The test process for the flashing LEDs is interesting, I wonder if the leds have a first-power-up test mode that does that 3-die test and then burns a fuse to disable it perhaps.
That's an interesting thought. From what I've seen, a lot of those self-flashing LEDs go through a fairly quick sequence at startup anyway before they go through other cycles, which could be a production optimization? I've gone looking for RGB LEDs that specifically have a slow color cycle for a silly little lamp project, and they're hard to find.
Considering the LEDs start off on a leadframe, and some customers might want to receive them that way, I'm a little surprised to see them being bade in short strips rather than as a continuous process. Maybe it's too hard to get that reliable, as any jam could hold up the whole line
Seems like it would be hard in a lot of ways, given the number of steps involved. You'd either need a big and very complex line to do everything in one go, which would be a ton of work to set up and change over between products, or you would need a way to handle and store big reels of partly-assembled LEDs between steps without damaging them (especially between wirebonding and encapsulation). The little trays of strips seem to be a good compromise between throughput and handleability. I'd be curious how much volume companies like the one in the video do compared to some of the big LED brands, in terms of quantity per PN. Maybe a big continuous line makes more sense if you run batches of a million, but I would guess companies like we see here are set up to be a bit more nimble than that.
We try very hard not to use through hole LED's at all, preferring SMD wherever possible.
I was wondering if that might be a reason this video is available now. Perhaps the writing is on the wall for this type of LED so they're not worried about some competition using the info in the video to start up and undercut them - maybe fairly soon they will be the sole, or one of the very few, manufacturers left.
I doubt there's any real secret sauce in the video. There will be a ton of invisible details that are critical to the process, like specific recipes for preparing the encapsulation and phosphor materials, tuning the speeds/forces/quantities used in the automation equipment, etc. It's kinda like trying to replicate a moderate-complexity electronic device from a photo of the PCB: you can kinda see what parts are involved and, with some knowledge, fill in the gaps about how everything is connected, but it will still take some work to figure out what specific parts are suitable, detailed layout considerations, etc.