Noctigon KR1AA is available

[jon_slider]

@ToyKeeper

Could you make sure they have the same product number, same firmware version, same configuration, and same ramp level, and then measure the output ripple

Good idea.. instead of using same lumens I tested at same ramp level this time:

Both have Firmware 0161-2025-07-07
Both have the same ceiling 110/150
Both are in Advanced Smooth Ramping
Both outputs are at floor level 50/150:

different LED type, CCT, and Lumens:
KR1AA w SFT-70 5000K, @ floor level 50/150 = 42 lumens,
D3AA w DD 519a 4500K (gives 3400K), @ floor level 50/150 = 19 lumens

(white balance locked at 5000K)

I think the drivers are also set to different Voltages.. 6V for KR1AA and 9V for D3AA. Maybe that factor affects the results?

[cannga]

@ToyKeeper

@jon_slider

lower Flicker Index

It’s strange that there’s any measurable difference. Regardless, a 17% ripple at 37 kHz is way beyond the limits of human visual perception.

I have not kept track of which light, but have seen and wondered about these shallow modulation in lights with regulated driver.

I could think of 3 possibilities:

1. Real; if so the question I have is are such shallow modulations used as a way to control brightness?
2. Unintended electronic noise of driver causing flicker.
3. Error of Opple.

[ToyKeeper]

@jon_slider

I think the drivers are also set to different Voltages.. 6V for KR1AA and 9V for D3AA. Maybe that factor affects the results?

Good point. They won’t be comparable because they’re different voltages, so the regulator will behave differently even with the same firmware settings.

Your results demonstrated how much it can change from one ramp level to another though, even on the same light with the same battery.

@cannga

I could think of 3 possibilities:

Most regulators work by quickly setting an output level, then reading the result, then setting a new output level, then reading the result, and so on… so they naturally tend to oscillate. Most have configurable settings to controll the oscillation speed, because faster oscillations look better but use more power for the control chip, while slower oscillations are more likely to be visible but allow the control chip to use less power.

Additionally, the regulator has only limited resolution for the steps it can set the output to. So it’s unlikely it will be able to hit the exact level requested. It just tries to get as close as possible by oscillating between nearby levels.

[Dc38 (The Glutton)]

Not exactly..i couldn’t see it per se, but i noticed it.

[Dc38 (The Glutton)]

First run only showed abour 4.5 days…probably a fluke.

For the sft 70, 1/150, second runtime test on a full 2500mAh NiMH. 1.38V per 8c corrected values. 0400, 2/6/2026.

[LedTed]

I did try the LT FW1AA deep carry clip on a N-E KR1AA.2.

Original KR1AA clip left (black torch). FW1AA clip right (gray torch).

The locum clip got more springy than when installed on its intended FW1AA, but still spun with little effort.

This combination would not deep carry when clipped to the thin fabric of my pajama bottoms.

When I put the ring portions of the clips together to compare thickness, I could discern none by eye or feel.

[Dc38 (The Glutton)]

The mischiveous lil bastid in the background, lol…

[Firelight2]

The stock KR1AA clip will never have great retention as it is too thin. However, the retention it does provide CAN be improved.

Just remove the clip from the light. Then using pliers:

• Bend the ring downwards then release. It will spring back most of the way, but should still have a slight tilt downwards that wasn’t there before the mod.
• Bend the clip, from the tip to the middle of the curve, downwards. The clip should spring back almost, but not quite, all the way back to its original position after bending.
• Don’t overdo it. Try bending maybe 20 degrees first and see what it looks like. If there is no effect, you might need to bend it more.

The effect of this mod is to increase the pressure between the tip of the clip and body of the light and it just takes a couple minutes. If done properly, it doesn’t affect the look of the clip.

Note: it might make tailcap battery changes slightly more awkward as the clip will tend to want to slide upwards when the tailcap is removed.

Also, use caution when bending, especially if it is your first time modding flashlight clips. If you overdo it you could damage the clip.

[jon_slider]

@ToyKeeper

If the battery is weak, it does 3 flashes to warn the user… and limits the maximum power.

My Raw Aluminum KR1AA V1 from Jackson does this 3 flashes error very often when I load the battery through the head. It almost never does it when I load the battery through the tail.

btw posted a POVD bug here

I think it is a good idea that Hank glues the head of the V2 model as that will mostly prevent this Anduril weak battery detect error.

in other news:

Im happy to report I just got my Shipping Notice for a Polished Silver KR1AA V2 w NTG50 2700K and magnetic tailcap, that I ordered from Hank on Jan 20. Woohoo!

[eckyeckypikang]

Got my shipping notification this morning, too, for my KR1AA & a single-channel D2 with NTG35 & SFT25R in 50000K five-thousand Kelvin.

I was kinda holding out hope that the longer the delay, the better the chance of tweaking the clip and improving the finicky connection problems. But I’m looking forward to getting my hands on it regardless…

[Dc38 (The Glutton)]

50000k? Do you heat your soup with that thing?

[eckyeckypikang]

Lol

How else does one expect to signal the aliens?

oops

[horzuff]

Dang, my order from 14.01 is still MIA, no shipping notice yet. Maybe I ordered something problematic

[gravelmonkey]

… Actually…

Googled it because I couldn’t resist…

50,000 CCT is a nice typo as it seems to be about as hot as stars get. According to scientists, this is typical for a young white dwarf star, and would have peak wavelength of 58nm (Vs, for example red at 620nm and long wave UV at 365nm). Apparently 10-120nm is referred to as Extreme UV.

Given the atmosphere absorbs UV radiation, I’d assume the 50,000K light would get attenuated quite quickly and therefore be no good for signalling aliens…

Source on stellar spectra

[eckyeckypikang]

Lol - this is why I love this hobby!

[sp5it]

My KR1AA with SFT-70 3000K 95CRI

[ToyKeeper]

@jon_slider

btw posted a POVD bug here

I answered on Reddit too, but I’ll copy it here. BTW, Reddit is not a good place to post bug reports. That’s what GitHub is for.

@ToyKeeper on Reddit

It’s not accidental. It does that on purpose.

In aux-leds.c, a “pattern” of 0/1/2 means off/low/high aux:

    #ifdef USE_POST_OFF_VOLTAGE
    // use voltage high mode for a few seconds after initial poweroff
    ...
        // use high mode if regular aux level is high or prev level was high
        #ifdef USE_AUX_THRESHOLD_CONFIG
            // always high if configured for high aux
            // otherwise 0/1/2 depending on recent main LED brightness
            if (pattern != 2)
                pattern = (prev_level >= cfg.button_led_low_ramp_level)
                    << (prev_level > cfg.button_led_high_ramp_level);
        #else
            pattern = 1
                + ((2 == pattern)
                   | (prev_level >= POST_OFF_VOLTAGE_BRIGHTNESS));
        #endif
    ...
    #endif

A high standby aux brightness explicitly overrides all other logic for choosing how bright POVD should be. Because it makes sense to do bright POVD for a moment before turning aux to “low” or “off”, but it doesn’t make sense to do “off” or “low” POVD for a moment before bumping the aux up to high for long-term standby.

To help clear up confusion about this, I added a commit to the manual to document this better.

[Dc38 (The Glutton)]

So, theoretically, it’s darn near technically perfect?

[jon_slider]

@ToyKeeper on Reddit

it makes sense

thanks for your reply

I would prefer POVD options 3 and 4 to work consistently at any Standby Aux level

[Dc38 (The Glutton)]

Is there a way to have options for the POVD fade in and out gently as well? The sudden blap, blap of the aux lights is kind of disorienting in the dark…