I too connected an output to toggle once per scan, but instead of an oscilloscope, I connected a data acquisition instrument to it. (the advantage is I can review what I recorded.)
Anyhow, the scan time matched the S? bit (SLC5/04), but I discovered something interesting in the process. The scan rate was steady at 4msec, rock solid. However, when I went online with RSLogix500, every tenth scan or so would be 11msec. I attributed this to the the frequency at which RSLogix was being updating over the comm port.
For my application this was found to be a problem as it introduced variation into the load cell measurement being performed once per 50msec.
Congrats on using the oscope. It seems so few do. You've provided some degree of proof. Data driven decision making; your boss should have given you more credit, even on the side.
CeCo3