BLF A6 FET+7135 Light Troubleshooting and Mod thread

989 posts / 0 new
Last post
DB Custom
DB Custom's picture
Offline
Last seen: 15 hours 48 min ago
Joined: 01/13/2013 - 22:28
Posts: 20623
Location: Heart of Texas

Just remember, heating up the pad at the base of the spring is also heating up the switch. Pretty much as easy to kill the switch here as soldering directly to the switch.

Try removing the spring from the board, soldering the wire on in the middle, then run the wire up through the spring and solder the spring back on around the edges. Cut the wire a little long, coil it inside the spring slightly so it’s got easy room to compress, then solder the top end of the wire to the top of the spring. I like folding the wire over the very end of the spring then soldering it. Even if the solder lets go of the spring the wire will stay in place.

Edit: A through-pcb bypass for high amps…

djozz
djozz's picture
Offline
Last seen: 6 hours 50 min ago
Joined: 09/07/2012 - 17:04
Posts: 16712
Location: Amsterdam

Do I hear a lathe in the background? Smile

Btw, I stand by my method doing the spring bypass on the outside, it is a bit theoretical in my opinion that the solder joint at the base of the spring would break, it is stiff one piece with the base under the spring. 

Everyone has their own style Smile

DB Custom
DB Custom's picture
Offline
Last seen: 15 hours 48 min ago
Joined: 01/13/2013 - 22:28
Posts: 20623
Location: Heart of Texas

Style notwithstanding, what happens if the top lets go and the wire… still connected at the pad… touches the side of the pill? Direct short, of course. Once you’ve had this happen, you develop more of a safety conscience than a style. Wink

Also, keep in mind, we are advising many people around the world here with unknown soldering skills. Err on the side of safety.

cabfrank
Offline
Last seen: 7 hours 17 min ago
Joined: 11/19/2010 - 17:25
Posts: 3025
Location: healdsburg, california usa earth

My soldering skills are known. They are known to be somewhere between bad and nonexistent.

hank
hank's picture
Offline
Last seen: 2 hours 45 min ago
Joined: 09/04/2011 - 21:52
Posts: 8768
Location: Berkeley, California

So does anyone make a short little quick-blow fuse with wire leads, that could serve as a spring bypass?

Seems like a thoughtful redesign of the whole switch notion could pay off considerably if there’s a way to separate the conductor from the spring in a way that would negate these low-likelihood, high-boom direct short possibilities.

SawMaster
Offline
Last seen: 6 months 1 day ago
Joined: 05/03/2015 - 17:18
Posts: 1802
Location: SC USA

Perhaps you could make your own by cracking open a glass automotive fuse, or by melting the plastic on a plug-in type with solvent?

Phil

dudunphy
dudunphy's picture
Offline
Last seen: 2 weeks 4 days ago
Joined: 07/28/2015 - 12:22
Posts: 406
Location: Colorado

DB I’ve add your video to the op. Along with some explanation as to why it’s safer. I hope that helps a lot of people! Thanks!

Dustin

DB Custom
DB Custom's picture
Offline
Last seen: 15 hours 48 min ago
Joined: 01/13/2013 - 22:28
Posts: 20623
Location: Heart of Texas

Let’s say “potentially” safer, we want people to think on the side of safety while they’re modifying these lights, right? I mean, reckless abandon is well and good when with a consenting adult but sometimes thoughtful care and consideration are appropro…

dudunphy
dudunphy's picture
Offline
Last seen: 2 weeks 4 days ago
Joined: 07/28/2015 - 12:22
Posts: 406
Location: Colorado

Yes there will always be tree burning. And triples you can weld with. Right? Wink

Dustin

DB Custom
DB Custom's picture
Offline
Last seen: 15 hours 48 min ago
Joined: 01/13/2013 - 22:28
Posts: 20623
Location: Heart of Texas

Wait til you build your first small triple, it’s so exhilarating, scary, and just outright fun you’ll wonder what took you so long.
The first time I saw 12A in a smallish light, man, what a trip! Then I saw 16Amps, 17Amps, holy cow, I was afraid to turn it on! And now I have a quad Titanium X6 pulling 21.3A from a single cell, yeah, it’s just freakish and fun and mindblowing and dangerous and and and…

You’ll see! You’ve got the bug! Silly

ToyKeeper
ToyKeeper's picture
Offline
Last seen: 46 min 11 sec ago
Joined: 01/12/2013 - 14:40
Posts: 10248
Location: (469219) 2016 HO3

If you want a fuse, it’s pretty simple to do. Just don’t bypass the spring. The spring will act as a fuse.

I know this because I’ve destroyed springs that way. It actually works pretty well, and prevented anything else from getting damaged.

djozz
djozz's picture
Offline
Last seen: 6 hours 50 min ago
Joined: 09/07/2012 - 17:04
Posts: 16712
Location: Amsterdam

ToyKeeper wrote:
If you want a fuse, it's pretty simple to do. Just don't bypass the spring. The spring will act as a fuse. I know this because I've destroyed springs that way. It actually works pretty well, and prevented anything else from getting damaged.

After a night's sleep I understand the safety concern of an outside spring bypass a bit better: I know I make solid solder joints but not everyone will make a good job of it, and you don't want to have it come loose. I think what TK says is a pretty good point: the driver spring of the BLF-A6 is not so long, so the gain of a bypass is a bit less so just leave it out. You can still do the tail spring bypass any way you want it because when it shorts there it is not a direct battery short but the light just switches on (in my case in moonlight Smile ), much less of an issue for something that will not often happen anyway.

scotlarock
Offline
Last seen: 1 day 12 hours ago
Joined: 03/04/2013 - 12:35
Posts: 220
Location: Michigan

Last night I did a spring bypass on both springs. Testing with a 25R freshly charged with 4.19v yielded 6.7a on my Fluke clamp meter…wow! Is anyone else seeing that many amps? Hopefully that poor little XPL can live at that power level.
The driver on this light is awesome;
Great UI
Nice moonlight
Lots of top-end output
Many thanks to everyone who worked on this project!

djozz
djozz's picture
Offline
Last seen: 6 hours 50 min ago
Joined: 09/07/2012 - 17:04
Posts: 16712
Location: Amsterdam

scotlarock wrote:
Last night I did a spring bypass on both springs. Testing with a 25R freshly charged with 4.19v yielded 6.7a on my Fluke clamp meter…wow! Is anyone else seeing that many amps? Hopefully that poor little XPL can live at that power level. The driver on this light is awesome; Great UI Nice moonlight Lots of top-end output Many thanks to everyone who worked on this project!

wow, that is more than anyone got yet, you must have had a very low voltage XP-L by chance. I have had XP-L leds that went to over 9A en ones that blew at 7A, so you may just be lucky with this one Smile

DB Custom
DB Custom's picture
Offline
Last seen: 15 hours 48 min ago
Joined: 01/13/2013 - 22:28
Posts: 20623
Location: Heart of Texas

Good work Scot! You are now solidly addicted. Silly

I have one that’s doing 6.2A on a Samsung 30Q, an HI variant actually, but most don’t go that high. Like Jos says, it’s the luck of the draw.

scotlarock
Offline
Last seen: 1 day 12 hours ago
Joined: 03/04/2013 - 12:35
Posts: 220
Location: Michigan

Hey Dale, since switching to your method of using my clamp meter to Measure amp draw, my readings have gone up, a lot. Here’s another point worth noting, I took the battery directly out of the BLF light after testing and placed it in the A6 I won in your giveaway, and amp draw dropped to around 5.4. Edit 6.1a

DB Custom
DB Custom's picture
Offline
Last seen: 15 hours 48 min ago
Joined: 01/13/2013 - 22:28
Posts: 20623
Location: Heart of Texas

Likely to be an emitter binning difference, each step on the bins is approx a 20% gain in output. We did see some dramatic shifts in the equilibrium when Cree made the last changes and high amperage is not common anymore, while output can very well be the same.

At any rate, a 500-600 mA difference is fairly normal between 2 identically built lights. It’s always difficult to get em exactly the same.

I’ve also built a light, ran the tests, wrote down everything as it happened only to come back later and find new tests give different results. The cells lose capability with age and wear, the emitters break in differently, whatever causes it, it’s a fluctuating set of parameters that befuddles me on a regular basis.

Tom E
Tom E's picture
Offline
Last seen: 2 hours 21 min ago
Joined: 08/19/2012 - 08:23
Posts: 12708
Location: LI NY

I'm loosing faith in my 14 AWG DMM lead setup now, after hearing these crazy #'s... Wow! Is there any reliable way you have to measured tailcap amps on the same setup you used the clamp meter on?

I'm seeing the same as you Dale. Generally months later, #'s go down, not always though. My triple X6 that was ~4200 is now like 3900 or so, for example. Cells sure do lose it. I'm think'n I'm not smart bout storing them lower - I tend to keep my cells charged up, which is probably not helping.

dudunphy
dudunphy's picture
Offline
Last seen: 2 weeks 4 days ago
Joined: 07/28/2015 - 12:22
Posts: 406
Location: Colorado

burnsd wrote:
I have size issues.

I tested my order: one of each tint. The 3D didn’t light up. The others did.

I pulled the 3D tailcap and did the screwdriver bypass and it lit fine.

The 3D tailcap on the 5A didn’t light. On the 1A it will only if I really crank down.

Putting a paperclip spacer in the 3D tailcap causes it to work on any light.

So, the 1A lights on any tailcap. The 5A and 3D light only on the 1A or 5A tailcaps.

The 3D and 5A tubes are visibly shorter to the naked eye. I can easily pick them out (1A is longer, 3D is way off center).

So as I understand it, I have two short battery tubes (from 3D and 5A), and one short tailcap (from 3D). And I need to email that to heyanqing1@banggood.com

Is that accurate?

I also wanted to state that while the differing tube lengths are a crappy ordeal esp. for any protected battery people (although springs on both ends helps) that the tube length in itself shouldn’t affect power through the light. The problem it seems is the length of the threads and/or how deep the caps are threaded. You’re running out of thread before the un-anodized portion of the tube is seated properly. Also be sure that your retaining rings are tight as they need connection to their respective boards! As I and DB have mentioned be sure your tube is not reversed (clip slot near tail). I am already guessing you are thorough enough to have checked that. Wink Please be aware that some of my posts are for anyone reading in the future as well.

Dustin

scotlarock
Offline
Last seen: 1 day 12 hours ago
Joined: 03/04/2013 - 12:35
Posts: 220
Location: Michigan

Hi Tom, just measured the A6 that hit 6.7a on my fluke clamp meter. It reads 5.4a on my Klein DMM with 14” 12g a leads.

DB Custom
DB Custom's picture
Offline
Last seen: 15 hours 48 min ago
Joined: 01/13/2013 - 22:28
Posts: 20623
Location: Heart of Texas

Occasionally I will interrupt the negative lead on the star and put a loop between the lead and the pad to clamp onto, thus getting an actual emitter amperage reading. I keep the loop as short as I can fit the clamp meter on comfortably, usually about a 3” lead.

12 ga solid wire from Romex in about 3” lengths and folded over on one end to fit in the DMM give very close to the same readings as my clamp meter. You just have to be careful not to be measuring something with more than 10A draw. My clamp meter reads up to 40A.

Tom E
Tom E's picture
Offline
Last seen: 2 hours 21 min ago
Joined: 08/19/2012 - 08:23
Posts: 12708
Location: LI NY

Hi Scott - hhmm, that's interesting. What exactly is the technique you are using with the clamp meter? It could be a true 6.7A, but maybe you won't see that high # when all is assembled and buttoned up? Usually a light box/sphere is the ultimate test, if you can test output while doing the amp measuring method, then test it fully assembled. I've seen my tail amp measurements reading low sometimes: output is lower when jumpered with the meter, compared to fully assembled (bypass's of course).

I'm thinking you are doing it with a heavy gauge wire at the tail end of the battery/light?

bdivi
Offline
Last seen: 1 year 3 weeks ago
Joined: 02/16/2015 - 09:07
Posts: 38
Location: Sofia, Bulgaria

Clamp meter and short thick wire is the only accurate method to measure currents in that range and with voltages arround 4V. Regular multimeters have too high internal resistance and introduce large error.
I measure arround 5A with my Fluke187 and above 6A with the clamp.

djozz
djozz's picture
Offline
Last seen: 6 hours 50 min ago
Joined: 09/07/2012 - 17:04
Posts: 16712
Location: Amsterdam

bdivi wrote:
Clamp meter and short thick wire is the only accurate method to measure currents in that range and with voltages arround 4V. Regular multimeters have too high internal resistance and introduce large error. I measure arround 5A with my Fluke187 and above 6A with the clamp.

I use another way that works well. I have a 0.01 Ohm resistor (1% accuracy, rated for 20A) mounted on a piece of brass with the leads soldered into a terminal block. Not on the picture but I use two 10cm pieces of massive copper wire soldered to solid copper clamps for connectors. I measure the voltage over the resistor, current is 10mV/A.

bdivi
Offline
Last seen: 1 year 3 weeks ago
Joined: 02/16/2015 - 09:07
Posts: 38
Location: Sofia, Bulgaria

That is indeed more accurate than a multimeter but still there are 10mOhm plus two wires added into the current path (let us assume 20mOhm total including the connecting wires, connectors and soldering).

20 mOhm multiplied by 6A produces a 120mV voltage drop.

This may not seem much but with the exponential VA curve of the LED 120mOhm will give you a difference in the measured current of at least 0.5A. See the attached Cree XML diagram for example.

For comparison I use a wire for my clamp that is less than 2mOhm.

djozz
djozz's picture
Offline
Last seen: 6 hours 50 min ago
Joined: 09/07/2012 - 17:04
Posts: 16712
Location: Amsterdam

bdivi wrote:
That is indeed more accurate than a multimeter but still there are 10mOhm plus two wires added into the current path (let us assume 20mOhm total including the connecting wires, connectors and soldering). 20 mOhm multiplied by 6A produces a 120mV voltage drop. This may not seem much but with the exponential VA curve of the LED 120mOhm will give you a difference in the measured current of at least 0.5A. See the attached Cree XML diagram for example. !http://i285.photobucket.com/albums/ll50/_JustinCase_/XM-LVfcurve.png! For comparison I use a wire for my clamp that is less than 2mOhm.

I'm not sure the connection set-up for the sense resistor gives another 10mOhm (I hope less), I did not check that but I can do that and report here, I'm curious too (not before tomorrow btw). The 0.5A drop at 120mA is also correct for the XP-L at 6A btw.

Gunga
Gunga's picture
Offline
Last seen: 1 day 21 hours ago
Joined: 11/28/2014 - 16:56
Posts: 1851
Location: Vancouver, Canada

So. All the modes on the light work except moonlight. It is basically off. Do you think touching up the solder on the 7135 will fix this?

dudunphy
dudunphy's picture
Offline
Last seen: 2 weeks 4 days ago
Joined: 07/28/2015 - 12:22
Posts: 406
Location: Colorado

Battery charged? You could check out post 52.

I’d touch em all up… The first four modes run through the 7135 no fet. Then 7135 + fet on 5 and 6. If you have any modes at all with your light then 7135 is working. Others may chime in…..

Programming?

Dustin

Gunga
Gunga's picture
Offline
Last seen: 1 day 21 hours ago
Joined: 11/28/2014 - 16:56
Posts: 1851
Location: Vancouver, Canada

I’m a modder and we are all more patient but I really don’t have a ton of time for troubleshooting. I’m not very happy about this.

It’s a freshly charged new LG HE2.

DB Custom
DB Custom's picture
Offline
Last seen: 15 hours 48 min ago
Joined: 01/13/2013 - 22:28
Posts: 20623
Location: Heart of Texas

Actually dudunphy, that’s not quite right. ToyKeeper posted a list of how the two channels work together earlier in this thread…

The modes are:

Group 1___7135 power___FET power
Moon_____0.8%…………….0
Low________8%…………… 0
Med 1______43%…………. 0
Med 2_____100%………….2.7%
High 1_____100%………….22%
High 2_____100%………….54%
Turbo______0………………100%
Group 2___7135 power___FET power
Low________8%……………0
Med_______90%……………0
High______100%……………35%
Turbo______0……………….100%

So, modes 1,2,3 work on the chip with no mosfet(or 1 and 2 in Group 2), all the other modes except turbo have a blending of the two. Turbo being 100% Mosfet for maximum output.

Point being that it might seem to be working “fine” on all modes save moon, with that one being off apparently. This is showing up as a common problem on this light, either they had issues with soldering the 7135 or they got some bad chips, which can happen.

Pages