Best Thermal Paste According to Independent Studies?

Yeah i mentioned that, but on copper it works! Plus, AS5 issue that i mentioned earlier…have to test it though

My point here is that PK1-2-3 or even Gelid wont do any good here, vise-vesra, they are worst that a cheap pure sillicon paste, the emitter in our case is always working at its max( most of the times at least), temps are above 85 deg, those CPU pastes arent designed for that

Yet…AS5….unfortunately my CPU is permanently metal- glued, and the LED die dont have build in thermal sensors

Have to ask someone from Xtreme if he tested with AS5, they do experiment alot

Ive always had great luck with arctic silver 5

There is very little difference among all of them. Even the best thermal paste will only give you a 3-4 Celcius better than the cheap one, which doesn’t make meaningful difference IMHO.
It’s more about the peace of mind, knowing you have used the best.
I’m not suggesting that you go with the cheapest one, but I’m also not seeing the point why anyone would spend $12 for a 4-gram syringe.
I’ve tried possibly all the best thermal paste in the market for comparison, from Shinetsu, Gelid, Prolimatech, Zalman, and my conclusion is that they are not worth the money you spend.
Even the infamous Shinetsu 7783D ($8-9 3 gram) only lowers my CPU 3C better than the cheap Servisol thermal paste ($2 25 gram tube) at full CPU load.
I bought Servisol 5 years ago, and I still have half tube even now. It performs the same like 5 years ago.
If you want to save your money, go aliexpress & search for Halnziye HY-810 or GD900, which I suspect to be the same thing.
It’s so much cheaper than those “fancy” pastes from Shinetsu, Tunix, Gelid, etc for a much more reasonable price.
You can get GD900 for $7.70 for 3 syringes @15gram, which you can use for a lifetime. They are non-bleed pastes. Viscocity is medium.

I have used Halnziye HY-810 & GD900 in the last 2 years. I bought HY-810 for only $10 (20 syringes 1gram) , and $ 4.5 for GD900 (30 gram tube). They give the same performance as the stupidly expensive Shinetsu, Prolimatech PK3, you name it (but I actually use Servisol for my flashlight). Unfortunately HY-810 is a little more difficult to find in retail package now, but you can still find it in Ebay for $1 a tube 1 gram.
Or simply buy GD900 in aliexpress, which you can buy much cheaper in a tube of 30gr ($5).
Both HY-810 & GD900 performs exactly the same in my overclocked i7, and I actually suspect they are the same thermal paste produced by Halnziye, only re-packaged.
Even Tunix has considered selling HY-810 under its own name as a successor of TX-4, and has asked puroverclock.com to evaluate their investment.

Here is the result:
http://www.pureoverclock.com/Review-detail/hy810-and-hy880-compound-review/

Excuse me, but i wont agree with you: reasons

- Firstly, the temp differences arent 2-3 deg but they can reash >10

- Every deg lower DO MATTER when we speak of extreme OC

- When the LED or the CPU are well OCed that thermal paste difference is even bigger

  • The main problem::: most of them DO DEGRADE pretty fast in time, and by ” in time” i mean 1-2 weeks when we are talking about extreme OS, believe me, i have a huge headaches about that matter before we cleared that issue out in Xtreme sys forum

Well, a causal user wont bennefit from an ultimate thermal interface, its true

Fujik sold by DX is the worst I’ve ever used, usually use Arctic Allumina.

In VERY high temperature and small footprint die’s the thermal coefficient/conductivity really matters, especially when it’s attached to an air cooled (or water cooled) radiation device, the faster it can move the heat from a small surface area to the larger surface area heatsink yes…those very detailed charts do come into play

With lower power direct soldered emitters and large footprint copper stars, the temperature coefficient might not be that important, due to the fact that thermal saturation occurs because the heat really doesn’t have anywhere to go once it bleeds off from the emitter to the star, to the pill, to the body of the flashlight

Cost effectiveness vs thermal effectiveness comes into play in this area

Computers vs flashlights = apples vs oranges

I use silicon and epoxy (and sometimes super cheap artic silver on the threads of the pill)…and the thermal management seems to do just fine, dropping $20-$40 on a tiny tube to put on a <$20 flashlight…well you see the conundrum?

Using good thermal compound is always good…but in flashlights will it really make that much of a difference?

just to add some info

I used Gelid GC extreme on GPUs, it has excellent performance in beginning, but after few months, it degraded to have absolutely terrible performance, worse than using no paste at all, I think it actually became a pretty good thermal isolator

The thing with Coolabs is it doesn’t degrade. That’s the key. Everything I tried before that- well maybe not AS5 that always lasted but it had a awful long cure time. But other pastes fail. They work for a time but dry up and sure enough over a couple months you would see the temps basically skyrocket and everything would fall out. In flashlights it may not be critical, but no thermal transfer over time is something I’d rather not deal with, it can only come back to blow up in your face :slight_smile:

What you see now is not what you get down the road… If the emitter is screwed down it may not suffer too much, and if the light is cheap then I use the bucket tim… But I’d definitely use it in lights you are going to push if possible (I know the aluminum thing is something we never had to deal with- always copper or nickel)

Personally I’ve never seen any thermal paste (even the cheapest one that I got free from a $2 heatsink) that causes >10C difference, unless when compared to tooth paste (as shown in Tom’s Hardware graph above ). The only exception might be Coollaboratory liquid metal, which also has pros & cons (I wouldn’t explain this here).
It’s true some super cheap paste degrade pretty fast, perhaps several months. If it degrades within 1-2 weeks, then I think it must be the lowest of the lowest quality (which is why I said I don’t suggest to buy the cheapest paste, but at least a decent one).

Even in extreme OC, when a couple of degree makes the difference betwen stability vs instability, then I believe it has been pushed too far beyond the limit because it can crash anytime. The room temperature itself can go up & down several degrees throughout the day. I wouldn’t call it a good extreme OC. I also believe when we are talking “extreme OC”, then it is already beyond normal thermal paste (watercooling, liquid metal pro, etc).

I agree with what Warhawk said, that is why I only use the cheap Servisol for my flashlight, which has been proven to perform almost as good as the expensive paste. This cheap Servisol also has the right bleed consistency/density to be applied to the flashlight, better than my other expensive thermal paste.
I’ve been using this Servisol for more than 12 years. I used it mostly for friend’s computers, and it doesn’t degrade for years. I have many friends still using the same paste which I applied 5 years back, and their CPU temperature is still relatively normal (apart from dust cleaning on the heatsink itself). And it is only $2 for 25 gram tube.
I have applied HY-810 & GD900 to my overclocked i7 desktop @4.8GHz & all my family’s notebooks since 2 years ago, and all seems to be working very well till today.

This! Read it carefully

I think you havent read my whole post, pls did it

Is Prolimatech Pk1-2-3 lowest quality? 15 euros per 5 grams? OR maybe Gelid? 21 euros per 5 grams!
Its not a matter of months dude, its a matter of weeks, especialy on a naked CPU die! Believe me, i passed through this, and i am just sharing my experience, for the others NOT TO SUFFER the way i did, and wonder wft is going on
There is a Xtreme forum topic about this matter

Arctic Silver Arctic Ceramique 2 on the other hand is extremely cheap, and lasts for years, as user “”fellfromtree”” explaind it well, it doenst dry that fast…maybe ina year or too yes, havent tested it much
I use this one widely on my LEDs

I ordered AS5, and will torture my wifes PC :bigsmile:
WIll post results eventualy, on theory it shoudlnt degrade that much, but again, the silver can oxidise, who knows, have to see it

P.S
On my PC setup, the difference between Gelid( tested it second) and Arctic Silver Arctic Ceramique 2 on the WC setup and under 5 gigls load was exacly 10 degrees, compared with CL liquid metal its 15-17 degrees
All this with the Intel default thermal glue removed

I assume that's the same stuff I have, they've changed the packaging since mine was bought ~10 years ago. The tube mine came in split open and died long ago, it's since been scooped out and now lives in a repurposed Tylenol bottle. Still works like the day I bought it (whenever that was). There is no paste that will work any better in any conceivable flashlight application, the heat loading just isn't high enough. It also doesn't degrade over time either in use or in storage.

I use this same stuff on my 4.4GHz i5 3570K, been assembled for over a year and still runs the same full-load temps as when new.

Interesting… Well you would know on the flashlight end for sure Comfy- I’ll bookmark that for later

Yes I'm a cheap bastard, but more important than the price is whether it's a good value or not. If it's cheap but doesn't work, it's not good value, it's wasted money just the same as if you overpaid for something that does work. Sometimes the better stuff is truly worth the extra money but a lot of times it's just not. If the best stuff is indistinguishable from the cheap stuff, that's just throwing away money. Flashlights and light meters, unlike the human brain, are not subject to the placebo effect.

A quote from a welding instructor dude: "Sometimes 'good enough' really is good enough, and sometimes 'good enough' is better."

And sometimes you feel like a nut, sometimes you don’t.

I now use Cool Lab Liquid Metal Pro with all my copper pilled lights (and brass pilled). This includes several of my own build (the copper pill that is) and the Sinner lights with their copper pills.

Is it better? Don’t really give a rat’s hiney. I’m not wealthy, but I don’t count pennies either. The amount of the liquid metal I use is so small that the price of the syringe is redundant. With many burgers costing $8-10 these days, how does it really matter? I like it because it’s as good as it gets. And unlike a direct reflow onto the copper or brass pill, it’s far easier to remove the star if I have/want/need to replace the emitter.

I have also been known to use Arctic Alumina Thermal Compound as a potting agent, using big globs of it at times. Yeah yeah, JB Weld Water Weld Stick sets up as fast and would hold up better under temp extremes. Need to get some more of that, don’t get out much and can’t remember I need it when I’m out there.

If it’s copper or brass, I use the liquid metal pro. If it is self centering with no problem, I use Arctic Alumina Thermal Paste. If I need to ensure it doesn’t move, I use Arctic Alumina Thermal Adhesive. On the more rare occasion that my pill is a huge chunk of copper and I am pretty sure I won’t be messing with it again, I re-flow it.

I’m using the cheapest paste I could find, some 25g tubes of Ceramique 2 that I picked up on sale for about $2 a piece, and it performs within spitting distance of the more expensive premium pastes that I’ve got. I’ve thought about trying some of the “boutique” solutions like CL Liquid Ultra, but for the $10-$15 I can essentially buy two or three more emitters and cherry pick the best one of the lot instead. Plus, I don’t have to worry about getting conductive goop on the driver or emitter as I would with Liquid Ultra.

I’m a little curious about the discussion of thermal pastes degrading quickly when used on naked dies. Back when the single core Athlon processors were popular, I had a 3200+ running smoking hot with a semi-passive heatsink for years without issue. Heck, even two months ago I replaced the thermal pad on a cheapo Gigabyte 785G motherboard northbridge with some ancient RCA branded thermal paste to give it a little more overclocking headroom, and there hasn’t been any sign of the paste deteriorating despite running constantly at around 65C.

It’s not the first time I heard about thermal paste degrading in several weeks on naked die. I believe their stories, but I just don’t understand why it happens to some & not others.
Even MX4 was accused the same in a computer forum when it always degraded in just 3 weeks on a GPU. If I remember correctly, the user solved the problem after changing to Arctic Silver 5.

4 years ago, I used the cheap Servisol white goop on Asus N61JQ laptop with i7 720QM & ATI HD5730, both on naked die. The i7 720QM is one of the hottest CPU ever known to man. I often did several hours of video encoding on it, and the CPU usage was 100% on all cores & temperature reached 90C easily. The CPU idle at 58C & 62-67C when browsing & youtubing. It worked for like 12 hours like that everyday. I didn’t bother changing to better thermal grease even when I had at least 10 better grease in my fridge, from Shinetsu G751, 7753D, Prolimatech, Arctic Silver IV, MX2, MX3, Coolermaster Premium, Gelid Extreme, STG2, etc because I didn’t care less about the difference anymore. I have played around with all of them on the first week I got the laptop to tame down the devil CPU, but I only got around 2-3C better among them. I used the cheap goop for almost a year without any problem until I had regular BSOD problem. First I thought “Hhmmm it must be the goop drying” (even when I didn’t detect any anomaly on CPU & GPU temperature). Understandably, the first suspect was the cheap Servisol on that scorching hot CPU, so I changed to Shinetsu……still BSOD. After testing further, it wasn’t even the CPU or paste to begin with. It was one of my 4GB DDR3 going bad.
The laptop died just 5-6 months ago. It was the GPU problem showing heavy artifacts, blank screen, BSOD. It’s either the dedicated VRAM or the GPU core itself. Oh well, that laptop has served me so well in the last 5 years. Rest in peace , my friend.

I put coolabs on copper last night and took a pic… You have to rub it in first for it to perform right. Then put a little in the middle of the pill and mount strait down.

Your practically soldered then :slight_smile:

Best overall is MX-4, spread, and performance.

Yeah have to disagree… MX4 is probably one of the worst offenders of failing over time. I’ve used them all- even the diamond pastes. They go on- anyways already explained in this thread :slight_smile:

I got some of that stuff comfy recommended… Silicone heat transfer compound. It’s metallic oxide and silicone oils- mine came open though they squeezed some out… Going to file a complaint but I still used some. I like the applications it’s used in though, maybe the best for flashlights

It seems legit at least. Says does not separate on the tube, but found there was oils in the top- but it was opened too so who knows. Using it on the threads between the sink and the pill

I’ve always just used Arctic Silver for CPUs. Sometimes boxed CPUs come with a little thermal paste, which I use when I build non-enthusiast machines where the people will probably never know or care as long as the machine doesn’t overheat.