The experts say it can be done!
Is there anyway to plate copper onto aluminum and if so how would one go about accomplishing this? I have aluminum heat sink extrusion but would like to solder the components onto the heat sink. Aluminum obviously can not be soldered to so i want to plate a thin layer of copper to it and solder to that. Is this possible or am i goign to end up with a multi-piece heat sink.
You could well have a bit of difficulty soldering onto a heatsink....
Make the heatsink out of copper?
The link mentions using chemicals you would not normally want to use (cyanide)
We could vacuum evaporate copper onto an aluminum plate.
Safer IMHO but you would need a vacuum evaporator.
If you have the tools laying around would be easy to build a simple evaporator.
Still easier & cheaper to make the heatsink out of copper.
My concern would be the different coefficients of thermal expansion. Attempting to solder the copper plating would result in strong shear forces at the Cu/Al interface sufficient to delaminate the plating locally. This in turn would greatly increase the thermal resistance (deg C/W) and render its heatsinking properties ineffective.
Good point Mariss, besides evaporating cu onto al would result in a very thin ~5k angstroms thick.
Very easily melted while soldering besides delaminating issues.
I go back to making the heat sink out of cu in the first place.
Or they sell metal paste that might work?
Not sure if you can solder to this thou.
You might want to put the electrical contacts into the paste?
Ideas just bouncing around now
Good luck with what you try.
If you are in S. California and would like to try the evaporation send us the part and $180.00 I will put some copper on it for you with no guarantee that it won't fall off
An open mind is a virtue...so long as all the common sense has not leaked out.
I would not be too worried about the state of the copper plating....what about the state of the component after it has been roasted by the Real BIG soldering iron for some considerable period?
Oh, I know...put a Real BIG heatsink on the component leads....
Keep in mind solder dissolves copper. A 5k angstrom Cu thickness may very well completely dissolve into the melted solder like sugar in water. My apologies to Dr. Pete; I misread him as the originator of this thread.
Heat will not damage the components if they are surface mount rated. Transistors are specified at 260C soldering temperatures; Pb63Sn37 (standard leaded solder) melts at 178C, Sn96Cu3.5Ag0.5 (lead-free solder) has a eutectic temperature of 228C. The common mistake is to use an undersized (too low Wattage) iron cranked up to a very high temperature (>300C) instead of using an adequate sized iron set to 210C and 260C respectively. A hot, undersized iron cannot heat things up quicker.
Keep in mind this is a heatsink so it's mission in life is to prevent you from doing what you have in mind; soldering to it. Find its rated thermal resistance, divide it into your delta-T. That's how many Watts of heat you will need. Example:
Thermal resistance = 0.5C / Watt
Delta-T = 235C this is 260C need to solder minus a 25C ambient.
Power = 470 Watts! (235C / 0.5C per Watt)
That is one hefty soldering iron.
Next, find the thermal time constant for the heatsink. Mutiply it by 3 to find how long it will take to reach 95% of the applied temperature. Example:
Say the thermal time constant is a very realistic 10 minutes. It will take 30 minutes to reach the desired temperature using your 470 Watt iron. That also makes you a very patient man hefting this substantial iron. Can't wait that long? A 1,500 Watt iron gets you there in 10 minutes, a 15kW iron gets it done in 1 minute.:-)
Put the heatsink on your kitchen hob, and get it up to say 260 Centigrade .
Quickly place your flux paste, component, and solder. Clamp the assemply and chuck it into a very large container of liquid nitrogen. Then pray, (while you are being engulfed by clouds of nitrogen gas that may asphyxiate you) that the thermal conditions experienced by the component do not fall beyond the top and bottom limits.
It might work...