Not this time. However, if you get Narsil flashed onto it, you can use its mode-group UI. It works almost exactly like you described.
HAHAHAHA 
Well, I got a programmer and a clip and dove head first into firmware flashing…
I actually managed to get Narsil onto my D4! :partying_face:
It took a fair bit of head scratching and about 15 browser tabs, but I got it working on the first try!
TK: Thanks for the detailed review, updates on D4 UI tweaking, and your helpful flashing info. :+1:
TomE: Thanks for Narsil. I can practice with the D4 until the Q8 launches! 
Narsil is nice, no?
Be careful about heat though; Narsil doesn’t have the right kind of thermal management for a small hotrod light. That’ll probably happen soon though.
Yes, Narsil is sweet!
Thanks for the heat advisory. I set the stepdown temperature to the minimum by clicking just after the 5 second mark, but it still gets blazing hot while stepping down.
Yeah, Narsil’s thermal algorithm is calibrated for a much larger light. It doesn’t really work on the D4.
I tried an XP-L HI at 70 C … and nothing exploded.
Compare to the same thing with a limit of 45 C:
(I cut this test short because it wasn’t as interesting)
At 70 C it was certainly painful to touch for more than a few seconds, even with cold fingers. But it didn’t actually damage anything, and it maintained what appears to be about 1500 lm for the duration of the test until LVP activated.
The 45 C test looks like it stabilized at about 800 lm or so.
Oh, and earlier I stuck one in the freezer on top of some ice. After just 2-3 minutes the ice had stuck itself to the aluminum and formed itself into a matching shape… and I didn’t even have the light on. I was just cooling it off to check what the temperature readout mode would say if it was below zero. As expected, it told me the temperature was 252 degrees C. 
(that’s fixed now; it now treats below-zero as zero)
Edit: Remind me not to trust my phone’s light sensor. I tried the 70 C test in my light box, and it appears to have stabilized at about 2250 lm, not 1500 lm.
Madame, very interesting light your're into. Thumbs up. A GB would have a good success imho....Can we have your avatar engraved in the green version?? ;P
What kind of inaccuracy does the phone’s light sensor have? Is it indicating something along the lines of 40% when the output is actually 60%, or were you trying to treat the phone’s “lux” as actual lux?
I think she meant that actual values doenst matter, graphs are made to show how does thermal step-down works, without precision OTF lumens measurement.
Phone sensors can have restricted lux range, they can work well in normal conditions (for example 10-10000 lux) but over this accurancy may be bad.
I’m asking whether there is an inaccuracy affecting the shape of the graph.
The consistency of measurements with the ambient light sensor of a phone are helped a great deal by placing a diffuser in front of it, easiest is a sticking a piece of copier paper in front of it. Of course then you need to calibrate the reading (use one of the many ‘luxmeter’ apps), but you need to do that anyway.
(via my sigline you can find some basic testing that I (and others) did on that, it is link3 in the luxmeter tests)
(and someone has made a BLF luxmeter app, when I have time I will try to find the link)
Great work TK!
Zak made the ceilingbounce app and it’s been pretty good for me once I got it dialed in. I don’t use it for scientific lumen testing, but it’s great for runtime graphs/ballpark testing and will show differences in brightness consistently with the use of a diffusion shoe box.
I’ll reflash it with your updated D4 firmware when available. Will it be on the repository soon?
I’m guessing that 70c graph is as you have fixed the UI thermal step down now, this isn’t as it was out of the box originally,
I know I had to have answered my own question here just based on personal experience because if that’s not the case then you’ve done the test all wrong TK,
Proper way to test this light out of the box would be, simply walk out to your car, truck, motorcycle or horse, double click light, drop into front pocket, climb in said mode of transportation and buckle up for safety of course and then drive, ride or gallop very quickly just one block, Edit: Definitely no further than 600 meters, and then give me the results please, Now that’s testing the poor man way, low tech but extremely effective and unforgettable
Nice work TK, You are the Goddess of BLF, but what happens if you are offered a paying job.? Late nights.?
Also amazing observation with this light is how fast it actually does cool down, unusually quick so I believe…
Edit: Sorry TK forgot where you live so my test idea may not be practical or safe…
I have not seen a “BLF” luxmeter app, but Ceilingbounce is specifically designed for testing flashlights:
- It has a calibration factor to convert the raw “lux” read by the sensor to lumens in combination with an integrating sphere/pipe/shoebox, with peak and 30 second readings.
- It has a throw testing mode that gives candela and FL1 throw at peak and 30 seconds (plus 10 so you can sweep the light over the sensor and ensure you got the brightest part) given an “effective distance” calibrated with lights of known intensity.
- It makes runtime graphs, calibrated to the 30 second output = 100%. It also saves a CSV.
And more features are coming when I have time.
Thanks that looks interesting. Also T18’s shrink wrap idea. I totally agree I’d prefer if it had some knurling.
Here’s another idea for grip, Vape Bands!
Very grippy silicone rubber around 23mm dia. they fit the S2+ great.
Just one of these near the tail end of the tube works great for a tactical grip.
The D4 has a pretty thick tube: 23.9mm(spec from TK) I tried one on a flashlight head that measured 23.5mm and it fit no problem.
You can get a lot of different colors, and different widths (thin or thick) FT has tons, search “vape bands”:
Ceilingbounce works great for what it does. It has been immensely helpful for this project. But if I want to measure lumens, I have a calibrated light box.
It definitely beats trying to estimate lumens based on the relative placement of pixels in a graph.
In any case, I got 1500 by guessing 4300 for the peak value (not the 100% value) and scaling it by the relative position of the stable state on the graph. I got 2250 by putting the light in my light box during a similar test. The main difference was that the graph was fan-cooled, while the light box test wasn’t cooled at all. And one used a calibrated and reasonably trustworthy sensor, while the other used whatever cheap thing was good enough to auto-adjust screen brightness in sunlight.
There’s nothing wrong with Ceilingbounce. My phone sensor probably just has a non-linear response curve.
hey what’s it take to flash this thing with your new firmware? or preferably buy one with the blink at 100% and the memory?