Homemade XP-G R4 Drop-In

7 posts / 0 new
Last post
brted's picture
Last seen: 1 month 4 days ago
Joined: 01/12/2010 - 19:44
Posts: 2371
Location: Atlanta
Homemade XP-G R4 Drop-In

Recently I got in a group buy from Cutter Electronics in Australia so that I could get a couple of Cree XP-G R4 LED's with a neutral 5B1 tint. I also bought the new Shining Beam 3-mode 1400mA NANJG ROHS 101-AK driver. I bought a smooth reflector empty drop-in from DealExtreme to put them in, but I wound up using an orange peel reflector.

NANJG 101-AK Driver

I really like this setup. The driver has 3 modes and memory, but, like the NANJG 112A driver, the memory is based on how long the light is on, which I'm not crazy about. However, it draws exactly 1400mA from a freshly charged 18650 lithium ion battery. The modes are separated well and the low is pretty low. I noticed a little bit of an audible squeal from the driver on Medium sometimes, but not on low. The driver already had LED leads soldered in which worked out just fine for me since I was starting with a new LED. Everything went pretty smoothly.

The LED itself is nice because it has a neutral tint, but is still an R4 in brightness, just one step below the brightest XP-G available, R5 which is only available in cool white. This is my first neutral white anything. I thought my Fenix had a fairly neutral tint, but the pictures below will show that isn't the case. People say colors look more natural with a neutral tint, but then others say Cree's LED still don't have that high a Color Rendering Index (CRI), so it doesn't matter. It definitely looks different, not like any of my other LED's.

To compare brightness and tint color, here is a series of shots at 1/25th second exposure with the XP-G R4 on the right against the Fenix L2D on the left. First, on Low:

Fenix L2D vs. XP-G R4 neutral on Low


Now with the L2D on High and the XP-G on Medium:

Fenix L2D vs. XP-G R4 neutral on Medium


Now with the L2D on Turbo and the XP-G on High:

Fenix L2D vs. XP-G R4 neutral on High


Now a fade series of the same two lights on highest, except now at 1/200th second:

Fenix L2D vs. XP-G R4 neutral on High 1/200th second


And now at 1/1600th second:

Fenix L2D vs. XP-G R4 neutral on High 1/1600th second


This is more of a fair comparison. Here is the XP-G vs. my new Ultrafire WF-502B with XP-G R5 from DX. It is 1 mode only, so here is a fade series at maximum, on 1/25th second exposure. The reflector of the R5 seems made for the small XP-G LED while the reflector for the R4 is made for a larger XR-E, so the beam isn't as concentrated. But you can still see the contrast in the two tints.

XP-G R5 vs. XP-G R4 neutral at 1/25th second


And at 1/200th second:

XP-G R5 vs. XP-G R4 neutral at 1/200th second


Now at 1/1600th second:

XP-G R5 vs. XP-G R4 neutral at 1/1600th second


Outdoors I take pictures with a 4-second exposure to make the pictures looks like what you see in real life. I aim the hotspot at the fence posts in the distance which are 120' away. The forked stick is 25' away. Here is the XP-G R5, which cranks out a lot of light and has pretty good throw:

XP-G R5 outdoors


Now here is the 504B host with the XP-G R4 neutral. I'm not seeing much difference in tint in this picture, but I think that could be the camera adjusting or something. That's too bad, because there is a substantial difference:

XP-G R4 neutral outdoors


Now the R4 on Medium:

XP-G R4 neutral outdoors medium


And on Low:

XP-G R4 neutral outdoors Low


Since those pictures don't show the contrast between the two lights that well, here is another pictures with the XP-G R4 neutral on the left and the XP-G R5 cool white on the right:

Outside XP-G R4 neutral vs. XP-G R5 cool white


I don't notice any pulse width modulation (PWM) where the driver flashes the light on and off so quickly that the eye just sees dimmer light. But on Low and Medium when I was taking the wall shots, I was getting some resonance flickers from the camera screen's own sampling rate. So I know it is doing PWM, but it is a very high frequency. Here I am slinging the light on Low as fast as I can on a timed exposure, and the beam still seems smooth.

PWM test shows smooth beam


Here are the results of the runtime test. Because the driver uses linear regulators, the draw is remarkably consistent until the it falls out of regulation. The way it works with a linear regulator is that it provides a certain voltage to the LED and the maximum amperage is based on how many 7135 chips are on board, with 350mA per chip. This has 4 chips, so 1400mA max. Any voltage in excess of the target is burned as heat by the regulators so with higher voltage it is burning off a lot of extra voltage and becomes more efficient as the voltage supplied approaches the voltage out. The light gets very warm, almost hot (good heatsinking!) and stays that way for most of the time. I can't measure light output, but I would guess it is pretty constant. I will try to do some measurements of efficiency (current in vs. current to LED) later on. But here is the runtime test, stopping every 10 minutes to measure at-rest battery voltage and tailcap current draw. The driver stayed in regulation for almost an hour and then the current draw started to drop. I'm impressed. The overall battery life isn't that great, but given the higher current draw, it is decent. This is a protected gray 2400mAh Ultrafire battery.

Time Volts Current (mA) Med (mA) Low (mA)
0:00 4.19 1420 440 70
0:10 3.99 1400 420 60
0:20 3.97 1400 420 60
0:30 3.91 1390 430 60
0:40 3.86 1380 410 60
0:50 3.81 1380 400 60
1:00 3.74 1280 360 60
1:10 3.59 1150 300 50

Efficiency readings. I measured efficiency based on the battery voltage under load and the current at the tail (V in and I in) and then the voltage drop across the LED and the current drawn by the LED (V out and I out). I only have two meters, so I measure these two sets of readings at different times, but on the same battery. I use four lithium ion batteries at different discharge states. For some reason during these tests (maybe bad soldering on my part leading to some kind of short on the board) on my fully charged battery the driver would not hold Medium mode and slip to High, so I was not able to get readings for Medium on a fully charged battery. These efficiencies are better than when I first did an efficiency test and got efficiency of 58-68% on High, but they are still kind of low for a linear regulator. When measuring efficiency on a 3x7135 driver (the NANJG AK47) at the same time with the same batteries, I got efficiencies of about 85-90%. I think part of the reason is this driver is delivering less voltage, closer to 3V while the other driver was giving 3.38V. That test was done on a XP-G R5 cool white LED and this is on a XP-G R4 neutral, but I wouldn't think the bin would have much to do with it. Also, since this driver is delivering a higher current, it seems to fall out of regulation faster, so the current starts falling off at a battery voltage below 4V with this driver whereas the 3x7135 driver is level to below 3.8V.

 V rest (V)3.653.834.064.18
HighV in (V)3.543.683.883.92
 I in (mA)950118012901380
 P in (W)3.364.345.015.41
 V out3.
 I out780106013301380
 P out2.423.224.104.20
MedV in3.623.794.00n/a
 I in260320350n/a
 P in0.941.211.40n/a
 V out2.542.522.60n/a
 I out250300300n/a
 P out0.640.760.78n/a
LowV in3.643.824.044.15
 I in40505060
 P in0.
 V out2.362.342.282.28
 I out30405050
 P out0.
Edited by: sb56637 on 02/14/2012 - 08:18 Reason: sb56637 replaced budgetlightforum.cz.cc with budgetlightforum.com via Scanner Search and Replace module.
alfreddajero's picture
Last seen: 1 year 3 months ago
Joined: 01/14/2010 - 00:28
Posts: 1798
Location: Virginia Beach VA

You get better color with the neutral white thats for sure......nice  pics as usual.

With Darkness, there will always be Light.



Don's picture
Last seen: 1 year 3 months ago
Joined: 01/12/2010 - 16:32
Posts: 6617
Location: Scotland

It is amazing how much more three-dimensional the world seems when lit with lower colour temperature (Towards the red end of the spectrum) light. With the rather blue light from cool white LEDs - especially the older ones - the world seems like a set of cardboard cutouts. All "white" LEDs are blue LEDs with appropriate phosphors to give enough variety of colours of light to make it seem white to our eyes.


I have no idea of the mechanism for the lack of depth perception with blue light and enough (or, better still, more than enough) light will normally fix it regardless of the colour temperature but it is odd.


The numbers from my light tests are always to be found here.


FlashPilot's picture
Last seen: 6 hours 32 min ago
Joined: 05/10/2010 - 16:07
Posts: 4829
Location: USA

Thanks for the detailed review. I just got the 1400 mah driver from shiningbeam for my (Q5 XR-E) MRV mod to XP-G R5 and am quite pleased with the results. Now Id like to mod another with a 5B1 neutral tint R4 XP-G. To bad i missed Cutters group buy.

Here's a look at the color temp scale:

sb56637's picture
Last seen: 3 hours 31 min ago
Joined: 01/08/2010 - 09:29
Posts: 7141
Location: The Light

Very nice FlashPilot! Did you make that chart?

Budget Light Forum ...where Frugal meets with Flashlight!

FlashPilot's picture
Last seen: 6 hours 32 min ago
Joined: 05/10/2010 - 16:07
Posts: 4829
Location: USA

No, its from Cree's site. Wink Its quite useful if its accurate.


okwchin's picture
Last seen: 8 months 3 weeks ago
Joined: 01/07/2011 - 04:41
Posts: 1164
Location: Australia

I have a 5C1 from cutter (the currently available 5 tint group from cutter) and its smack bang on the BBW line but slightly warmer. Beautiful tint too. 

"like everyone else - I’m looking for my next “last” flashlight" -  ohnonothimagain