Positive LED slug epoxy trick

[NYLYTE]

It seems that a number of people doing upgrades and mods have been bitten by the fact that the SSC (Seoul Semi-Conductor) emitters have a positive slug much like the red and R/O (Red Orange) luxeon emitters and stars.

I have seen advice in the forums to cure a thin layer of Arctic Alumina epoxy to the slug first to provide proper electrical isolation. Here is a quick tip that should help in accomplishing this goal.

Epoxy does not adhere to low density polyethylene plastic (LDPE). This is the cloudy clear plastic of quart and pint size takeout containers, and is usually clearly marked LDPE for recycling classification.

Simply mix your two part epoxy on part of a slick LDPE takeout lid and find a nice clean spot on the same lid for a properly sized dollop of the mixed epoxy. Gently press your SSC or red/R/O emitter (or star) onto the dollop of epoxy and make sure not to squeeze out all the insulative goodness (not too thick, not too thin).

When the epoxy has cured (wait several hours, or preferably overnight), just flex the lid and both your epoxy mixing mess and the LED will easily peel free from the slick LDPE surface. The insulating layer will have a surface as slick and shiny as that of the lid you used.

I hope this helps people with their mods and upgrades. I have spelled out all terms to facillitate searches in the forums.

Have fun,
//NYLYTE

 

[yellow]

just noticed this design flaw of the SSC some weeks ago, no way I will use them instead of the better alternatives.

Adding anything that worsens heat transfer, especially there at the primary slug, means no way the efficiency, output and lifetime will be achieved.

 

[doc_felixander]

if you have a massive heatsink with a diameter that increases quickly and doesn't contact the body immediately at the upper edge, it's more efficient to have direct contact at the slug and to isolate within the heatsink where you have a much bigger contact area by cutting the heatsink in to halves (or whatever), polishing and making the AA gap there.

 

[chris_m]

Design flaw? White Luxeons have a negative slug - is that a design flaw?

I'd love to know what you plan to use instead which is a "better alternative". Also why having a positive slug means it's not possible to get optimum heat transfer.

 

[doc_felixander]

the Crees ARE an alternative. if better or worse, that's a matter of taste.
i guess i won't have to list the pros and cons here.
technically it does have a better thermal design.

EDIT: he's talking about the SSC's positive slug. only few people here would deny that it's a major design flaw, and think of other applications, e.g. automotive - same issues there.
i really wonder what they were thinking.

 

[IsaacHayes]

NYLYTE: GREAT idea!!! How come I didn't think of this?

 

[Norm]

What IsaacHayes said. I had been scratching my head about what to do with a quad P4 mag I'm almost finished this sounds like a nice reliable simple method.
Norm

 

[chris_m]

Given the Cree is 8 degrees/W junction to slug and the SSC is 6.9 degrees/W junction to slug, then can you explain to me how exactly the Cree has a better thermal design?

Meanwhile in the real world it's dead easy to solve the "problem" of the positive slug either by using an anodised heatsink or by using a star mounted one, which has 8.5 degrees / W from junction to star base, and given the likely better thermal transfer from a star than from the slug of a Cree, the chances are the overall heat transfer will still be better than a Cree.

I suspect they were thinking that the most important thing is to get heat out as efficiently as possible - a target they seem to have achieved. If that means the slug is not electrically neutral, then that is an issue that can relatively easily be solved by integrators.

Note: I have no affiliation to SSC (or any other manufacturer or distributor). I just hate to see misinformation being proclaimed as gospel. I am using both XR-E and SSC P4 LEDs in my homemade and modding projects - each has its merits for different situations, but there is certainly no "design flaw" with either that precludes its use in all applications.

 

[Dawg]

I think it is downright scary that so many people know so much about LED's.

 

[yellow]

Doc Felixander got completely right, what I meant.
By now the "usual" way to go, with SSCs, seems to be to add some nail polish, or a gap filled with glue between emitter and heatsink. You can not disagree to my opinion that this worsens any specs (output, efficiency, lifetime) of the led, maybe by a really noticeable degree in high power devices (overheating)
Also most ppl here use the body of the light as the negative contact (at least I do). Even most circuits have a common negative input / negative output to the led. The positives have to be isolated.
So its easy to use the body as negative, the emitter directly onto any heatsink, directly put into body - without any more thinking.

SSC means: adding thermal barriers --> directly at slug, of from star to heatsink (which is also an additional thermal barrier, dont forget this) or to somehow isolate the heatsink to/from the body, which is even more difficult. And bad also, why use aluminium lights perfect for acting as final heatsink but decreasing heat transfer? Coulda use plastic lights instead.

Looking at the specs, Seoul and Cree are not this different (SSC on paper, or at tests (like newbie`s) is a tiny bit better). So in reality they will be worse, at least after a bit of use with added degradation from too much heat (than necessary).

If it were not the larger beam degree, offering better focusing ability, noone would use the SSC.
Then: "they use the same emitter". Why is the slug positive then? Should be possible to make it totally neutral, as Cree does, not?

 

[StefanFS]

I have used up 10 SSC P4 emitters for modding so far, some in lights without anodized heatsinks (Fenix & Liteflux). I never really saw this as being a problem, I just made sure I had a thin layer of Arctic A. epoxy under the emitter in the lights without anodized heatsinks. Heat transfer works very well in these lights with AA epoxied emitters, even when I am using LiIon. When I install Cree X-RE there are other concerns. But thats life, different types of emitters need different ways of installing!

 

[leeleefocus]

I agree with most of what yellow says.

 

[liveforphysics]

The reality here in flashlight applications is simple. We tend to use case-negative designs because it is tradition to do so. From the standpoint of useing more efficient and reliable NPN junctions to switch the control current, useing a case positive flashlight design and ground plane makes more sense.

As far as LED design goes, the most important part of the LED to keep cool needs to have a solid thermal path to the sink. The most cost effective design to permit this is a solid metal connection between the sink contact and the part that requires cooling.
If your anode has greater thermal energy transfer needs, by all means give it cooling priority with the direct metal pathway to the sink.

The objections here are caused by traditional views of flashlight design and retro-fit related inconvience. The SSC has shown to have superior cooling abilities than the CREE and all the lumiled case-negative designs. Embrace it, and the solution the poster of this thread has is an excellent solution for the retro-fitting related troubles you may be experienceing.

Best Wishes,
-Luke

 

[Curious_character]

On every star I've seen, the mounting pad is electrically isolated, so no insulation is required under the LED. There's only a problem with LEDs mounted directly on an electrically conducting pad.

I use Arctic Siver compound rather than Arctic Alumina because it's twice as thermally conductive. I can't find a spec on Arctic Alumina epoxy, but I'm in the process of making some measurements on various compounds and mounting methods. I wouldn't recommend materials like fingernail polish because of their poor thermal conductivity. Adding just 0.001" of Kapton (polyimide) tape (adhesive removed) under the LED resulted in about an 18 degree C additional temperature rise of my test Luxeon III LED running at 1 amp (3.8 watts).

c_c

 

[MrAl]

Hi,

I'm always against using any form of electrically insulating material but sometimes
it just cant be helped, and i have seen in the past that the trick is to simply
keep the layer of material as thin as possible.
I have been running two white Luxeons that have been tacked down to a nice
sized aluminum heat sink with a most unlikely substance: PC11 epoxy, and they
run just as cool as those done with Arctic Alumina (although i prefer the latter
because it's *made* for thermal designs). Thing is, after i applied the layer of
epoxy the Luxeon was pressed down hard and slid around on the surface a little
to get the absolute thinnest layer possible, knowing that very thin layers of almost
anything will conduct heat well.
The two have been running for several months now.

I've often wondered what would have happened if i had applied a layer
of thin plastic food wrap between two layers of epoxy to hold the Luxeon
down to the heatsink. This would provide good heat transfer and electrical
isolation, although i dont know how well the food wrap would fair in the long
term run (years). Perhaps a layer of mylar tape between the two layers of
epoxy (Arctic Alumina), although i dont know how well the adherence would
be. Perhaps a small circle of mylar tape with extra epoxy around the edges
to help nail the Luxeon down.

 

[riffraff]

As liveforphysics says, it really is just traditional and retro-fit concerns that bias us to a negative case. There's really no good reason why a positive case wouldn't work just as well, or even better. You'd never even know the difference unless you went through the flashlight with a meter.

 

[chris_m]

What yellow says is all very well if you're a modder, but hardly a "design flaw" (given despite the impression you might sometimes get here, we are actually pretty irrelevant). In any case, the one light I've modded with an SSC P4 has an anodized heatsink and no electrical path through the casing, so it's wrong to assume it's even an issue in all cases. You also dismiss using a star as giving a poor thermal solution, despite me having already posted thermal specs for SSC stars, which show that the thermal resistance using one of these is negligibly different to using a Cree emitter (and actually likely to be better as an overall solution, due to the greater surface area of the star). Finally you obviously don't understand LED manufacture, and haven't read the other technical threads on here if you have to ask why they can't make it neutral like Cree - it's because the die is only a small part of the LED, and Cree have an insulating ceramic layer between that and the slug which provides electrical isolation, but also a bit of thermal insulation. SSC obviously decided it was better to get the heat out the best way possible, and let the integrator worry about electrical isolation (if even necessary) - not a "design flaw", simply a different way of doing it, and not even better or worse, simply different.

 

[MrAl]

Hello again,

I thought the argument for the negative case was because in automobiles
the ground is usually negative ground, and we would probably bolt the
heatsink to the metal of the car. This would only work with one Luxeon
anyway though, as two Luxeons in series would have to be isolated one
way or another no matter what polarity the ground (metal of the car) was.

 

[chris_m]

Just one final point

And bad also, why use aluminium lights perfect for acting as final heatsink but decreasing heat transfer? Coulda use plastic lights instead.

You're comparing apples and oranges. The difference to heat transfer a plastic case makes is a whole world different to the tiny thermal resistance you're introducing by having to electrically isolate the slug (if you do at all - other better solutions if you can be bothered, but from what you say it seems you can't and just want everything to be nice and easy).

 

[chris_m]

I've often wondered what would have happened if i had applied a layer

of thin plastic food wrap between two layers of epoxy to hold the Luxeon

down to the heatsink. This would provide good heat transfer and electrical

isolation, although i dont know how well the food wrap would fair in the long

term run (years). Perhaps a layer of mylar tape between the two layers of

epoxy (Arctic Alumina), although i dont know how well the adherence would

be. Perhaps a small circle of mylar tape with extra epoxy around the edges

to help nail the Luxeon down.

Not as good heat transfer doing this as you seem to think, as the food wrap or mylar is a good thermal insulator, even in a thin layer - you're much better off using the original suggestion on this thread of an extra layer of thermal epoxy if you must do that.

Meanwhile I don't believe negative earth on autos has anything to do with it, as people here aren't generally installing LEDs on autos. It's simply that some (many?) flashlights come as standard with negative cases, and people are trying to mod them. Not an issue at all if you're designing from ground up, and not necessarily even if you're modding and can swap casing polarity. Meanwhile I'm sure those people who put LEDs in autos for a job don't have any trouble finding easy ways to electrically isolate the LEDs if necessary.