Based on the two provided marking codes, the actual part of the marking that is likely to be useful will be 41 or 418, with the Q1 or Y2 suffix being related to manufaturing location/date code/batch/whatever.
No idea on the actual part, but on first glance to me with that resistor divider I'd assume a P-Channel MOSFET controlling power to some parts of the circuit.
The reverse engineered schematic by fripp doesn't make much sense and so may have some errors. For example pin 1 of this SOT23 device (potentially the P-Ch gate) is shown connected to two 10k resistors R30 & R32, but we can only see one in the photos. That would be R32 which would thus turn on the P-Ch FET when U3 optoisolator is on. But the power for the LED in that opto comes from the same supply which apparently originates through this same unknown part?
Another weird section of the schematic is around Q1 which seems to be a zener diode based regulator for the PWM controller power, but the supply comes from Vdd and it outputs also to Vdd. Clearly that wouldn't be right.
I guess this could be part of a latching power enable, but it is hard to work out with potential schematic errors and unknown connected circuits on the 3 pin headers.