Copper plating AIXIZ modules *GRAPHS ADDED*

[piferal]

Re: Copper plating AIXIZ modules *PICS ADDED

Nice work :beer:

The copper layer placed in the modules, is approximately as the same thickness as that of nickel removed?

 

[lasersbee]

Re: Copper plating AIXIZ modules?

^^ The idea here is to make an improvement of an existing product. Even if the
improvement is small, it better than no improvement at all..

So I plated some modules last night and will add a video shortly. The process to
copper plate is fast and simple. After playing around with it, I ended up finding
that it takes about 1min in the solution to produce a nice layer of copper over
the brass.

I first stripped the nickel plate off the modules with 30% sulfuric acid. This acid
was purchased from UAP NAPA auto parts supplier, but the same results can be
had with acid from a car battery..

I placed the modules in a plastic container and let sit for about an hour until
no nickel could be seen on the modules.

I then gave them a rinse in water mixed with baking soda to neutralize any
remaining acid.

For the copper solution I made use of used PCB etching fluid which was a nice
deep blue from all the copper I had once removed from copper clad when making
circuit boards. A great way to recycle this fluid. I have to find the bottle for
the acid that was used in this, but im sure any etching fluid saturated with
copper will do the trick. Keep in mind there are other ways to copper plate,
any of which would also work well here.

One thing to keep in mind, if you use an acid solution holding the copper,
immediately submerse the newly plated part in water mixed with baking soda
to neutralize and remaining acid on the part, then wipe dry with a clean cloth
and dry.

The power supply used was an old PC P/S running off the 3v rail.. Here are
a couple of pics of the bare brass after stripping the nickel, and beside
the newly plated copper one. I started with old modules that had dead diodes
pulled out, this is where the marks in the metal came from...

Will be making a quick tutorial video for the process. Very simple to do, taking
about 1min per module to apply the plate..

Good job WB.... I like the fact that you explained the process..

When I saw the change to the Thread Title I was expecting to
see the pics on the 1st Post...


Jerry

 

[JLSE]

Re: Copper plating AIXIZ modules?

Wow. I dont know if that will do anything, but those look great

+1 if I can

Im going to try and test this to see what results I get. Im thinking of using a brass
mount, test a LD, then copper plate and re test the same LD. The idea is to look
at the output graph and operating temp of the LD itself..

Nice work :beer:

The copper layer placed in the modules, is approximately as the same thickness as that of nickel removed?

Id say its close if not more.. The nickel plate on these things come off pretty easily
with light sand paper, some modules are better coated than others so its hard to
say for sure..

Good job WB.... I like the fact that you explained the process..

When I saw the change to the Thread Title I was expecting to
see the pics on the 1st Post...


Jerry

Done

I will add the video I took to the first post as well.. I was going to try for silver
plating, but after some reading it turns out that silver plate requires nasty chems
like that of gold plating.. So I stuck with copper... for now :eg:

 

[Benm]

Re: Copper plating AIXIZ modules?

Cu, Al, and brass all offer excellent conduction. It's JUNCTIONS that need to be eliminated.
...
In real-world scenarios, the choice of copper over brass or copper over aluminum will not provide nearly the improvement that most of you seem to expect of it.

I would totally agree here. If you do the math for heatsinking, the thermal conductivity of any of these metals is not likely to be a significant factor. There are two things that need to be overcome, in this order:

- thermal resistance from ambient to heatsink
- thermal resistances between the various parts (diode->module->heatsink).

For the first one the sole solution is using a bigger heatsink (or forced airflow), and for the second you will probably resort to thermally conductive pastes and such.

The only way plating these modules would be of any use is when the plating is softer than the base metal was. This would help improve contact between surfaces under mechanical pressure, and i can see how it could work between diode and module.

 

[JLSE]

Re: Copper plating AIXIZ modules?

I would totally agree here. If you do the math for heatsinking, the thermal conductivity of any of these metals is not likely to be a significant factor. There are two things that need to be overcome, in this order:

- thermal resistance from ambient to heatsink
- thermal resistances between the various parts (diode->module->heatsink).

For the first one the sole solution is using a bigger heatsink (or forced airflow), and for the second you will probably resort to thermally conductive pastes and such.

The only way plating these modules would be of any use is when the plating is softer than the base metal was. This would help improve contact between surfaces under mechanical pressure, and i can see how it could work between diode and module.

So what would be the reasoning of gold plated copper diodes? I have only seen
a few non plated cmounts in CNI lasers, but everything else down to 5mW aixiz
supplied LD's have gold plate?

If it were for corrosion resistance, there are other metals that could do the same job
at far less of a cost.

 

[Cyparagon]

Re: Copper plating AIXIZ modules?

If it were for corrosion resistance, there are other metals that could do the same job
at far less of a cost.

The layer is so thin, and the surface area is so small, it is trivial when compared to the cost of the "guts."

See? It's cheap.

Why gold? Google says "The use of gold in electronics is based on the combination of its electrical conductivity, its ductility and its total freedom from corrosion or tarnishing at either high or low temperatures. Its near perfect corrosion resistance means gold provides an atomically clean metal surface which has an electrical contact resistance close to zero, while its high thermal conductivity ensures rapid dissipation of heat when gold is used for contacts. "

 

[Benm]

Re: Copper plating AIXIZ modules *PICS ADDED

Indeed, the amount of gold is so tiny it doesnt really affect the product cost.

The downside of copper is that it -will- oxidize over time, leaving you with a relatively bad thermal conductor on the surface. They gold plate the leads on these diodes for the same reason: copper would work fine, until it forms a thick oxide layer and becomes difficult to solder.

These gold platings are micrometers thin and micro/nanograms in weight. A gram of gold may cost $20 or so, but even applying a milligram to something like the feet of a processor costs only a few cents for the gold.

 

[Things]

Re: Copper plating AIXIZ modules *PICS ADDED

I also believe gold is a lot easier to solder the die to with indium solder. May be wrong here.

The bottleneck we have is the thermal transfer between the laser die and it's mount. Maybe if they designed laser diodes with contact pads on all sides.

But, even then our other main issue with high currents is COD, which no amount of heatsinking will fix either.

 

[JLSE]

Re: Copper plating AIXIZ modules *PICS ADDED

I also believe gold is a lot easier to solder the die to with indium solder. May be wrong here.

The bottleneck we have is the thermal transfer between the laser die and it's mount. Maybe if they designed laser diodes with contact pads on all sides.

But, even then our other main issue with high currents is COD, which no amount of heatsinking will fix either.

Agreed, COD if im not mistaken is the result of optical flux within the cavity. At
least this is what I remember with the PHR's for example.. So its not a heat issue.

One thing stands out and that is an LD will put out more power the faster
you can pull the heat off it.. Better overall contact is the goal even if
you cant put any more current through it safely..

Funny you mentioned the relation to indium and gold, had not thought of that
possibility

Mentioned above:

What if you plated just the back of an aixis module, with the diode already in it?

Im not sure if it would work like this but Im picturing the copper (or whatever you use) filing in the cracks between the diode/module fusing them together and maybe making a flat seamless surface on the back between the diode/module.

I think this idea may be on to something. There is a large amount of real-estate
on the back of the diode not in contact with anything. Im going to try a few
tests and see how it graphs with all things equal,

What I have in mind is, same heatsink (AL) same diode, same current 1A, but
after the first test of bare brass, I will plate in copper, then plate with the diode
in place.

Im curious to see if it affects how many mW are lost in the first 30 seconds
in each condition.

I would like to measure temp, but will have to figure something out for that.
The LD mount I have in mind is not a press fit, but has a threaded collar,
and is all brass.. Unfortunately I wont be able to compare to a nickel plating
but it should give an indication to weather or not there is any difference,
slight or not....

 

[Seoul_lasers]

Re: Copper plating AIXIZ modules *PICS ADDED

:gj: Whatever.. the long and short of it... you will have increased the modules conductivity and any help to get heat away from the diode quickly helps with overall efficiency of thermal transfer from the diode.
It might be in 1% range... Good idea.

Keep experimenting, and use an IR thermometer to see what difference it makes.

I wonder what a pure aixiz copper housing might do in an aluminum sink?

 

[JLSE]

Re: Copper plating AIXIZ modules *PICS ADDED

Well the results for the following tests are in

At 1A I tested the same diode in the same mount, same position and distance
from the sensor, and the lens focused to the same point...

Though I saw an odd slope is the plated tests, im seeing around 100mW more
when using copper plated brass with arctic silver paste over bare brass
with the same paste. I also graphed dry runs where there was no paste at
all..

Next test is going to be a diode mounted in a copper plated aixiz and
after the first test im going to plate the rear of the module with the diode
still in the mount...

I still have to try for the nickel vs copper plate, but will have to make a
removal tool so the diode doesnt get damage when removing.

Either way the results do seem to show an improvement and 100mW
is nothing to sneeze at..

BARE BRASS

BRASS + SILVER PASTE

COPPER PLATE NO PASTE

COPPER PLATE + SILVER PASTE

 

[Isoleucine]

I wish to add to the forums...

High Quality Sulfuric Acid in excess of 18M in some cases, may be purchases as drain cleaner. At home improvement stores, these acids are so strong, they must be placed in plastic bottles, and then in plastic bags.

 

[JLSE]

Another update, mounted the same LD in a copper plated aixiz with arctic
silver paste. Took another 45sec graph as well as after plating the base with
the LD already mounted..

The plating didnt take as well as I didnt give it a soak in sulfuric acid
prior to the plating and only cleaned away the excess arctic silver with
99% alcohol. I will have to repeat this one again on a cleaner diode
which should still give the difference..

I noticed that the aixiz mount was slightly lower in output @1A than in
the brass mount used in the fisrt tests. Im assuming that the first
mount held the LD via threaded ring applying pressure and more importantly
contact to the rear of the LD where in an aixiz mount there is no contact.

Though this didnt account for much of a difference between the 2 mounts
it may be attributed to different grades of brass. Hopefully the next plating
to the rear of an LD in the aixiz will show more definitive difference..

There was only a 10mW difference in favor to the plated LD base on this run..
Yet a difference none the less. Same conditions again of focus and positioning
in relation to the LPM sensor and orientation in the AL heatsink.


COPPER PLATED AIXIZ + SILVER PASTE

SAME COPPER PLATED AIXIZ + SILVER PASTE + PLATED LD BACK

 

[Benm]

Interesting stuff!

Its unfortunate that you are not able to measure the temperature of the laser diode though, just the power output doesnt give that much information about it. Considering the curves "COPPER PLATE NO PASTE" and "COPPER PLATE + SILVER PASTE" i wonder what is actually going on there. I dont see how the dip can be explained.

There are indirect methods of measuring the diode temperature though: You can look at the wavelength if you have a spectrometer, but even if you don't, you can look at the forward voltage.

The forward voltage should come down in the order of a mV for every degree increase. You will require a very precise current source to use this property, as the forward voltage will change more than a few mV if your current source is off by a small amount.

 

[Somedude]

Thank you , i was just getting into this myself and you saved me alot of time.:wave:

 

[benmwv]

Cool :gj:

I have something else you might test, if you don't mind.

Whenever I build a high power build I usually solder on the wires then I put a blob of thermal epoxy (arctic alumina) on the back of the module. My thinking was that it at the very least holds the pins secure, and hopefully helps it dissipate heat better since it contacts the large open area on the back of the diode.

Also I was thinking that on my next build I would take the back barrel and cut it off maybe 1cm past the threads then screw that little ring onto the module (with thermal paste in the threads). That should give it better heat dissipation because in a module without a back barrel the threaded part (which is in line with the diode and is on the main heatsinking area of the module) never touches the heatsink so the heat has to travel up into the thin walls of the focusing part of the module to transfer to the heatsink.

I think if you do both of those things it should give much better heat dissipation.