Hi,
I am simulating a VCO based on transmission lines propagation delay with Spectre. I sized the transmission line so that it runs at  approximately 5 GHz. Running a transient simulation with the moderate setting give me the right frequency but with the conservative I have some oscillations at 9,8GHz! In a second experiment the transmission line has been increased in length so that the oscillation should go down but then whatever the accuracy the oscillation is at 9.5GHz. I checked with Eldo and the same netlist runs at about 4GHz.
I was thinking that the conservative simulation would increase the parasitic effect of the circuit but apparently it is not the case.
The transmission line models are broadband (until 30GHz) and generated with fastcap+fasthenry (netlist of vcvs and R L Cs).
Does anyone have a suggestion on the setting I should use to get reliable results? What is the difference between Eldo and Spectre for the transient analysis?

Thanks for the help,
Greg