Ran the 219c D4 on the 100% 7135 mode overnight.
To put the efficacy in perspective, the Astrolux S41 (219B) gets about 60 lm/W on the ~100 lumen mode according to my old measurements with an 18350 Efest.
what emitter was this done with please?
Tint graphs for the 219c D4.
Here’s the tint shift in different parts of the beam on 100% 7135.
Tint in different brightness modes.
My unexpected Meteor measurement got me curious, so I ran some tests…
On turbo with full cells, it measured 4848 lm in my light box. Significantly less than the 7450 lm it’s rated at for XP-G2 3D emitters. Then I did a runtime test with it on turbo in cold water.
Afterward, I recharged the cells and ran the same test on my D4-219c. Would have used XP-G2, but that one is my dev host and is basically a floating head with the brains floating outside its body. So, 219c was a lot easier to test. I ramped it to 1212 lm in my light box (1/4 of the Meteor, since it has 1/4 the battery capacity), then put it in cold water and started the test. Well, as close to 1212 lm as I could get it quickly, but it was actually about 1250 lm.
Here are the two runtime graphs:
I didn’t let either one run long enough to shut off completely… just long enough for LVP to give a warning. After each test, the cells probably still had enough power to keep going for quite a while before the light would actually shut itself off.
Voltages:
- M43: 4.17V start, 3.10V end
- D4: 4.15V start, 3.17V end
Edit: I’ve been meaning to make a thing to combine graphs… so here are the two merged.
Note, this is nowhere near maximum power on the D4. It’s set to be 1/4 as bright as maximum power on my Meteor. It also isn’t directly comparable since they used different emitters. The Meteor’s XP-G2 emitters are significantly more efficient, and its driver should be too. But I hope this gives an idea of how the runtime graphs compare at a similar amount of lumens per cell.
EMISAR D4 change to 219B V1 CRI/R9:9080
moon
Very small
219C 4000K 9050、219B 4500K 9080、XP-G3 5000K 90??
I would like to show you that the triples and quads from Carclo have an element that is present on each optic as it seems some people are under the impression their optic has a defect, it is the way the optics are being manufactured. This can also be seen in the datasheet.
Running above the 100% 7135 cuts down on the efficacy 20 %.
That’s odd. You first had 2818 lumens at 3.83V. There’s some funky stuff going on here…
Doesn’t this depend on the output? The higher the output (you chose 370 lumens), the bigger the PWM contribution of the FET, so efficacy should go down further?
thanks for the review and for your work 
I agree on that, basically it’s a linear combination between the 7135 and the fet, so I expect the efficiency to lower continuously down to the direct drive efficiency
Thanks for this!
Thanks to all involved in the creation of this little monster …
Arrived today:
I will definitely use it with Sanyo GA( Is more than enough for me) . It is a hand burner in 15 seconds, even after the step down …. I will have to reset the “thermostat” of the flashlight 
I love it !!
D4EVIL inside indeed! I like it
I think the light’s dark front area may have eaten some lumens compared to a light which doesn’t have that. After recalibrating for that effect, I think it may have been doing 3241 lm instead of 2818.
The graph a few posts ago (with 4848 lm) was done after recalibration, and both lights were measured with the same setup to keep things consistent.
I wanted to see how the shape of the output curves compared at a similar level of power per cell when thermal regulation wasn’t a factor. It seems the Meteor is partially regulated at first, then drops off starting at a fairly high voltage. This continues until it gets down to a pretty low level, and then it blinks to indicate low voltage. The D4 has more of a typical direct-drive (ish) curve until it detects low voltage, at which point it blinks and steps down to a stable regulated level. Although not shown in the graph, it then stays until voltage is low again, and repeats the blink-and-drop. After doing this enough times to reach the lowest level, it finally shuts off. Based on other tests, it looks like the first stepped-down level would probably last 30-40 minutes, extending off the right edge of the graph, and then the next level might take even longer, so I didn’t wait for those.
The Meteor has more area under its curve in this test, though it’s not clear how much of that is due to a more efficient driver vs more efficient emitters (D4 used 219c, which is significantly less efficient in this configuration). But while the Meteor was running at maximum output the whole time, the D4 was at only 1/3rd its maximum power. It’s a bit like comparing a commuter car to a race car. They are good at different things.
Got mine and a couple others today.
Man this thing gets hot in literally 2 seconds. May have to get the Xp-g2.
They look good together.
Highest it got was 128f.
figuring those O-Lights aren’t getting a whole lot of attention
I’d say they look absolutely fabulous together.
I honestly can not stop touching the D4. Fiddling, fiddling, fiddling… fiddle… fiddling, fiddling, fiddling… Hold on gonna get some coffee.
(1 minute later) I’m back. Fiddling, fiddling, fiddling…
I do like the simplicity of the charging. Fit and finish is great. Good UI. HAS NOTHING on the power of the D4 though.