BLF A6 FET+7135 Light Troubleshooting and Mod thread

Thatā€™s not a tube length issue, thatā€™s a bad or badly-soldered 7135 chip on the driver. Itā€™s the chip with 3 legs on one side and only 1 on the other.

I finally acquired one which acts this way so I can test and debug it, but as far as I can tell itā€™s entirely a hardware issue. I think someone already reported that they fixed one by re-soldering the chip, but I still need to try that.

You know you donā€™t have to actually click the button to change modes, right? All it needs is a gentle half-press like what one might use to get a camera to focus without taking a picture. A full click is only needed to turn it on or off.

In any case, I have no idea why it would behave that way. The code doesnā€™t have anything even similar to ramping built in. Itā€™s like itā€™s running totally different code.

Ugh! Well, thanks for letting me know.
Unfortunately I have close-to-zero electronic skills, let alone the time to figure how to re-solder a 7135 chip on the board.
I think I will contact Neal and point him to this discussion to see if I can get a replacement.

+100 to both statements

I would check the retaining rings for looseness (won't help if threads are not long enough) but if you're cranking it so hard that you force a ring to make contact that may be the issue. Wiggle the springs and see what happens they should be tight. +1 contact Neal for new head if you can't solder.


Done and done. ;)


I really don't want to steal too many things without giving credit. You guys please, please, please let me know if I step on any toes!!!

One of my samples has no turbo stepdown and also, the head does not get hot at all but the mcpcb looks to have been forced in and bent up the edges of the mcpcb so its obviously not touching the shelf. Will require mods but no turbo stepdown is fine by me :wink:

I knowā€¦ Just did It that for the video so you could hear when changing modes, because in mode1 & 2 sometimes it flickers when ramping is finished, didnā€™t want it to look like a mode change.

Already contacted BG, do you know if they can send out spare drivers?

It might also be that the step-down isnā€™t noticeable, especially if itā€™s not shedding heat the way itā€™s supposed to, or if itā€™s using a low-amp cell. Either one could cause the top two modes to look pretty much the same.

The lumen output curve for LEDs tends to rise, peak, then actually go back down as current and heat increase. If it canā€™t shed heat it could be past the peak, and a power step-down could then move it from the right side of the peak to the left side at roughly the same lumen level.

Or it could be something else entirely.

You are correct Tk, I just ran both side by side, the ā€œproblemā€ one has a barely perceptible step down. Thank you.

OK, I got three, 2 neutral, one warm.
Just ran one of the neutrals for several minutes at highest brightness ā€” not uncomfortably hot in the hand but definitely getting past warm to hot.

At this point I couldnā€™t control the UI consistently, because I couldnā€™t make a short enough pulse consistently.
Several times trying for short pulses I got 1-2-3-4-5-6-7-6-5-4-3-2-1 ā€” guessing it was at a temperature almost confused, and by the time I got it up to brightest, it no longer would recognize a short pulse as a short, but read it as a medium.

Would someone ā€” TK? ā€” be able by now to set out a test procedure to follow ā€” like letting it heat up for X minutes at turbo then doing steps ā€” that would show problems if theyā€™re going to happen?

Or is it time I opened it ā€” well, them ā€” up and looked at the solder?


And if I try to fix solder problems ā€” is there a small solder sucker available anywhere?
And with a temperature-controlled (Hakko) iron, what temp. should I set to be poking at this stuff?

The temperature sensitivity bothers me, too. I think using an X7R capacitor should help, as well as looking for better specs on the resistors. Hardware isnā€™t something Iā€™m good at though, so input from others would be very helpful. :slight_smile:

If I understand correctly, it should be possible to get the performance to about 85% (maybe 90?) with good hardware, instead of the 50 itā€™s getting now. But at least itā€™s better than the 10% it had initially. (or, measuring the other direction, timing adjusted by 1.15X when frozen, instead of the 2X it gets now, or the 10X it got before)

Unfortunately, wight left before the temperature issues were discovered, so there hasnā€™t been an electrical engineer involved to help fix it.

I've heard people have been using that small braided copper. Solder will follow heat. The copper has other helpful uses as well.

http://www.mtnelectronics.com/index.php?route=product/product&path=25&product_id=172

Hi, about the high-pitched whine some people reportedā€¦ I found a really easy way to measure it today.

The method is simple: connect an LED to a DMM, put the DMM in frequency mode, and shine a light at the LED. It works in reverse, converting photons into current.

I measured PWM speed on two production BLF-A6 units, and got the following: 15.49 kHz, 13.35 kHz (nominally 18.75 kHz)

So, there is a lot of variation and some of it is easily within the audible range.

Also, I measured my new attiny25 test driver running the blf-a6 code: 31.74 kHz (nominally 31.25 kHz). Thatā€™s plenty fast, and I might step it down to a slower clock speed to get it to 25 kHz.

Question about contacting the ā€œProfessional Guyā€ ā€”- will they listen to me if Iā€™ve opened up the driver to look at the solder?
Or should I contact them first about the temperature sensitivity ā€” or is that just one of those things for us in the first round to get used to living with?

If itā€™s the parts used, I guess Iā€™d expect to get used to it.
If itā€™s bad soldering, Iā€™d rather know that ā€” and be able to ask them to replace it, or try to fix it and then ask them to replace it ā€¦. or, heck. it gets really confused at this point about this issue.

Banggood canā€™t fix the temperature sensitivity. Thatā€™s more of a medium-term project, with research and testing and samples, and eventually new drivers.

The temperature thing was 80% solved before production, but the remaining 20% is quite a bit more difficult. (initially, when frozen, the medium-press time was 10x as long as normal, but on the production units itā€™s about 2x insteadā€¦ and the best we can hope for is about 1.15x)

I donā€™t know much about circuit design or hardware in general, but DEL has been very helpful so far. :slight_smile:

Now that is a creative solution, I had remembered that the LED will create current, but did not think to use it as a sensorā€¦ Geniusā€¦

The light you shine is the one you are testingā€¦ so this can tell for other lights as well. What brand of multimeter are you using?

Yes, as soon as I confirmed it worked I measured a bunch of lights. :slight_smile:

Iā€™m using a Fluke 8846A. Itā€™s mostly doing automated tests for work, but I can use it for other things between tests.

Just to let you know about customer service responsiveness, I wrote three days ago an email to the address I had of Neal, which is zhanghuaihong [at] banggood [dot] com and no answer so far.

For those who are having the "7135 problem" & have a soldering iron;

Try touching the hot iron with a small amount of solder on it to the ground tab of the 7135 for a second or two. You should see the flux on the board around the 7135 bubble a little.

Just doing this was enough to reflow the 7135, & fixed my driver issue.

.....Of course, I took the opportunity to add thicker wire & spring bypass while I was at it :)

Some people have talked about getting a different clip and some have asked about Olight clips specifically. I checked my Olight S20-L2 clip last night and it does definitely fit the A6 host. However, there is a problem with the Olight clip. It doesnā€™t hold on very well. I had always wondered why if I bumped into anything, there was always a good chance the light would drop. Now, I think I understand. Comparing the Olight clip with the A6-SE clip, the Olight clip doesnā€™t wrap as far around the circumference. So, a bump can cause the clip to let go of the light. Iā€™d forgotten about that until I bumped the door of my car getting in last night, and the A6-SE fell to the ground! :open_mouth: