KC1 keychain light is available now

Agree,with firefly mode

It already has a firefly mode. Set the aux LEDs on low mode.

Not much redesign would be needed.

Anduril 2 already supports tint ramping. You might need another contact pad on the driver for the extra wire to the star, but that should be the extent of changes to the driver.

A new quad star would need to be made with 2 of the main LEDs on each channel. But making a new star shouldn’t be difficult. Such stars already exist on the Sofirn IF25 and Lumintop FW4X.

It should not be necessary to make any changes to the aux board, though it might be cramped connecting the extra wire around it. However, if that became a problem I’d rather have tint ramping than the aux led board anyways. Just ditch the aux board. The light could still have aux LEDs in the switch to make it easy to find at night.

Of course it needs hardware changes… it would need two regulated power channels, currently it only has one. K9.3 already had two + two led channels, hence the firmware change only.

That’s not the extent of changes needed for the driver. It needs an entire second power circuit, isolated from the first one… and it’s already a tight fit with just one.

Ok. Good to know.

Still… if Lumintop can fit it into an FW4X with similar sized driver, and if Mecarmy can fit it into a PS14 with a narrower diameter driver, it should be possible to fit this into a D4V2.

Plenty of space on the 22mm driver, especially since the DD FET channel becomes unnecessary, the wires would be more problematic but still doable by replacing the relatively thick aux silicone wires with thin Teflon wires, the main negatives wires could lose a gauge and use Teflon as well. I mean even in the current D4v2 it’s a bit of a mess, personnaly when modding one I replace the aux wires with Teflon, much more room to work with.

The K9.3 currently has two sets of firmware:

  • Tint toggle
  • Tint ramp

The tint toggle method was the original, where 3H simply switches from one set of LEDs to the other. It’s like having two separate lights, with only one powered up at a time. This is instant, and effectively uses a different ramp for each set of LEDs because one is a 5A linear and the other is a 9A linear plus a DD FET.

The tint ramp method is brand new for this light, and is a little more complicated…

  • Ramp levels 1-130 use smooth tint ramping between both sets of LEDs, with only the linear power channels. This has variable tint resolution depending on how bright it is, so high levels are very smooth but the lowest levels are less smooth when ramping “sideways” between tints. However, the up/down brightness ramp is smoother than before, since it uses the new dynamic PWM thing from a few weeks ago.
  • Ramp levels 131-150 smoothly turn off both linear channels and power up the DD FET instead. So turbo is always the 9 main LEDs, no matter what tint is active elsewhere in the ramp.
  • Switching from the 9 main LEDs to the outer 3 LEDs takes about 4 seconds because of the smooth ramp between. This allows all the in-between blends to be used, but isn’t as fast if the user wants only the endpoints.
  • At the far ends of the tint ramp, past the end, there is an auto-tint mode. It goes from one set of LEDs to the other, based on brightness.

If you like to use one set of LEDs or the other, but not both at the same time, the toggle method is better. If you mostly want to use a blend, the smooth method is better.

There are also some other recent changes included:

  • added an option to select turbo style / 2C style: Anduril1 style, Anduril2 style, or no turbo. This is configurable for both simple UI and advanced UI, so people can get turbo in simple mode now if they want.
  • added an option for whether hold-from-off should ramp up or stay at the floor
  • made smooth ramp speed adjustable (can be slower)
  • enabled auto-lock in simple UI (but it requires going to advanced mode to turn it on)
  • made the ramp auto-reverse window a bit shorter, so it’s easier to ramp up, stop, then ramp up again
  • made it easier to enter large numbers in config menus… click for +1, or hold for +10
  • added a configurable “jump start” to get moon level to turn on faster on slow drivers (needs to be configured per-light, though, since hardware varies so much)
  • added dynamic PWM (a.k.a. PFM, pulse frequency modulation) for more precise brightness resolution at the bottom of the ramp
  • fixed audible whine on DD FET modes on some lights
  • fixed corner cases in the hybrid memory timer, so it resets reliably to the saved level

Basic usage is still the same though… click for on/off, hold to change brightness. All this extra stuff is just improving things people asked for.

Oh, um, before anyone asks… the “tint toggle” and “tint ramp” firmwares are not going to be merged into a single runtime-selectable option any time soon. They use totally different code, and the toggle method uses a bunch of custom overrides which aren’t compatible with anything else. These two don’t even use the same hwdef file. The differences between them are deeper than it might seem.

Here is a very quick component placement for a dual linear driver (no tracing, no layout optimizations)

Plus as bonus, dual current sense for a 1 000 000:1 dimming range (i.e. super low moonlight), which would also solves this :

That is if HDR/dual sense is supported, and supported for tint ramping (which looks like it would need a lot of changes to the tint ramping code).
Alternatively two small FETs + resistor channels could be used, that means supporting 4 total channels and 2 channels per LED tint ramping.

Anyhow putting the high dimming range aside, it would have 8 components on the back, only one more than the D4V2.5 driver. In term of cost, probably similar or if more, not by much : smaller cheaper linear FETs since only 2 LEDs per channels, no DD FET, since the 2 linear channels should be capable of close to direct drive performance. The op-amps being the most expensive parts, doubled, but slightly less precise and cheaper ones than in the D4V2.5 can be used if dual sense or resistor channels are used since thoses give much more dimming range.

Hank, in anticipation for your next project, i cobbled this together. Thanks for selling me just the driver and switch for this mod!

Nice, by the way, Nichia 219BT-V1, R9080 Warm White, 2700K is available

Are the Nichia 219B 2700k new or refound. Any measurements?

Hank, can I order KR1 with 219B?

A quicker option is to email him directly

I did but no response. That’s why i stired the water here.

I doubt it. I’m not sure, but I think the KR1 driver’s regulated circuit only goes down to 5A as the lowest turbo level. This is fine for the W1/W2 LED options, and for XP-L and SST-20… but Nichia 219B can’t really handle that much power. It’s rated for 1.5A and really should be kept at 3A or less. So Hank would need to make a lower-powered driver for it, or use special firmware which never goes above 50% or maybe 60% power.

For some lights, that’s worth doing… But in a thrower, 219B doesn’t really make much sense.

nice! 2700k is much more exciting than 3500k for me! Thanks Hank!

Thx TK for explanation!

Hank just emailed me saying to be patient, they are soon releasing small HI CRI thrower…

Yeah. I don’t know much about it, but it sounds interesting. I sent him a quick test firmware just to help verify if the driver works, and to check if I’ve got the right pinouts… but I haven’t heard any more details yet. Just some initial hardware-enablement type stuff.

Meanwhile, I’ve been enjoying a KR1 W1 in 18350 form. It’s small enough to lose in my purse until I need it, and throws way farther than such a small light seems like it should be capable of. I leave it in lockout mode, with a manual memory level set somewhere practical, so I can just pull it out and use 2H for momentary at a useful brightness. The beam is similar to an Emisar D1S, but the light itself is only a fraction of the size.