please help in justifying these things in microwave filter

Hi, I designed band pass microwave filters using Agilent ADS for my project, and the simulations were perfect, but when I did the measurements, The whole results were relativily good but not like the simulated ones so I put the following questions that appear after analyzing the results.

why there is not a good reciprocity Snn not to some extents equal to Smm?
what cause the insertion loss at my resonate frequency instead of starting from zero,it started from -5db?
what causes my bandwidth to be narrower?
what type of losses I have to take into account when dealing with filters?
(mismatch, coupling, parasitic elements)
how can I recognise these type of losses by using just only the s11,s12,s21,ans s22 data file from VNA,which for which.

thanks

Hi

How is the filter realized? Harpin, Combline...etc?

Regards

Hi
It seems to be calibration error

also don't forget ur dielectric material , and conductor finite resistivity so it , also how accurate ur PCB realization ?

also the connectors , and soldering

u need also to be sure how accurate ur calibration of the VNA ?

khouly

my microwave filters consist of four planar lumped filter and one planar distributed filter(parallel coupled filter).my substrate RT/Duroid with dielectric constant 10.2,tanD=0.0023,thikness 0.635mm, my frequency range is 6-14 GHz