Does heat paste needs changing?
- Thread starter YazzN9
- Start date
Many of the pastes use silicone oil. Silicone oil is somewhat volatile and tends to evaporate over time. I've seen many, many cases where the oil has disappeared completely. Some thermal conductivity remains even with the oil gone. Pastes with ester-based oil are more stable in terms of loss of the oil. Silicone oils are loathed and despised in some equipment because after evaporating from the hot parts they condense on things like switch and relay contacts and can cause serious problems (poor contact for low-level switches, burning to silicon dioxide (sand) on high-power contacts).
Most of the thermal conductivity comes from fine particles suspended in the oil. The particles might be zinc oxide, aluminum oxide ("alumina"), boron nitride (high performance but very expensive), other ceramics or even metals such as aluminum, though because metals are electrically conductive they aren't suitable for many applications. Most of the oxide or ceramic materials are white. Aluminum is silver-grey. I think there are a few around with colored pigments that contribute nothing to performance added. I've used thermally conductive adhesive that is blue. Many thermally conductive electrically insulating "washers" are grey, I don't know if the grey comes from the filler or the silicone rubber binder. Many of the gap-filling soft sheet materials are blue, and I'm pretty sure the blue is deliberately added. But don't judge by color. If you need high performance you must consult the datasheets.
Most of the thermal conductivity comes from fine particles suspended in the oil. The particles might be zinc oxide, aluminum oxide ("alumina"), boron nitride (high performance but very expensive), other ceramics or even metals such as aluminum, though because metals are electrically conductive they aren't suitable for many applications. Most of the oxide or ceramic materials are white. Aluminum is silver-grey. I think there are a few around with colored pigments that contribute nothing to performance added. I've used thermally conductive adhesive that is blue. Many thermally conductive electrically insulating "washers" are grey, I don't know if the grey comes from the filler or the silicone rubber binder. Many of the gap-filling soft sheet materials are blue, and I'm pretty sure the blue is deliberately added. But don't judge by color. If you need high performance you must consult the datasheets.
#12 at #5
I have never seen any data on this that I can recall. I suspect it would be hard to find because the research would take years to complete. It is the kind of thing that might be published by the IEEE. Unfortunately, accessing their pubs is expensive if you aren't a member and don't have institutional access.
I haven't looked for data for industrial type thermally conductive pastes for a long time, since I've used mostly the sheet type materials. I have used some interesting phase-change stuff that is sold as a rub-on stick - it melts and flows at some moderate temperature, rather like a wax. I believe it was an Aavid product. I don't remember what prompted me to try it (I've designed a lot of industrial power electronics, so I've experimented with a lot of weird stuff).
Most of the things I've seen that had lost a significant amount of oil have been power semiconductors that had been in service for several years - voltage regulators, thyristors, power transistors in audio amps or motor drives, etc. I've repaired things where I have strongly suspected that degradation of the thermal compound had resulted in component failure, but it's tough to be certain about that. I've also seen power semis that have melted their own solder joints and survived!
I may have some books or other ref material with at least a little info on this. I'll try to check in the next few days and come back here if I find anything interesting.
I have never seen any data on this that I can recall. I suspect it would be hard to find because the research would take years to complete. It is the kind of thing that might be published by the IEEE. Unfortunately, accessing their pubs is expensive if you aren't a member and don't have institutional access.
I haven't looked for data for industrial type thermally conductive pastes for a long time, since I've used mostly the sheet type materials. I have used some interesting phase-change stuff that is sold as a rub-on stick - it melts and flows at some moderate temperature, rather like a wax. I believe it was an Aavid product. I don't remember what prompted me to try it (I've designed a lot of industrial power electronics, so I've experimented with a lot of weird stuff).
Most of the things I've seen that had lost a significant amount of oil have been power semiconductors that had been in service for several years - voltage regulators, thyristors, power transistors in audio amps or motor drives, etc. I've repaired things where I have strongly suspected that degradation of the thermal compound had resulted in component failure, but it's tough to be certain about that. I've also seen power semis that have melted their own solder joints and survived!
I may have some books or other ref material with at least a little info on this. I'll try to check in the next few days and come back here if I find anything interesting.
hi,
I like 'ebp' have repaired/replaced semiconductor components that appear to have failed due to over heating.
The thermal paste 'dries' out over time and loses it thermal conductivity.
It is important when replacing semiconductors to clean the mating surfaces and apply fresh thermal compound.
E
I like 'ebp' have repaired/replaced semiconductor components that appear to have failed due to over heating.
The thermal paste 'dries' out over time and loses it thermal conductivity.
It is important when replacing semiconductors to clean the mating surfaces and apply fresh thermal compound.
E
Last edited:
It certainly dries out on Ignition modules on vehicles, & causes over heating of the module, First sign is engine stops when module over heats, & will start when cools down. If you don't replace the thermal paste, it will soon result in a module failure. Very common here in Australia with Ford Falcon models EA to EL.
recklessrog
- Joined May 23, 2013
- 985
I was involved in the development of re-breathing equipment, and to pass certification in California and India, (soon adopted universally) we could not use any silicon based thermal compound in the electronics because of it being volatile and evaporating. Conversely, clear silicon grease was allowed on non nitrile "O" rings and threads in the areas that only passed air. I think that a lot of the restrictions were because of the elevated O2 levels present in the system.
Also I think that I am right that after the Apollo 1 fire that killed the astronauts, a lot of research revealed many problems with components subjected to high levels of O2. (with enough pressure and temperature, anything will burn in 100% O2)
During the 90's we repaired many hundreds of power supplies used in plasma t.v's. A common problem was failure of switching power Fet's and shottky power diodes mounted on substantial heatsinks. One major manufacturer issued a a service bulletin that required that if any of the power Fet's failed during warranty, we had to remove all the others, clean off hardened silicon based thermal compound, replace the thermal pads and use a silicon free compound they supplied us with. It was in a large white tube with no markings so don't know what it was, but no t.v's that had this treatment came back with similar problems.
One point worth mentioning, when tightening the fixing screw that secures the Fet, which were mostly into threads in the aluminium heat sinks, it is wise to run the equipment for a soak test then re-tighten the screw, some settling takes place in the grease and insulating pad being squeezed, and some stretching of the aluminium heatsink threads which would otherwise leave the power Fet too loose.
Whilst on the subject, With power Fet's, thermal runaway is not supposed to happen as increasing temperature raises the Rds on resistance. BUT! you can get a sort of mechanical thermal runaway due to the way they are mounted to the heatsink. The leads are soldered directly through the pcb. The continual expansion and contraction pulls on the solder, weakening the joint until it becomes a poor contact with enough resistance to cause heating which travels up the leg into the device adding to the heat that has to be radiated away by the heatsink. End result is everything runs hotter, excess heat builds up in the local area, (Not good for electrolytic capacitors causing them to dry out and raise the ESR)
All this carries on until the joint starts to arc because it is passing high current, high voltage pulses, then BANG!! blown Fet's and other components
all caused by the initial inadequate mechanical design!
Also I think that I am right that after the Apollo 1 fire that killed the astronauts, a lot of research revealed many problems with components subjected to high levels of O2. (with enough pressure and temperature, anything will burn in 100% O2)
During the 90's we repaired many hundreds of power supplies used in plasma t.v's. A common problem was failure of switching power Fet's and shottky power diodes mounted on substantial heatsinks. One major manufacturer issued a a service bulletin that required that if any of the power Fet's failed during warranty, we had to remove all the others, clean off hardened silicon based thermal compound, replace the thermal pads and use a silicon free compound they supplied us with. It was in a large white tube with no markings so don't know what it was, but no t.v's that had this treatment came back with similar problems.
One point worth mentioning, when tightening the fixing screw that secures the Fet, which were mostly into threads in the aluminium heat sinks, it is wise to run the equipment for a soak test then re-tighten the screw, some settling takes place in the grease and insulating pad being squeezed, and some stretching of the aluminium heatsink threads which would otherwise leave the power Fet too loose.
Whilst on the subject, With power Fet's, thermal runaway is not supposed to happen as increasing temperature raises the Rds on resistance. BUT! you can get a sort of mechanical thermal runaway due to the way they are mounted to the heatsink. The leads are soldered directly through the pcb. The continual expansion and contraction pulls on the solder, weakening the joint until it becomes a poor contact with enough resistance to cause heating which travels up the leg into the device adding to the heat that has to be radiated away by the heatsink. End result is everything runs hotter, excess heat builds up in the local area, (Not good for electrolytic capacitors causing them to dry out and raise the ESR)
All this carries on until the joint starts to arc because it is passing high current, high voltage pulses, then BANG!! blown Fet's and other components
all caused by the initial inadequate mechanical design!
