It just unscrews. You’ll need something to put into the dimples. Some people use tweezers, but circlip pliers work just as well.
The retaining ring on these F13’s is glued or torqued very tight, so undoing it first time is difficult, but just needs some elbow grease.
aka “the hork” 
I got my nose plier and finnaly managed open it and to move the + to the other + , now its better but not better than my Convoy c8 at stock,
do you think the F13, with a 3A driver ( + 0.380mah*2 ) can handle 3.8A without problems ( of course with a good 26650 ) ?
It can handle 3.8A no problem.
Thanx for the fast reply .
So its ok for 3.8A what about 4.1A or 4.5A ? and should i add some thermal paste behind the led ?
Is the GB still active?
No, and the code do not work anymore 
Just got mine. Ordered 7/8. Took over 5 weeks.
I’d put as thin a layer of thermal goo (Arctic ceramic, as I had it handy) on the black anodized flat part, then placed my replacement 20mm Nichia star there and turned it back and forth a bunch of times, thinking I was spreading the goop evenly. Soldered the wires, made sure the reflector pressed down around the emitter when tight, used it a while, then the color went wonky — the dome and phosphor fell off the LED.
Oops. So I took it out.
Well, lookit that.
The black anodized “plane” is a shallow dish shape,not flat — it still had goop covering it evenly, but the contact with the star was only around the edges. Looking at the back of the star, there was no contact anywhere except around the edges.
Put in a different emitter (Noctigon this time) using more thermal paste.
But I guess I need to get out some fine sandpaper, or something, to make the emitter’s back match the flashlight’s front.
Maybe take a look at pflexpro’s thread, “Flattning a P60 pill”. I haven’t tried that technique myself, but it seems like a winner to me. ImA4Wheelr seems to be suggesting in that thread that aluminum should be annealed before hammering. I don’t know much about it it, but it seems to me that regular old 6061 is probably soft enough to let you get away with it without annealing.
Although annealing would make the aluminum softer and less likely to crack, I don't think it is a viable option in most lights. The temp needed to anneal is near the melting point. So it would change the anodized finish color and would also soften threads and such.
Haha, I see. I guess it’d better work w/out the annealing then eh?
It seems like manufacturing a little set of anvils which are convex, concave, have the star pflexpro showed, etc would be neat. I should get motivated and give it a try (although only with simple ones for now).
Or putting the star on a drill with some grit under it and using it as the grinder, making the star and the surface on which it mounts grind to matching curved shapes.
Or, heck with it. Maybe it’s just this one flashlight I got that has this problem?
Anyone else checked whether there’s a flat plane on which to mount the LED, and good contact, in their lights?
I’ve got a second one to check here.
I should have specified not viable in lights that don't have removable pills. I think it would be fine for typical pills, but the softening threads could be an issue.
I think wight is correct that it will probably be fine to do without annealing. I like the idea of an "anvil" set for flattening pills. Probably would be very popular item around here. I find pills are not flat much more often than flat.
> an “anvil” set for flattening pills … would be very popular item around here.
> I find pills are not flat much more often than flat.
How do they do this sort of finish work in reputable machine shops?
It’s commonly done — cutting holes into a piece of metal from opposite sides,
and ending up with a flat plane surface in between the two holes.
Does it take a couple more passes with a grinder and finer abrasive, before anodizing?
Or put it into a hydraulic press and anvil, stamping the piece flat where it should be flat?
This can’t be a new problem in metalworking.
Is it something they’re not doing in China yet?
Or are we getting the discards, and some other buyer getting the properly finished work …?
Reinventing the wheel, reinventing the plane, reinventing everything. What’s the problem here?
Easy enough to check for proper contact, just slip a bit of carbon paper in and rub the parts and look to see where they did and didn’t touch. (Does anyone still make carbon paper?)
^ Good question. Maybe a lathe guy or machinist knows the answer.
I think you just reduce the cut depth for the facing operation. A lot.
So for this light we’d
Remove the LED
Remove the driver
Get an “anvil” with a good enough flat face that would fit up inside where the driver came out, and rest against the plane surface.
And put it in a press with another “anvil” held at exactly the same orientation down through the front
And squeeze.
That’d work if the metal flows enough under pressure (if it’s close enough to even thickness all across, just ‘dimpled’)
But if both surfaces are a bit concave — it won’t.
I think I’m just going to goop on more goop….
No need for it to flow. If both sides of the shelf are concave, we use a convex anvil on the bottom and a concave anvil on the top (LED) side. Right?
it’s way beyond anything I can manage.