The durability of thermal pads :D

[bluevd]

This made me laugh quite a bit. Figured someone might also find it funny.
Original source: http://www.tglobalthermal.com/faq-durability-of-thermal-pads.php

[sasa]

I've never seen a single silicone based thermal pad/grease/PCM that don't age. Usually, silicone oil will seep out and the thermal material will become flaky and crusty. Even silicone polymers (curable encapsulants) age over time at elevated temperature, let along silicone oil.

I have actually intended to open similar thread about. I have actually finished complete teardown of my 10 years old 360W Chieftec PC PSU and I have actually found dozen of them attached on power transistors between two large heat sinks.

Interestingly, they are all in quite good condition, I would say excellent - not a single one is decomposed even I bend it in half... It is just behaves and feel as fresh tin rubber...

Between transistor thermal area and pad was actually thermal grease, still behaves as grease.

Is it really possible to last that long on quite high temperature as in old PSUs may be?

[bluevd]

Usually, silicone oil will seep out and the thermal material will become flaky and crusty.

Interestingly, they are all in quite good condition, I would say excellent - not a single one is decomposed even I bend it in half... It is just behaves and feel as fresh tin rubber...
...
Is it really possible to last that long on quite high temperature as in old PSUs may be?

Actually, the temperature is not the biggest concern regarding aging. However, air and contaminants are! That's why a good surface contact for newer pads (from ~2000 onward) ensures high stability over time. I've been researching the manufacturing technology and was quite surprised in the quick evolution from 2000 till 2002 (nearly 170% increase in breakdown temperature resistance for some types), very low chemical reactivity, thermal conductivity, etc. As to the grease, I've seen mixed recommendations from manufacturers: some say it's a big no-no, others say it helps with sealing against air. My conclusion however, after about 7 months of research, is that the recommendations for pads is like switch debouncing: varies from source to source and it's hard to find accurate, repeatable measurements that would define proper installation instructions (though in 90% of the cases, any of the mounting techniques work adequately).

[cdev]

I've had small aluminum heat sinks on thermal pads on a Conextant BT878A video capture card migrate downward in a snail-like manner under high heat conditions. They did not fall off but given that this was inside a computer - had one fallen off and then caused a short the result could have been an equipment failure or even a fire.

[Ian.M]

That's fairly obviously an inappropriate adhesive used on the pads, or self-adhesive pads used outside their recommended application. e.g. the adhesive is intended as an assembly aid only and the heatsink must still be mechanically fastened.   

Any cheap-ass machine shop can mill small heatsinks out of aluminium scrap left over from another operation, stick self adhesive thermal pads on them and sell them for a considerable markup, but the odds are they'll slap the cheapest possible thermal pad on there - if they don't simply use thin double sided tape!.   An extrusion specialist that produces heatsink profiles *SHOULD* know better but it may be reject profiles with the good bits chopped up on a ghost shift, and the pads pure black market chinesium.

[wraper]

I've had small aluminum heat sinks on thermal pads on a Conextant BT878A video capture card migrate downward in a snail-like manner under high heat conditions. They did not fall off but given that this was inside a computer - had one fallen off and then caused a short the result could have been an equipment failure or even a fire.

Those are not silicone thermal pads. Also don't confuse normal thermal pads with adhesive pads.

I've never seen a single silicone based thermal pad/grease/PCM that don't age. Usually, silicone oil will seep out and the thermal material will become flaky and crusty. Even silicone polymers (curable encapsulants) age over time at elevated temperature, let along silicone oil.

I think you confuse silicone thermal pads with silicone-free type (higher operating temperature and better thermal conductivity) which is not elastic/reusable. Also silicone thermal pads are very different. Some are very strong, other are very easy to damage even when new.

[SeanB]

To stick those small heatsinks on i just use 2 part epoxy, a thin layer on a nicely sanded base of the heatsink ( 220 grit waterpaper on a good flat surface, rinsed well afterwards with alcohol) and then a nicely cleaned and sanded top of IC is as good in most cases as any thermal epoxy. My go to is Pratley Quickset Steel, though Araldite Steel is just as good, though it takes 15 minutes to gel instead of 3. A small amount ( match head size) in a thin layer works well.