Emisar D18 introduction

LED Flashlights – General Info 18650 Flashlights
361

I’m not sure, I suspect this is way better. Just the 2S2P clunky battery carrier on the MF01 compromises it, as well as the size and weight, which is on the limit of usefulness.

I say that as an owner of an MF01, which has given good service, despite it’s quirks and some essential fixing-up required (fortunately mine was not glued up). This could be a worthy replacement

Just quickly slip 3 (or fewer, in an emergency) parallel cells into this, 350 or 650 long, All three cells perfectly balanced from the charger, and when removed.

No unwieldy heavy brass carrier to struggle with, three parts to drop, plus cells, only two hands, and multiple opportunities to get them in the wrong way around, in the dark, all those springs, studs, plastic rings and multiple contact points in the way, nevermind trying to prise out the cells from it without damaging the wrappers. I do not like battery carriers.

It even looks coat-pocketable with 18350 tube, balanced by some spares in the opposite pocket. The MF01 isn’t really, more a rucksack pocket device. Or a large brick when swinging from a wrist lanyard.

The fundamentals look good to me. Will be following early-adopter reports with interest particularly how well the new Anduril thermal control firmware performs in use.

362

Agreed, carriers suck. Milled tubes are much better and typically a solid piece with the head for much better heat displacement.

According to the GB it seems that the MF01S will also run Anduril so they’re even in that dept

The MF01S has a clear single piece optic with auxiliary LEDS which is a big benefit many want

If hank does produce the 18350 tube that would be a smash hit with BLFers

I definitely trust the quality of Emisar over Astrolux. Astrolux is good but Emisar is great

363

Texas_Ace has said the MF01S isn’t responding to Andurils thermal stepdown properly so it looks like it will use NarsilM only.

364

+1 for the D18

that makes more sense. as I wrote my previous msg I initially wrote D18 Anduril and MF01S not but then double checked and was like… But I totally thought… where’d I read that…

thanks for confirming!

365

Don’t get me wrong, NarsilM is great compared to the crap they have on my MF04 but it’s no Anduril

366

So you don’t mind it’s slight turn off delay?

367

Deleted.

368

After much thinking, I finally made up my mind on not buying this flashlight, because I have already seen, how much heat can be made by 9 pieces of SST-20 emitter. The head of D18 is just a bit bigger, so I cannot expect much more sustained output from this flashlight.

Still, it is very good to look at this flashlight (it is perfect craftsmanship), so if I wouldn’t have been purchased ROT66 earlier, maybe I would consider buying it - but… I want this with much bigger head (20mm Yajiamei TIR optics for each LED /45kcd for 95CRI SST-20 is not that much, 100kcd+ would be better/, more separation between LED plate and battery compartment, and bigger fins), preferably with a buck driver, and with 4-6pcs of 18650 battery instead of 3pcs :slight_smile:

That hypothetical model (let me call it D18S) would lure me into buying it :partying_face:

369

If the sensor is calibrated correctly, 70 C. But the user can miscalibrate the sensor on purpose to get a higher temperature limit.

I set it to a pretty conservative level in order to make sure the thermal regulation would have something to do… since that was the point of the test. During actual use, people may want a higher limit. The level it stabilizes at depends on a lot of things, including the ambient environment, and yesterday was the warmest day so far this year. On a colder day it would generally run at somewhat higher power. I’m also guessing it’ll be somewhat different on a real production model, since what I have is missing parts.

Perhaps I should test a M43 too, for comparison. It’s a bigger light so I expect it’ll stabilize somewhat higher, but it’ll still be interesting to see. Charging some cells will take a while though, so don’t expect it today.

370

Just ordered one.
Went with sand color and the 5K emitter.

371

Here are the graphics from the distinguished Maiden666:

https://www.forolinternas.com/viewtopic.php?t=13622

372

I tried to do a similar thermal test on my Meteor M43. The results were … odd. It seems to max out at about 4800 lm, and it wasn’t very stable. Also, it got a lot hotter than the D18 even though it’s theoretically set to the same temperature limit… the manual says 50 C. There’s no way to calibrate it though, so to get a comparable result, I’d have to re-run the D18 test with a higher limit. I think I’d have to set it about 15 to 20 C higher to get the same temperature.

As far as I can tell, the D18 was actually about 45 C. It was configured with a 50 C limit, and it regulates inside a 10 C window below that, so the average should be about 45 C. The Meteor, however, seems to have been about 60 C or 65 C. It hurt to touch. I’m guessing this is due to the usual attiny85 random calibration issue.

Anyway, here was the Meteor’s result:

I think the average was about 3300 lm… but it was much hotter than I wanted.

373

:smiley: I’ve loved Anduril since I got it! :open_mouth: I didn’t even notice the turn off delay until you mentioned it. :weary: Now I have a new, not too significant but annoying aspect of Anduril…

I still love it but that is strange and annoying now that I notice it everytime!

374

Which would bother you more?

  • A 0.4s delay to turn off after a single click
  • Turning off and back on every time you do a multi-click action (such as click-release-hold to ramp down)
375

Is the UI supposed to be able to find an output level that is sustainable or does it simply drop to a pre-determined output once thermal regulation kicks in?

I got my PL47 recently and it appears to be the latter. I used a temp gun to calibrate the temp properly, set the thermal reg to 50C and ran a test on turbo. My PL47 does indeed reduce output at 50C but that’s all it does and the output it changes to isn’t anywhere near sustainable. Rough numbers:
initial temp 24C
1:20m/s on turbo it reached 50C and stepped down
Rather than stabilize temperature continued to climb significantly
1:00 later on the step down output it was reaching 64C before I shut it off

I see no way through Anduril to choose the output level thermal reg drops too.

376

A slight turn-off delay is soooo much better than a double-click flash.

A slight turn-on delay is soooo much better than trying for moonlight and getting level 11 instead.

377

Please can you elaborate on this?

378

The latter would be thermal precaution, not regulation. And there’s not much science behind that. Almost as simple as timed stepdown.

379

When it’s at room temperature, does the temperature check mode blink out the correct temperature?

If the limit is set to 50 C, it should do that initial ramp-down before it actually hits 50 C. It triggers that as soon as the predicted temperature gets that high, not the current temperature. If it’s waiting until it actually hits the limit, something isn’t right.

Both, ish.

There are three sections of the ramp:

  • Top / Paranoid: When the predicted near-future temperature reaches the limit, it quickly ramps down to a “sane” level. Usually this happens significantly before the body temp actually reaches the limit.
  • Middle / PID: Normal PID regulation, but it won’t go above the “sane” level or below the “safe” level.
  • Bottom / None: No regulation. Not bright enough to need it.

I generally set the “sane” level by measuring what the light can sustain at about 50 C, and then doubling it. So, if a hot-rod light can sustain 2000 lm at 50 C, the “sane” level would be about 4000 lm. During use, turbo would ramp down quickly to 4000 lm, then it use PID to continuously search for an ideal level.

On some lights, the “paranoid” part isn’t there, so it only has a PID zone and a no-regulation zone. This is the case for anything with a reasonably low power-to-mass ratio. For example, the D4 build has a “paranoid” zone, but the D4S does not… because the D4S doesn’t heat up fast enough to need it.

The PL47 has all three zones, so it is expected to do a quick ramp down from turbo, and then it should do normal regulation. However, Fireflies never provided a dev host suitable for thermal testing, so the settings are largely based on guesses instead of measurements. I don’t even have a production model of the PL47. And for the E07, I wasn’t even involved… it just re-uses the PL47 driver without changes, so it hasn’t been calibrated at all.

The D18 development was done more cautiously though… like, tested and measured on actual hardware, adjusted, measured again, adjusted, measured again, etc.

380

From the factory, the attiny85 isn’t calibrated very well. Its temperature readings vary within a range of about 25 C. So one light may regulate to 45 C, and another might regulate to 70 C, and they both think they’re doing the same thing.

This is the reason why Anduril has a temperature calibration function. It’s necessary in order to work around the hardware’s lack of factory calibration. That way, after the user calibrates it, it should be within a degree or two of the actual temperature, instead of being off by up to 25 C.

The Meteor’s manual says it should sustain 50 C, but mine seemed significantly hotter. I’m guessing this was due to the lack of factory calibration combined with the lack of a user calibration function.