On Air
PIN Switches in a Microwave Control Chain
How reflective and absorptive PIN switch topologies work, what isolation and insertion loss mean in practice, and how driver timing shapes a microwave control
Pieces in this section
-
How an emergency alert interrupts a song
The header burst, the two-tone attention signal, the relay chain that predates the internet, and why the system is deliberately built to survive the failure of everything modern.
13 minute read4 notes
-
How far a station is allowed to reach
Height above average terrain, the class table, the protected contour, and why a low-power station on a hill can beat a high-power station on a plain.
13 minute read4 notes
-
Why the AM band changes at sunset
Ground wave by day, sky wave by night, and the whole apparatus of power reductions, directional patterns and sign-off times that exists because of it.
12 minute read4 notes
-
What is actually riding on an FM carrier
The baseband, layer by layer: mono sum, pilot tone, stereo difference channel, station text, and the digital sidebands that sit outside all of it.
14 minute read5 notes
A solid-state PIN switch is a diode whose junction behaves as a current-controlled resistor at RF, and the topology chosen for it decides where the energy goes when the path is closed. Reflective designs send the blocked signal back toward the source; absorptive designs terminate it in an internal 50 ohm load. The choice between the two, and the isolation, insertion loss and switching time that follow from it, sets the behavior of the whole control chain.
What are reflective and absorptive topologies in solid-state PIN switches?
In a reflective switch, the off arm presents a mismatch. A shunt PIN diode biased into conduction looks like a few ohms to ground, so the incident wave reflects. The same is true of a series diode in its off state, which looks like a small capacitance and reflects most of the incident power. The reflected energy travels back down the line, and in a receiver front end that can be harmless or damaging depending on what sits upstream. A reflective SPDT feeding an antenna is common; a reflective switch feeding a sensitive low-noise amplifier is not, because the reflected power returns to the amplifier output.
An absorptive switch replaces the reflection with a termination. In the off state the signal is routed into a matched load, usually a thin-film or chip resistor integrated into the module. The port then looks like 50 ohms in both states, which keeps the impedance environment stable for the stages on either side. The trade is added loss and a load that must dissipate the incident power, so thermal design enters the picture at higher levels.
Both families are built from the same building blocks: series diodes, shunt diodes, or a combination in a pi or T network. The arrangement, not the diode, determines whether the switch reflects or absorbs. A useful reference on how these arrangements are specified and compared is the treatment of reflective and absorptive PIN switch topologies published by Control Line Review, which covers the solid-state switch, attenuator and detector sections of a control chain as one subject.
How does isolation affect solid-state PIN switch performance?
Isolation is the attenuation between input and output when the switch is commanded off, expressed in decibels. It is not a single number. It depends on frequency, on bias current, on temperature, and on the impedance of the source and load the switch sees. A datasheet figure of 60 dB at 2 GHz is a statement about a specific test fixture, and the same part in a different board can measure several decibels worse.
The measurement method matters as much as the number. Isolation is normally taken on a vector network analyzer after a full two-port calibration, with the switch biased to its off state and the unused port terminated in a matched load. The reference plane must be at the switch ports, not at the connector faces, or the fixture loss is folded into the result. For high-isolation parts, the noise floor of the analyzer becomes the limit, and a common technique is to measure two identical switches in series and divide the result, which recovers roughly twice the dynamic range.
Isolation also degrades with bias current in a way that surprises people. A PIN diode needs enough forward current to fill the intrinsic region with charge carriers; below that threshold the diode behaves as a lossy capacitor and the off-state attenuation falls. Driver current, not the diode alone, sets the achievable isolation. Temperature shifts the carrier lifetime and therefore the switching behavior, which is why isolation is often quoted over a range rather than at a single point.
What insertion loss can I expect from a solid-state PIN switch?
Insertion loss in the on state is the sum of the diode's forward resistance, the matching network, and the connector and housing losses. A single shunt diode SPST at 1 GHz typically shows 0.3 to 0.6 dB. A series-shunt SPDT in a small module commonly lands between 0.8 and 1.5 dB across an octave. Multi-throw switches with more junctions add loss with each stage, and a broadband design covering several octaves pays for its bandwidth in flatness.
The loss is not flat. Series inductance and shunt capacitance form a low-pass response, so insertion loss rises with frequency. A part specified at 1 dB at 2 GHz may be 1.6 dB at 6 GHz. Designers who need flat response over a wide band often accept higher absolute loss in exchange for a smaller slope, or add a compensating network. Return loss in the on state is the companion figure, and a switch with good insertion loss but poor match will still disturb the chain.
Power handling interacts with loss. The forward current needed to keep the diode conducting under a large RF swing is higher than the small-signal bias current, so a switch rated for 10 W may need a driver that can supply several times the current quoted in the small-signal test. The same driver current that improves isolation and power handling also raises the dissipated power in the diode and the module.
How fast can a PIN switch be driven?
Switching time is set by the diode's carrier lifetime and by the driver circuit. When forward bias is removed, the stored charge in the intrinsic region must be swept out before the diode reaches its high-impedance state, and that interval is the dominant part of the turn-off time. Typical beam-lead and chip PIN diodes used in control components have lifetimes in the tens of nanoseconds to a few microseconds, which maps to switching times from a few nanoseconds to a few microseconds depending on drive.
A fast driver does two things: it sources a well-controlled forward current during the on state, and it provides a reverse voltage and a low-impedance path to pull the stored charge out during turn-off. A driver that only removes bias leaves the diode to recombine on its own, and the switch is slow. A driver that applies reverse bias through a low-impedance switch can shorten turn-off by an order of magnitude. The driver also sets the on-state current accuracy, which feeds back into isolation and power handling.
Timing is usually specified as rise time, fall time and propagation delay, measured with a detector and an oscilloscope rather than a network analyzer. The detector's own video bandwidth limits what can be resolved, so a switch specified at 10 ns must be measured with a detector and amplifier chain that can follow it.
What sits beside the switch in the same signal path?
A control chain rarely contains a switch alone. Voltage-variable attenuators and digital step attenuators follow the switch to set level, and their specifications overlap with it: attenuation range, frequency flatness, accuracy, monotonicity and the control law that maps voltage or a digital word to decibels. A step attenuator with 0.5 dB steps and 0.3 dB accuracy is a different instrument from a voltage-variable attenuator with a smooth but loosely specified curve.
Detectors and logarithmic video amplifiers close the loop. A detector converts RF to a video voltage, and a log amp compresses a wide input range into a manageable output span. The figures that matter are the transfer curve, logging accuracy, tangential sensitivity, video bandwidth, VSWR, and the losses and calibration that a VNA measurement imposes. Tangential sensitivity, usually quoted in dBm for a given video bandwidth, sets the smallest signal the chain can see.
These parts share a signal path and a set of assumptions. A switch with 1 dB of loss changes the level reaching the detector; an attenuator with poor flatness changes the shape of the transfer curve; a log amp with limited video bandwidth cannot follow a fast switch. Reading the three sections together, as a control chain rather than as separate components, is how the numbers stay consistent from specification to measurement.
The same discipline of narrow questions applies away from the bench. A PIN switch datasheet states isolation and insertion loss under named conditions, and a driver timing figure holds only for the edge it was measured on; a bench run that ignores those conditions proves little. Readers who want the parallel case in computing, where memory tests and SMART readings each answer a narrow question, will find the limits of that method set out in plain terms.