coaxial feeding printed antenna

hello;
im going to design a printed antenna withcoaxial feeding resonants at 3 and 5.5 GHz at the same time.
i modelled this antenna with cst and hfss respectively. simulation results are fitting. so i fabricate this antenna and get the measurement results. but there are differences between simulation and measurement. i dont know what is wrong?
can you help me

Looks like a considerable difference in substrate Er or another parameter that influences the effective wavelength.
Also the impedance matching is apparently affected.

sorry;
i atteched wrong result.
this is the true one

Hello,

results are close but higher frequency resonance is off by considerable margin. Howz the manufacturing quality of antenna?

the manufacturing seems ok, but i dont know why are the results so different.
as you see, i modelled coax in simulation and get good results. in practical can i get these results by coax or must i use a balun structure? can you say anything about this?
thanks for the help

Do you have photos of the manufactored sample? Especially detailed view of the feeding connections. This might be helpful to analyse what is going on here.

here are the photos af the manufactures antenna

the blue structure is supporting the substrate
its epsilon is nearly 1. it doesnt change the resonance frequency