Tool/software:
I connected a PWM GPIO to the input and connect the output to a 50R oscilloscope input:

When the PWM (70KHz) is falling done (falling edge from 3V3 to 0V), I have a 1V gap for 50nS. Do you know where it is coming from?
Note I have the same issue if if I connect or not the scope on the F_50R. So it looks, the gap is done internally by the THS3091.
Yellow is TP8
Green is PWM GPIO
Blue is F_50R connected to a 50R scope input.
This one is when the output is not conned to the scope
This one is the screen shot when the ouput is connected to a 50R scope input.
Thanks
Regards,
Damien