robbak Yes,it doesn't really matter. The power levels will be less - To roughly calculate it - to get the power you can expect, multiply your supply voltage by 0.7, square that, then divide by your speaker impedance. So if you are using a 19 volt notebook PSU like most of us, that's ((19*.7)^2) ÷6 , or 29.5 watts

robbak No, the circuit is powered from Vi+ and Vi-, via a 1k resistor and controlled by ZD1. There is no isolation - G is connected directly to Vi-, which raises the question of why an opto-isolator is part of the circuit. As for how much power - the chip and opto's LED is powered through a 1k resistor, and the mosfet's gate through a 10k resistor, so it is not going to be very much!

robbak It uses two AP3050D class-d amplifier chips, with a small stick-on heatsink. I doubt it will ever do 100 watts RMS - they really are getting that rating from doing P=v^2 over R on the power supply voltage through a 4-ohm speaker. Pushing 6 amps through the chips's mosfets are going to overheat them. If you are doing that, you'd need a better cooling system, and pushing a 24v peak sine wave through 4 ohms doesn't get you 100 watts RMS! Mind you, class D amplifiers with a H-bridge mosfet array are a little bit magical, with what they can do with a single-ended supply.

robbak No. These are simple cyan LEDs and do not produce any ultraviolet. Germicidal UV LEDs are expensive, have a distinctive appearance as they have a metal construction with a clear fused quartz window, and create no visible light. Selling a cyan LED lamp as germicidal is simple fraud.

robbak You do not need an expensive UVC LED to harden glue. A simple and cheap near-UV purple LED will do fine for that. Indeed, UVC is far too energetic, and will damage the glue and your eyes.