Review and disassembly: Sofirn C8F

Please don’t keep on mentioning Thorfire. These are Sofirn. And it seems not great. They have gone their own way, seemingly without understanding first principles.

Whereas Thorfire are (IME) absolutely solid. Every one. Unless you know differently ? PS: dipping the head in white spirit is hardly a fair test, not at all surprised that you damaged some things trying that.

That’s all.

I'll soon be testing the C8T, Tom Tom, and I'm more or less sure it'll work right. The issues with the C8F are a combination of different emitters (XP-G3 vs XP-G2) plus an inadequate MCPCB, this latter latter being the most hurtful.

With regards to my TA13, for some reason at least one of the inner o-rings is becoming damaged and I don't know why yet. We'll see.

Cheers ^:)

Okay, I just got mine with the XP-G2s (or -G3s, whatever). Hmm, looks like the whole square is phosphor-coated, so -G3.

Anyway, wow! Nothing wrong with mine. So, nope, I ain’t gonna be taking it apart. :smiley:

Seriously, youse can dissect yours to your hearts’ content, but unless there’s some pressing reason to do so, and risk shearing off any LEDs, I ain’t a-doing it.

Gotta admit, this is my first triple, and I was expecting 3 separate but overlapping beams, loads of artifacts, but this is nice. Looks like one tight beam for what’d be 3 grouped 20mm reflectors.

Okay, due to the G3, I got the expected color-shift in the beam. Neutral hotspot, cool spill, strong yellow corona. Would G2s be a better choice?

Nicely lubed tailcap threads, buttery smooth, not scratchy at all. Nice firm click of the switch. Guessing it’s easily over 1000lm with a 30Q, via ceiling-bounce tests.

On high, it gets very warm, very fast, so it feels like there’s a good thermal path in there. Anyone check for thermal goop? I forgot to look at earlier posts before writing this…

Anyway, I’m suitably impressed. I was expecting a not-as-tight beam as a regular big-reflector C8, but this is… nice! It’s kinda like a domed XM-L/-L2 in a C8. Quite a bigger hotspot than my XP-E2 C8, of course, but it’s still very usable. For more throw, I’d stick with an XP-L HI (or extremely high-strung XP-Gwhatever), but as-is, this trip XP-G3 is quite good.

Best of all, the UI doesn’t piss me off. I almost cringed flipping it back to low, almost expecting the blink-on-low to switch 3/5 modegroups, but… nothing! Low, medium, high, yeah, doubleclick for strobe, but usable modes!

Damn, I can’t find a single thing wrong with it. :smiley:

(Well, okay, aside from the misbehaving G3s as far as color-shifts.)

If you don’t have a Nichia 219C flashlight yet - get them.
4000K or 5000K doesn’t matter, these LEDs produce such a nice light.
Just read through the MF-01 Group-Buy thread - all of the 219C owners are blown away from that beautiful light.

I will make this and put a BG FET+1 driver in it to power it

L4M4, the onboard driver in this torch is quite good, I'd just swap the emitters and be done with it.

Cheers :-)

Just got my C8T and it’s a beautiful machined without flaws, beam is also great even better than Convoy XP-L Hi. But the best is how powerful it is !

4.2Amp on HI with LG HG2 and 3,89A after 30 seconds
Quick lumen reading shows
Moonlight 1,5 lumen
Low 67 lumen
Medium 430 lumen
Hi 1320 lumen !!! X

At 5 metters I got 71,5 kcd for 534 meters of throw.

I don’t have any plans for modifying this one but maybe AR lens ,and stainless bezel :+1:

Haven’t you used the light before disassembling? Did it work then?
You took the light apart and reassembled it with a bezel from a different light and manufacturer. I won’t be surprised if this bezel - due to different construction - is not able to press the MCPCB against the shell. You mentioned you couldn’t screw down the bezel completely because of a thick o-ring. But this might also be caused by different or too short threads.

XPG3 loves heat i have XPG3 doing almost 9amps and they get really toasty. This would be why they un-soldered them selves the stock board can not handle the current. The stock board may not be making proper contact with the light its self if its not fully flat on the bottom.

Just to inform XP-L hi version or C8T have DTP copper PCB 16mm with nicely spread silver thermal grease underneath,and two screws for fixing in place. Tint is also amazing around 5700K maybe V3 2B I would say, much better than U6 3A .

Even if the board isn’t perfectly flat - we are just talking about 2Amps per LED and evenly spreadheat transfer paste - that shouldn’t be a problem

XP-G3s are just……special. As are the XHP50.2s based on them. It seems like around 25% of them are just “weak” while the rest work great aside from the tint shift. At least that has been my experience.

I’ve had probably 6 XP-G3s fail, and 2 XHP50.2s. All of them the same failure mode, they work fine for a little while, then they go blue all of a sudden and cut out. Sometimes a reflow fixes them for a bit, but they have always failed again shortly after.

I’ve even caught a few as they went blue, shut the light down and powered it up and they worked again for a bit, but again, they always fail shortly after, so I don’t trust them.

All of my experiences are with emitters I have reflowed myself on DTP boards, and all showed good reflow upon replacing them. None of them were driven all that hard either, I’d say under 4A, usually under 3A.

After they fail they will usually still light extremely dimly if given voltage, and sometimes will kick back on if pressed on, so I suspect something going on with the internal bonding to either the thermal pads or the annode/cathode.

I want to like the XP-G3, but between the tint shift and the fact that they are the only emitters I have ever had fail without serious abuse, I’m probably going to mostly avoid them in favor of 219C.

Are those problems you speak of common to all of the newer generation Cree emitters, Jensen567? If so, failures with XP-L2s and XHP70.2s should also be common, I think.

Let's tell Cree we love bond wires and clear domes. :-D

Cheers ^:)

The dome is clear, just looks yellow from the top. No, seems to be isolated to the XP-G3 size die, I haven’t had a single XP-L2 or XHP70.2 fail like that.

I used it only for about 5 minutes, it was day and no point in using it longer.
And I used call that can’t give more than 3,5A because I had that near me.
Later i put HG2 inside and it died.
Bezel has even longer threads than original because I can screw it down completely with original o-ring. I used thicker one because I wanted even more pressure on MCPCB.

That’s exactly what’s wrong with mine.
1 works good and 2 dim or good when pressed.

Yeah, I’m not entirely sure what is causing it, but it seems to be fairly common with the XP-G3 and XHP50.2.

The fact that I have seen the exact same behavior from both of those leads me to believe there is some sort of manufacturing flaw with this die that has gone unnoticed by quality control at Cree.

Since they sometimes recover with pressure, my first guess would be something to do with bonding the flip-chip die to the external package.

Thankfully it seems isolated to only those 2 models, as the XP-L2s and XHP70.2s I have used are all going strong, even with some serious abuse in a few of them.

I own heaps of XPG3s never had this issue. I think i own 10 XPG3 emitters they all still work fine even after being driven really hard. What boards are your XPG3s on?

Various, but all DTP and mostly noctigons. I have probably more than 30 of these emitters I have used in various builds for myself and others, in that process I have had 6 failures. I have had 3 XHP50.2s fail out of probably 12 or so I have used. That is a relatively small sample size, but they are at least from various batches. I have also seen quite a few posts from other users here with issues using these emitters.

The ones that do work can take a beating, but some just die even with a relatively light loading. These are purely my experiences, so YMMV.

Looks like, after all reported issues, Sofirn decided to remove C8F from their store…

They are going to release fixed version soon.