So feel free to call me crazy or what have you… but today I was thinking about how good the “old” XM-L2s were and dreaming about a gen3 with 3 bond wires again (or more J) )
I know in the past we’ve mentioned solder reinforced ideas but that ended fairly quickly.
My thoughts are as follows…
Dedome XML2
Seal phosphor with LEDseal or equivalent
Silver solder(for fear of unknown with Pb in this scenario) on a single wire to the bond wire side of the bare chip
Proceed submerging in plating bath and what not…
Anyway I’m thinking about trying it. I’ll mix up some plating solution at work this week and then try some madness next weekend. Since I don’t have a lumen measuring setup yet, my thoughts on determining this experiment a failure or success with hinge on A) does the LED still work? B) are there any gross side effects that make it not useable? C) did the Vf drop from unplated state and a series of current levels? D) can it take 10A (max on my power supply)? :evil:
It is doubtful that silver plating would have any impact on Vf or maximium current before failure. We are talking about DC. There is no skin effect, so what counts is the conductivity of the bond wire, and that is a function of the material, and circular area of the conductor. The relationship between the resistance and circular area of the conductor is linear. It is unlikely that you could plate enough silver onto the bond wire to materially the area, and thus alter its resistance. If you could drive the LED with 100Mhz AC, then it would make a significant difference. The skin depth at 100mhz in copper is literally a few microns, so even thin silver plating has a significant impact. (It is common for microwave and high power VHF/UHF transmitting tubes to have silver plated anode connections for just that reason). Silver also is very soft, so it will contribute very little to the physical strength.
For the record, the skin depth at 60hz is about .3 inches, which is why you see some high voltage transmission lines use a 3 conductor bundle instead of a single larger conductor. At 60hz additional diameter beyond about 1 inch increases the conductivity very little, but greatly increases the cost, so 3 smaller conductors have better conductivity and lower cost than a single large conductor.
I was actually thinking of trying this awhile ago, but with copper. And actually have a copper plating bath working now so could give it a shot but part of the reason I lost interest is that XP-Ls don’t seem to have this problem and you can easily start to overheat the phosphor to the point of negative returns on current already. Now at the same time I always liked the idea of using the XM-L2 due to it’s larger thermal path so maybe it could be pushed a little higher doing this. A lower Vf would be nice though but I have no idea if it would change that a significant amount, if lower resistance wires and springs can than seems like the same would apply.
Also a little worried about getting good coverage, as in if you cover any of the bond wire with the sealer or even if some is already covered by the phosphor coating that could cause a hot point that may still kill it.
And yeah I agree the thickness you can deposit is just how long you want to run it, I’ve deposited a layer of copper over 1/4in thick and as long as you have the right bath composition you really don’t have too many issues. Also why would you need to solder anything, can’t you just connect to the MPCB or led pad?
I respectfully disagree with the part about not making a difference. I think we need to consider the scale we are dealing with, and my desired outputs.
The bond wires on an XML2 are aprox. 30μm. A 3μm coating would reduce resistance by 20, 10μm by 80, and 15μm by 125%. That’s not considering the conductivity of Ag vs Au.
Also, my goal here would not be to dramatically increase lumens or surface brightness, but rather lower the IR losses in the bond wires and improve Lm/W. This would make higher Amps on direct drive for any given cell, and also improve time in regulation for those drivers.
I do not know how he has done it, but I have a custom aesphiral light (costomizer is Vinz from Southern Germany) with a dedomed XM L2 running at just above 7A…………he also has a process to reverse the green light output commonly seen in dedomed XMl-XM-L2 leds.
No one knowns how he has done that either though……
This light is so stable it can run at the highest mode until the cells are drained.
So, still very worthwhile to find out how it can be done, especially since dedomed XM L2’s have a higher output that the XP-L HI leds.
If you’ve never done plating before it may be but I never looked into silver plating. Really I have next to no experience plating and one of the reasons for learning was this very project, but am now more interested in electroforming heatsinks using scrap copper. On such a small scale you might find using one of the “safe” solutions of acetic acid may work which I played around with, you have to keep the current way down and if you try to plate too much on you get treeing. Right now I’m just using a standard sulfuric acid bath with the recommended amounts of HCL, copper sulfate, PEG as a carrier and using taurine as a leveler. The taurine is a total guess if it actually works as a leveler, it’s just a similar compound to conventional levelers that are very hard to acquire so my thought is it might have the same polarity thing going on to act as a leveler. Currently my plates are very rough but I’m plating at a very very high current density for copper recycling. At lower densities the deposits are smoother. Also have a small pump for circulation.
Talking about this makes me want to give it a try, I have a XP-E with a broken bond wire I can use as a test to see if the phosphor is damaged and what have you.