I agree with Matt on that Tom. Also makes a huge amount of sense to get the Noctigons with emitter already reflowed, while I don’t mind reflowing 2 or 3 I sure wouldn’t want to do 70!
Oh - only got qty 10 bare. I have about 25 (or more) bare copper stars, both SinkPAD's and Noctigons in 16's and 20's so I'm looking to use some of them up some how, but Hank makes it difficult when for a mounted emitter on quality copper, he charges less than others charge for the bare emitter -- it's a good problem I guess... I've given up waiting for the elusive XM-L2 U3's. If they do show up, they will initially be pricey and who knows what tint, and if anything like the XML U3's, they will appear shortly before the next generation is released anyway...
TIR optics? Illumination supply carries the clear and frosted for the 20mm 3up. I have used them many times. They should fit fine on the copper stars. Same pattern.
They also carry a 20mm star and I have used them many times as well.Aluminum, but I think copper is overrated when you are talking about 3 leds on a 20mm star. It's going to get hot as hell quickly. More important to have a good/large copper heat sink behind it.
I disagree. As you said, its going to be hot as hell quickly. You want the emitter temperature to be as low as possible. Copper MCPCB helps with this. There are so many graphs/tests where copper mcpcb show a significant gain.
There have been several people who have used copper pills. I have never seen anyone show any significant measurable gain over say, aluminium (when copper mcpcb is used).
I believe Tom E did some testing in a C8. Aluminium vs copper pill. Copper pill did basically nothing if I remember correctly (someone please correct me if Im wrong and show me some numbers).
Even with 3x-XP-Gs driven at 3-4,5A divided on them. That is still huge amount of heat despite not being "overdriven", I have a feeling there will be decent gain with copper. Or maybe I should say. Lumen sag will not be as much with copper when used for several minutes.
I think Old lumen is trying to say that having a large heatsink is important to add more thermal mass.
I personally also think copper is almost unnecessary for most cases (although I may still use them just because).
The heat is generated across the board on 3 LEDs, so I believe the heat issue is already less of an issue. Example: running a triple nichia at 3A = 1A per LED. This means each LED is producing 1/3 of the heat at a different point on the board. Compared to an XML running 3 amps where all of the heat has to pass through 1 LED footprint.
Also I think copper bonds are shown to be beneficial once you start pushing LEDs past typical limits. Most of these boards are going in small hosts that can’t handle high heat anyways (hence the importance of high thermal mass) so it is likely you won’t reach currents where a copper bond is beneficial.
However, if you are trying to hit max lumens, say in a triple XPG-2 set up, then yes these will help.
I don't see how the copper triple will not make some sort of difference, even for 3A total, 1A per emitter. The emitter tests I did I think referred to are here: https://budgetlightforum.com/t/-/20318#node-23492, with a stock SS T08 running about 1.95A, so it appeared the copper star made a 8-10% increase in lumens. I would suspect even at 1A there is a measurable difference, specially for a smaller emitter like the XP-G2. Not only that, you have the 3A of total heat. Of course it's not only the aluminum to copper, it's the direct thermal path vs. the dielectric layer on the IS board, I'm assuming it has one. Also the Noctigon is spec'd at 1.75mm which is pretty thick - the IS aluminum one is not spec'd - I'm thinking it's even thinner.
Match kind of proved the most critical heat sinking junction is the closest to the emitter. After he tested the SinkPAD's and seen how they did with and without additional massive heat sinking, he kind of came to that conclusion. Now, I think additional heat sinking is all about longer, cooler runtimes but has little effect on short term total output.
OL is right. I've said the same before and if you look at Matchs graphs of - for example - the XPG2 here:
Emitter Test Results pt4: XML2, XP-E2, XP-G2, and Nichia219! (along with sinkpad tests) (I’ll just paste the pic)
and then take a look at the graphs at 1A.. the difference is negligible. Maybe there is a 5% increase at such "low" currents if you use 3 LEDs. An XML at 3A gains about 10% on a copper Sinkpad.
So.. are copper boards cool? Yes. Are they the messiah everyone thinks they are? Not unless you drive your LEDs really hard (4.5A and up for a triple).
Maybe I would still use them just because they are cool, but on the other hand.. reflowing one LED is hard enough, I dont want to reflow 3 at once. :D
Small hosts can handle high heat much better now, thanks to copper boards. It was a theory I had when the copper boards appeared, my main enthousiasm for copper is based on that. And since then I have seen it happening in a fair number of mods I did. It works as follows: the heat path from led to body shell is 'wider' with a copper board, so the temperature difference between the led an shell is lower at any time, therefore the led can be driven harder, causing the shell getting hotter without the led overheating. And at last, the higher temperature difference between the shell and air makes even a small host sufficient to dissipate the increased heat .
Well for me, I'd probably be looking for at least 3.5A to 4.5A total, so at 1.4A shown on the chart, there's maybe a 5% improvement which I'll take - I spend more for an AR lens to get the same bump, but the copper has more advantages than just lumens. We won't really know unless someone built identical triples with a difference of aluminum and copper, and do all sorts of tests. Hard to tell what the impact of the heat of 3 emitters on one star will be on these 2 different designs, but for all the time and expense going into a custom build like this, I sure would choose the copper board, even for 3A total.
Hi,
I was really worried about re-flowing emitters and avoided it for the longest time… even just 1-at-a-time, but in the last month or so, I first did some practice single emitters, and then several 3ups, with XP-G2 and XP-E2 with the stars from IS, and despite being concerned about it at the beginning, I actually find it kind of enjoyable (and easy), even with the triples, which are no harder than the singles. I just put the solder paste across the pads, drop the emitters onto the pads, and then I use a small torch under a copper plate (with the star on top of the plate). Wait until the paste turns shiny, and then remove the heat. That’s really about it.
you can buy the boards dressed if your happy with cool white or 3c tints.
I’m getting some to “improve” some mr16 drop ins I’m not over awed with, but I’ll be using 90 cri xp-g2’s, so I’ll have to reflow, I’m going to try the hot air method I think, unless the soldering station proves as good at this as it does on single boards.
Around 10% increase at 1,5A to the emitter is a good gain. And then you only have around 5W.
I have a feeling that when you start pushing 15W through that one mcpcb (instead of 5W), you will most likely see even more than 10% increase from aluminium to copper mcpcb.
If people think these are small gains, then I don't see the point to bother with copper pill instead of aluminum pill. Id much rather take a copper mcpcb and aluminium pill, than aluminium mcpcb and large copper pill.
If I get 5% increase at 10W. Then its worth it to me.. Either way, gotta love the "coppa".. :love:
:)
I agree. I just think in general most people won’t be pushing 1.5A per led in a triple of this size. The host will just be too small to handle the heat. And if you do push it that far, these boards will give you a gain in lumens but thermal mass (size of light) becomes an issue too.
I personally only run 3A to a triple nichia due to both heat and high Vf. At 1A each, I don’t think the increase in lumens will be noticeable. At the same time I don’t want to increase the Vf that is typically seen with these boards. Another thing is, I typically don’t even run the drop-in at 3A. Most of my usages are at the 1A level = .33A per LED.
These boards do give you a gain a small gain at higher currents, but depending on your usage, it may be a bit unnecessary.
I guess these triples will mainly be used in tube type lights because of their size. I’m waiting to see what kind of drivers everyone will be using.
Yea, I think a triple nichia in a xiaozhi or something like that would make a nice EDC.
It's less than 5% even at 1.6A.. your feeling? Sorry, but that's not scientific. :P
With 3 emitters, the effect of the dielectric doesn't change, it's still the only real bottleneck. Heat transfer of aluminium is by far good enough.
I have a triple Nichia run at 3.4A in a Xeno S3A, that might be the single 18650 light with the least mass available and it gets hot real quick - with a regular PCB.
On the graph at 1,6A im seeing about
475 aluminium
510 copper
That is more than 5%, although, I was off with the numbers in my previous post..
My statement is still the same. Id take copper mcpcb instead of copper pill. This was where I disagree with Old-Lumens.
Do you mean brass pill?
I’m shooting for 5A, maybe a bit more, in a Solarforce L2P host, solid copper pill, with 7 levels to adjust if/when heat becomes an issue. Works for me. Scientific or just plain Like It Like That!
LED > solder > copper trace > dielectric layer > aluminum base
LED > solder > copper base
The smaller the transfer area, the more critical good conductivity is. With AL the LED has to transfer all its heat through a dielectric layer, essentially fiberglass, the same area as the LED pad itself.
If you're worried about too much current producing too much heat, you should look into the fancy new drivers they have these days - they have more than one mode, if you can believe such technological wonders are possible - and you're no longer stuck with just 100% or nothing! It's amazing!