Emisar D3AA is available now

I can see paracord wrapping the tube of this light being a popular mod. Could double as insulation to bump up the temp limit, too.

Here you go, sizes are of course somewhat limited for the switching (boost and buck) options but should fit many larger head lights.

Not that they theoretically couldn’t be such, but none of Simon’s drivers are FOSS. I’ll pay double for that alone.

Perhaps the added cost is due to Hank using higher precision parts to allow for ultra-low moonlight levels. I’ve noticed that Convoy lights usually have multi-lumen “moonlight” levels.

Another factor is the increased human eye sensitivity to light in the middle of the visible light range. 5700K should have most of it’s spectrum in the range the eye is most sensitive to. Warmer CCT’s have increasing proportions of their light further out into red end of the spectrum, where the eye is less sensitive.

Since the lumen is defined based on the human eye sensitivity rather than radiant intensity, this also results in a lower lumen rating at warmer CCT’s. This is intentional, because the idea behind the lumen unit is to convey the intensity as we perceive it, rather than in more typical physics units.

AFAIK the current Emisar-Noctigon FET+“X” and CC drivers, as well as the Lume1 all use Attiny1634. Don’t know if that’s true of the former’s boost drivers, but consider that there were issues with the very-low levels on the linear CC drivers until the firmware improvements that TK made.

Overall not really a huge deal to me, but will be waiting for some runtime/efficiency/stabilization tests before forking out the cash on a B35A or XHP35 HI version.

As my best walking light, this will be held for long periods of time (once it arrives, B35AM 3500k green-green); I wouldn’t mind the extra safety, or just being able to go hands free in an instant. I personally can’t really feel the original d4v2 lanyard hole unless I death-grip it.

Anyways… easy enough to attach a lanyard to the clip.

Is there enough material to drill a lanyard hole in the tailcap as in the d4v2?

The lowest level on this light is about 0.46 lumens, which is lower than any other boost-driver light I have used. Interestingly, the moonlight “jump-start” is much too strong on this light and causes a bright flash, even at the lowest level of 1/50. However my light is not using the latest firmware (2021-10-04) …I had to order another USBasp and can’t reflash right now, so perhaps things have changed slightly. It’s also very unstable on the lowest level, with a strong flicker that can even be seen in the video.

The max output at startup (1,640) is an average of multiple tests, the highest I recorded was 1,683 (immediately at startup). @ 30 seconds it is 1,610 lumens. Those numbers are not guaranteed to be super accurate though.


Also yes, the light was indeed calibrated, but the test was ran at the default 45C. I think the default temperature is the most meaningful, as it still gets pretty hot in the hand, but I’ll do tests at higher max temps as well. Bear in mind too the runtime test is literally done in a bookshelf @ 21C, so thermals are not best-case scenario.

And yeah… this light should have had a lanyard hole

I think 1600+ lumens at a CRI of 9080 in a throwy beam coming out of a single LED is amazing.

Funny though as I think most people still prioritize quantity over quality. Maybe this is subjective, but at one point I think 20% more lumens is a minor increase compared to the jump you get in clarity going from CRI 70 to 90+. There are a lot of bad tinted high CRI options out there, but enough good ones that personally don’t see the need to ever go back to a low CRI light.

On mine, I measured ~0.56 lm to ~1300 lm with a 50E cell. Moon was stable, but turbo wasn’t as bright as I expected. However, I haven’t measured with a higher-amp cell. It could just be a weak battery.

I see the moon pre-flash here too, and I wasn’t able to get rid of it. I think it’s a hardware quirk, because it shows up even with jump-start at level 1 (which is default, and effectively means no jump start at all). I should still double-check though, to make sure it still happens even when jump-start isn’t compiled in. As far as I can tell though, simply turning on the boost circuit, without even sending any PWM to the control pin, causes a brief flash. It’s sending a flat zero for the brightness level, and it still flashes.

For thermal tests, what I’ve found works well (simulates real-world use, approximately) is to rest the light pointed downward on a wire shelf, with a fan on “low” blowing gently toward it from about 1 meter away. I place a phone under it to log data, and I adjust the position so turbo mode reads roughly 50% to 90% on the light’s sensor. I’m using an old Nexus 4, and its sensor goes from 0 to 10,000, so I typically try to make turbo read roughly 8,000. This gives good resolution without clipping. Hooray for analog gain staging!

For narrow lights, I place a spare lens on the wire shelf, and put the light on the glass. That’s only if the light is small enough to fall between the shelf wires though.

The startup current overshoot is something that I have in all the boost drivers I’ve made, but it’s only really problematic with the MP3431 based ones, Loneoceans’ boost driver had this, Mike C’s too, he used a small FET to pull down the Op-Amp inverting input for a short duration during startup to solve this, when using a dual sense resistor topology for very low moonlight like I do this can be mitigated by waiting for a couple of ms before turning on the high power range FET.
Both of these solutions don’t have support in Anduril though, and it would need a hardware revision… depends how problematic it is in the case of the DM11, how bright is this flash ?

Edit : I missed Luxwad comment, so ”bright” it is.

For floody purposes, absolutely. For a thrower to lighten up something 100-200m away, I’m not sure. FF E07x pro, I would love some high cri, if it was 6000 lumen plus, but that generates heat. Ahrgh, it’s all a circle jerk with compromises!

Oh true, it’s weighted by a curve that peaks in the green region.

Hank, what currents are you using for the newer e21a tint ramping drivers? Are you still using only 1.25A/led by default as with the original 5A driver?

Can we chose to use 1.5A/led - 1.6A/led? What about with the upcoming DT8? Thank you.

The tint ramping D4V2/KR4 use 1.9A per LED, so, total 7.6A.
DT8 should be available in 2 weeks time.

Hank, How are you guys holding up? I guess that you met some pressure with the release of the tint ramping lights, plus the DM11. You must be working your asses of 24/7! Hope you get some rest from time to time! :slight_smile:

Hank, do you plan to offer this flashlight with more floodier optics, like mention in this review.

Thank you

You can always buy the floodier optic by sending $5 to contact@intl-outdoor.com, we will add it to your order.

Hi Hank, I’m asking out of curiosity… why don’t sell products on Taobao? I know you have a store that you can take Alipay but you don’t list any products. It seems like you have a fan base in China as well but they might have trouble figuring out how to order.

I’d be happy to recommend your lights to Chinese people I know but it’s too complicated for most.