Transmission Line PCB

I wanted to verify all my different assumptions and tool assisted calculations before I do further work on PCIe, SGMII* and Etehernet PCBs.

So I calculated the impedances with AlterPCB Transmission Line Simulator (also a nice view of the electric field regarding EMI etc…)
I’ve choosen multiple transmission line derivats:

• Coplanar (Single / Differential)
• Stripline (Single / Differential)
• Microstrip (Single / Differential)
• 90° corner with a coplanar waveguide
• Impedance controlled via

I designed a layout with simple sma edge connectors and the transmission lines according to the calculations.

Measurements & Rework

And … the measurements showed ‚horrible‘ results. Let’s have a look at the coplanar waveguide:

On the right hand side one can see that the impedance variation is for (calculated 50 Ohm) pretty okay +7/-3 Ohm, but at the start and the end there are huge impedance discontinuitys.
I thought so long about the transmission line itself, that I completely missed out on the connection.
As a no-brainer I slapped some SMA edge connectors on there and didn’t care further.

The main issue here is this region:

Those SMA edge connectors (or at least this footprint) are mainly designed for 2-layer PCB designs where the impedance for 50 Ohm is roughly this pad size

I evaluated my guess by cutting the pad size down to the width of the trace

Here a before & after screenshot of a S21 measurement with the HP8560A up to 2.9GHz.
The blue trace is the PCB as designed; the pink trace the measurment with the rework.

And here the TDR comparison of the 11801C:

Simulation when it was too late

I tried to simulate the results with Qucs I’ve got the following results for the reworked and original