Wait.
Don’t use arctic alumina adhesive!!!
It sets too fast for this kind of build. And the strength of the bond gives you no room for error incase you don’t center it right on your first try. Use a non-adhesive compound. It is more forgiving, and it will also probably perform better.
Also, bonding the pcb to the pill before soldering the wires makes soldering the wires more difficult. The pill acts as a heatsink (because that’s what it’s supposed to do) and wicks heat away from the solder pads and it takes longer to get them up to temp. It’s better to solder the wires on the pcb before you mate the pcb to the pill.
If you’re still trying to put a 10mm pcb into the drop-in, I think you should give up on it. The hole in a XML P60 reflector is about 7mm wide and the rim of the hole has to sit flush or nearly flush with the pcb for it to focus properly. The solder pads on a 10mm pcb are barely even 8mm apart. There’s just no way that using a 10mm pcb with a XML reflector will give you good results.
If the wires are connected to the driver, there’s not really any other way to have the PCB, other than sitting in the pill, unless you are using unusually long wires. Then trying to smear thermal paste under it will result in white paste getting all over everything. I’ve never once had any problem soldering the wires while the PCB is pasted onto the pill.
It’s certainly a difficult one, but can be done. You need to get the solder connections as clean and low-profile as possible to allow it all to screw down properly, and ensure you cover them with something (thin) to avoid a short…
Can you tell me what is a “non-adhesive compound”? Is Fujik such a thing? Or something else.
I do have one other emitter+PCB in-hand now, a “Cree XP-E, 16mm Star MCPCB, R2, WG tint” one, but earlier, I think that folks said that the driver I have (the original one from the drop-in) would source too much current for the XP-E emitter.
Actually, I think that the same was said about the Nichia 219 (on the 10mm PCB), but I had gone ahead and tried it, because it seemed that, electrically, it’d have a better chance of surviving with the driver that I have.
So I guess I’m still a little confused:
- Attach emitter PCB to pill 1st, or not?
- Attach the emitter PCB to the pill using Fujik (which I don’t have yet), or with something else?
BTW, I was able to get the reflector screwed into the pill more today. I had to use an angle pliers to hold the pill, and turn the reflector by hand. Once it got past the 1st couple of turns it was ok, probably because the “lip” of the pill is what got messed up when I was trying to get the emitter off of the pill before.
And, BTW, that was with the original black insulation disk between the emitter and the reflector.
Now that I have the helping hand, it should make soldering a little easier :)…
Later,
Jim
Remove the old wires from the driver and use new wires.
Soldering the wires on the pcb first makes it a lot easier to get clean, low profile joints on the pcb. This helps minimize problems with not having enough room to clear the reflector. It makes it tougher to get clean joints on the driver, but those joints don’t have to be as clean.
Also, it’s how Dr. Jones does it, and many other experienced builders do it that way too.
You’re right, but those would have to be some extremely low solder joints and extremely flat wires.
If the product description says “thermal compound” or “thermal grease”, it’s usually non-adhesive.
Some common examples of non-adhesive thermal compound are:
Arctic Silver 5
Arctic Cooling MX-4
Fujik Thermal Compound
There are dozens on dozens of similar compounds.
If the product description says “thermal adhesive” or “thermal epoxy” or “thermal glue”, it’s almost always adhesive.
Some common examples of adhesive thermal materials are:
Arctic Silver Adhesive
Fujik Thermal Glue
The only difference between the names often comes down to just the last word, Compound vs. Adhesive/Glue.
Hi,
Ahh. Ok,
I may have not explaining what I tried, but the two things that you suggested,
(a) soldering the wires to the emitter PCB first before positioning the emitter PCB into the pill, and
(b) using something like arctic silver thermal compound
are EXACTLY what I had done the 1st time/originally.
I had mentioned that when I did solder the wires to the emitter PCB, I had a hard time doing that, because the pill kept scooting around the table, because I didn’t have anything to hold it. That caused the solder joints to be kind of “tall”. But, in any event, when I tried to screw the reflector back on to the pill, the emitter PCB kept moving around, because there was nothing holding it to the pill surface (just the arctic silver thermal compound.
As a result, the emitter kept ending up off-centered, which is at least one factor causing the really ugly beam I mentioned.
I think that the other problem was that the Nichia emitter is physically tiny, compared to the original emitter, and, in particular, it’s “shorter”, so when I have the reflector screwed down, the Nichia emitter doesn’t protrude into the reflector, but is more “below” the bottom of the reflector. I think that this latter “shortness” problem would even be a problem if I had gotten an emitter mounted onto a 16mm PCB. I’m not sure what to do about this one, maybe the Nichia’s because of their physical height, need a completely different reflector?
Edit: Here’s a white wall shot of my current “repaired” drop-in, in a 502B host (about a meter, on high mode, with the Nichia 219 emitter):
Is that ugly, or what :(?? It actually looks worse in real-life, plus it’s kind of dim.
Jim
Reflow that Nichia on to a 16mm pcb, and everything will be better.
Actually, I’ve been trying to get (backordered) one on a 16mm PCB.
But, I also just found this:
and maybe I’m having too high expectation for the Nichia (see post #9). According to those test, the flux with the Nichia was way low, compared to the others he tested.
Plus, although I didn’t mention it, and you probably can’t see it from the picture, the beam is very very yellow/orange. I just got a light with a Xenon lamp in it yesterday, and to my eyes, the Nichia color was just a little less warmer than the Xenon, which makes it still quite warm.
So maybe what I’m seeing is about the best that I would be able to do with a Nichia, and possibly should try some XM-L (L2, or L U2/3) instead (which I’m still waiting for). In other words, maybe it was my bad choice to try the Nichia, not just because of the size of the PCB, but just the emitter itself was probably not the one that I should have been trying.
They are B10 (100-110 lumens at 350mA) brightness bin emitters being compared to B14 (140-150lm), K1 (130-140lm), R4 (130-139lm), and R5 (140-149lm) emitters. Of course they won’t be as bright.
Their similarity in color temp and CRI to a natural light source is their main appeal. They’re the wrong emitters if you wanted max output and a cool color temp.
It can look a lot better than that. Not on a 10mm pcb though.
Sounds like it.
I appreciate ALL of the answers and the patience!!
This has all been quite a learning experience, and, even though what I ended up with, even to me, looks terrible, I am still amazed that I was able to resurrect this drop-in, and even improve it a bit from when I started working on this.
It probably doesn’t sound like a big deal to you all, but it IS, to me :)!!
Thanks,
Jim
Hi,
Going off in a different direction, because I’m impatient: I also got an XP-E:
and was wondering, if I replaced the 219 Nichia with the XP-E (on a 16mm star), what do you think?
I did some testing last night, powering the XP-E directly off of a Li battery, and it looked really bright. How would that work out in the P60 drop-in I have?
Also, assuming that’s a “go”, still a little confused about fixation vs. no fixation of the star. Seems like I got feedback both ways. I do have thermal compound now (arctic silver), but no Fujik or Arctic Alumina yet.
Also, I got my helping hands in yesterday, so soldering should be better, I hope :)!!
Thanks,
Jim
Hi,
FYI, I went ahead and replaced the Nichia on a 10mm PCB with a XP-E on a 16mm PCB. I eventually did this because I accidentally “sliced” the Nichia dome when I was messing around, trying to center the emitter better.
You all were right, it was a heck of lot easier to do with the 16mm PCB!! Not only that, but with the 16mm PCB, the original black insulation cover was able to fit. So, I now have a working drop-in with the original driver and an XP-E emitter :)!!
I will, for now, declare this a semi-successful operation :)!!
Here’s a white wall shot with the “new” drop-in:
I’m not quite sure what those horizontal lines are in the picture. I’m wondering if they could be because the insulation cover is not correct for an XP-E emitter (the hole in the cover is a lot larger than the XP-E emitter, and I think that XP-E dome should still be raised a little)?
Horiz lines in a beamshot like that is usually PWM showing up. I assume you had the light on high during that shot? You usually don't see PWM on high (but sometimes do).
-Garry
Yes, it was on high mode. I kind of thought about PWM so I confirmed that I was on high mode. Strange, huh :(.
Would over-driving the XP-E maybe cause something like that? The original drop-in had an XM-L emitter, where I now have an XP-E in it, so I’m just throwing out stuff, because I can’t explain it (not that that means much :laughing:.
I don't think overdriving an LED will cause that. Have you checked tailcap current with this new XP-E setup? Curious how much you are overdriving it.
-Garry
No, I haven’t yet.
I mentioned this in another thread, but when you all say measure tailcap current, is the tailcap actually in the circuit?
What I mean is that when I measured current, I connected from:
the tail end threads of the light
=> meter
=> - end of battery =battery + end of battery
=> (the spring on the) driver in head
i.e., the tailcap was not actually in the circuit.
When you measure “tailcap current” do you connect it differently than above?
Thanks,
Jim
garry,
I was about to try another white wall shot, and checked that the light was on high mode, and, just looking at the screen (I was using the camera from my tablet), I could see the lines… kind of like when you look at a TV through a camera.
Weird!!
Jim
Hi,
Tried the current measurement both ways:
- The way I described above (no tailcap in circuit): ~1.57 amps
- With tailcap in circuit: ~.8 amps
Edit: So does this mean the tailcap switch is ~5 ohms (V/I = ~4 / .77 = ~5.19 ohms?
Is that high? Or low? Or normal?
FYI, the host was a SolarForce “Blue” L2x with the default forward clicky switch.
Jim