TK's Emisar D4 review

LED Flashlights – General Info 18650 Flashlights
383

I swapped the XPG3 emitters and stock star for one of Richard’s offset starts with 4000K 90+ CRI Nichia 219Cs tonight. Since the new star is thinner I used a piece of copper sheet cut to size between the new star and the shelf with AS5 thermal grease on both sides.

Initial impressions:

  • The Nichia tint and CRI look really nice. Better than the S42B emitters the light came with.
  • It gets hotter a lot faster on the Nichia. Not too surprising given that the Nichia has lower vF than XPG3 and can pull more amps. Instead of lasting 20 seconds before rampdown, now it’s more like 10 seconds or less. If I decide to keep the Nichias in I’ll need to recalibrate the thermal stepdown so it turns on faster.
384

I didn’t get the impression it was harsh. I responded in kind of a hurry on my way out the door though, so maybe what I said sounded harsh? I hope not; at least it wasn’t intended that way.

Maximum cell length: 67 mm. It won’t work with anything longer.

Cell type: Someone measured 20 amps on turbo. A protected cell can’t handle that. In the good case, the battery will shut itself off. In the not-so-good case, a high-resistance protected battery could potentially overheat and die.

Cell configuration: Only one cell. If you put in 2x18350 or 2x16340, it will kill the driver.

Use only the recommended batteries.

It’s worth noting that a 3500 mAh high-capacity battery won’t generally run longer than a 3000 mAh high-current battery when used at high current levels. Effective capacity drops as current increases, and high-current cells are better at dealing with this.

The light has low-voltage protection, and will progressively step down then shut itself off when voltage is low. It tries very hard not to let you drain a cell far enough to damage it. However, LVP does not shut it down while in a blinky mode or tactical mode. So, don’t leave it in beacon mode for weeks at a time.

385

@ToyKeeper,

It’s worth noting that a 3500 mAh high-capacity battery won’t generally run longer than a 3000 mAh high-current battery when used at high current levels. Effective capacity drops as current increases, and high-current cells are better at dealing with this.

1) I like to use the same battery in the D4 as I will do in my USB charger, so a 3400-3500mAh rating for a high current battery will make sense.

Can you recommend any such 3400-3500mAh “Highest Possible Output” batteries?

2) Does this mean leaving it on in blinky or tactical mode could break the battery and turn it on fire after some weeks? :laughing:

386

It might not be able to step down much faster. The duration of the initial temperature peak mostly depends on how long it takes for heat to go from the emitters to the driver. It doesn’t depend so much on the thermal calibration value.

The extra layer of copper might delay the response by delaying the heat transfer a bit. I haven’t tested that configuration, but usually any interface point will slow down heat conduction.

One possible way to improve it would be to put a thin heat-conductive material between the MCU and the shelf, to replace the air gap. Another possible speedup would be to run a wire from the MCU to the MCPCB. Or possibly solder the two layers of copper together to reduce the interface effect. Or remove the extra copper piece and use a thicker O-ring, to eliminate the extra interface entirely. Or mod the firmware a bit by bumping up the prediction strength (but this might cause oscillations, and might make it act like a skittish child at a horror movie, jumping at every hint of the scary heat monster).

I don’t know what would work best.

387

What do you guys say about the “Keeppower IMR18650 NH1835 3500mAh” battery for driving the Emisar D4?

It’s 3332mAh, gives you “pulse discharge” current of 20A, and sustained 10A.

I don’t know how long a pulse is but if at 3800 lumen the Emisar D4 will step down in 3-4 minutes anyhow, then maybe the battery’s pulse discharge will be sustained for 3-4 minutes too so it’s a perfect match??

http://lygte-info.dk/review/batteries2012/Keeppower%20IMR18650%203500mAh%20(Black)%202015%20UK.html

388

If they are based on the Sanyo GA. They are on the Richard's list for recommended - runtime cells for this light. I would then buy directly the Sanyo. Unless you are looking for some kind of protection... But again beware the extra length added to the cell.

389

TheBo, ah you mean the Sanyo cell by the name “Sanyo NCR18650GA 3500mAh”!

That’s a good point, Test of Sanyo NCR18650GA 3500mAh (Red) shows it’s an equally good battery.

It can be bought here, yey!: http://www.hkequipment.net/product-p/ncrgax2.htm

And also here 1 Stück Sanyo NCR18650GA 3,6 V 3500 mAh 10A Wiederaufladbares Lithium Batterie Sale - Banggood Deutschland sold out-arrival notice-arrival notice .

Okay, I’ll go with those. I have only one last question - is it not correctly understood, that actually this “Sanyo NCR18650GA 3500mAh” will have as big output as the batteries listed under “Highest Possible Output:” on “Highest Possible Output:” also have??

Or is its output capacity relative to the Emisar D4’s need of 15-20AMPS actually lower than that of some of the batteries listed under “”Highest Possible Output:“”. Tried to compare charts at lygte-test.dk and couldn’t really see any difference.

(By the way for a side-comment, the Efest IMR18650 3500mAh is listed under “Highest Possible Output:” too, but the test on http://www.mtnelectronics.com/index.php?route=product/product&path=80&product_id=855 / Test of Efest IMR18650 3500mAh (Purple) 2016 reports “hardware problems” in those batteries. So stay away I guess.)

390

Ok. i hope i'll explain my self.

Richard has divided suitable batteries for this light in 2 groups.

  1. maximum performance
  2. best ratio between very very good performance and duration

NOTE: the batteries in maximum performance are IMR type, usually battery with the highest possible current delivery have low (or nowadays is more correct to say a bit lower) capacity than normal batteries. Since few years ago they were around the 2500mAh mark (LG HE2, Samsung 25R, Sony VTC5 etc..). Differently high capacity batteries at the time were Panasonic NCR18650 which were 3400mAh but the current they delivered was much lower. And they were suited for let's say for max 4-5A, maybe less. If you ask a battery more current than what it is designed for, battery can act in different ways, probably, voltage will sag (a lot, current will still be capped, so you wont' go any higher anyway, and you end with a lot of disappointment. Now since the 18650 is the most interested by evolution, the highest power delivery batteries are 3000mAh (LG HG2 and Samsung 30Q). You should find these in richard'sm maximung performance along with the previously mentione d 2500mAh type. This is nice because now the difference is not a lot compared to time ago with the highest capacity ones. Beware that there's also another important thing to notice. When you ask for a lot of current a battery will drop is voltage directly correlated to how much current you ask "her".

See this graph http://lygte-info.dk/review/batteries2012/Common18650comparator.php if you play selecting and deselecting currents you'll see that SAnyo 18650GA behave the same as Samsung 30Q at 1 A but Sanyo runs much longer, if you select 15A you'll see this behavior changes a lot .

Also the capacity a battery has varies with the current you ask "her" so ...a Sanyou 18650GA could well have real 3250mAh at1A draw, and a samsung just real 2900mAh, but the sammy retains almost 2700mAh at 15A draw, almost like the SANYO, but if you look at the 30Q the voltage in the samsung is much highe, thus giving you much power (power = Voltage * current)

Batteries in group 2 are still very nice because of what i just said, but since batteries with high current delivery have also less internal impedance you will squeeze till the maximum lumen just with group 1 batteries, althought with group 2 you will run longer, i don't know how much visual difference you will get to your eyes, maybe negligible, maybe not negligible... it's up to you, if every minutes count or not....what you use YOUR light for and what you planned to use YOUR batteries for...i think this is the fun stuff with this, and that's why people on this forum is so addicted and enjoys themselves.

Hope you got your explanation.

For the others, SORRY GUYS for this post.

TK , hank and Richard ..thanks for the light... I'm a bit under money now (tew thousand euros for the dentist) and i'm struggling with the monkey on my shoulder that slaps my face telling me to buy this in grey nichia....

391

I don’t think the copper spacer under the star makes a difference. If it weren’t doing a good job conducting heat, the result would be the head should heat up slower rather than faster. Instead, it’s actually heating up faster with the Nichia and the spacer than it was with the stock XPG2.

I’m not too concerned about the LEDs getting too hot. They can take the temp. However, I would like to be able to get a reasonable period of burst before the light ramps down.

I agree that probably the only way to speed the rampdown maybe to increase heat conduction to the heat sensor. The sensor is in the MCU? Then maybe some of that heat conductive foam (that unfortunately I don’t have) squeezed between the MCU and the back of the shelf. Alternatively, I suppose I could do something to lower the turbo output. Maybe replace the driver wires with thinner ones.

392

For those looking for a nice compact USB powered charger, try the Xtar MC1+ or the ANT version for charge level indicator.

A cheaper alternative is the Lii-100, not as well built and a larger in size.

I’m ordering the Nichia version and for daily carry I will be using it with a GA instead of a VTC6, because other than absolute max output there is no benefit in using a high drain cell with low Vf emitter. Also helps to control the heat on turbo runs by limiting the output.

393

Any chance for the high CRI R9050 version of the 5000K chip making it in anytime? This thing is almost perfect. A more recessed switch and a high CRI 5000K Nichia would get it there. I’d get 3 or 4!

394

@battery, is your point that there would be four CRI R9050:s, hence yielding 4 * 1200 = 4800 lumen?

Except for the additional luminosity, why do you like the idea of using CRI R9050:s?

395

Where did you pull this 4800lm number from for R9050?

Why would you want to use protected cells in direct drive lights? To trip the protection circuit at random time and to make sure the output is as low as possible on max?

396

Re R9050:
I read somewhere that they could be pushed that far, but I can’t find it again now, the numbers I find now are 200-700 lumen. Anyhow still curious why the original poster re R9050 suggested using them.

Re protected cells:
I just did not know the dangers of using unprotected cells. But it seems to be totally OK, not many fires have happened becaue of them (presuming the electronics used are of good designs and there’s no violence to the batteries).

So that was all, protecting myself. Since I learned they cap at 3A or so and hence are useless, and given the right electronics protection adds no benefit anyhow. So i’ll buy unprotected batteries!

Thanks.

397

Not even a higher brightness bin from Nichia in R9050 will output more lumens than those 219C in R8000, the higher the CRI the lower the brightness anyway.

Good ones can trip at like 6-7A. Anyway protected cells are not for quads and triples in direct drive where depending on the flashlight you can expect 12A to 20+A

398

Can someone explain how the Emisar D4 with quad nichia 219’s and direct drive can do 3600 odd lumens and the astrolux S41 with same emitters and direct drive is only capable of 1400 lumens.I know driver resistance and springs will make some difference but what else is going on.

399

The main reason is probably bad driver design with cheap components. But everything (thin wires, springs with high resistance, not enough copper on the boards etc.) adds up. The LED:s are probably not the best bin either.

400

The S41 is using 219b and the D4 219c leds which are supposed to be brighter.
However the S42 also uses 219c leds and is dimmer even on DD… so you’re right that there must be something else to the story. I’m also curious to understand what can make such a big difference.

401

D4 vs S41, the differences… this is what I’ve been doing for several years now, using top components and reducing resistance everywhere I find it. This is where my hot rods excel, just putting in the time to get em maxed out. So, now that people are finding out what a true direct drive hot rod can do, it’s somewhat of a mixed bag is it not? Shorter run times, much hotter temps, cell damage, possible burns… flashlight top fuel dragsters take more precautions than the stuff that comes carded at WalMart…

I’ve built em like this knowing how I use them. Now that hot rods are becoming more readily available it wouldn’t surprise me at all to start hearing of more issues. Play carefully people! :wink:

402

In between the failed S42 Nichia 219c (which is dimmer then the S41 219b) and the D4 with the same (?) emitters there seems to be a 100% increase in output on turbo - the D4 is twice brighter. That’s a lot! Can springs, tube contacts, driver components, wires or else make that much of a difference?