TK's Emisar D4V2 review

No. Changing the ramp ceiling sets the double click from off ouput, but not the double click from on (turbo) output.

As far as I know, changing the turbo output requires a small code change.


went back and found this from the OP:

Official Specs

Aux LEDs: Red, yellow, green, cyan, blue, magenta, rainbow aux LED options is selectable through config UI, as well as the voltage indication function.

agreed, on intl outdoor there’s no mention of a white, but it is mentioned a couple times in the OP. Guessing the pictures/video/gif shows there are only 6 colors, not 7.
If that’s the worst thing to pick at on the light, not too bad I’d say.

With the target market for these being giant flashlight nerds, I think the tint bins should be listed on the order page.

Off may be a Color too?

Blacker than Black. :wink:

Like with Video.

Apologies for the long comment… lots of questions to answer.


—— memory ——

TBH, I find manual memory annoying. It means the brightness resets every time I turn the light off/on, instead of remembering where I was a moment ago. But it was requested a lot, so I added it.

Yes, I think so.

I’m thinking of maybe changing some button mappings at some point though. Like, maybe 5 clicks for config modes and then 4 clicks would be available for things like manual memory. Not sure. This isn’t likely to happen soon though.

Okay, that makes sense. Things started out that way, but a lot of people complained. So I changed it to not remember shortcut-moon and shortcut-turbo, and people in general seemed happier with it.

Basically, people mostly voted against “true memory”… so it was changed to something fancier. I don’t think anyone has asked to get the old method back before, but reverting it back to true memory is a pretty small code change, if you are able to reflash it.


—— reflashing ——

Yes. It’s literally the only way to flash firmware on this light. There’s a picture of it in the review.

A couple of Lexel’s earlier drivers didn’t allow reflashing without unsoldering anything, but the issue was limited to only his drivers, and he fixed it after I complained. But Emisar’s lights are completely unrelated to any of that.

Not yet. :frowning:

I’m really hoping someone will make these commercially available, but it hasn’t happened yet. I’m not sure why. Someone really needs to sell some reflashing kits.


—— configuration ——

The ceiling level is configurable, but double click while on goes to full turbo. To change that level, you’d need to modify the firmware.

On the Nichia 219 model though, turbo is limited to 75% power. This is to reduce the risk of the light damaging itself.

Yes. Go to the mode and brightness you want, turn off the light, then activate momentary mode. That’s the mode and brightness it’ll use during momentary.

The supported modes include ramps (smooth or stepped) and the entire “strobe” group.


—— aux LEDs ——

Originally it was:

  • Purple: full
  • Blue: nearly full
  • Cyan: high
  • Green: medium
  • Yellow: low
  • Red: very low
  • Off: critical

… but Hank felt that was too complicated, and the color mixing wasn’t very good on the prototype, so he asked me to simplify it to just blue/green/red to match how the Meteor’s voltage display worked.

It’s not very white. It’s red+green+blue, or … kind of an ice blue. Ish. I’m not totally sure, since the resistors on mine are different than production lights, so my color balance is a bit different.

The animation shows the rainbow mode… which includes the first 6 colors but not “white”.

These are the “colors” available:

  1. red
  2. red+green (yellow)
  3. green
  4. green+blue (cyan)
  5. blue
  6. blue+red (purple)
  7. red+green+blue (white-ish)
  8. rainbow (cycles first 6 colors)
  9. voltage

Regardless, if you want a white aux LED mode, I’d recommend just using moon mode instead. It makes more light and uses less power.

Only one of the aux LED channels has PWM. The others would have to be emulated. This makes smooth color-mixing pretty impractical.


—— misc ——

Nope, the bezel threading changed.

The tailcap seems to be swappable though.

Ah. Yes, it has a lot of extras. OTOH, the only things people need to know to use it are: Click for on/off, hold to change brightness. Everything else is optional.

This light is the definition of a pocket rocket.

There are more safety features in this than the original D4, like setting a default ceiling level of only half power instead of full power… and muggle mode, for lending it to others. But I’d still recommend being careful with it. In particular, lock it if you’re going to carry it anywhere the button could get pressed.

The original D4 lockout function was 6 clicks, and then it blinked slowly to confirm. It took a few seconds to lock or unlock, and it was necessary to unlock the light to use it.

The D4V2 lockout function is 4 clicks, and the confirmation blinks are much faster. It now takes about 1 second total instead of a few. Additionally, it gives access to two low modes during lockout, so it doesn’t need to be unlocked for quick tasks.

Or a slight turn of the tailcap can lock it too.

To be clear, it’ll still get pretty hot in turbo mode before it settles on a sustainable level. People mostly have wanted to make sure turbo will run for at least a few seconds before ramping down, even if that means getting hot… so instead of making turbo step down after like 2 seconds, it stays there for a little while.

But in general, aside from that initial peak, it should stay at a pretty comfortable temperature.

That’s going to take a few hours to measure. I suppose I should do it though… charging up a cell now.

DEL made the driver. He’s not active on BLF any more, but he has made the drivers for several of the most popular lights around here. For example, the BLF GT driver is pretty fantastic.

You certainly aren’t wrong about the light activating too easily. I would prefer a recessed switch too.

Where I disagree is that I don’t think a recessed switch would make sufficient difference. My D4S has a recessed switch and it still gets turned on in my pocket occasionally. I have had pretty much every light I’ve ever carried do that at least a few times, which includes the FW3A, ZebraLight H600Fd, Fireflies E07, Thrunite TH30, and many others. The only lights I really trust never to do that are twist-on types. (And I never to get stuck using a modern UI with a twist-on switch!) Some switches are more prone to this problem than others, but with any light that really heats up, even once is a serious safety hazard.

So I have come to the conclusion that some form of lockout is necessary no matter what kind of switch is used. Given that, the lack of a recessed switch is a minor disadvantage, but not a critical one.

Just a different perspective.

I have a D4 v.1 with Nichias (and no screws) where one has desoldered but will light up a bit if I take off and replace the optic, for a moment. I have a replacement LED although removing the MCPCB and reflowing the existing ones may get it up and shining. I have a daunting task. I just don't want to put it away in a drawer. I've reached out for help to some here in the forum but any recommendations are welcomed!

Gratitude to Hank, TK, & DEL!

Thank you so much for all of the time you’re putting into this thread, in addition to your amazing work on the UI!

Based on the responses to my post there are at least a few of us who want to buy a pogo-pin adapter. I will keep my fingers crossed and hope that someone will take pity on us and build a few for us. :-}

It does seem that more lights are going the pad route for flashing. I can hope that there will eventually be some kind of commercial product to do this. I just don’t want to wait that long.

In my experience, almost every sideswitch EDC light will occasionally turn on in the pocket. However, improved design reduces this occurrence. My experience:

  • Emisar D4v1 stock - accidental activation maybe once every few weeks. The switch protrudes, but usually requires a fairly firm press. Note however, that my D4s’ switches are inconsistent. On 1/3 of them the microswitch seems slightly damaged and activates with a much softer press.
  • Nitecore Concept 1 - accidental activation occurs immediately, 100% of the time when the light is placed into the pocket. (worst offender).
  • Zebralight SC63/64 series - deeply recessed button greatly reduces chance of accidental activation. I think I got an accidental activation about once per year with these.
  • FW3A - averaging an accidental activation about once every few weeks. Big exposed switch which can activation from multiple angles is the problem.
  • DQG Tiny IV 18650 - no accidental turn ons. Tiny flush button is quite resistant to accidental turn on. Button is hard to find by feel though.

So do I find lockout necessary? … no, except for the D4 v1 and the Concept 1.

  • Occasional pocket activation on non-turbo isn’t a big deal. I just turn it off when I notice and I’m good to go.
  • On the D4 v1, the risk is the button getting pressed and held by my pocket stuffing. Even if I turn the light off in moonlight it will ramp up to turbo and burn through my pocket. I roasted 2 pants pockets this way. Fortunately, the D4v2 features Anduril so I can set max ramp just below burning pocket level.
  • The Concept 1’s ridiculous protruding lever switch is so sensitive it triggers 100% of the time when put in the pocket. Definitely unusable without lockout.
  • For the FW3A, Zebralight, and my modded Emisar D4v1s with Anduril I do not consider lockout necessary. Even if they accidentally turn on in the pocket max ramp is set below “burning pocket” level. I think there’s little to no chance double-click turbo will spontaneously activate inside the pocket after it accidentally turns on.

I would have preferred a ring around the button on the D4v2 to reduce the frequency of accidental activation. I will likely mod the switch on mine to add a metal switch button and/or raised ring around the button to reduce this issue.

But my goal isn’t to 100% eliminate accidental activations…. just reduce the frequency.

Reflowing can be done on a hotplate :slight_smile:

The wires on the D4v1 aren’t so bad to remove though, which is nice.

[/quote]
Reflowing can be done on a hotplate :slight_smile: How to reflow solder an LED emitter on a PCB or MCPCB. Cree XML2, XHP70, XHP50, XPG2, XPG3. - YouTube The wires on the D4v1 aren’t so bad to remove though, which is nice.
[/quote]

I would like to add mounting screws too but perhaps keeping it simple would be fine.

Thank you Oweban!

Okay, so it wasn’t very interesting. The sustainable output at 50 C was higher than I expected, so it basically dropped to ~800 lm and then just slowly descended as the battery drained. The only interesting part was that it bounced around a bit before settling at that level. So I’m not really happy with how the firmware handled it, but the hardware did a nice job.

Also, it took 6 minutes to heat up enough to even respond. And that’s with a totally full battery, which is the worst case for heat. A lower battery wouldn’t make as much heat. So during normal use I doubt this will really be an issue.

In any case, it looks like I still need to make a thermal simulator so I can improve the algorithm without spending days at a time just running tests.

Great review with a lot of information. And great job as usual! :+1:

This D4 V2 has it all in such a compact shell. I know i won’t be able to refrain very long from ordering yet one more Emisar light.

Regarding EDC and accidental activation with the inconveniences it could involve, i’ve carried a 18650 D4 since it came out (in v2) and got used to lock/unlock it with six clicks. It is kind of a minor pain though and i’m glad this new one allows a somewhat faster ritual.

BTW: i thought the D4 was already in its second revision… so unless i’m mistaken that should be D4 V3?

v1 did undergo some minor revisions over its production course, including:

  • widening the center hole in the shelf for the wires
  • adding screws to hold the star down
  • very slightly shorter overall length
  • (in late v1) driver switch to use different MCU with pads for flashing.

However, all of those were still v1. The D4 still looked the same and was marketed the same.

This new version has external differences and more changes. It is advertised as v2 and says “D4v2” on the tailcap.

They should be SST20 J5 FB4.
Just take the 3000k version, the green is less noticeable mixed with the yellow. :smiley:

I have the impression the large MCPCB helps with heat storage, no?

That might be why it handles a bit more heat.

I hope the wait for the D4s-v2 is next and D1S-V2 too

Hank told me the sst20’s have been different batches since they use a lot so YMMV