For glueing heatsinks I use this stuff with very good results:
Arctic Alumina Thermal Adhesive
Please note that removing a heatsink after glueing will mean that the chip comes off together with the heatsink ! Anyway, who wants to remove the heatsink of "throw away" electronics ? There are other thermal glues out there and I have tried quite a few. They have a few things in common: they are all expensive, take ages to cure after having been stored for a while and they're often very poisonous. This Arctic Silver Alumina is relatively cheap, mechanically strong and it cures fast (also after having been stored for a while).
Arctic Alumina Thermal Adhesive
Please note that removing a heatsink after glueing will mean that the chip comes off together with the heatsink ! Anyway, who wants to remove the heatsink of "throw away" electronics ? There are other thermal glues out there and I have tried quite a few. They have a few things in common: they are all expensive, take ages to cure after having been stored for a while and they're often very poisonous. This Arctic Silver Alumina is relatively cheap, mechanically strong and it cures fast (also after having been stored for a while).
Last edited:
An american member once sent me a few variations of JB Weld. It is very good stuff for its purposes and it is not expensive but it is less good for this specific purpose compared to Alumina. JB Weld is only used because it is cheap and it is available at many stores. There is texture that makes heat transfer less optimal. People that choose to use this epoxy only have success if they use it together with heat transfer compound which makes applying a specialty job.
A real thermal glue meant for heatsinks has a very fine texture and fills the gaps between heatsink and heat emitting device. As always some use whatever they have at hand "because they can" and claim that it works perfect. Some even claim superglue (cyanoacrylate) also is good... However, in the world of thermal glues some are more perfect than others 😉
A real thermal glue meant for heatsinks has a very fine texture and fills the gaps between heatsink and heat emitting device. As always some use whatever they have at hand "because they can" and claim that it works perfect. Some even claim superglue (cyanoacrylate) also is good... However, in the world of thermal glues some are more perfect than others 😉
Last edited:
Your shown alumina thermal clue also isn't "perfect" in heat transfer. They even don't show any parameters on the product page for calculations. As this is based on oxide, heat transfer is less effective.
With lapped down / polished surfaces, performance of heat transfer paste/compound is almost negligible.
With lapped down / polished surfaces, performance of heat transfer paste/compound is almost negligible.
It was a reference to "Animal Farm" so a metaphor. As an engineer one knows nothing is perfect as that is here lesson number 1. I should have used the word "equal" but it was too late to edit.
I use the stuff to apply heatsinks on various chips and it works better than JB Weld (even with no published numbers). That was the message. In the stuff I encounter I mostly see heatsinks that are glued with very thick layer of glue, heatsinks that were not applied with some pressure, heatsinks that are mounted with adhesive film (worst there is), heatsink that are glued with RTV etc etc. Compared to this all Alumina works way better. In the world of number fetishists there will be a better stuff but it probably is not obtainable by normal human beings, extremely hard to find in small packages, very expensive or it is poisonous as hell. This Alumina costs less than 10 Euro and performs good. Please compare to, for instance, Fischer WLK5 (which is delivered in a chemical resistant box when ordered at Farnell !).
If some like to lap and polish DIP and BGA heatsinks I wish them a lot of fun 🙂
I use the stuff to apply heatsinks on various chips and it works better than JB Weld (even with no published numbers). That was the message. In the stuff I encounter I mostly see heatsinks that are glued with very thick layer of glue, heatsinks that were not applied with some pressure, heatsinks that are mounted with adhesive film (worst there is), heatsink that are glued with RTV etc etc. Compared to this all Alumina works way better. In the world of number fetishists there will be a better stuff but it probably is not obtainable by normal human beings, extremely hard to find in small packages, very expensive or it is poisonous as hell. This Alumina costs less than 10 Euro and performs good. Please compare to, for instance, Fischer WLK5 (which is delivered in a chemical resistant box when ordered at Farnell !).
If some like to lap and polish DIP and BGA heatsinks I wish them a lot of fun 🙂
Last edited:
What do you think ? I work with thermal glues for at least 20 years. I only briefly tested JB Weld in 2 versions. Good but not as good as the specialty stuff.
Why are we debating the obvious ? Is it that important ? We are dealing with "throw away" boards of a few Euro and most of them lack a heatsink or it is not applied like it should. For less than 10 Euro one can re-aply the same or a large heatsink and so with 20 or way more boards. With a DIY board it can also be used if one already has the stuff for these purposes.
Why are we debating the obvious ? Is it that important ? We are dealing with "throw away" boards of a few Euro and most of them lack a heatsink or it is not applied like it should. For less than 10 Euro one can re-aply the same or a large heatsink and so with 20 or way more boards. With a DIY board it can also be used if one already has the stuff for these purposes.
Last edited:
Sorry for the OT..
I'm contributing all the time, didn't noticed? Are you in a bad mood today?
It was just a simple question. 😵
Last edited:
Mood is just fine. Good morning my friend !
I just give an alternative for a cheap and cheerful solution for applying heatsinks a good way for DIYers. Or "well enough" if you like, compared to other solutions. I always use my tools to verify if what I did was effective. The stuff does what it is supposed to do for a low price. I have used pro stuff but toxicity is awful with those. I would not advise to eat thermal glue but I have used thermal compounds that have strong cancer inducing content while this seems a relatively safe glue so more suitable for the audio DIYer.
http://www.arcticsilver.com/PDF/AATA_SDS.pdf
I just give an alternative for a cheap and cheerful solution for applying heatsinks a good way for DIYers. Or "well enough" if you like, compared to other solutions. I always use my tools to verify if what I did was effective. The stuff does what it is supposed to do for a low price. I have used pro stuff but toxicity is awful with those. I would not advise to eat thermal glue but I have used thermal compounds that have strong cancer inducing content while this seems a relatively safe glue so more suitable for the audio DIYer.
http://www.arcticsilver.com/PDF/AATA_SDS.pdf
Last edited:
Cheers! 🙂
I use "amasan T12" from "Jürgen Armack GmbH".
http://www.uni-kl.de/elektronik-lager/488201
I use "amasan T12" from "Jürgen Armack GmbH".
http://www.uni-kl.de/elektronik-lager/488201
Entirely different product ! That is a heat transfer compound and not a glue. Fine if one has a CPU that needs a layer between removable heatsink and CPU.
There are quite a few of those and some are better than others. I haven't tried Amasan T12. The manufacturer is not throwing around useful numbers as well 😉 Most encountered misuse in general that I see is that too much of it is used. However if one has chips or CPU's that lack the possibility to clamp a heatsink a glue is a good (albeit definitive) solution. In all other cases I would indeed use a removable heatsink with a thin layer of good thermal compound. No doubt.
There are quite a few of those and some are better than others. I haven't tried Amasan T12. The manufacturer is not throwing around useful numbers as well 😉 Most encountered misuse in general that I see is that too much of it is used. However if one has chips or CPU's that lack the possibility to clamp a heatsink a glue is a good (albeit definitive) solution. In all other cases I would indeed use a removable heatsink with a thin layer of good thermal compound. No doubt.
Last edited:
Oh, i missed that you asked for clue - sorry. I don't use glue at all. I have some high performance sticky thermal pads i use. The heatsink on your TPA3116+BT module is attached with it.
Was attached by it.... Many adhesive thermal pads are in the shadow of both a clamped heatsink with good thermal compound or a heatsink attached with good thermal glue. Often their thickness is what hampers a more optimal heat transfer. You see I have a clue about glue 😀
Last edited:
LMAO.
To be fair, what DoctorMord said was correct. The CPU guys lap the chip and the heat sink for a tight fit. Then apply weight (e.g., 30 lbs.) to insure the JB Weld is as thin as possible while curing.
Since this is a 2nd source of heat sinking for the TPA3118 (the solder pad being 1st), I would think it is an acceptable acceptable solution, especially for a one-off need.
If I had a routine need for heat sink glue, I would probably procure the Arctic Alumina based on your recommendation.
Last edited:
A thin layer and some weight are always necessary either with JB Weld or Alumina. The pc guys that overclock go in great lengths to get the maximum out of their overclocked CPUs which is kind of Formula 1 performance work. They use stuff out of spec and even overvolt the CPU so every possible gain is needed otherwise the results are not good and the pc will be unstable. CPU's have some more Watts to dissipate as well, a fan or water cooling are almost mandatory to get the most out of the overclocking. Glueing a CPU of a pc is somewhat too definitive IMO as it can never be removed again. With Class D power efficient amplifiers and media players and the like the situation is somewhat different. With cheap Ebay boards that cost a few Euro the manufacturer throws the smallest heatsink he can get his fingers on attached with some RTV on the chip. Often just some slight force is enough to get the heatsink off. It can be a nice improvement to add a larger heatsink with a good thermal glue in that case.
Very good cleaning of both surfaces and leaving the device with the weight for enough time to get the epoxy fully cured is a recommendation too. Dust, a hair or fingerprints are enough to deteriorate the heat transfer.
* AFAIK JB Weld is slightly conductive which makes matters somewhat difficult too. I find it hard enough with non conductive glue to apply it without it oozing between pins of IC's. If it happens I remove all glue and clean everything and start over again. Cotton sticks leave off small particles which makes the situation not optimal.
Very good cleaning of both surfaces and leaving the device with the weight for enough time to get the epoxy fully cured is a recommendation too. Dust, a hair or fingerprints are enough to deteriorate the heat transfer.
* AFAIK JB Weld is slightly conductive which makes matters somewhat difficult too. I find it hard enough with non conductive glue to apply it without it oozing between pins of IC's. If it happens I remove all glue and clean everything and start over again. Cotton sticks leave off small particles which makes the situation not optimal.
Last edited:
What concerns would I have about JB Weld "oozing out" and making contact with the chip's leads?
I have noticed batch inconstancy on JB weld but its ok overall.
I would not glue a heatsink anyway.
From my experience with the class AB chip amps, the thermal issue is instantaneous Tc increase on for example a high dynamic range fast crescendo.
Aluminum seems to be 'slow' as I saw over T shutdowns with it.
I like a first layer copper spreader because is 'fast' then second slow layer any material.
On the 3118 with a copper spreader on power pad as shown in previous pic then pcb final sink everything is peaches.
No experience yet with 3116.
.
I would not glue a heatsink anyway.
From my experience with the class AB chip amps, the thermal issue is instantaneous Tc increase on for example a high dynamic range fast crescendo.
Aluminum seems to be 'slow' as I saw over T shutdowns with it.
I like a first layer copper spreader because is 'fast' then second slow layer any material.
On the 3118 with a copper spreader on power pad as shown in previous pic then pcb final sink everything is peaches.
No experience yet with 3116.
.
Sorry sir, i'm not a native english speaker..
But i'm pretty sure, "with it" is right, because:
http://forum.wordreference.com/threads/attached-in-to-with.2934398/#post-14833440
Last edited: