Just did a little test underwhile browsing BLF. I put a fresh 30Q cell in my D4 and using my sphere I set the output at 550 lumen (my D4 now has 3500K 90CRI 219Cās so it is comparable to the stock 219Cās at about 600 lumen). I started hand-cooling the light. I must say I had a rough time, I had to swap hands every minute and it was still an unpleasant temperature, but it really was sustainable, after 12 minutes I decided that I could just go on like that and stopped the test. The output had dropped to 500 lumen then.
I did it again (start at 550 lumen) with a more depleted 30Q (3.7 V at start) and now it was a bit easier, but not much.
So yes, 600 lumen as the maximum sustainable temperature inside at room temperature seems about right. But without hand-cooling it will be lower (TKās temperature feedback makes it about 300/400 lumen, as read in CRXās data)
It were 4 - x mm longer and had a heavy shelf behind the emitter? (would a longer light be a problem?)
Is this telling us that this light is a great candidate for real cooling fins?
Not an expert here but are there other drives that can run the light with less heat (or brighter for the same heat) thermal efficiency?
I understand the meteor m43 has a more efficient driver arrangement, could something like that be a āgoodā option for the D4?
And My personal soap box, I say we need a ājust rightā sized light between the 4x18650 MeteorM43 and the 16340-18650 D4.
Iām thinking 26650 with 6-9 emitters. How much meat and fins and with what type of driver should it be specked with?
āBasically, thermal regulation is rather unpleasantly like trying to steer a fast-moving cow around a hairpin turn on an icy race track. While wearing someone elseās glasses.ā
I have a dumb question/idea.
If it is so hard to manage (thermal management of the X where the sensor is placed near the emitter) and as the gating interaction seems to be the ability to hand hold the light (as opposed to frying the emitter). What would happen if instead we just thermally managed the body/head of the light.
When Light gets X hot (subjective to the hand) then use that temp as the target temp. My hunch is this will make the management and both easier and more accurate.
Then all we have to do is try to steer a fast-moving cow period and (āaround a hairpin turn on an icy race track. While wearing someone elseās glasses.ā) goes away.
Hard to tell whatās optimal, but that is definitely an option.
What it DOES NEED is active cooling of some sort, so it can maintain an output of at least 2000+ lumens indefinitely.
Make it as small as possible and donāt forget high CRI.
I really like your idea. I always though that 3x18650 fit much more comfortably in the hand. The number of emitter I leave it to the experts, but 3x3 is 9. Sort of like an in-out burger (for those who donāt know their best seller is a ādouble-doubleā) so maybe a triple triple is what is needed for the Goldy Lux light. (pun intended)
I have always thought that 4x 18650 light were just a bit large to actually want to carry (sure it can be done) but out of our flashlight actual carry minutes what percentage is from 4x18650 lights? My hunch is it is out in the long tail of the bell curve.
Not my idea Mr. Ronin. This light will see daylight soon. How soon, that we donāt know yet. This light is the D7, and will be presented after the D1S IIRC. 3x18650 with 7 emitters it is, yay. :partying_face:
Better quality, (DQG) seems to have issues from what Ive read.
Why so small, just a bit longer and maybe real cooling fins and the light does not need to be throttled back.
more emitter choices, it just kills me to buy a new light and then spend a bunch of money and effort to go and change the emitters. better to pay for that up front but only once.
So yes A DQG but with Zebralight/Hank Build quality. Super well thought out (and responsive revisions) TK software not trying to be smallest (why do that?) I say shoot for most usable aka the light we grab whose use falls under the big part of the bell curve. And while Iām at it, in a real sand/earth tone (not gold or almost gold) color.
I see you know how to say the right things to the boys but how do I know your not just a tease?
Like they say, photos or it did not happen. But If Hank said that a D7 was in the works, now that would be interesting. and is the rumor of 3x18650 settled or is a more moderate/carryable 26650 light possible (Or is that a D6) Part of my reasoning for the 26650 is greater size/weight flexibility. With a sleeve or different tube (yaaa custom tubes like the D4) we can have a light that could run 1x 26650, 18650, 18350, 18500, 14500, 16430.
So the light gets designed with just āenoughā mass/fins (seems like the D4 might be a nearly perfect light but short on mass/fins to really use the HP under the hood) and then choose your power/weight configuration of the day.
It depends on the timing and whether the driver was in sleep mode when power was disconnected. This is entirely an analog circuitry issue. If the MCUās charge drains far enough, it will reboot (which counts as a battery change). If it doesnāt drain far enough, it wonāt notice the tailcap was loosened. The firmware canāt do anything about that.
After using Crescendo for a while, I kinda want different actions to ramp in different directions. Likeā¦ hold to ramp up, click-then-hold to ramp down. But that wasnāt something I changed in the D4. Itās something which will have to wait until I have a completely different code foundation written to make multi-click events easier to process.
Also not something I changed. It was specād that way. My preference would be that momentary mode is permanent so the only way to exit it is to disconnect power. Iād also prefer that it use the memorized level instead of always doing turbo. Maybe in some later UI.
On a FET+N+1, Iād probably want a reference blink at both channel boundaries. Also, some other input mappings. Like, I prefer battcheck on four clicks.
I suppose I really need to get that new code written. Itās relevant for several projects now, and Iāve been ā¦ distracted by other shiny things.
Yes, though itās a mess. Look in the crescendo directory. I was originally doing this for crescendo before the D4 thing happened.
Nice. I wonder if RMM might be able to stock something like this. Ideally Iād want something more like the Olight Baton clip, deeper-carry than the B205, but the B205 seems to be the best fit for the D4 out of the ones listed on that page.
Nope, its arms are too wide to fit in the clip groove.
Itās a slightly bigger light, and its highest mode without thermal regulation is 312 lm. So, Iād say the D4ās ~600 lm steady state (at a moderate thermal ceiling) is pretty fair. If you want it brighter though, you can still set the ceiling higher or give it a better heat sink. Itās funny to watch what happens when touching a block of ice to the D4. Even just getting it wet makes a huge difference in the level it stabilizes to.
Some are trivial, others are not. Iām working on some new code to make all these things trivial, and to make new UIs easy to create, but itās not ready yet.
After getting the different current readings on my 20700 fitted D4, I changed it up some. The tube threads at the head are pretty short, so I moved the copper spacer to the tail end and made the necessary adjustments for it to all work. Now, with a better ground that is consistently repeatable, Iām seeing 17.6A out of the freshly charged Sanyo NCR20700B cell. This allows for 3915.75 lumens out the front, at start. The 219C is a greedy emitter, itāll take whatever you give it, up to a point.
I charged up and tested one of TomEās Sanyo NCR20700A cells, wow! 21.14A at the tail for 4391.85 lumens from the Emisar D4 with Nichia 219C emitters. Yowza! Do it get hot? Is it noisy at Mardi Gras?
Thatās basically how the thermal calibration mode works now.
The sensor is not at the emitters though. The sensor is behind the MCPCB, the shelf, and an air gap. Its measurements are rather lagged, which is the main reason itās difficult to steer. Putting the sensor directly on the MCPCB would make things much easier.
Infrared thermometers are cheap and good for various things, like cooking. For longer range measurements, look for a bigger distance to spot ratio. Otherwise 1:1 ones can be had for a few bucks shipped.
Thanks for the reply on the SC600. Mine is the Mk II cool white, so H2 is still around 670. No matter though, thanks for the confirmation. Still this D4evil is impressive.
but how do I know your not just a tease?
