ISO3082: PWD received pulse deflection tested is twice as large as datasheet

Hello Hui, 

Thank you for reaching out. I am not sure I understand your question. Based on your description, it sounds like the Pulse Width Distortion (PWD) is twice as large as the value shown in the datasheet. Would you be able to share a scope shot of the waveform showing a higher than expected PWD?

yellow is ISO3082

and green is output

The datasheet guarantees a PWD of less than 200 ns. As far as I can see, the propagation delays are symmetrical. Please zoom in further and measure with cursors.

(The ISO3082 has two output pins. V_OD must be computed as A − B. The propagation delay is measured at the 50 % points of the edges, without a capacitive load (i.e., without a cable). See figure 7-4.)