Unfortunately my program doesn't show complex impedances (or I just don't know how to make it do that) but according to the passband/retloss charts, plus the Smith chart (after 40 years, still trying to get my head around that one) it appears to be 50 in/out...might want to plug the values into your program and see what you get. ?I just halved the component values and a nice pattern appeared centered on 14mHz...all the hard work has already been done designing the original filter.?
PS I've found that tinkering with the input/output capacitor values (moving them up slightly higher than 50 ohms) seems to give you a much better return loss curve.
