[Runtimes] - Emisar D4V2 Luminus SST20 2700K 95 CRI

Today I publish runtimes for my Emisar D4V2.

Version I bought:

I bought the light in November 2022 directly from intl-outdoor.com

Please note that my shoebox is only roughly calibrated, so don't take these numbers as the absolute truth. Unless you know where to find a Maukka flashlight

18350 tests were done with an Epoch 1100mAh 9A IMR battery.
18650 tests were done with a Sanyo NCR18650GA 3500mAh battery.

All tests are done under cooling condition. Let’s specify it, it’s not just a fan. I push compressed air directly on the flashlight. It’s winter time here and the tests are done in a heated shed. Ambient temperature is 12°C (~54°F).
My gold was to compare the Emisar D4V2 with my Zebralight SC600w Mk IV Plus 18650 XHP50.2 NW.

Besides the size, weight and heat management, output/runtimes should not be considered in competition because one is a 2700 K light and the other a 4500 K light. I was definitely expecting more efficacy with the Zebralight. I matched the D4V2 output with the SC600w outputs the more accurately I could. This was not easy and never 100% perfect.

Let’s start with the 18350 at Turbo:

Turbo is… turbo, so no direct match with any SC600w level. This light does not manage heat very well. My compressor is not industrial and is limited to 2.6 SCFM@90 PSI. I needed to push it to the maximum, but it was not enough to eliminate the heat factor completely. Started at 1723 lumens. Still 1723 lumens after 30 sec. Gradual step down to 1357 lumens after 6 min then a major step down to 531 lumens.

18350/adjusted to 1296 lumens:

Starting at 1296, the light quickly steps down to 1270 lumens after 20 sec. Then 1270 until the 5th min, then 1244 lumens until reaching the 9th min. Major step down to ~500 lumens.

18350/adjusted to 706 lumens:

Besides the little hiccup at the end, the output was stable at 706 lumens for 25 min.

18350/adjusted to 346 lumens:

346 lumens, flat for 64 min, then steps down.

18350/adjusted to 154 lumens:

154 lumens, flat for 157 min (2.5 h), then steps down.

Now using a 18650 at Turbo:

This looks a bit messy. During the first test (in blue), I temporarily pushed the air to the maximum with the pistol and realised I was not cooling the light enough so the output went up again reaching 1600 lumens for a few minutes. That’s when I new this light was not managing the heat the same way the SC600w is capable. I did and uncooled test after (see further down). The second test (brownish line) was done at ~2.6 SCFM which was still not enough, but that’s the maximum . Tested #2 started at 1703 lumens. Reached 1629 lumens after 30 sec. Gradual step down to 1372 lumens for nearly 19 min, then a major step down to ~525 lumens until the 38th min.

18650/adjusted to 1289 lumens:

Started at 1289 lumens for 1.4 min then 1263 until reaching 29th min. Major step down to ~500 lumens until reaching 43th min.

18650/adjusted to 701 lumens:

701 lumens, flat for ~78 min, then steps down.

18650/adjusted to 347 lumens:

347 lumens, flat for ~182 min (3 h), then steps down.

18650/adjusted to 157 lumens:

154-157 lumens, flat for ~490 min (>8 h), then steps down.

Let’s compare the heat management: Emisar D4V2 vs Zebralight SC600w @ Turbo (no cooling):

Because a 2700 K emitter is normally less efficient then a 4500 K one, the light may produce more heat for the same level. Still, the difference is astonishing. For the SC600w, I also noticed that the cooled test has less runtime compared to the one without cooling. A 30 min difference. My electric car is also more efficient when the battery is hot.

Let’s play with graphs:

Next runtimes tests with my SC64w HI 18650 XHP35 Neutral White.

Thank you for doing the work. Looking at the graphs, I was wondering how normal D4V2 managed to regulate the high lumen output. Then I took another look at the first image and realized that the light has boost driver.

Tactical Grizzly’s review finds that boost driver is estimated to have about 25% longer run time at sustained output when compared to the normal (linear+FET) version.

Your results seems to indicate that Zebra light has about 40~55% more run time than D4v2 with boost driver where the difference in run times are larger for higher output.

Combining the above two results, it seems that the Zebra light would have about 80% more run time than linear+FET D4V2. I still very much appreciate the benefits of Anduril UI and high CRI LEDs, but it is very useful to know what the costs are in terms of efficiency and run time.

1 Thank

Very interesting graphs, thank you for the work you’ve put in!

1 Thank

My novice takeaway is that the boost driver significantly improves/lengthens runtimes, notably at turbo and high levels….CORRECT? I’ve stayed away from Emisar hot rods because of the huge output dropoffs……but these graphs look much better?

The booster is needed for regulated output. Maybe the constant current driver was able to, but Hank told me that the booster option is more advanced than the cc driver. What does that means in details ? I don’t know. No mather the config, if the temperature is building up, the protection will kick in and you loose any regulation.

I looks like there are some limits. Turbo is an exemple. For the SC600w, no matter how much air I am pushing, the output will decrease at the same time the battery voltage is decreasing.

Most runtimes you find around are done without cooling or just a fan . I did not want to show something you can find elsewhere. Both test (cooled, no cooling) are important IMHO. I mostly use my lights as ECD and for short period of time so there is no heat build up. I like knowing that after lighting my lamp 25 times for 1 minute, the intensity will still be X lumens. If you plan to use a D4V2 (or SC600W) for long period at Turbo, the heat will be a problem. Even in front of a bike in winter, the output will drop because of heat. 71 g of aluminum is no enough. No perfect light exist. Only compromises :slight_smile:

Perhaps I’m missing the point of tests that involve such cooling methods? Even tests using a fan for cooling aren’t very realistic in comparison to what sort of times a person could expect in real world use (e.g. holding it in their hand.)

Every kind of test can tell something different from a light. I'm not sure I want to hold a flashlight for 2 hours waiting to see the result though

Selfbuilt, once elegantly explained why he was using a fan to cool the lights during a runtime test. Really worth the reading:


From what I gather, if you buy a light with a boost driver, you will get:

  • Lower turbo brightness
  • Higher sustainable brightness
  • Higher efficiency
  • Brighter moonlight mode.

I’m really confused by the cooling method here too.

Compressed air is hot, wet and dirty… and not in a good way. 100% humidity and the compressor is like a space heater. It’s running for 2 hours?? I mean there are situations where compressed air for cooling makes sense but not here.

Why not just use a fan? Hundreds of CFM from a few watts, instead of the other way around.

Sorry if you’ve addressed this somewhere else already, I know this is a year ago. I just stumbled upon this thread and was like- wait…he’s doing what?? Had to say something lol

Yes, it’s an air compressor. A don’t know any fan that would produce enough CFM to cool the D4V2 on turbo. I like to see how regulated a level can be. “What would be the runtime graph if heat was not a problem ?” is my basic question. When I use a light on turbo, I never use it long enough for the heat protection to kick in so that’s why I like to know this. But I also like to know how good or bad the flashlight will handle the heat (compressor off) compared to others. The compressor has a min/max PSI so it’s not running all the time. I reduce the CFM everytime I test a lower level.

But 2.2 CFM is like a 1 watt fan

My compressor can only push 2.6 SCFM@90 PSI, but this air flow comes out of a little hole. Diameter is maybe 0.062 inch. This pressure is aimed at the flashlight head only half an inch away from that head. A fan area is way bigger and this pressure would be lost. A lot of the air flow would go elsewhere. The fan would do great at most levels though I’m pretty sure.