Same as Southern Man suggests, and especially that the lights staying on in run 3 confirms that power collection is adequate.
I would therefore give priority to checking the connections from where the two wires to the motor brushes are connected to the board, the connections to the brush terminals on the motor, and the state of the brushes (if visible). It only needs a poor connection that breaks intermittently to cause the hesitations you are seeing in run 3.
(The hesitations and light flickering in runs 1 and 2 are typical of a mechanism that hasn't run for a long time, power collection unreliable until the moving contacts in the circuit begin to polish up against each other and conduct better. Possibly also mechanically a bit gummed up with dried lubricant in the drive trains, which has been pretty much sorted by run 3. I don't hear any mechanical distress when the loco pauses in run 3, which suggests that the problem is loss of power to the motor.)
There's likely going to be more improvement possible once you have reliable running.
I would therefore give priority to checking the connections from where the two wires to the motor brushes are connected to the board, the connections to the brush terminals on the motor, and the state of the brushes (if visible). It only needs a poor connection that breaks intermittently to cause the hesitations you are seeing in run 3.
(The hesitations and light flickering in runs 1 and 2 are typical of a mechanism that hasn't run for a long time, power collection unreliable until the moving contacts in the circuit begin to polish up against each other and conduct better. Possibly also mechanically a bit gummed up with dried lubricant in the drive trains, which has been pretty much sorted by run 3. I don't hear any mechanical distress when the loco pauses in run 3, which suggests that the problem is loss of power to the motor.)
There's likely going to be more improvement possible once you have reliable running.
