Four years of work have resulted in the creation of a material that can fill the same role as conventional solders, but with twice the thermal conductance of current state-of-the-art TIMs. https://m.phys.org/news/2018-08-supersolder-unparalleled-thermal-properties.html
LPF Donation via Stripe | LPF Donation - Other Methods
Links below open in new window
ArcticMyst Security by Avery
Super solder
- Thread starter steve001
- Start date
- Joined
- Jul 4, 2008
- Messages
- 2,499
- Points
- 113
Wow! looks like an interesting material. I'll have to see if they are up for getting out some samples soon. 
- Joined
- Sep 20, 2013
- Messages
- 17,387
- Points
- 113
Huh. That's very interesting. I expect this will be used inside of devices instead of being a solder one would use on a circuit board. It must have a very high melting point or it wouldn't be useful.
kecked
0
- Joined
- Jun 18, 2012
- Messages
- 948
- Points
- 63
I wonder if it solved the tin whisker issue
https://www.maximintegrated.com/en/app-notes/index.mvp/id/5250
in case one is interested
https://www.maximintegrated.com/en/app-notes/index.mvp/id/5250
in case one is interested
Last edited:
- Joined
- Sep 20, 2008
- Messages
- 17,622
- Points
- 113
Would be the ideal solder to use on TEC P/N
assemblies.
Jerry
assemblies.
Jerry
Last edited:
- Joined
- Jun 12, 2015
- Messages
- 7,609
- Points
- 113
Interesting stuff. That was a good read. :thanks:
Yep, better TEC's could be a possible application.
Yep, better TEC's could be a possible application.
Benm
0
- Joined
- Aug 16, 2007
- Messages
- 7,896
- Points
- 113
I think it would be a very nice application thing though.
Really, in what situations is the thermal conductivity of the solder a -limiting- factor in practical applications?
Obviously it's useful to have thermally conductive solder when you roll an SMD board and use surface area as a heatsink, but that tiny thin layer of solder probably will not be your limiting factor - the size you have on the pcb is far more likely to be limiting.
From the article:
Electrically insulating solder? I wonder what application that would have really.
Solder mosfet onto board, find drain flange electrically insulated... so routing a thin trace to the middle pin of the device to fix that, which dissipates more power than the short part through normal solder from the flange did to begin with.
Really, in what situations is the thermal conductivity of the solder a -limiting- factor in practical applications?
Obviously it's useful to have thermally conductive solder when you roll an SMD board and use surface area as a heatsink, but that tiny thin layer of solder probably will not be your limiting factor - the size you have on the pcb is far more likely to be limiting.
From the article:
Electrically insulating solder? I wonder what application that would have really.
Solder mosfet onto board, find drain flange electrically insulated... so routing a thin trace to the middle pin of the device to fix that, which dissipates more power than the short part through normal solder from the flange did to begin with.
- Joined
- Sep 20, 2008
- Messages
- 17,622
- Points
- 113
I can see a possible faster response time when
using a TEC in the Seebeck Effect configuration.
Jerry
Benm
0
- Joined
- Aug 16, 2007
- Messages
- 7,896
- Points
- 113
I doubt that could be significant. Surely (normal tin/lead or leadfree) solder is a pretty bad thermal conductor, but as long as you don't bridge very large gaps with it, it doesn't have that big an effect.
If you had, for example, a TO-220 package component and a copper heatsink, i think the resistance would be lower if you soldered it to the heatsink compared to mounting it with a bolt, nut and paste.
With a bigass heatsink that would be very impractical, but with a small heatsink or when using PCB area for heatsinking it's quite viable.
I've also seen it done in older equipment with metal can transistors (like 2n2222 in to-39 packaging) where they just soldered some flanges/lips on for extra heatsinking.
I doubt this practice keeps the transistor within it's absolute maximum thermal ratings, but then again, if it works, it is even done in mass produced items. It was fairly common to do stuff like that in radio electronics too - you'd solder a emitter-case-connected transistor like a mrf237 straight to the PCB or even metal enclosure
If you had, for example, a TO-220 package component and a copper heatsink, i think the resistance would be lower if you soldered it to the heatsink compared to mounting it with a bolt, nut and paste.
With a bigass heatsink that would be very impractical, but with a small heatsink or when using PCB area for heatsinking it's quite viable.
I've also seen it done in older equipment with metal can transistors (like 2n2222 in to-39 packaging) where they just soldered some flanges/lips on for extra heatsinking.
I doubt this practice keeps the transistor within it's absolute maximum thermal ratings, but then again, if it works, it is even done in mass produced items. It was fairly common to do stuff like that in radio electronics too - you'd solder a emitter-case-connected transistor like a mrf237 straight to the PCB or even metal enclosure
- Joined
- Sep 12, 2007
- Messages
- 9,399
- Points
- 113
Increasing the copper thickness of the PCB would be a far cheaper, simpler, and probably more effective way to increase thermal performance.
Benm
0
- Joined
- Aug 16, 2007
- Messages
- 7,896
- Points
- 113
Well, not for continuous power really. When using PCB area as a heatsink the major issue is getting heat from the board into the ambient.
Going for double thickness copper does help a bit in spreading it, but does nothing for the board-to-ambient thermal resistance. What they sometimes do is have heatsinking area on both sides of a board with a shitload of via's punched through so you get effectively double the surface area, which is fine if you have to board space to do that.
For short peaks having thicker copper does help by adding thermal mass, and in that case you may want to have that solder coated as well - it doesn't add that much but is essentially 'free' when having the pcb produced by most factories.
Going for double thickness copper does help a bit in spreading it, but does nothing for the board-to-ambient thermal resistance. What they sometimes do is have heatsinking area on both sides of a board with a shitload of via's punched through so you get effectively double the surface area, which is fine if you have to board space to do that.
For short peaks having thicker copper does help by adding thermal mass, and in that case you may want to have that solder coated as well - it doesn't add that much but is essentially 'free' when having the pcb produced by most factories.
- Joined
- Sep 20, 2013
- Messages
- 17,387
- Points
- 113
You can still sink the board to your module, if it has a place for it. Or, to another part of the host. Seems to work well for some of the drivers we use here.
- Joined
- Sep 12, 2007
- Messages
- 9,399
- Points
- 113
And you think changing the solder type will fix all that?
He thinks this proprietary copper nanotube solder will be "essentially free". That's cute.
Benm
0
- Joined
- Aug 16, 2007
- Messages
- 7,896
- Points
- 113
I think you're confusing things here:
If you do a design and need some heatsinking, it sometimes makes sense to just have a large copper pad on the PCB for that. This does not involve any additional material, it's essentially the same as filling the PCB with a ground plane in terms of manufacture.
Putting some via's through to use both sides may add a bit to production cost, but that's usually nothing major.
Also leaving the copper on the PCB actually can reduce cost for the manufacturer: any copper that says on the board doesn't have to be etched away saving on the chemicals required to do that etching.
If you have any doubts you could always request a quote for a board design, and if it has large copper areas on it it will usually not cost more. Exceeding a certain number of via's for a given board side might though, really depends on the manufacturer.
If you do a design and need some heatsinking, it sometimes makes sense to just have a large copper pad on the PCB for that. This does not involve any additional material, it's essentially the same as filling the PCB with a ground plane in terms of manufacture.
Putting some via's through to use both sides may add a bit to production cost, but that's usually nothing major.
Also leaving the copper on the PCB actually can reduce cost for the manufacturer: any copper that says on the board doesn't have to be etched away saving on the chemicals required to do that etching.
If you have any doubts you could always request a quote for a board design, and if it has large copper areas on it it will usually not cost more. Exceeding a certain number of via's for a given board side might though, really depends on the manufacturer.
- Joined
- Sep 12, 2007
- Messages
- 9,399
- Points
- 113
So you're equivocating larger copper pads with heat sink fins? That has some merit, but you can't always rob surface area from the rest of the circuit. Ground planes and signal traces need some love, too. The point is this magical wanky solder claims to have higher conductivity. And assuming I take them at their word, increasing the copper thickness INSTEAD would also increase conductivity with many more benefits and fewer drawbacks.
Benm
0
- Joined
- Aug 16, 2007
- Messages
- 7,896
- Points
- 113
Using copper planes isn't a replacement for heatsinks in most cases, but in ones where are you are just on the border of design limits it can be a cheap way out.
As for the effect of increasing the copper thickness i have my doubts: For short power bursts it would help, but for long term heatsinking the effect would be pretty small.
Most PCB material uses 1 oz/ft2 or 1.4 mil thickness copper traces. You can have it manufactured in 2, 3 or 4 times that thickness, but the area remains basically the same. Ramming the via's through only serves to use both sides of a board for heatsinking.
A downside of using thicker copper layers is you can't do small details any more due to under-etching and such, so the 1 oz/ft2 has become more or less the default.
If you really want to get a lot of heat out through the board nowadays you'd probably opt for aluminium-core boards that transfer heat from things like led's to the case very well. Downside of that is that they conduct heat so well it becomes difficult to solder onto them
As for the effect of increasing the copper thickness i have my doubts: For short power bursts it would help, but for long term heatsinking the effect would be pretty small.
Most PCB material uses 1 oz/ft2 or 1.4 mil thickness copper traces. You can have it manufactured in 2, 3 or 4 times that thickness, but the area remains basically the same. Ramming the via's through only serves to use both sides of a board for heatsinking.
A downside of using thicker copper layers is you can't do small details any more due to under-etching and such, so the 1 oz/ft2 has become more or less the default.
If you really want to get a lot of heat out through the board nowadays you'd probably opt for aluminium-core boards that transfer heat from things like led's to the case very well. Downside of that is that they conduct heat so well it becomes difficult to solder onto them
