RF Signal Routing Transmission Lines and Connectors

Keep the line short
Those long looping pieces of insulated hookup wire common to low frequency prototyping aren’t a good idea at radio frequencies. Their impedance is random as they move and couple to nearby objects and they often behave more like antennas then signal pipes. A general rule of thumb for RF connections – shorter is better. At UHF frequencies and above, if you can see the leads, they’re too long. Transmission lines are needed to provide controlled pathways for RF signals – microstrip within the board and coaxial cable to connect different boards.
Within the Board
RF signal traces within the board are made from a type of transmission line known as microstrip – flat copper strips connecting various signal destinations. Microstrip impedance is a function of line width and other factors. In general, wide lines have low characteristic impedance with the value rising as the width narrows. Line impedance within the board is generally less important at lower frequencies (HF) but becomes more critical at UHF frequencies and above. At microwave frequencies, sections of microstrip are used to synthesize capacitors and inductors for resonating and impedance matching, making both length and shape of each trace critical. A good rule of thumb for routing signals in fiber epoxy board is the 1/8 inch line width (0.105 – 0.125 inch depending on board dielectric) which represents a Z0 of about 50 ohms. Tables and software are available for calculating exact width for a desired impedance in the high end board materials .
Microstrip lines can be produced either by etching one side of a double sided circuit board ( copper is left on the other side to form the ground plane ), or by cutting and pasting adhesive backed copper foil tape on top of a single sided circuit board. Figures 2 and 5 show examples of microstrip layout. One important point – RF doesn’t like sharp corners. When making a 90 degree turn in microstrip, bevel the outside edges at 45 degrees .
Microstrip bend.
Entering and Leaving the Board – Launcher Blocks Signals entering and leaving the board generally do so via coax to microstrip transitions. In prototyping work, a fixture known as a “launcher block” is often used for this purpose. One can envision the incoming signal leaving the tip of the coax connector and being “launched” as a wave into the microstrip. Shown in figure 2, this fixture consists of a drilled and tapped rectangular aluminum block designed to be attached to the ground plane side with machine screws and mounted flush with one end. A coaxial connector is attached to the block such that its center pin just rests on the microstrip. It’s important for the connector block, connector, and board to fit together snugly and squarely in order to minimize signal loss during the transition.
BNC and SMA connectors, shown in figure 3, are probably the most popular for low power RF assemblies. BNC connectors are suitable for frequencies up to 1-2 GHz, while SMA connectors are useful well into the microwave range. Ideally, a coax connector and launcher block should be used for each RF signal which enters and leaves the board. For short cable runs and module interconnections and frequencies below about 100 MHz, carefully soldering the coax directly to the board can be a practical alternative.
Coax Cable
During initial development, it’s a good idea break up circuits into functional modules with each module on a separate circuit. When routing signals to and from test equipment or other modules, it’s important to use high quality coaxial cable. Use the type designed for RF (not audio). It should have a continuous braided shield and, ideally, a loss of less than 3 dB per 100 ft at the frequency of interest. Cable designations and loss factors are tabulated in a number of handbooks. Values for some of the more common cables are listed in Table I. 50 ohms is normally chosen as the standard cable impedance, although other impedances are used in specialized applications, such as cable television.