I’ve got a Small Sun ZY-T08 that I dedomed the emitter. Am I going to have to take material off the back of the reflector so the emitter goes in deeper for the best focus?
That or add copper under the emitter’s MCPCB.
My XM-L2 dedomed direct drive ZY-T08 is very well focused, it does 148kcd with a very tight spot and no donut hole even at a few inches but the reflector and pill are both pretty extensively modded, looking back it was a lot of tedious work, if I ever built another XML one I’d just stack a sanded smooth Noctigon under the emitter’s Noctigon, probably solder them together to help heat transfer.
Not only did I shorten the base of the reflector but I modded the bottom lip of the pill so it could screw way farther up into the head (note this alone took a ton of work- I had to machine material off the battery tube and add a 1/4” tall copper contact ring to the driver board)
This is all assuming you have a metal pill, if your working on a plastic reflector light you got it easy.
i had a hard time focusing it, my problem was flattening the soldered cables, and replacing the stock insulator for a flatter one (i´m talking about the OL group buy light, with aluminum reflector)
before replacing the insulator I had a big donut hole, now its fine
So if you have a “donut hole” modifying the reflect to allow the LED to sit deeper will help that?
With a good cup of coffee, donut holes don’t stand a chance with me. 
The further in the reflector you get the emitter the shorter of a distance the donut hole will disappear. Mine now doesn’t have a donut hole at anything over 2”, stock with a dedomed XML it took about 6’ to focus the hole out.
The new aluminum ZY-T08 reflector is a piece of (insert favorite word here.) It doesn't fit right, it's too short, and even sanded/modified it isn't as good as the original plastic one. I spent (wasted) my fair share of time trying to make it work right and while I got it pretty good, it still isn't as good as the original. I'm sitting on 10 lights that I don't even want to sell in stock form because it just isn't right. Trying to get some replacement plastic reflectors as we speak, but it looks like I'll have to pay for them.
I ended up reflowing two MCPCBs together (do this even with the plastic reflector) and radiusing the corners and flat portion of the reflector so it could sit as far up as possible. I opened up the emitter hole on one reflector so I could a good centering ring, but even with the ring sanded paper thin the focus wasn't great. I ended up using the centering ring to center the MCPCB while the arctic alumina set, then I removed the centering ring so the reflector could sit deeper.
Even with all of that, the beam still isn't as good as the plastic reflector even though the donut hole is gone. The shape IMO isn't as good.
I’ve been struggling with this same issue. The problem as I see it is the relationship between the emitter surface and its position relative to he focal point of the paraboloid reflector. Instead of trying to raise the LED higher up into the hole, I just milled off the back end surface of the reflector. So much so that I began to enlarge the hole. This effectively got the emitter surface closer to that mathematical sweet spot. The amount of reflector lost is really insignificant, and I couldn’t measure any differences on ceiling bounce tests with or without the machined reflector.
That's right, which is also part of the reason why I opened up the hole a bit. I had one sacrificial reflector that I kept on modifying, testing, modifying, testing... Dedomed emitters can be a pain, but when you get them right the increase in throw is substantial.
Richard I’ll get some pic’s of my reflector today (not home now). I also opened the hole up but more than “a bit”, I actually removed the entire bottom of the reflector so the entire Noctigon16 can be seen when looking straight down in the head. I feel the reflector is [after being modified] almost identical to the stock reflector in my new olight SR51, it took a couple times to get the star glued down perfectly centered but not only is the beam perfect there is NO possible way it can short out on the reflector.
I also modded the part where the pill & reflector interface and like I mentioned I modded the top of the bottom lip of the reflector so it can screw farther up into the head (and therefore also had to add a copper positive contact ring.
Ok here’s the pic’s, you can see my modded reflector vs a SR51, you can also see how far my pill screws up in the head vs a stock one.
Smallsun on left, olight right
How deep the pill sits
Don’t forget to clearance the body
Wow, I've ground the pill a bit to make it sit deeper before, but never that much! I'll try opening up my sacrificial reflector a bit more to see what I can see.
I’ve just started messing around with the fine adjustment of reflectors but it seems like what we need is a MCPCB with a raised boss that the LED mounts on.
Just the footprint of the LED die, raised about 2mm above the rest of the PCB, would also allow clearance for soldering.
I can't remember if it was ouchyfoot, Comfychair, Oldlumens, or some/all of them that I've seen do this before. The cut a small copper piece to use as a raised pillar for the thermal pad, then made the electrical connections. Now that you say that, I've got something to try... ;-)
When I can, I always prefer using thermal compound instead of adhesive but if you're cutting the hole out that big chances are you'll need to glue it.
It depends on the profile of the reflector, as to where it sits in relation to the led and it also depends on the led. MT-G2, XM-L2, XP-G2, Nichia all need a different height, as well as domed versus de-domed. Every reflector is different and every combo will have different needs.
It should be safe to say that in order to have the room, (height), needed to experiment, the led has to sit higher than the rest of the star. I usually cut copper pieces that fit between the led and the star pads, to raise the led off the star, allowing me to open up the reflector, so it will fit down over the substrate, (if needed). Then a custom centering ring is also needed.
Every build is different and it's best to experiment outside the light, by putting the star on a chunk of aluminum, powering it up and then trying different heights of the reflector in relationship to the led and dial in the height you need, for the nicest beam and/or highest lux reading, depending on if you want a thrower or a smooth overall beam.
I’m almost embarrassed to say this following a post by the master of Mag-Lites but on my 4D Mag build I filed down a Sinkpad to the point it fits within the hole in the reflector. Fine tuning the focus is as simple as a twist of the head.
Another reason to love Mag-Lites.
What sort of heatsink setup did you use that allowed the reflector to cam? I have my first ever mag 2D sitting in query right now waiting for me to decide what to do with it and a focusable light would be awesome.
ETA:
I just found a nice trick to not blind yourself when looking for the donut, hold an anti-static bag in front of your face and look threw that. Pretty useful for anyone like me who doesn’t care for lower modes on my throwers.
Easier than the time when a MT-G2 I have dedomed its self and I had to use my welding helmet set on the darkest setting to check for any damage to the die.
I purchased the heatsink from Moderator 007, think I got the last one he made. You could do the same thing with one from H22A over at CPF.
You have to cut the tower off the original switch to make room for the heatsink. Once that and the wiring were done I filed the sinkpad down below the size of the hole in the rebel reflector. I mounted the sink so the reflector just bottoms against it when the head is screwed down all the way.
All assembled just unscrew the head to the proper height for focus.
A few photos here.
This is the metal reflector light. So you moved the pill up about 1/4”? I could machine a piece of copper or aluminum to mount to the top of the pill and machine the reflector down to make it fit.
Focus was much easier on my dedomed zommie
