Three Oslons tested: 1) latest gen. Oslon Black Flat 2) SSL80 4000K 92CRI latest gen. 3) SSL80 4500K 96CRI 1 gen. before latest

The only danger even if the thermal pad grounds out is not having modes, same as shorting the negative lead, so it’s not really that big a deal overall anyway. Thanks though, I know to take precaution so it should work out all right. Thinking about using a similar light and putting 9 of em in. :smiley:

I like the Arctic Alumina Thermal Adhesive because it can be removed through application of heat. :wink:

Is the center pad just connected to ground? seems like there was too it then that, although I have not looked into it that close.

The negative pad and thermal pad are connected.

What are your thoughts on this? http://www.thermal-grizzly.com/en/products/13-minus-pad-8-en
I already posted this in another thread. I bought a 0.5mm sheet of this and will be testing it in a Tool AA with an Oslon. You think it’s too thick for good thermal contact? I think with a screwed down MCPCB it should become very thin, as it’s very elastic. At least it has the same thermal conductivity as Arctic Silver 5.

I would not want to use that under the mcpcb, it has horrible thermal properties compared to thermal paste.

You think the layer stays too thick? Because Artic Silver 5 is 9,0 W/mK and this is 8 W/mK. Don’t mix this up with the crap they put under cheapo heatsinks. Just by the description, this is some next level thermal pad. :smiley: :+1:

Edit: They go up to >30 W/mK but really cost an arm and a leg.
Edit2: http://forum.notebookreview.com/threads/thermal-paste-roundup-performance-price-recommendations-update-thermal-pads.796820/ OK maybe they still suck. :frowning:

I have seen the new carbon nanotube thermal pads that are promising, is that what this is? In the picture it looks like the normal silicone based thermal pad.

In which case you are correct, they are almost always way too thick. It says the thinnest version is 0.5mm thick for example. Most thermal paste would be a fraction of that.

Graphene thermal pads are conductive.
Those only work well when surfaces are very flat and parallel, they perform similarly to thermal paste.
These pads are not that, it is some different custom materials, much better than just regular silicon thermal pads.
They are also pretty expensive, more W/mk = more cost.

If you need to apply a layer of thermal adhesive that is insulating you need to use some sort of spacer like a piece of electrical tape on an edge in order to keep the two surfaces apart while the epoxy dries.
A piece of electrical tape will be like .2mm or so.
Using these high performance thermal pads will still be decent for a low wattage LED like the black flat, especially since the pad compresses, only problem is that you still need to secure the PCB down somehow without using screws that will create a conductive path.

Anodized screws maybe? May get problematic when the anodizing wears off.
I just realized there are even screws completely made from plastic, well…

Yeah, or ceramic screws which are super expensive.
Another way is to use some sort of hold-down plate with screws outside the MCPCB area.

The easiest way is to use thermal epoxy and just remove the screws once it is dry.

Small insulating fiber washers + insulated sleeves work for other applications. There is no reason they shouldn’t work to insulate screws from an MCPCB.

How do you keep the threaded shaft of the screw from contacting the MCPCB inside the hole? The walls of the screw hole? Easiest to use a non-conductive screw. Or none at all.

“+ insulated sleeves”

Better if insulated washer and sleeve are one unit, but probably hard to find.

…or get some Laird thermal sheet stuff.
Various thicknesses available.

I tried putting the Black Flat in a Shooter 2X today. The driver I had in it is an original MaxToch Buck driver for 2 cells at 6A. Poof! went my little jewel. So I heated up the 32mm Noctigon, inside the light, (difficult!) and removed the fried one to put a new one in it’s place. Worked fine in testing, was in the light with the 32mm board glued with Arctic Alumina Thermal Adhesive. I reduced the output of the driver to 4.22A pushing an old XM-L T6 on a test board with 2 18350 cells. Got it back together and working, tested 424Kcd in the 2X, but wanted to try a centering ring to see if I could dial in focus. Centering ring ripped off the emitter. UGH! Managed to remove the 32mm MCPCB and clean it up, re-flow the emitter properly and it worked fine on my bench. Put the light back together with new thermal adhesive, everything checked out fine, assembled it, fired it up…. pooof! goes a second little jewel. :person_facepalming:

To insulate the non isolated board I use this method successfully:

- Non conductive thermal paste, I use Arctic MX-4

- Lapped he MCPCB on #320 grit SiC to give some grittiness to the base of the MCPCB. Too shiny and most of the thermal paste will squeezed off the base.

- Apply thin MX-4 to the base of the MCPCB.

- Sprinkle the pasted MCPCB with #320 grit SiC powder. SiC is a relatively good conductor at 20W/MK and a very effective insulator. Much better than AlO3. This SiC powder serves as the thickness guide and separator while still provide decent thermal conductivity in between the paste. The same technique is used by some manufacturers by adding micro glass beads to their TIM.

- Gently, blow the excess powder.

- Place it on the heatsink. Move it SLOWLY in XY direction while firmly press on it. Make sure you spread the paste and SiC powder well.

  • I use PEEK screws to fasten the MCPCB. It’s a very strong plastic screw with service temp up to 270C. It’s not cheap but still far below ceramic screws.

Note:
SiC powder is extremely cheap. I got mine for $3 for 1 kg, last for years. If you don’t have some, you can get some by soaking abrasive paper in the warm gasoline.
PEEK screws are available through online stores such as Ebay, Aliexpress. I paid $50 for 25pcs (M3 x 5)
MX-4 on the other hand is relatively expensive but works extremely well. Also very easy to apply. It’s rated at 8,5W/MK I guess.

- Clemence

I really like your story… Weird things happened to me before and they will probably happen again. Weird things like driver fail after I tested it, emitter burnout without any special reason driven on low current draw, shortcuts when reflector managed to peel 2 layers of kapton tape, UFO specks of dirt on emitter etc.

Plenty of weird things can happen… Maybe because of that some of us like modding stuff?

Oh! Another good one… Once I mixed to much flux into solder paste(some guys were giving this kind of advice for soldering paste refreshment :slight_smile: ) - I tried that same paste for reflowing emitter - I had put emitter on top of that, reflowing started and puff! Emitter flew away over whole room and of course it did not survived.

:beer:

I also had that flying LED moments with my home brewed solder paste. If the flux you’re using is too thin and has low boiling point, it will pops the LED off the board. Slow heating can prevent this. But the better fix is to use proper tacky honey like consistency flux.

- Clemence

Moisture in flux is cause, it starts to boil over 100C and expand rapidly > parts flying.

Is it bad that I still don’t use flux to reflow LEDs after all that years of modding ?