Nichia 219B in Klarus Mi1C Ti - Issues

While hunting for my ultimate EDC flashlight, I came across the Klarus Mi1C Ti.

It is small, convenient, has an acceptable interface (long pressing to turn off is a bit of a nuisance but not a deal breaker), lots of little perks etc.

The only thing I would want different, was a warmer, higher CRI emitter than the Cree XP-L HI it comes with… But since this little gem is easy to open, I swapped the emitter with a Nichia 219B sw45.

The beam patern is right for me. A lot of spill and a decent hotspot. All nice and gradual.

If only the world was so good… Alas! The flashlight is behaving a bit odd. It steps down in a few seconds when I turn it on. If I step it back up a few times, it works! When it gets warm enough, it works fine. I tried a different (micro usb rechargeable) battery, by Fenix, and it seems a bit better, but not much. The light sometimes flickers, which is strange as it is supposed not to have a PWM driver.

Is it possible the driver cannt handle the Nichia 219B? Does anyone know why this is happening and how it can be fixed? I haven’t tested much, but I think when the battery voltage starts to drop a bit, the flashlight works fine. It definitely works fine when it is hot.

Any insights will be very much appreciated, as this is, in any other way, a very good EDC light for me.

I have CU version with XPL hi and it also flickers like crazy on turbo. I’m still waiting for vapcell 16340, maybe it will be a little better.

That is putting my mind at ease, somewhat.

I should have tested it with the XP-L HI LED before I modded it but I was too excited :slight_smile:

Anyway, I have charged the Fenix cell using a micro USB cable and it seems to be working fine now. Previously I had charged it with my OPUS 3100 charger, at 500ma and maybe that was too fast for the cell.

I will test with an Olight micro USB rechargeable and an Olight non micro USB rechargeable, when I get home.

Unfortunately, I am still experiencing random step downs from high to middle, which go away once the flashlight warms up. There is also random flickering, which also disappears when the light warms up.

Could it be a faulty LED or is it because of the differences between the original Cree XP-L HI and the Nichia 219B, which the driver cannot handle?

Could the driver be adjusted to accomodate the differences?

Is there any one knows electronics and could guide me to maybe replace the correct resistor?

Here are some pics of the driver

My guess is your batteries have to much voltage sag to run in turbo for any length of time. What batteries are you using and how old. They will also have to be fully charged in order to stay in turbo any length of time. The turbo step down I assume is looking for a set voltage (in firmware) from the battery while its under the load to trigger it.
Most 16340’s will not hold a voltage very well under a 2 amp or more load, a few can. The Trustfire 16340 does pretty well but the new Vapcell 16340 is the best I have tested.
I have tested several different 16340’s and the top 3 would be

  1. Vapcell - the most capacity with the least voltage sag
  2. Olight (ORB-163C05-10C has to be modified to work) - lower capacity but holds voltage well
  3. trustfire (red-gold) - a little less capacity than olight and slightly lower voltage under load
    The Vapcell has the highest capacity with the least voltage sag under a 3 amp load.
    .
    I changed the current on a Mi1c and ran into the almost instant step down until I used the trustfire 16340. Now I use vapcell for all of my high drain 16340 lights.
    1100 lumen 16340 klarus Mi1C Resistor Mod
    This may not be your problem but if you changed nothing else besides the emitter, it has to be either a bad connection or battery sag. The driver is sensing the voltage to low so it steps down.
    As li-ion’s are used under a decent load they heat up internally which as they get warmer are able to give more current with less voltage sag.
    .
    Just my 2 cents :wink:

moderator007, thanks!

I have not modified the resistor as in your post, to provide more current. All I have done is change the LED. Are you suggesting that had a similar effect?

Could I just use another resistor, higher than 0,05Ω to limit the current, so the flashlight will work with my existing batteries?

I have tried the included Klarus one, Fenix and Olight (but not the 10C which needs modding).

I would be happy to try the Vapcell as well. What is the exact model you have tested?

maybe it is unrealistic to expect a 16340 to maintain 600 lumens for more than a minute (and only on a full charge):

I dont think the problem is the LED, nor the battery
imo the problem is that 600 lumens is not sustainable from a 16340
For example, the Olight S Mini also cannot maintain 600 lumens for more than 90 seconds…

I use those lights on Moon, Low and Medium primarily… I do not expect more than a brief flash of Hundreds of Lumens.

The situation here is not the driver stepping down after a minute.

It steps down randomly within 5 seconds. If I step it back up, maybe once, maybe a few times, it will then stay on high (600lm with the XP-L HI LED, less with the Nichia) until it gets hot.

I agree with moderator 007 that it must be the battery not providing enough power. In his case, the reason is the resistor mod, which drives the XP-L HI with more amps. In my case, it is the Nichia LED, which is obviously drawing a bit more power than the nominal 1.5A of the flashlight, enough to cause issues with normal 16340 batteries.

The way I see it, I either go for a vapcell, and keep the light as it is, or mod the R050 resistor to something with a bit higher value, to decrease the amps within the abilities of most 16340 batteries. I will try the vapcells first, as I prefer more output.

By the way, I did knock a couple of SMD components from their place while removing the black protective cap, but luckily I was able, with some effort, to solder them back on. I re-assembled the light and it works fine.

the input and output cap of this driver is laughtable
seriously 0603 size?

also it seems one 0402 cap near than OPAmp seemed to moved off its designated pads during reflow,
my guess is that this causes the problems

Well check that out, Lexel you got eagle eyes :+1: . I didn’t even notice that until I zoomed in on the pic.
Fix the capacitors and see if that helps.
.

As I said in my previous post, two capacitors had moved, either by me while removing the black cover, or in manufacturing. They have both been fixed.

The light still flickers from time to time and it also steps down. Remember, it did this before I ever touched the driver, so unless those moved components were a factory defect, they did not cause the issues.

I am still using the Nichia 219B LED, so this may be the cause, as it is probably drawing more current than the Cree XP-L HI.

I have ordered 4 VAPCELL 16340 from AliExpress. When they arrive I will be able to tell if it is the battery causing this.

I also had a lot of trouble running N219b in a light similar to the Klarus (Utorch S1 Mini)
Clemence said it was because the driver is linear
I dont know what that means, technically

but the E21a 2000k, and the stock XP-L, will run longer on maximum, than the N219b 4500k in my Utorch S1 Minis

I get lots of unreliable behavior if I try to use Max with the N219b, flicker, step down, very dim output

my conclusion is the circuit does not like LEDs with higher vf
I dont know the cure
I blame the driver design

here is a thread in which I tried to discover what was wrong

but at the end of the day, Ive had to accept that I simply cannot use maximum with the N219b, with that driver

I had hoped the issue would not exist in the Klarus, sorry you had to be the guinea pig to discover that the Klarus, like the Utorch, has a circuit that is not N219b friendly

maybe there is a fix, and if so, it probably would work on the Utorch too… I hope you figure it out

the driver is definately not linear

Well, problem solved, thanks to moderator007.

Even though I have not modified the driver, the lower fv of the Nichia 219B in combination with this particular driver is causing an increase of the current from 1.5A with the Cree XP-L HI to about 2.8A with the Nichia 219B.

This is a significant increase and not many 16340 batteries can provide this kind of current.

I tested the following, with the same results (random, non thermal step down and flickering):
Klarus 16340 700mAh (micro usb rechargeable) - included with the light
Fenix ARB-L 16-700U 16340 700mAh (micro usb rechargeable)
Olight ORB-163P06 16340 650mAh (5C)

This makes sense as none of the above are rated to provide 2.8A

I then tried the Olight IMR16340 550mAh (10C) which is rated to provide up to 5.5A. This worked great. It can also be used in the light without any modification, as the tail spring will contact the positive without issues. You will need to charge it in a compatible Olight flashlight or use a small magnet, with most chargers, if you do not want to mod the battery, which will remove its ability to charge in Olight flashlights.

I have ordered the new VapCell 16340 for a more elegant solution.

excellent info!
thanks for figuring that out :slight_smile:

am I understanding correctly that:
with 219b, the circuit increases Amperage by almost double, and
the increase in current reduces runtime, and increases heat, making the light less efficient?

is there a way to modify the driver so it does not increase current to 2.8A when using the 219b?

I think you are referring to this cell:
Olight IMR RCR123A / 16340 550mAh 3.7V Protected Lithium Ion (Li-ion) Button Top Battery for S1R II (ORB-163C05-10C)

and as far as the

Vapcel INR16340 800mah 7A

is that battery Protected, like the Olight IMR?

That’s the cell, made for the S1R II. Olight 10C 16340 3amp Discharge Graph (Now Vapcell is the best 16340 I have tested)
You should be able to increase the feedback resistor to about double and bring the current down to half the 2.8 amps. Change the value from R050 to a R100 and see how that does.
If you lower the current the other batteries should work then.

It depends on how you look at it. The increase in current makes the light brighter, which is great. It compensates for the fact that the Nichia 219B I replaced the Cree XP-L HI with, is less efficient and would produce less light, had it been fed the same Amps. As it stands, the light looks to me to be a bit brighter than it was before. Of course, I have no means to measure that objectively, so take it with a grain of salt.

On the other hand, the increased brightness and the usage of a less eficient LED will definitely have an impact in thermals and consumption. The light will start thermal throttling after 30-40 seconds, with a small visible (flicker like) step down and then gradual reduction in output. Nevertheless, it is still bright enough to use without any extreme plunges.

So all in all, it is a matter of preference. Some prefer a brighter high mode for less time and some prefer a less bright high mode for a bit longer. I think I am happy with what I have at the moment. I love such brightness from a High CRI LED in such a small light and it does not get uncomfortable hot nor does it throttle too soon.

As moderator007 said, replacing the R050 resistor with one of higher value should limit the Amps going to the LED. Thus the brightness and heat will be lessened and the runtime will be increased. The light will also be usable with batteries that can provide less Amps, such as the included one.

That is the one.

I do not know for sure but the facts that it says “End-of-discharge Voltage: 2.5V” in the specs and it has a visible neck a few mm below the positive pole, suggest that it is.

That is exactly what I think, but I am not sure I want to do that. I think I would rather use high power cells, like the Olight and the Vapcell and keep it as it is, with higher output.

On that, jon_slider, moderator007, anyone who cares to comment:
What do you think? Would you keep the light as it is now, with a higher output and battery consumption, requiring specific high output cells but giving out more light or temper it to be compatible with all RCR123 batteries and have a longer runtime, but produce less light?

Protection is a High Priority for me.
I do at times like to live dangerously, and I have been known to use UNprotected LiIon.
However, I do not recommend my risk taking behavior to others.

IF I could confirm that the Vapcel has built in overdischarge protection, and I wanted to EDC the light myself, I might keep the resistor as is, for now.

I might even carry it with UNprotected cells, but that requires more presence of mind than I may have available at times.

(I did read that the minimum discharge spec is 2.5v, and I share your wishful thinking that it has overdischarge protection. Trust, but Verify)

In the grand scheme of things
I would prefer if the light worked with the provided Protected 16340 cell.

Extensive testing has revealed occasional flickering.

I think the specs for the control chip are for 2A max, if I read it correctly, so understandably 2.8A is outside its comfort range. Switching power supplies tend to ripple in these situations so, I guess, hence the flickering.

I have ordered some R082 and some R100 resistors to replace the R050. This should put the power supply within its comfort zone and make the light work with any battery.

Using a single unprotected cell should be fine in this light as long as you know what to look for. You can run a little test to see where you realize the cell is low and not producing as much light.
Try using the light on high and when it steps down to medium keep using it to drain the cell preferably when it’s dark. As you use it, at some point you should notice the light getting dim. When it becomes obvious it’s losing lumens turn it off and pull the cell. Wait a few minutes and measure the voltage with a DMM and see where it’s at. If it’s above 3.0v then I’d say your good as long as you pull the cell when you see the same loss in daily use. The light may even step down from medium to low that be another good indicator if it does. If you mostly use it in low mode then you probably will not get any indicator until the cell is flat.
.
I prefer to have a big blast of light on hand and not use it much than to need it and not have it. I just use a lower mode for most task to extend run time.
It’s like sticking Nitrous Oxide on your daily driver. Just for the times when you need to merge into that relentless traffic. :smiley: