# Characteristic impedance of RF PCB when narrowing the trace width to fit pad

I am working on a PCB design that should be matched at a characteristic impedance of 50 Ohms. The circuit down converts a 2.4 GHz signal to a signal in the 2 Mhz- 100 Mhz range using a VCO and a mixer. For the board I will be using, I calculated the required trace width to be 1.04 mm but the pad width of the VCO and mixer is 0.30 mm and 0.45 mm respectively. I have used the trace width of 1.04 mm and only narrowed it down to the respective width when connecting to the VCO and mixer. I wanted to know if this will affect the characteristic impedance of my trace and thereby degrade my signal or since the narrowing of the trace only happens at the end of the trace when connecting to the pad, the characteristic impedance will not affected that much by the narrowing?

• You will be fine. We are taught, in absence of SmithChart work, the critical distances are 1/5 or 1/10 or 1/20 wavelength. The 2.4GHz wavelength of 12cm allows 12cm/20 or 6mm of tolerance. Your "narrowed down" region is just 2mm, right? Apr 18, 2017 at 13:08
• You can examine crosstalk, as a fidelity risk. Efields form capacitive dividers: the aggressor/source/Transmitter is one plate, the victim/destination/Receiver is another plate, and the GND traces or GND plane forms the 3rd plate. These 3 plates form 2 capacitors; use parallel-plate C = E0*Er*Area/Distance to model coupling from trash generators into your signal chain. Apr 18, 2017 at 13:13