J18 with the KD beast Driver upgrade -- DON'T BUY!

Flashlight Modding and DIY Parts
1

Update May 13, 2014: There have been many reported failures with the KD 5/7/9 LED drivers. My two here are ok on little runtime, but there have been many that fail after just a few minutes of runtime. Please Don't Buy!!

--------------------------------------------------------------------------------

I know this is a rather dated light but I still like it's small form factor and ability to run on 3 26650's. There's been much discussion on it, but I haven't seen (maybe missed it) on details of the full mod of this light. Well, I went ahead and finally installed the KD 7 LED 3 mode beast driver, used Fujik epoxy to pot it, copper braided the springs, and got pretty decent results: 5,280 lumens @30 secs with fresh TF 26650 Flames, and it seemed to be holding up the output well, considering, after the potting of the driver. The tailcap amps was 3.6A stock, and 6.18A with the KD beast driver.

So, I went ahead and upgraded the XML's to XM-L2 U2 1C's on SinkPAD's.

Before, with 28 gauge wires:

After with 22 24 gauge wires, 20 gauge driver wires, AS5, 2500 GRIT sanded:

After buttoning it all up and tested, was very disappointed.

I'm getting 4.96A at the tail initially, and it drops steadily, and measured 5,865 @start, 5,162 lumens @30 secs with it dropping and at a good rate.

All LED's are lit up, appear to be even. I'm thinking the higher voltage demand of the XM-L2's may be the cause? Maybe the driver can't support that voltage well?

Maybe I should try the KD 9 LED version on this J18? Maybe I should try wiring out the center LED and try it with 6 XM-L2's?

I have a spare J18, so thinking of swapping the beast driver into the stock light just to check if it behaves the same as it did before with the stock emitters.

Any ideas? Thanks!

Conclusions (for now - Jan 11th)

  • comfy noticed the max rated output voltage was only 23v and I measured just under that, so, I removed the center emitter, so with just a six-pack of XM-L2 U2 1C SinkPAD's, amp draw shot way up and also appears to be more stable (no dramatic drops)
  • Best #'s on fresh SONY 26650 cells (unprotected, low resistance):

tail: 6.57A, lumens: 7,208 @start, 6,834 @30 secs, throw: 76 kcd (measured at 5m)

  • With good protected cells (KeepPower 26650's using KK's):

tail: 7.37A, lumens: 6,970 @start, 6,691 @30 secs, throw: 75 kcd (measured at 5m)

  • With another stock J18 on hand, I ordered a 9X LED 5 mode beast driver from KD and will see if I can get the full 7 XM-L2 LED's working well, probably will get higher output

Updates (Feb 20th)

Tried 4 cells as suggested by GBD - 3 SONY's and added a fresh KK button top. Took the last tube section off my other J18 and added to the 9X driver J18. Did a tailcap measurement - heard a crack as I touched the lead to the tube threads, holding the other lead to the battery -- nothing blew up, and it read over 7A. Yikes! That's like 112 watts!

Ok - it survived the tailcap measuremt, now for the 30 sec lightbox test. Here's the #'s:

9,452 lumens @start - 8,976 lumens at 30 secs

and it survived... so far. Ummm, the KD driver is not rated for 16V input, just 6-12V, so I suspect I'm in dangerous territory.

I finally beat the Chinese lumens printed on the light!!


2

I dunno, but I have never gotten TF flames to do over 4.5 amps. In fact, that is exactly what I got out of mine, where the Sony US26650VT would do nine amps in the same light. That's with MT-G2 leds.

3

Actually I was surprised the TF 26650's went that high too, but also tried with KeepPower 26650's and straight KK unprotected cells - all same results.

Those TF 26650's though did over 6A even when down to 4.03v, so maybe I got good ones - they are maybe a year old, maybe before they were counterfeited.

Was really hoping to see close to 6,500 lumens out of this, because the driver wass doing 2.6A per LED, maybe little less with loss's, with the XML T6's on aluminum. Interesting - the stock 16mm stars were about 1.55mm thick, thicker than most. I was concerned because I sanded down the back of the SinkPAD's to get rid of the depression, so resulting thickness was 1.42mm. The stock driver wires (2) were 20 gauge but rated at only 150C, so still upgraded the driver wires to the better 20 gauge I had, rated at 200C.

4

I would try some higher drain batteries and see how they do

5

Check the output (both volts & amps), just the tail numbers might not be the numbers you need to figure it out.

6

I'll have to try that. How do you check output voltage though? Never did that. Just measure voltage across the + to - driver wires with everything connected? Not sure what to look for - never actually measured LED voltage before, specially on multi-emitter lights.

7

I have been getting Power-izer from batteryspace.com. You do have to pay flat rate small box, but you can get quite a few batteries in it. Seems like at least equal to the Sony US26650VT, maybe a little better. Feedback, please? And I been getting LG 188650s there, and they do seem better.

8

Yep, just the voltage between LED+ and LED-, can be either at the driver or on the two MCPCBs that host the connections from the driver. Whichever is easiest depending on where you break into the circuit to take the current reading.

I have the 9x 5-mode version and bench tested it with one of those '100W' arrays (10s10p) and it'll only do around 1.4-1.5A, I suspect because it can't put out quite enough voltage. The 9x drivers will work with your 7x setup, it'll only boost to the voltage required to do the rated current.

9

Thanks comfy! With around $45 invested in LED's only, $110-$120 total, I do want to get this light working to the level it's capable of . Wouldn't mind spending another $20 for the 9X driver. Thinking if it is the higher voltage demand of the XM-L2's, I'd order the 9X LED driver, and in my 2nd J18, I could build it out with XML U2's and U3's (got plenty spares) on copper and use the 7X driver in that light. The copper alone should help big time.

flo - yes, got a couple of Powerizer 26650's - best there is, think little better performing than the SONY's but higher capacity - win/win.

Like I posted earlier, the lower amp readings I'm getting now are even with 3 KK unprotected fresh cells, so the KK's can definitely handle these amp levels, though the TF 26650's did too with the XML's.

Edit - I can do the volts/amps readings before I pull the driver. I want to do a sanity check of the driver in the other J18 to re-check it. There are really 2 issues:

- initial amps is lower, from 6.18 to 4.9A

- seems to drop amps and output as if the driver is detecting over-temp (thought the potting solved that)

10

You can also remove one LED from the chain and compare that output to what it does with all 7. That'll tell you if it's the voltage not able to boost high enough to do the rated current.

11

Yep, maybe try that too! Oh boy, dunno if I'll have time this evening... Gotta get my monthly invoice in for my job tonight - need money to support my habit .

12

Tom, not sure if you noticed the specs on my X60vn. In particular the voltage. It’s actually dumping 25.2V, 6 x 18650 in series, into the 5 XML2’s. I measured that out of the carrier myself. Suppose to be running at about 4.5a, but I never checked that.
But that 25.2V - sitting voltage of course, might be what helps the XML2’s perform nicely. 6400 otf lumens split between the 5 LED’s is 1280 lumens each.
If you can get that 7 LED TRJ18 driven that hard or harder…….
That’d be a 9000 lumen monster. Very, very cool for a light with a moderate size.

13

Yes, that X60 is a big time light with mass and 6 cells. The J18 is borderline pocketable, exact same host as the J12 . The KD driver should be doing 2.6A per LED, but not quite getting there, and not sure I'd go higher. I tried opening up the J18 head but couldn't budge it with heat and strap wrench's. Really needs extra heat sinking but can't get inside. Other option would be cutting it open and building something - beyond my feeble capabilities . Really thought I'd hit about that 6,400 lumens or so.

14

I know the TRJ12 you built for brother is very impressive for its size. The light surprised me hom much power comes out of that little head it has. I think it ended up around 4600otf or so.
Those TRJ’s when driven hard are very cool flooders. Everything in front of you gets lit up. Unfortunately the small heads don’t appear to handle all that heat for too long. But that’s to be expected with their size.

15

comfy - I finally did the measurements, here they are:

voltage: 22.15v measured across the +/- wires at the driver, LED side -- 3.16v per LED then?

amps: 2.23A for all 7 LED's, 3.2A for only 6 LED's

So, the voltage is too low, I'm thinking? Thought it would be 3.6v to 3.9v maybe? Also, the 2.23A is what I expected, considering I'm measuring 4.96A at the tail.

But for 6 LED's -- Wow! 3.2A? That's pretty high - driver is rated at 2.6A - 3A.

So what do you make of all these measurements? Do you think the 9X driver would be worth a shot?

16

Yes, current increased when the voltage requirement dropped. The 9x version should have enough overhead to run at what it's supposed to (and as you saw with the 6 LED test, they'll only boost to what's required).

Did it do the same 22.15v with both 7 & 6 LEDs (I'm gonna guess that with 6x & 3.2A it was a bit lower, say 20-21v)? That 3.15v each with 7x is pretty weird, that voltage should put it around only 1.5-1.6A.

17

you should try the 12x driver :slight_smile: :slight_smile: :slight_smile:

MOAR POWER.

18

Id' have to wire it up for 6 LED's only to test voltage. Maybe I'll do that and test/measure it in this configuration - should be over 6,000 lumens for 6 LED's at 3.2A each. I'd like to see if it's stable - if driver can maintain output better than with 7.

19

Hmm, with all those nice copper stars maybe the leds run too cool as well, not good for the voltage, does the output climb as the body heats up quite bit?

20

Well with the 7 XM-L2/SinkPAD's, it dropped quite steadily - looked like amps and lumens dropped at the same rate.