Hi,
why aren't you using your scope in combination with a low valued resistor to determine the current sourced by your PWM source, as suggested in post #10 ?
By investigating a waveform with a scope you should use a low Volt/Devision setting, which gives more insight by means of filling the screen in y-direction as much as possible (e.g. 500 mv/Div instead of 5 V/Div). Further, according to your recorded waveform it looks like your PWM never reaches the off-state (0 V). Also your PWM freuency is seems to be pretty low, is it your actual implemented freuquency?
BR