I have a Jax-Z1 flashlight and I would like to optimise throw. So far I have only played around with de-doming the LED on copper Noctigon heat sinks. I have tried de-domed XM-L2 and XP-G2. The lux readings at a distance of about 26 ft (8 meters) are:
XM-L2 U3 de-domed: 1760lux at 8m or 112klux at 1m
XP-G2 S2 de-domed: 2080lux at 8m or 133klux at 1m
Notice that the increase in throw is only 13. All LEDs are used with the same driver and the nominal current for the Jax-Z1 driver is 3A. The XM-L2 has an emitter size of 4mm2 and at 3A produces about 1000lumens resulting in a surface brightness of 250lumens/mm2. The XP-G2 has an emitter size of 1.92mm2 and at 3A it produces about 800lumens resulting in a surface brightness of 416 lumens/mm2 (see emitter test results . Since surface brightness determines throw should the XP-G2 at the same current throw approximately 66 more than the XM-L2? Why I am seeing only a 13% increase?
I also checked another thread with a similar mod, a Jax-Z1 with a de-domed XP-G2 and a 5A driver and the throw is 228klux@1m. I think my throw with the XP-G2 is underestimated. Would the stock Jax-Z1 driver be affecting the results? Am I doing something wrong?
The Aspheric is nice AR coated one but the only downfall it has is the focal length is really big. That is to say the LED focus sweet spot is with the Aspheric really far far from the LED. You could try other Aspherics with shorter focal length but that’s a huge task and expensive one too. Also because of this your OTF lumens would be around 50% less than say using a reflector but keep at at and you could be rewarded with 300K lux and 150 OTF lumens:)
You should be seeing at least 150kcd with the stock driver and dedomed XP-G2. Mine have always come out at least this high or higher. 200kCd if you add 2 R330’s on top of the sense resistor.
*Edited for the correct number of resistors to stack on the sense resistor.
You can solder the springs with copper braid or wire to lower the resistance and use an imr battery to get the most amps out of the driver.
If that doesn’t work then you’ll have to play around with centering.
I tend to believe it is more of a centering problem than driver amp output. According to this mod:
he used an XP-G2 to XM-L adaptor to get the LED nicely centered. I am working on having an adaptor there and see if that helps. I will also check the amp output but I don’t see why the amp output would be different if I only change the LED and keep everything else identical.
You know, you’re about 20% off of my numbers. Any chance you could give us the numbers from a couple other lights you’ve measured?
It’s possible that your meter just reads low compared to some of the rest of us.
Edit: Can you please take a picture of the zoomed in image of the die on a wall? The quality of the die image will probably help us tell you if you have have a centering problem. Generally it gets smeared or a weird corona if off center.
I think 133kcd for a dedomed XP-G2 driven at that current is a bit low, and knowing more information would be helpful.
You are using the stock driver, right? So did you run it with one or two batteries? The Z1 stock driver is a buck driver, which mean the driver has fairly high resistance. If you run it with one battery I am pretty sure that the LED current is nowhere near the “rated” 3A - Probably you will get 2.6A or lower than that at the LED even with high drain battery. To run the buck driver at its full potential (3A in your case) you should use two batteries in series. The input voltage of two batteries in series is 8.4V, which would help to overcome the high resistance of the buck driver and giving constant 3A to the LED.
2. In real world, current measured at the LED is usually slightly (or more than you think, in some cases) lower than what you have measured at the tail due to the energy loss at the driver, and maybe other places as well like the spring or switch.
3. Is your lux meter measuring accurately? Do you have other “known lights” with “known lux reading” to compare with?
4. As FYI, dedoming an LED can potentially increase its Vf during operation, which means that your battery will struggle more to keep up its (input) voltage to the increased Vf of the LED to provide the current required consistently.
5. Have you done the “spring mod” as suggested by some members above? This can significantly reduce the resistance of the spring. The “spring mod” should be done on both the switch spring and the driver spring.
6. As for the not being perfectly centered part - IMHO I don’t think that will affect your lux reading that much, provided that it is not off-center by too much.
7. Did you use good quality high drain battery in it while taking lux measurement? Was it fully charged at 4.20V?
8. What gauge of wires do you use to connect the LED with the driver? For example I always prefer to use 22awg wires in my high current builds, as they can deliver high current and at the same time not being too thick to physically interfere with other parts (i.e. reflector, retaining ring…).
There are a lot of variables in low output problem like this. I would like to suggest that it’d be better to check the current reading first.
You sound like a very knowledgeable person in this field, I really appreciate your response. I’ll answer some of the questions you asked.
1. I used one 5200mAh 26650 unprotected battery that I got from mountainelectronics. I made sure the battery was fully charged before I ran the tests.
2. I measured the current today. It starts at 1.9Amps and slowly rises to 2.2Amps after 1-2minutes. I do all the light measurements within 1-2 minutes of a fully charged battery so it would have never risen to the rated 3Amps. At 2.2Amps the lux readings are more reasonable.
3. I use a Suche HS1010 Lightmeter. I don’t have any known lights for reference or calibration.
4. I use the standard gauge wires that came with the light to connect the driver to the LED. I haven’t made any changes to the driver yet.
I will look into the “spring mod” and the perfect centering of the LED. However what I don’t understand is why is the XP-G2 only 13% better than the XM-L2 at the “same” current (2.2Amps)? Should not the XP-G2 be much more bright as I explained in the beginning of this thread based on die emittance values?
Here are some pictures I took today, the dark vertical lines are from the door pattern, not the LED.
I haven’t measured any other lights but I will try and do. In any case thought the light meter I have (Sunche HS1010) has been giving me consistent readings from different days. Even if it underestimates lux values why is the XP-G2 only 13% brighter than the XM-L2 at the same current (which is actually 2.2A, not 3A)?
The XP G2 has a much higher forward voltage then the XM L2, so the standard driver probably isn’t able to deliver the voltage and thus the amps the led needs to perform well.
My Jax Z1 has a dedomed XP G2 on Noctigon, bridged spring and a LD01 driver, delivering 5A.
It’s measured at 215Kcd @ 1 meter (with Sony Konion VC5, and a tad less with good 26650 such as Keeppower 4200/Efest3500 purple).
I didn’t do this mod myself but a friend on the German TLF forum, Kenjii.
The 26650 battery you use is a good one, so no problem with that. Did you try to use two batteries in series and measure the current draw? This is important to try, and let us know what is the amperage you measure at tail.
I forgot what gauge if wires does the driver come with, should be 24awg or larger number, which is not really good for high current application. You should at least replace the wires with 22awg later especially if you want to run it with only one battery. If you are going to run your light with two batteries (in series) then replacing the stock wires may not bring any significant effect at all.
As the other has explained the dedomed XP-G2 seems to have higher Vf compared to dedomed XM-L2. There are some threads around here discussing about this issue. Beside that does your dedome come clean and the die is not damaged?
I only have one good 26650 battery and then two cheap chinese 18650 batteries. I will try using one of the 18650 batteries and check the Amps with two batteries. I will be planning on using mainly one battery so I will replace the wires with 22awg and also do the mod with the tail cap spring.
I measured the lux with domed XP-G2 first and I got a reading of 64klux at1m. I then used three different de-domed XP-G2, two de-domed with gasoline and one with heat, and they all gave me a reading of around 130klux at1m. From what I have read around de-doming should double throw and in my case it does so I believe the de-domed LEDs are good.
I also read on another thread that the measured 2.25-2.30A at the tailcap of a flashlight should mean about 3.80-4.00A at the emitter with driver losses. If that’s correct then in my case the 2.2A at the tailcap should mean 3A or more at the LED?
If you plan to run it with only one battery in the future I would strongly recommend you to replace the stock driver with a good linear driver, for instance the popular and inexpensive Nanjg-105C, or some direct-drive drivers like the BLF17DD… etc. You can order these excellent drivers from our fellow member RMM. The reason I recommend you to run it with a linear driver is mainly because it has much lower internal resistance as compared to the buck driver, as a result you can push higher current through the LED more easily with one battery.
It is good to know that your dedomed XP-G2 is good, at least we can eliminate one possible factor now.
No, in this case the input voltage is significantly higher than the LED Vf and a buck driver is used, so we need to calculate everything based on wattage to estimate how much current the LED is actually receiving. Since he was running his light with two 18650 in series, the input wattage is calculated as below: Input W = 8.4 x 2.3 = 19.32W
And we know that the input wattage is always equal to the output wattage, and I assume that the LED Vf at that point is 3.5V (just a wild guess), hence Output W = 3.5 x ? = 19.32W
The “?” is the current, so 19.32/3.5 = 5.52A (at the LED, theoretically)
Yes, 5.52A at the LED is just a theoretical value - an ideal world value, because in real world the buck driver is not 100% efficient, especially with the lousy stock driver that comes with the 1405. In this case I’d assume again that the driver efficiency is only 60%, which will give you the real world LED current as 0.6 x 5.52A = 3.31A only!
Sorry I am poor at explaining thing like this, and if you are confused with what I said earlier feel free to ignore it lol. Hopefully other members can chime in and explain this in a more straight forward way for you.
For your JAX Z1 case, if you measure 2.2A at the tail (with one battery) then I’m pretty sure that the LED would see lower current than that.
OK, I have another update which I think will make sense. I replaced the XP-G2 with XM-L2 and took an amp measurement at the tail cap. With the XM-L2 it starts at 2.5Amps and after 1 minute it rises to 2.9Amps. So basically the same driver, same battery, same wires give:
XP-G2: 2.2Amps
XM-L2: 2.9Amps
As somebody explained this is to be expected because the XP-G2 has a higher forward voltage than the XM-L2. With these numbers of current draw my Lux measurements are consistent. So from your adivces so far it I want to perform a good mod of the Jax Z1 I should:
