Luxeon Circuit Isolation

Slick,

I just mentioned this in another post. You're OK bonding emitters directly as you are now as long as you are using single emitters, and not clusters of them.

You'll start getting serious and weird grounding problems when you work with 2 or more luxeons sharing their mount in a non-electrically isolated manner and most notably when you use Xitaniums to drive them.

The problems manifest themselves in severely dimmed luxeons, dim flickering and unwanted light activation when the metal mount is touched (ie. grounded) by hand.

Been there and done that - believe me, it might be too late to save an epoxied luxeon once you start noticing these problems. So, if you're thinking of working with multiple luxeons, always bear in mind this very important bit in the datasheets.

Make a piece of metal that has the same footprint as a TO-3 transistor, epoxy the luxeon to the metal, and mount to heatsink just like a TO-3 transistor(use isolated mounting kit).

Good suggestion, star. The footprint for a luxeon/O and /C is damn close to a TO-3. The luxeon screw holes match the TO-3 mounting holes perfectly - only the luxeon is wider in the other dimension. I have a large TO-3 HS that one fits in perfectly. As soon as I get my order, I'll show a pic of a luxeon mounted in it.

-Darin

Thanks Guys!!! I'm really glad I asked this question. This makes perfect sense about how a single emitter would be OK to touch the ground plane, but not multiple emitters.

From this information, it's easy to see how this would be a problem for multiple emitters (touching ground) that are connected in series.

But since a single emitter is OK to touch ground, I wonder if this would be a problem for multiple emitters if they are connected in parrallel?

Does anyone know the thread title or date for Hotfoot's discussion of this topic? Thanks again!!

Hmm,

My experience has been that Artic Silver will reach over 20 meg isolation in 24 hrs. It makes a very good insulator.

Anybody experience/measure something different?

With all due respect, Lambda, I wouldn't doubt the electrical insulation of AS, and I have used the stuff in the past with great success too in the past. The insulative properties of the epoxy isn't in question here.

From my work with mostly multiple-emitters, the danger lies in portions of the emitter slug/mounting base contact where the emitter slug actually touches the base metal and is not insulated by a thin layer of AS. This could happen easily since AS recommends that pressure be applied on the bonded surfaces such that as much direct contact as possible is achieved, leaving AS to merely "fill-in-the-gaps", as it were. This direct contact would constitute an electrical path.

McGizmo and I have had an offline discussion about this and suggestions ranged from using a very thin layer of lacquer/epoxy to prep the surface prior to mounting, or inserting a thin piece of absorbent paper (like facial blotters the ladies use) beneath the slug or using a fine mesh material instead of paper.

Sorry, Slick - my previous mention of this on the CPF wasn't a discussion or thread, merely a post on Wayne's heat DIY calculation disccussion thread. Sure looks like we've got a good discussion going here, though!

It is a good idea to isolate the luxeon if the power supply uses a resistor in the ground line to regulate current.
If the luxeon shorts to the case, you get a runaway condition(power supply thinks there is no current, so it supplies more and more current until something blows up).

Details of the Mod:

The plan calls for 3 white Luxeon emitters in the head of a 3 D Mag. I'm using a copper disk 1/16 thick Arctic Silver'ed to an aluminum disk 1/8 or so thick. This heatsink sits on a ledge in the Mag's head and is the proper thickness for me to "float" the collimators (without using any holders) to mount them.

The light will be direct drive, most likely using "day-old" NiMh's and ocasionally alkalines for some supervised intermittant use. The Luxeons will be wired in parallel.
__________________

The Plan:

I had originaly intended to use my Arctic Silver epoxy to mount the 3 bare emitters to the copper side of the heatsink. I'm certain that they will make enough contact between the copper and the slug to leak current if possible. The slug extends out of the LED body far enough that it's what makes the entire footprint.

If the LED's are prone to current leakage via the slug but can operate safely mounted singularly (sitting on the negative lead), I wonder if this would hold true for parallel connecting them? I can see that this would NOT work if they were connected in series, since leakage would effectively "ground out" one (or more) of the LEDs in the circuit.

So has any one else built multi-emitter "Hydras" where they mounted their emitters to a common ground plane? - which in my case will also be the negative power pickup..

I have been reading over the Lumileds application briefs for the Luxeon emitters and noticed that they say the heatsink that the emitter is epoxied onto needs to be electrically isolated from both the positive and ground circuit.

This has me a bit worried since I have been using Arctic Silver to epoxy my emitters directly onto the copper heatsink. This copper heatsink sits directly on top of the flashlight barrel and also serves as the negative power pickup.

This method has been working just fine but I'm concerned that there may be a problem because I'm not conforming to Lumileds design recommendations..

Does anyone know what the purpose of this "isolation" is? Is there a problem with how I've been building my mods?

Slick,

The Hydras I built have the heatsink electrically isolated from the flashlight body, with the emitters AS'ed to the heatsink.

Hootfoot,

I've yet to ever see a condition where the emitter shorted to the heat sink with AS. I supose it could happen, but I just rechecked several AS mounted Luxeons, and all show greater than 20 meg between heatsink and either terminal.

Guess I'm just not pushing that hard.......

Originally posted by lambda:

I've yet to ever see a condition where the emitter shorted to the heat sink with AS. I supose it could happen, but I just rechecked several AS mounted Luxeons, and all show greater than 20 meg between heatsink and either terminal.

Click to expand...

<font size="2" face="Verdana, Arial">The relevant test would be to measure the resistance from the slug to the heatsink. The slug is difficult to reach once mounted. Nowhere have I seen in the Luxeon literature that the slug is common to either electrical lead on the device, only that it is not electrically neutral. Someone which a bare emitter could check this for us, measure slug to leads. I suspect [but don't know] that this will be some type of semiconductor junction, not a simple resistive path.

Originally posted by lambda:
I've yet to ever see a condition where the emitter shorted to the heat sink with AS. I supose it could happen, but I just rechecked several AS mounted Luxeons, and all show greater than 20 meg between heatsink and either terminal.
Guess I'm just not pushing that hard.......

Click to expand...

<font size="2" face="Verdana, Arial">I know - McGizmo can tell you too that I once felt that way too. It may have been something like mounting more than 100 luxeons in this manner before I noticed something was amiss. Mine were situations where I had 3 emitters Artic Aluminaed directly onto a common electrically conductive base/heatsink. The luxeons connected in series and are driven by a Xitanium.

Your mileage may vary since you're driving them parallel, but I'm pretty dead sure you'll run into it once you start doing more multi-emitter setups in the configurations similar to what I'm doing.

Or maybe *I'm* pushing them in *too* hard...

I like Doug's idea of measuring the resistance from cathode/anode to slug, but he's probably right in saying that it could be more complicated than that - and may even vary from example to example, just as color does.

Yikes - yet ANOTHER luxeon quirk.

someone showed me a diagram (a hand drawn one) of LS internals when i had asked this question awhile back. the diagram, if i remember correctly, had the positive lead going directly to the die and the negative lead going first to the slug and then from the slug to the die.

i had a bare amber HD LS sitting around before and measured 0 ohms between the _positive_ lead and the slug, open between the negative and slug.

Originally posted by papasan:
someone showed me a diagram (a hand drawn one) of LS internals when i had asked this question awhile back. the diagram, if i remember correctly, had the positive lead going directly to the die and the negative lead going first to the slug and then from the slug to the die.

i had a bare amber HD LS sitting around before and measured 0 ohms between the _positive_ lead and the slug, open between the negative and slug.

Click to expand...

<font size="2" face="Verdana, Arial">The discrepancy you note may be due to the green/cyan/blue/white dies being different from the red/amber ones. The diagram you saw may have been for the former while your measured sample was the later.

hotfoot,

where did u get your Xitanium in s'pore?
how much did it cost u approximately?

Originally posted by earwax:
hotfoot,

where did u get your Xitanium in s'pore?
how much did it cost u approximately?

Click to expand...

<font size="2" face="Verdana, Arial">hi earwax,

Please contact me via my PM or email if you are interested in the Xitaniums - my company is the local distributor for them. The smallest model retails at S$129.00 w/o GST and will drive up to 8 1W luxeons in series, scaling voltage automatically whilst regulating current at 350-375ma.