EC-22.7 Layout at High Frequency, and Designing a Link

What this is and why it exists

Everything so far assumed components connected by ideal lines. Here the line is copper on a board, with a dielectric under it and a connector at the end.

That change is where most radio designs fail. A carefully matched amplifier is undone by a via nobody thought about, or by a gap in the plane beneath a trace.

The second half of the topic joins the module to the rest of the area. A link budget starts at a transmitter's power and ends at a receiver's noise floor. Every figure in it has come from one of the preceding topics.

The vocabulary

• Controlled impedance — a trace whose geometry is specified to hit an impedance.
• Permittivity — the property of the board material that sets wave speed.
• Return path — the route the current takes back to its source.
• Transition — any change of layer, medium or connector along a signal path.
• Isolation — how well one part of a board is kept out of another.
• Link budget — the sum of gains and losses between transmitter and receiver.
• Effective radiated power — transmitted power multiplied by antenna gain.
• Fade margin — the extra received power allowed for conditions getting worse.

The mental model

A trace's characteristic impedance comes from three things: its width, its height above the reference plane, and the permittivity of the material between them. Once a board carries high-frequency signals, that impedance is a manufacturing requirement written on the fabrication drawing, not a hope.

The return path is the idea that saves the most trouble. At high frequency the return current does not spread out; it follows the path of least inductance, which is directly beneath the signal trace. A gap in the plane under a trace forces the current to detour around it, and the resulting loop radiates and picks up interference.

Transitions are where loss accumulates. Every connector, every via, every change of medium is a small mismatch. Individually each is negligible; a chain of them can dominate a link. Above a few gigahertz a via transition is designed deliberately, with ground vias beside it, rather than accepted as a default.

Isolation is the hardest practical problem on a transceiver board. A transmitter can be a hundred million times stronger than the signal its own receiver is trying to hear, on the same piece of board. Physical separation, shield cans and careful supply filtering are what make the two coexist. All three have to be planned into the layout from the start.

Then the link budget. Start with transmitted power. Add the transmitting antenna gain and subtract the loss reaching it. Subtract the loss through the air. Add the receiving antenna gain and subtract the losses on that side. The result is a received power. Compare it with the receiver's noise floor and required signal-to-noise ratio, and the link either works or does not.

Finally, two things that turn a calculation into a product. A link designed with no fade margin fails in rain, in a different building, or with a hand near the antenna. And what you may transmit, at what power, in which frequency ranges, is decided by regulation rather than by engineering.

What you should now be able to explain or do

• Say what sets a trace's characteristic impedance and where that is specified.
• Explain where the return current flows and what a gap in the plane does.
• Say why transitions matter and how a via transition is designed deliberately.
• Describe the measures that let a transmitter and receiver share one board.
• Build a link budget from transmitted power to received signal-to-noise ratio.
• Explain why margin is required and where the power limits come from.

Check yourself