all.

I also think ICD gets a lot more buzz than it deserves. It was the first thermal paste to be able to beat AS5 (although not all tests between them show this), and lots of people think it is the default best even today. But in reality I believe it has been superseded. But the gains we are chasing are ever smaller.

AS5 and ICD are just a couple of the handful of great thermal pastes we have today. You really can't go wrong with any of them. For this, I didn't vote in the poll.

 

Really? I have always thought that you need to burn in certain paste before it works in full effect

 

The cure time doesn't do anything to help with temperatures. It's all about the cooler in the end with that crap.

I use GC-Extreme and it keeps temperatures considerably lower than with Arctic Silver 5. I do the "blob" method when I paste too, and it always gets good coverage. I'll state that Arctic Silver 5 isn't BAD by any means. It's really really freakin' good for anything you put it on. Just know that there are better pastes out there despite them not being advertised 24/7 like the AS5 is.

For the Pentium 4, I wouldn't really bother with using ICD on that machine; AS5 is fine. Most of my retarded temperatures were due to my poor application methods in the past... >.<

Essentially you put enough of an amount on the die to spread it around on the inside in the shape of the surface area, but smaller, while leaving enough TIM (Thermal Interface Material) after that to let it spread over the CPU. I'll demonstrate this at some point.

When you spread it, you use the tip of the tube you're using and just bob it up and down along the applied area. It ensures that you have a larger initial application area and guarantees you perfect coverage.

The thing you gotta remember about TIM is that the contact plate for the heatsink and the chip itself must be squished together. Just think of it like you're gluing something together. You want a perfect seal, if a little comes out the sides, it's not a problem. If anything, the excess will allow for greater thermal heat transfer. That's why you see laptops with a GOB of crap on the CPU, because the manufacturers don't want to deal with a potentially overheating laptop. The dramatic decrease in temps when you repaste is because the stock crap they use is...well...crap TIM.

Also, for those of you who don't have any TIM...if you have regular Mayo (NOT FAT FREE) you can use that as TIM and it works BETTER than ICD. XD

The fact that it's still recommended EVEN TODAY after all these years really does say something about how good the paste has performed. I voted for the AS5 since it's cheaper to purchase in large quantity than ICD and ICD is a PITA to apply... >.< I only did it once and, while the temperatures were crazy cool, I won't do it again. XD Too thick for my tastes.

Another good TIM to consider is Shin Etsu X23-7783D. It solidifies the connection when not in use, I.E. when the PC is off, so it lasts longer when it's not being used. Supposedly the 7783D won't ever dry up and need replacing because of this. I had a tube but it was enough for like... 2 applications and was CRAZY expensive. >.< You can get it on FrozenCPU.com too...but don't get it unless you're planning on seriously dropping some cash. ( $4 for a 1.5g tube!!!)

On that note, if it matters to anyone who has a desktop, the closed-loop liquid cooler heatsinks like the older H50 and Antec 620 use X23-7783D as their stock TIM. Pre-applied. Most people are scraping like...$15 off their heatsinks when they use AS5 instead. XD

 

ICD is so overhyped. almost as AS5.
I've some considerations when getting TIM.
1-non conductive
2-thermal effeciency
3-price
4-easy to apply
also depending on the cooling, it may be required to resist sub zero temps.

currently I use coolermaster icefusion. better then AS5. not the best but great anyway. And very cheap. 200gr for $20! lasts forever.

if I just needed a few applications I would get arctic cooling mx-4.
for me the best is indigo xtreme. but imo not worth it.

 

Just to note Thermal compound is a function of volume not weight, confusion exists as people forget that Diamond weight is different than silver weight - 1.5 grams ICD is = .5cc /// 3.5 grams AS5= .5cc.

An application of compound is to fill voids (volume) in the joint so equivalent volume for both is required, so in your FrozenCPU example ICD is actually a dollar cheaper for the small tube comparison.

Liquid in a porous media they all dry out, it's in the nature of the beast as they say.

 

ICD is a pain to apply, it can also scratch the cpu and the cooler because it's so hard. possibly even voiding warranty.

get prolimatech pk-1 it's cheaper, performs better and it's easier to apply.

http://skinneelabs.com/2011-thermal-paste-review-comparison/

 

I need to touch on this point a bit...

from what I've seen there are NO conductive pasts out there today. I wasn't around much, but I know during the early 2000s there were a few that were.

 

AS5 is conductive...anything gold and silver is conductive

 

there are many conductive pastes around. more or less conductive ofc. but there are many that are conductive enough to do some harm.
I could give to so many examples. but basically, all liquid metal ones are very conductive, also besides gold and silver, copper based are also very conductive.

btw prolimatech has 2 new versions pk-2 easy to apply version and pk-3 nano version. starts to be selling this month.

 

Diamond is hardest but only incrementally so - aluminum oxide MOHS 9 - aluminum oxide you will find in arctic silver, MX4 and Shin Etsu and most others.

Many people do not understand the proper use of solvents and cleaners and lack patience so they squirt the cleaner on and start wiping immediately not letting it do it's work rather they turn it into a mechanical removal operation vs a solvent removal.

Good shop practice is to let the solvent re-liquefy the compound then wipe I recommend acetone as a better solvent.

I have a video almost complete on paste removal just have not had time to shoot the last scene we have been to busy automating some processes. I'll have to post it up on about four hundred forum threads to correct the misinformation when I get it done.


As for application we recommend a bead/pea/line on center then compress with heat sink, nothing challenging about that

 

I thought AS5 was sold in volume not weight..interesting

 

Wow, more AS5 is conductive posts.

LOL, anyone know how to use an ohm meter? Anyone bother to go to the AS5 site? Nope, lots of fan boys here? No?

I posted my results of my back to back AS5 vs IC7 re-paste on my UL30vt. As posted, my biggest difference was in proper cleaning or lack of I should say. I would use either in a heartbeat but would love to spend the time with other TIMs. I did try the old white silicone based stuff for grins, both AS5 and IC7 stomped it by 8 degrees.

I posted a site where all these compounds were tested, along with peanut butter but users were crying over methods of testing and whatever.

It would be interesting to get a respected site to do long term, and honestly funded, testing of what; 50 or so TIMs on the market now?

 

2011-2012 Roundup | Skinnee Labs
Thermal Compound Roundup - January 2012 | Hardware Secrets
Thermal Compound Roundup 2011 - General-Discussions - Overclocking
80-way Thermal Interface Material Performance Test | Thermal Interface Material,Thermal Paste,Heatsink Compound,80-Way Thermal Interface Material Best Thermal Paste Heatsink Compound Cooling Performance Comparison Benchmark Tests

 

Typical. Company representative spouting complete baloney. Diamond is FOUR TIMES harder than aluminum oxide crystal (corundum).

Lots of people complain that IC Diamond is harder to apply than most other pastes, so I think I'll just trust them on that.

No, AS5 is not conductive. Squirt some out and put an ohmmeter in it to see for yourself. You will get this: ---.

 

1.) ICD is harder, no ifs, ands or buts about it. No arguing that point only that with proper application and removal it is non destructive. There is no arguing that as well.
2.) ICD applies differently, it so no harder other than to get it out of the tube. It is harder to remove though. That is a good thing as heat will not just boil it out etc.
3.) AS5 is not conductive but earlier variants were. Some other non brand silver pastes may still be as well.

Now ICD is a waste if your CPU with normal paste rarely see's 80C. Put on the AS5 and just be done with it. Now if you are pushing 90C+ on a regular basis ICD MAY just be the ticket. Then again it may not!

Edit here; Some even with not to critical of an original heat issue may just prefer ICD anyway, I am one of those too.

The problem lies in the heat sink or heat pipe. If with AS5 you are running temps where the real world TDP of the chip is getting the cooling system to borderline failure ICD then could push it over the edge. Once over the TDP edge that is it for getting the chip cooler and it actually will get hotter as the cooler may never recover. This can be true of a cooling system that in reality is about to fail from being worn out (age etc.). Also with the greater heat transfer more vapor is created that has to cool and wick back down the pipe. If it can not wick fast enough you are again dead in the water with a possible folding of the cooler.

So in theory and practice ICD can actually be worse than AS5 for your system. If you apply ICD properly along with proper chip contact and get no improvement or higher temps than with AS5 then your cooler most likely is in trouble. It may be you need a new one or better one or maybe a cooling mod is in order. every systems needs are different and that is something you will have to determine.

 

interesting...well i guess i am wrong :/ Though I think ICD isn't vary hard to apply myself....you just squeeze it out and take a toothpic to pull the extra off so you dont make a mess...not hard


EDIT TANWare can you further explain how your cooler can age and die? I have never heard of this.

 

The cooler over time in the heat pipe can loose preasure/vacuum and this alters its calibration. Over time too it is possible to loose liquid itself from leakage. In a perfect world all heat pipes would last forever but this is not a perfect world.

You also have to take into consideration abuse bending and the like. Also over time the fins may get coated with oils or other stuck debrise even just blowing it out will not get rid of. This all can lower efficiency dramatically.

 

i wonder if my heatpipes are dieing then...i have had them for 3 years and they run at 80-90C for a long time and now using ICD and added a 920xm and using it to its limits lol

 

That's a possibility but the solution may not be practical due to cost and availability.

I read that most of the laptop pipes simply use water as the working fluid. As Tanware noted, these things could leak or lose efficiency over time.

I suspect the very same on my UL30vt which is pushing three years as well.

 

I actually had a leak from a heatpipe inside a notebook. it was a defective heatpipe. it developed a very very small hole.

but besides defective parts I never ever heard of heatpipes losing effeciency.
I've a server running for 24/7 since 4 years ago withn a scythe ninja cooler full of heatpipes and I never changed the TIM which was stock btw and I never ever had a single problem or it got misteriously hotter with time.

 

Mine has been acting weird lately...whether it is ICD going bad or heat pipes or what not sure or maybe fans RPM fluctuating. I got ym net heat pipe for my cpu today for my water cooling mod. Is there away to check if it has leaks or missing gas?

 

Not all pipes will go bad. I would guess over an extreme amount of time and abuse maybe all would but most hold up well. some are better than others and there is always manufacturing variances. Just know because it is a heatpipe doesn't mean it is a 100% garenty it will work 100% forever. They can go out, not they will but they can so it is something to keep an eye on and if there is an issue it can be suspect.

There is no real way I know of to check the vacuum inside the pipe. Ususally it is negative preasure so that the water vaporizes at a lower temp. Just remember the majority of our components are from china or asia and not always the highest quality etc.

Edit; if acting up about the best thing you can do is eliminate issues one at a time, be sure the vents are clear, be sure the fan runs properly, try another application of tim and then if all else fails get a new heatsink and pipe...............

 

I think Super Bee has used Ic diamond too thickly, because I hear only good things about it improving temps by 4 - 25C depending on application thickness.
Because Ic Diamond is really thick, to apply mine thinner, I have applied really thin lines about 1mm thick in an X and another X turned on the side (like a British flag), but to within 5mm from the edge of the cpu heat spreader. I have had excellent results that are rediculous with a nice big heatsink (Thermalwright macho).
With standard TIM and stock cooler, my cpu Intel quad core Q6600 was at 60C idle. After Ic Diamond about 30C, oh yeah!

 

Nice necro for your first post

 

No need to be a jerk about it.

Welcome to the forums.

 

Thanks, peanut gallery

 

Neither. Liquid ultra is the only way to go.

 

YES The only and lonely. All others far behind.

 

This, screw the other pastes.

 

It doesn't matter what paste you use.

When you're dealing with high-performance name-brand thermal pastes, they all perform within the same range of each other. Sure, you can run torture tests, benchmarks, etc that show that Paste A is 2 degrees better than Paste B. But in real-world practical use, that might as well be within the margin of error, and might as well be considered identical in performance.

BTW, you applied the IC Diamond incorrectly. The temp difference should be <3C difference. If you're seeing 10C difference, then there is more at work here than just the differences between thermal paste material

 

I.. suppose it could be about whether or not liquid ultra has better heat-transfer at high, stable temps. While grease paste, applied well at least, might perform better on the somewhat lower and irregular temps you typically get on a laptop. And that perhaps the increased price, and hassle with applying liquid metal might make an adhesive pad a better purchase in terms of ease of use and price, at least for a laptop.

I wondered about this in the guide I wrote.. but anyone who have used the Collaboratory products for a long time know how easily it corrodes, and if it doesn't like being heated up and cooled down a lot? And if an uneven seating (which you often get on the sub-par cooling assemblies on laptop) makes it stop working at peak very quickly, before it becomes useless? It essentially becomes an extra layer the heat has to go through, after all, so it seems possible that if the seating isn't perfect, it would reduce the contact surface area, if it shrinks or there's a gap, etc. ..or if it perhaps retains a stable heat-transfer over time, even if it won't work perfectly.. that sort of thing?

 

You forget the lifespan. This is just as important. Take forexample Noctua Nt-H1. Got lots of kind words... Good in the beginning but useless after a short time. Tested this on my extreme processor. After 2-4 days so was this paste useless. Gelid extreme was a little better but also this was useless after short time. The difference lies in the consistency on thermal pastes. The most soft varieties of thermal grease has a shorter lifespan regardless of high end brand.. So there is a difference in thermal grease that has been widely praised. It does not help with good temperature results if this only last a short while.

 

This is very true and often overlooked. Not only that but we've done a thread comparing all the TIMs against Liquid Ultra in the Alienware section and many people recorded double didget reductions in their temps under load both in heavy real world usage and benchmarking.

Also Liquid Ultra doesn't corrode if you have a copper heatsink which most people do.
Additionally it's not hard to clean (ultra), which is a wives tale... Not that you'll ever need to repaste anyway.

 

Dang, a few days?! How often should AS5 be replaced?

 

Something is really exaggerated here... Maybe you've got a defective one. Noctua NT-H1 is on my GT70 barebone's both CPU and GPU for six months and it still performs superb performance. Can't say "just like its first day" because the ambient temp then (March) was a lot cooler than now, so I would expect warmer temps. It is normal due to weather and NT-H1 is not the responsible in my opinion.