Since then, I completely rewrote the button input handling system. It’s now quite a bit smaller and also lacks bugs like the one you noted.
I still have a couple other bugs to hunt down, but they only affect lights with aux LEDs, so it’s not relevant for the FW3A. The other bugs are also incredibly rare, to the point that I can’t even make them happen on purpose, so it may take a while to confirm any fixes.
In any case, there is no more lock-up after long click sequences.
Perhaps since Lumintop only intended to test anodizing, they didn’t bother cutting actual threads? That’s why the tubes look different and why the parts couldn’t be threaded together.
Great, what version has been fixed? I am not sure it would be worth flashing again, Its not really an issue for me. I think it could be a handy feature.
It’s probably not worth reflashing unless there are specific changes you care about. The event system rewrite has virtually no user-visible changes. Mostly, it just … doesn’t hang any more, and the lockout mode’s momentary moon now lights up on every press instead of just the first three.
Other recent changes include:
Fancier “blinking” mode for aux LEDs. Instead of being on high for 0.5s every 4s, it makes more of a clock tick or slow heartbeat pattern. (…oOo……o……oOo……o……oOo……o…… repeating)
Moved off mode’s aux LED config to “7 clicks from off” instead of hiding it inside lockout mode.
Added a moon timing hint, to help people time the button release correctly to stay at moon from off.
Added support for some new lights.
Fixed a stepped ramping corner case when two steps were exactly 128 levels apart.
Fixed a small thermal behavior corner case.
Broke and fixed sunset mode.
Added more documentation.
Source code and build system improvements which don’t affect functionality.
Most of this makes little or no difference to the user interface.
Looking forward, most of the stuff which needs to happen is just hardware support… but more visible changes could happen too:
Tint ramping (on the lantern, and any other dual-tint lights).
Factory reset function.
Make the ramp auto-reverse in more modes, not just the main ramps.
Rewrite the thermal regulation… again.
Improve interaction between LVP and aux LEDs.
Configurable timer for sunset mode.
Possibly make strobe duty cycle configurable, or add a randomized strobe.
Maybe other new stuff, depending on what new features appear in new hardware.
More progress on grey anodizing… I think this was the second attempt to get a nice dark grey color:
About the parts not being together, Neal tells me: “not the production item, they will cut some screws before assemble”.
So the theory seems to be correct. It looks weird because the threads aren’t there. If I understand correctly, the ano factory took these pictures. And since the threads are cut after anodizing (so they can carry current), it wouldn’t be possible to put things completely together at this stage.
I think it may be caused by a small resistor in the circuit which is specifically designed to take the edge off each FET pulse. This was done to prevent excess voltage spikes and electrical noise within the driver, but it also seems to have the effect of increasing FET efficiency slightly.
Hopefully there is a cold blue steel tone to the original, similar to what is seen on my old Nokia 830 dumb phone.
If so, would some call it ‘gun metal’ ?