Code now public! BLF A6 FET+7135 Light. Short 18350 tubes and Unanodized Lights Available

BLF Special Edition Products
781

Yeah :bigsmile: sounds like everything is coming together nicely.

Now it is up to EE to produce this thing, at least it is so good that even if they fumble away a bit of performance on the top end, it would still be an EPIC light with all the rest of the features :bigsmile:

It is almost too good, all of their other models will look worse in comparison :wink:

782

That’s because you’re using different resistors on your driver than what I calibrated for. It thinks your battery is over-charged.

The battery-check blinks are:

  • 0 blinks: empty (< 3.0V), recharge the battery ASAP
  • 1 blink: 1% to 25% (3.0V to 3.5V), getting low and should recharge soon
  • 2 blinks: 25% to 50% (3.5V to 3.8V)
  • 3 blinks: 50% to 75% (3.8V to 4.0V)
  • 4 blinks: 75% to 100% (4.0V to 4.2V)
  • 5 blinks: over-charged (> 4.2V) or 4.35V cell

So, each blink is about 25% of a full charge.

We’ll nail down the exact resistor values and re-calibrate before the actual lights are made, but for now it seems our two prototypes are slightly different.

783

And with a fresh charged C5 at 4.21V it is still showing 5 blinks, so that’s consistent.

And 6.35A at 1608 lumens from this Sony C5. Earlier tests showed my LG HE-2 to be a bit hotter even. :wink:

784

It might be a bit brighter than the original firmware used for testing a few hundred posts ago… I discovered that turning on both power channels at full actually reduces the output, so the turbo mode is now using only the FET.

My guess is that the 7135 chip was trying to regulate power and actually just causing unnecessary resistance, so bypassing it for turbo makes it a little brighter.

However, the 7135 chip is enabled for all other modes, because it also makes the output more stable. So, the FET+7135 modes will still sag as voltage drops, but not as much as if it were using only the FET by itself. And, of course, the 7135-only modes stay pretty constant regardless of voltage.

785

I'm in for another light, tint NW for 2 total. RonMuller #277 on the master list. Hopefully we see #501 added to the list soon!

Is there going to be a master checklist or some other way of verifying that the final production light meets or exceeds the initial specs listed in the first post? Having said that this light looks amazing so far which is why I've increased my order to 2 lights. Thanks to all the people working hard on this!

786

This gets better, and better,and better, and better…

787

Thanks for the info update TK, very interesting.

@Dale, over 1600 at 6.35A :bigsmile: nice, i am curious what your hotter HE2 battery could pull, maybe up to that nice round number of 7A :smiley: or maybe it is just that the C5 have less voltage sag+no 7135 at max now that is making it do over 1600.

788

I have noticed that some flashlights for one 18650 cell have screw-in extension tubes available so that one or more extra 18650s can be used in series.

Will these extension tubes fit the A6, and is the driver designed to do such a thing?

Thanks

789

Maybe one of the “team” can answer some questions:

1. Does the convoy pocket clip fit on this light?

2. Is the reflector bigger/smaller/the same as a convoy S2?

790

No. Sorry. This light does not have the option of running 2 cells in series.

791

I have an extra clip. I want to know that answer too so I guess I will dig it up and try it.

The reflector is different than the S2..ie shorter..more like an S2+ but I believe deeper. I need to take a couple of mine apart and place them next to each other and take some shots.

792

Please take some measurement on the reflector when you do that, i want to see if there is a possible drop in smooth one somewhere to buy.

793

Sorry, this light can’t use two cells. That requires a special driver. Even if you found an extension tube, trying to actually use it would destroy the driver and/or the LED.

As for the pocket clip and reflector… not sure. I can’t really answer that. I have the driver and emitter hooked up in an air-wired test host, but I don’t have an actual EE A6 yet.

794

Count me in for 1 NW.

Thanks for putting this together. Maybe this will make up for my missing the last BLF SE. (X6)

795

Can someone please inform what mode will the driver start at? (I know it’s probably in here somewhere, but nearly 800 posts!)

7 modes seems like a lot to scroll through to get to the ones I’d use the most. On second thought I guess more is better and it will be much faster than the ramping UI’s.

Also waiting for someone to report on the fitment of the convoy/fasttech clip.

796

The driver starts in default moon mode, this can be configured to have memory by soldering a star or otherwise grounding a pin on the MCU. There is no star at this point, I think Wight is looking at the potential for a re-design such that a star might be put on the battery side of the driver for a user to be able to configure memory/no memory at will. Default is no memory.

I don’t have a clip for any of these so I don’t know. The reflector is of a different build, shorter (not as deep) as the Convoy S6 and S2 units. It seems to have a similar top lip as the X6 reflector that it’s o-ring fits on, but unless I’ve reassembled mine wrong the A6 o-ring goes between the lens and bezel. It’s deeply ribbed on the sides so it should help with heat sinking to some extent.

797

It’s not finalized yet, but what it does right now is:

  • Starts on the lowest mode, or on the last-used mode, depending on whether the config star is soldered. The star should be on the battery side of the driver where it’s easy to access, if wight’s design works out.
  • Short presses go to the next mode (low to high order), wrapping around from high back to low.
  • Medium presses go to the previous mode (high to low order). When already on the lowest mode, going backward like this will enter the hidden modes (turbo, tactical strobe, battery check (which also works as a beacon)).
  • The number of modes is either 4 or 7, depending on which mode group you choose. The 4-mode group is low/med/high/turbo, the 7-mode group is moon/low/med1/med2/high1/high2/turbo. Changing mode groups requires a bunch of uninterrupted fast taps and shouldn’t happen by accident.
  • With mode memory disabled, the following modes can all be accessed within one second: moon, low, med1, and turbo.
  • Turbo steps down to the next-brightest mode after 30 seconds to avoid melting any parts. IIRC, the next-lowest mode in the default config is still about 3 amps, so it’s still pretty bright afterward.
  • Modes are all spaced evenly on a visually-linear scale according to the cube-root model used by CPF’s “selfbuilt”.
  • The battery check mode uses a friendly “4 bars” model like the signal meter on a cell phone, calibrated based on several years worth of battery capacity data from HKJ. Each blink represents about 25% of a full charge.

Everything above is subject to change depending on what EE can actually deliver, and depending on what gets decided for how best to accommodate everyone’s preferences. So, still not making any promises… but everything above is already functional on the prototypes (except the solderable star, which the firmware supports but hasn’t actually been physically built yet).

798

“battery check (which also works as a beacon)”

Hi ToyKeeper, is that continuous slow blink or true beacon mode?

799

Neither. It blinks out the battery charge (0 to 5 blinks, depending on voltage), waits a couple seconds, then repeats. Each blink is about 100ms on a low-ish mode, with 400ms between blinks and 2400ms between cycles. This is not set in stone.

Personally, I like to use a “heartbeat” beacon which blinks twice (very short 5ms pulses) then pauses, then repeats, at about the same rate as a human heart. Or a dual-level beacon which “stutters” out a few bright flashes then rests at a lower steady output, then repeats, again at about the same speed as a heart. But there are only like 10 bytes left in the ROM so it’d be pretty difficult to add another mode. Extra blinkies were sacrificed to make room for configurability without reflashing.

The code will be fully open-source so people can change it and reflash if they want.

800

ToyKeeper, one word, fabulous!
Rammed into 1024 bytes. All hail BLF’s Moon Goddess!