Insertion loss is the drop in signal power that happens when you place a component, such as a cable, connector, filter, or switch, into an RF signal path. It’s measured in decibels (dB), and a higher number means more signal is lost. Every real-world RF component introduces some insertion loss, so understanding it is essential for designing systems where signal strength matters.
How Insertion Loss Is Calculated
The basic idea is simple: compare the power reaching a load before and after you insert a component. If the power before insertion is P_T and the power after is P_R, insertion loss in decibels is:
IL (dB) = 10 × log₁₀(P_T / P_R)
So if a cable delivers half the original power to a load, that’s about 3 dB of insertion loss. A component that passes nearly all the signal through might show only 0.2 dB of loss, while a lossy connector or long cable run could show significantly more.
In practice, RF engineers rarely measure raw power ratios directly. Instead, they use a parameter called S21, which captures how much signal passes from one port of a component to the other. When both ports share the same reference impedance (usually 50 ohms), insertion loss is:
IL = −20 × log₁₀(|S21|) dB
S21 is one of the scattering parameters (S-parameters) that a vector network analyzer displays. It represents the ratio of power delivered to the load versus power available from the source. An S21 of 0 dB means perfect transmission with no loss, and any negative S21 value indicates signal power has been lost.
What Causes Signal Loss
Three physical mechanisms account for nearly all insertion loss in RF systems:
- Resistive loss: Every conductor has some resistance, which converts a portion of the signal’s energy into heat. This is the most straightforward source of loss and gets worse with longer cables or thinner conductors.
- Dielectric loss: The insulating material surrounding a conductor absorbs energy as the electromagnetic field passes through it. Different substrate and insulation materials absorb different amounts, so material choice has a direct impact on performance.
- Impedance mismatch: When a component’s impedance doesn’t match the rest of the system (typically 50 ohms), part of the signal reflects back toward the source instead of passing through. This reflected energy shows up as part of the total insertion loss, though engineers often track it separately as “mismatch loss” or “return loss.”
Imperfections in connectors, solder joints, and PCB traces also contribute. Even small irregularities can scatter energy or create localized reflections that add up across a signal chain.
Why Loss Increases With Frequency
Insertion loss is not a fixed number for a given component. It changes with frequency, and almost always gets worse as frequency goes up. Two effects drive this.
The first is the skin effect. At higher frequencies, current flow concentrates near the outer surface of a conductor rather than using its full cross-section. This effectively shrinks the conductor’s usable area, raising its resistance. At 1 GHz the current is already confined to a thin shell near the surface, and by 10 GHz the situation is far more extreme. The result is higher resistive loss at every step up in frequency.
The second is dielectric absorption. The insulating materials around conductors absorb more energy as frequency increases. This absorption rises roughly in proportion to frequency at moderate ranges, then accelerates at very high frequencies as the electromagnetic field becomes more tightly concentrated in the dielectric material. Together, these two effects mean that a cable or connector rated at 0.5 dB of loss at 1 GHz could easily show several dB of loss at 10 GHz or above.
How Insertion Loss Affects System Performance
Insertion loss does more than just weaken a signal. For passive components like cables, switches, filters, and connectors, the noise figure equals the attenuation. In plain terms, a passive component with 3 dB of insertion loss adds 3 dB to the noise figure of the system. This matters most at the front end of a receiver, where any loss before the first amplifier directly degrades the system’s ability to pick up weak signals.
This is why RF engineers obsess over minimizing loss in the earliest stages of a receive chain. A 1 dB loss in a cable between an antenna and a low-noise amplifier raises the entire system’s noise floor by 1 dB, which can be the difference between reliably receiving a distant signal and losing it in noise.
Typical Values for Common Components
Knowing what “good” looks like helps when evaluating components or troubleshooting a system. High-quality waveguide electromechanical switches typically show 0.2 to 0.5 dB of insertion loss across their operating band. These are among the lowest-loss switching components available.
Connectors vary widely depending on type and quality. In a study comparing different SMA connector styles up to 8 GHz, the best-performing option (a solderless board-edge end-launch connector) kept insertion loss below 4.2 dB across the entire measured range, while a through-hole right-angle SMA connector performed far worse, with usable bandwidth of only about 1.5 GHz. Much of that variation came down to how well the connector’s impedance matched the 50-ohm system. The solderless connector measured 46.5 ohms at the transition, while the worst performer dropped to 21 ohms.
For cables, loss per unit length is the key spec. Short, high-quality coaxial runs at low frequencies might add fractions of a dB, but long runs at microwave frequencies can accumulate significant loss. Datasheets typically list attenuation per meter (or per 100 feet) at several frequency points so you can estimate total loss for your cable length.
How Insertion Loss Is Measured
The standard tool for measuring insertion loss is a vector network analyzer (VNA). The process starts with a calibration step that corrects for imperfections in the VNA itself, the test cables, and the connectors. Without calibration, measurement errors can easily exceed the actual loss of the component you’re testing.
After calibration, the device under test is connected between the VNA’s two ports. The analyzer sweeps across a range of frequencies and displays S21, the transmission coefficient, on screen. Since the VNA shows S-parameters in log format (20 × log₁₀|S21|), you can read insertion loss directly from the display as the magnitude of S21 at each frequency point.
Measuring very high loss values (above about 60-70 dB) pushes against the dynamic range limits of most VNAs. Specialized setups address this by inserting an amplifier in the signal path to boost the transmitted signal above the analyzer’s noise floor, while keeping an attenuator in the calibration path to protect the receiver and maintain measurement accuracy.
Insertion Loss vs. Isolation
In RF switches and similar routing components, insertion loss and isolation are two sides of the same coin. Insertion loss describes how much signal is lost through the “on” path, while isolation describes how well the “off” path blocks signal from leaking through. A good switch minimizes the first and maximizes the second.
Premium waveguide switches achieve both: 0.2 to 0.5 dB of insertion loss in the on state and 60 to 80 dB or more of isolation in the off state. In solid-state switches (PIN diode or FET-based), there’s often a design trade-off between the two. Achieving very low insertion loss in the on state can reduce isolation in the off state, and vice versa, because the same semiconductor junction properties that allow low-loss transmission also make it harder to fully block signal leakage.