ADS LineCalc vs Momentum: Microstrip Width and Length fo Zo and electrical length

Hello, any of you has compared results of a Microstrip 50Ohm line and electrical length 90o, when calculated Width and Length is done with ADS LineCalc, or with Momentum?

I wanted to calculate a 50Ohm and 90o Microstrip line in Arlon 25N H=0.508mm @1.6GHz

I calculated with LineCalc: W=1.14mm, L=28.85mm

Simulating with Momentum: Starting from LineCalc values, I realized that
Phase(S12)=90o if L=28.55mm, 0.3mm less that with LineCalc
Mag(Zo)= 50Ohm when W=1.2mm, a difference of 0.06mm with LineCalc.

Have you some measurements? What its better to trust, LineCalc, or Momentum?

Thank you

The chance for user error in Momentum is higher (thick metal <> thin metal, influence of meshing). I would trust Linecalc in this case.

Your results are very similar, 1% difference for the length and 5% for the width.
If you want to manufacture your design you have to check how accurate the manufactured line width is, the errors could be larger then the 60um difference from your calculation.

Precision should be better, as I use a LASER machine, its supposed to be very accurate. But even in case it wasn't, I'm actually curious about what to trust more.
Thanks for your answers.

Assuming that you have Momentum set up correctly (HUGELY IMPORTANT ASSUMPTION), I would trust it's answer more than LineCalc, since it takes into account all of the edge-effects, side currents, etc. For a starting point, LineCalc is very quick to calculate an answer, given that it's solving closed-form mathematical models to find the solution.

Momentum will take into account all of the interactions between the bits of metal in your circuit. This is especially noticeable when you tune up an interdigitated or gap filter using a simulation component, versus tuning the same filter using Momentum. When fabricated, the correct design should closely match the Momentum-tuned circuit.

Yeah, that's the key. Jim Rautio, one of the EM software gurus, once said: "All EM results are always wrong. The question is ... how much".

From what I have seen, the biggest error source is the user, who sets the mesh density and other model details which have a large impact on results. For a simple structure like the transmission line, I would think that well established equations for microstrip line impedance are trustworth, as long as you don't go to extreme aspect ratios. I fully agree that EM has great value for geometries where accurate models do not exist.