So I finally got around to modding my (perfectly functional) Saber 1A from XP-G2 to a Nichia 219. The mod went fine, but while testing tailcap current readings afterwards on 14500 I got a brief reading of 0.8 amps and from that point could only get a single 0.01 amp mode and 2 exactly equal 0.45 amp modes.
I thought it must be the 14500 so re-tried with an eneloop. Before the mod I would get 0.03, 0.12, 2.15. Now I get 0.03, 0.12, 0.12, so it seems high mode has simply disappeared…
I have approximately no idea why this has happened, so would appreciate anyone else’s help/experience with a similar problem. The only possibility I can think of is that there are three little solder coated points coming out of the back of the driver, maybe I shorted one of those when testing current. But I don’t know what they are, so don’t really know.
I can't say for sure what happened with your Saber but my T10 which I just converted to Nichia, while functioning perfectly, get's noticeably hotter when on high. No clue why that is but it is. I would hazard a guess that popping in the 14500 first burned out something in the driver. Did you run it on an AA or eneloop before running it off the 14500? If so, did it work then?
I had run it on both eneloop and 14500 before with no problems. One thing I had noticed was the relative inefficiency of the driver on the high mode, I would guess that was to do with a rise in forward voltage meaning more work for the boost circuit. With a Nichia 219 in (higher vf than XP-G2) the high mode had become even less efficient from my measurements (under 70 OTF for around 2.7 watts input), which is why I tried a 14500 again hoping that the higher voltage would mean better efficiency.
I would think that that inefficiency at higher vf and high currents could probably account for the added heat with a 219, but I don’t really know how it’s lost max mode, I’ve never come across anything like it before tbh.
I wouldn’t mind if the mode was just gone, but it’s weird having a low and then two identical medium modes.
I’ve just taken the driver out and connected it directly to an XP-E and done some random testing. The driver is a two tier one, each of the three pins going through both.
There is a second black wire (other one goes to LED negative) which was connected to the far inside end of the middle pin, with that as is the mode readings are 0.03, 0.11, 0.12 on eneloop and 0.01, 0.38, 0.40 on 14500. If I disconnect that or connect it to the left most one then no light. If however I re-connect it to the right most one instead then I get 0.03, 2.2, 2.2 on eneloop, meaning I have got two high modes instead of two mediums. On 14500 it goes a bit crazy though, pulling 0.01, 2.49, 2.49, when before the problem started the max it would pull on 14500 was 1.7 amps.
So, I’m thinking maybe it’s something to do with a fried resistor or connection to a resistor somewhere? Any ideas?
If I connect directly to the right most pin with the -ve from the cell I get the same modes as completing the light normally. If I connect directly to the middle pin I get 0.01, 2.2, 2.2 with eneloop, if I connect to either the right pin or the -ve on the driver then also the middle it automatically switches to 2.2 amps (max) if I was previously on a 0.12 (med) mode. Connecting to the left one shorts the circuit. So the pins definitely have something to do with modes.
With 14500 I get 0.01, 0.48, 0.48 connecting to the -ve on the driver or the rightmost pin, or 0.01, 1.5, 1.5 connecting to the middle. If I connect to the -ve or right pin and then also the middle one though it goes to 2.5 amps which is weird; especially as if I go to the middle pin first and then also the right or -ve on the driver it remains at 1.5… The right most one shorts out again.
Now a few minutes ago I connected to the middle pin and the right (at least that’s what I thought I’d done), it shorted and then I started getting 0.03, 0.12, 0.28 on eneloop, or 0.01, 0.48, 0.64 on 14500. I thought shorting it this way must somehow be fixing the problem as a higher mode was partially returning, so I did it again hoping to return it even further. Unfrotunately that didn’t work and it’s back to 0.01, 0.48, 0.49 and 0.03, 0.12, 0.12 again…
Does that help give any ideas to any of the electronics experts on here?
EDIT: Don’t worry, I messed with it a bit more and now the boost circuit’s dead and it pulls 0.04 with 14500 for the same half lumen (rubbish efficiency) and 0.24 for the med/high modes. I’ve emailed thrunite to hopefully get a new circuit for it.
Hi colleagues,
I have this flashlight and just modified it with Cree XM-L2 T5 4B0 (neutral white) emitter. I used this flashlight with 14500 and burned out the driver, so i made some investigations to repair it.
First of all i must note that driver seems to have output limited by voltage on Hi. Using XM-L2 led and new silver wires to connect it to driver instead of thin copper ones in original construction i obtained maximum current of 2A measured on battery. Using chart from CREE datasheet we can see, that voltage on LED is about 3,15V. In XP-G2 datasheet we can see, that on 3,15 V input we will have about 1,5 A forward current, and exactly this value was measured with the same 14500 when flashlight was in original state (with original XP-G2 LED). In Nichia 219 datasheet i see, that on 3,15V forward current should be about 500 mA, and it seems you received this value exactly.
Now few details about my investigations (they are fully described on Fonarevka but in Russian…)
Take a look to picture below. There are three resistors, that are related to currents in modes. I can’t clearly understand how they work (2 of them have large resistance than needed for usual current-sense), but shortening each from them push flashlight into Hi instead of corresponding mode (other modes are working correctly).
Red - Hi mode, default resistance 0,2 Ohm.
Yellow - Med mode, default resistance 2,2 Ohm.
Blue - Moonlight mode, default resistance is 2,7 kOhm.
After 5 minutes work on Hi with 14500 and XM-L2 led i’ve got Hi-mode resistor exploded. Also explosion damaged Med-mode resistor, so i replaced tham and also adjusted modes, because by default i had <10mA; 500mA; 1,9A in corresponding modes, so in light output there were huge difference between Low and Med, but Med and Hi was almost the same.
I used an SMD “shortener” (resistor marked with “000” with exact resistance 0,08 Ohm) instead of exploded Hi-mode resistor and 4,7 Ohm (marked “4R7”) instead of Med-mode resistor. As a result i obtained next currents (measured on battery):
Low - <10mA
Med - about 300 mA
Hi - about 2,1 A.
Now light output in modes is excellent for my purposes, also resistors become just slightly warm after 10-seconds work in corresponding mode (driver and led are not in flashlight tube, so there is no enough thermal dissipation from led to test it more, then 10 seconds). I tried to perform the same test with resistors with original values and they became really HOT in about few seconds.
Thanks for that detailed analysis, I think I might have fried a connection/switching mechanism somewhere, as I could get each of the three modes by touching directly to different components. But it would only give me Low/Med/Med or Low/High/High at one time. I still don’t really understand it to be honest.
I ended up turning it into a single mode (direct drive) with a Nichia 219, which draws between 1.5 and 0.8 amps over the discharge of a 14500 from 4.2v down to 3.5 or so. I’m happy with that, as after contacting thrunite to ask if I could get a replacement circuit they sent me a whole new Saber 1A free of charge!
I’m glad you got yours working how you wanted it as well
I used “shortener” with the same size as original resistors (0603 or 1608 metric), because due to it’s low resistance even at 2 Amps power dissipation is low. Med-mode resistor (4R7) has 1210 (3225 metric) size. It fits avaliable space precisely, but you should scratch lacquer near outer (closer to the edge of board) soldering pad and so extend this pad to solder new long resistor properly.
