A 50 ohm microstrip line (0.7 mils thick) on a 20 mil substrate (Er=3.7, Tand=0.008) calculates to be 42.5 mils at 1 GHz in Linecalc. A similar line constructed in momentum shows that 44 mils wide has much better return loss.

Similarly, a 50 ohm line on a 10 mil substrate calculates to be 21 mils wide in Linecalc, while a 24.5 mil line has much better return loss in momentum.

Both lines were 1 inch long, terminated in 50 ohms.

My momentum simulations for 50 ohm lines are not matching up with my linecalc calculations.

Which one should I believe?

Similarly, a 50 ohm line on a 10 mil substrate calculates to be 21 mils wide in Linecalc, while a 24.5 mil line has much better return loss in momentum.

Both lines were 1 inch long, terminated in 50 ohms.

My momentum simulations for 50 ohm lines are not matching up with my linecalc calculations.

Which one should I believe?

Momentum will therefore have a better return loss for the 50.3 Ohm line. Â Since LineCalc uses analytical models for computing impedances which are known to be approximations of varying accuracy, I would trust the Momentum results to give us a more precise answer.

I should add that the thickness is not taken into account by the Momentum port solver but should not have a significant impact in this case. There is an AKC document that actually shows how you could derive a more accurate Z0 taking into consideration the finite thickness, but I don't think that's worth doing.