Transmitter PCB: Design and Cost Factors

What a Transmitter Board Adds to a Radio

A radio board receives and transmits; a transmitter board is built around the transmitting chain and the problems that come with it. The receiver is a low level circuit that must be quiet. The transmitter is a high level circuit that must be clean, and the difference in the design is largely a difference in power handling and in filtering.

Three requirements dominate. The output must be at the right frequency and stable, it must not radiate energy at the harmonics and the spurious frequencies that the regulation forbids, and it must survive its own heat. Everything else in the design follows from those three.

The Chain

• Frequency generation. A crystal oscillator, a phase locked loop or a direct digital synthesiser, with a loop filter that sets the settling time and the phase noise. This stage determines the purity of the output.
• Upconversion and modulation. The baseband signal is applied to the carrier, either in a mixer or inside a transceiver integrated circuit.
• Driver amplifier. Raises the level to what the final stage needs, and often provides the gain control that sets the output power.
• Power amplifier. The final stage, which delivers the output and dissipates most of the heat on the board.
• Harmonic and band filtering. A low pass or band pass filter that removes the harmonics generated by the amplifier. Without it, the transmitter fails its regulatory limits even though the wanted signal is perfect.
• Power control and protection. A directional coupler or a current sense to measure the output, and protection against a mismatched antenna, which reflects power back into the amplifier.
• Antenna switch or duplexer. Where the same antenna is used for transmit and receive.

The Power Stage

The power amplifier is the component that decides the board. It has an input and an output matching network, a bias network, a supply decoupling network and a thermal path, and each of them is a layout decision.

Matching. The output match transforms the device impedance, which for a power device is a few ohms or less, to the 50 ohms of the antenna path. It is built from transmission line sections and discrete components, and it must be placed immediately at the device with the shortest possible connections. Every millimetre of extra line is an impedance error and a loss.

Supply decoupling. A power amplifier draws a large pulse of current on every radio frequency cycle, and the decoupling network has to supply it. The network is a set of capacitors of decreasing value placed as close to the device as the layout allows, with the smallest closest, and connected to the supply plane by multiple vias. The inductance of those connections is what limits the performance at the higher harmonics.

Bias. The bias network sets the quiescent current and therefore the linearity and the efficiency. It is decoupled heavily, because any signal that reaches the bias line is amplified and appears in the output as a distortion product.

Thermal path. The efficiency of a linear amplifier is low, so a large part of the supply power becomes heat in a small package. A thermal pad soldered to a copper area with a via array beneath it, and often a metal core base, carries that heat away. The junction temperature sets both the life and the gain, because the device parameters drift with heat.

Where the power is high enough, the amplifier is built on a metal core board or a ceramic substrate, and the driver and the control circuits sit on an adjacent FR-4 section connected to it. Our notes on PCB manufacturing describe the metal core and heavy copper constructions.

Stability and Cleanliness

A transmitter must put its energy in the wanted band and nowhere else. The two enemies are the harmonics of the amplifier and the products of any unwanted oscillation.

Harmonic filtering. The amplifier generates harmonics of the carrier because it is a non linear device, and the level of the second and third harmonic can be only twenty or thirty decibels below the carrier. The filter after the amplifier reduces them to the level the regulation requires, and it is designed as a filter rather than as a decoupling network: the topology, the component tolerances and the layout all matter.

Oscillation. An amplifier that is not unconditionally stable will oscillate at some frequency where the loop gain exceeds unity, and the symptom is a clean wanted signal with an unexpected emission elsewhere. The usual causes are a matching network with the wrong load at a harmonic frequency, insufficient decoupling on the bias line, or a feedback path through the supply or the ground. The remedies are a lossy element in the matching network, a resistor or a ferrite in the bias path and a layout that returns the amplifier current through a defined path rather than through the signal ground.

Grounding. The power stage needs a low inductance ground directly under and around the device. That means a solid copper area with via stitching to the ground plane, not a thin trace. Vias in the device thermal pad connect it to the ground copper below as well as to the thermal path, and both functions are served by the same structure. Our notes on PCB design and layout cover the layout techniques.

Materials

The material is chosen from the frequency and the power. A low power transmitter at a few hundred megahertz works on FR-4. As the frequency rises or the loss budget tightens, a low loss laminate with a stable dielectric constant is used, because the matching network and the filter are designed against that constant and their performance depends on it holding. At the highest frequencies and powers, a PTFE laminate or a ceramic substrate is used, and the thermal path becomes part of the material choice.

The finish matters at high frequency because conductor loss is a significant part of the total loss. A thick, uneven coating adds loss and detunes the matching network, so a thin metallic finish such as electroless nickel immersion gold or immersion silver is used on the radio frequency sections. Our notes on telecommunications PCBA describe the equipment these boards serve.

Regulatory Testing

A transmitter is a regulated product, and the test programme is not optional. The measurements that matter are the output power, the frequency error, the occupied bandwidth, the harmonic and spurious emissions, and the behaviour at the band edges. The tests are performed on the complete product with the antenna it will ship with, in a defined test setup, and the results are recorded.

The consequence for the design is that the filter, the matching and the shielding have to be final before the test, and any change afterwards invalidates it. Where the product carries a module rather than a discrete design, the module brings its own certification, which is the strongest argument for using one. Our notes on PCBA testing describe the test methods, and our notes on quality management describe the process control behind production.

Cost Factors

The board cost is driven by the frequency, the output power and the material. A low power transmitter on FR-4 is an inexpensive board. As the power rises, the material moves to a low loss laminate or a metal core, the copper weight increases, the finish changes to a low loss metallic coating and the tolerances on the matching network tighten, and each of those steps raises the price.

The test cost rises with the power as well, because the regulatory testing is performed on the product and the equipment needed to measure a high power transmitter is significant. Where the volume is low, that cost dominates the budget, and it is the reason a certified module is often the cheaper engineering answer even when the bill of materials is higher.

FAQ

Why does a transmitter need a filter after the amplifier? Because the amplifier generates harmonics of the carrier, and the filter reduces them to the level the regulation requires.

Can a transmitter be built on FR-4? At low power and at moderate frequencies, yes. Above that, the loss and the variation in the dielectric constant affect the matching and the filter.

What causes spurious emissions? Usually an amplifier that is not unconditionally stable, insufficient bias decoupling or a feedback path through the supply or the ground.

Why is a certified module sometimes cheaper? Because it removes the radio frequency design and the regulatory testing from the project, and the testing cost dominates at low volume.

Conclusion

A transmitter board is built around the power stage and the filter that cleans it. Match the device at the pin, decouple the supply with the smallest capacitor closest, keep the bias line free of radio frequency, give the device a low inductance ground and a real thermal path, and design the harmonic filter as a filter. Choose the material from the frequency and the power, and confirm the regulatory measurements on the complete product before the design is frozen.