Very well said, I think I even understand it a bit better now.
In looking at the above figure, the following may explain why you were seeing bubbles in the sight glass.
As the system head pressure falls below the set point of the headmaster, the valve starts to modulate - closing off port C and opening port R by the same ratio. This results in refrigerant backing up in the condenser as discharge gas begins diverting directly into the receiver.
Conversely, as the head pressure hits the set point of the headmaster, the valve starts to modulate once again - opening port C and closing port R by the same ratio....resulting in increased flow of refrigerant through the condenser into the receiver and restricting discharge gas from dumping directly into the receiver.
More than likely while you were adding refrigerant to the system, the discharge pressure temporarily exceeded the headmaster set point and therefore refrigerant was not allowed to stack in the condenser - so your system was sufficiently charged.....briefly.
Once you stopped adding refrigerant and the system was allowed to stabilize, the head pressure most likely dropped below the headmaster set point and refrigerant started stacking up in the condenser - and where you barely had enough refrigerant in the system before, now there simply isn't enough of a charge to both stack in the condenser and maintain a liquid level in the receiver.
Start adding refrigerant again and the cycle repeats.....until such time there is enough of a charge to allow the headmaster to do its job and form a liquid seal at the outlet of the receiver.
If you spend some time studying the figure and reflecting on how subcooling is derived, it should also become readily apparent as to why charging by subcooling simply is not an effective way to determine an acceptable system refrigerant charge.
I haven't clicked on JPSmith's link, but I would imagine he is directing you to Sporlan's step-by-step procedure for properly charging a system such as the one you are working on.