could not find sunset in a quick search, maybe TK has a video and I just didn’t see it in my quick search. Then again, probably not very interesting unless it is time lapsed.
Awesome, for anyone wondering I just timed the Neutron and the time is roughly 6 seconds from min ramp to max ramp so thats significantly faster on the FW3A. :+1:
Hmmm, I guess they’re sort of like blinky modes. They don’t look realistic to me, but I’m not sure what I was expecting. OTOH, they’re well done; I don’t see any sign of a repeating cycle.
As we discussed, it probably has less transmissivity but light reflected off it hits the reflector and has a chance to return - so overall it ended up with 6% loss.
If the second-chance emission hypothesis is true, TIR lights like FW3A won’t benefit from it because most of the light that bounces off the lens will hit flashlight walls eventually.
Sunset mode doesn’t look very interesting. It starts at low, gradually ramps down for an hour, then shuts off.
Candle mode was linked a few posts ago, though the video doesn’t show its auto-shutoff timer. It goes at regular brightness until the final minute, then dims, sputters, and shuts itself off.
Candle and lightning modes are completely random and don’t repeat. It uses noise from ADC readings as a source of real-world entropy, so the random number generator is actually random and never does exactly the same thing twice. It’s not crypto-grade randomness, but it’s pretty good for a 90-cent chip.
Sapphire is not a glass, it is grown as a single crystal of Aluminium Oxide.
It’s use in high-end watches is it’s extreme scratch resistance (it is nearly as hard as diamond). Actually quite easy to make, but not to shape.
Apple thought for a while that they would use it for the screens on top iPhones, but that project didn’t work, so they use Gorilla Glass like everyone else.
Except for the camera lense cover, where I think they still use sapphire for durability.
It’s added into the mix on every ADC reading, so 8 times per second. It’s not the only data source though. So the result is highly chaotic and genuinely random, with all bits constantly in flux. The lower bits are a higher quality of randomness than the upper bits though, so it’s recommended to do things like “rand() & 0b00001111” instead of “rand() >> 4”.