Split an RF signal two ways with a plain T-junction and you inherit three problems at once. The two outputs leak into each other, the input match falls apart, and reflected energy bounces back toward your source. A Wilkinson power divider clears all three. It splits one input into equal, in-phase outputs while holding high isolation between those outputs and a clean impedance match at every port. More than sixty years after Ernest Wilkinson first described the design, it is still the default choice for clean signal splitting and combining in radar, satcom, and test systems.
This guide covers what the device does, how the topology earns its isolation, how it compares to resistive and lumped element designs, and which specs decide whether a part will hold up in your system.
What a Wilkinson Power Divider Actually Does
Start with the basic question: what is a power divider? A power divider takes one input signal and splits it into two or more outputs of equal amplitude and matching phase, while keeping the impedance matched across the device. A divider with more than two outputs is an N-way power divider.
Power Divider vs Combiner
Run the same part in reverse and it becomes a power combiner: feed the output ports and their power sums at the common port. That two-direction behavior is why catalogs list these components as power divider/combiners. The hardware is identical. Signal direction decides the job.
The Wilkinson version adds the feature that makes it useful in real hardware: isolation between the output ports. A simple split lets signal on one output bleed straight into the next. The Wilkinson design suppresses that leakage, so a reflection or a fault on one branch does not ripple across the others.
One caveat applies in the combiner direction. When the signals you combine are not equal in amplitude and phase, the power that does not add up constructively gets dumped as heat inside the internal resistor. That lowers the safe power rating in combiner mode below the rating you would see running the same part as a divider, which is why combining networks for non-coherent signals need careful power derating.
How the Topology Works
The classic two-way Wilkinson divider is built from two quarter-wavelength transmission lines, one running from the common port to each output. Each line carries a characteristic impedance of the square root of two times the system impedance, so about 70.7 ohms in a 50 ohm system. A single resistor, set to twice the system impedance (100 ohms for a 50 ohm two-way), bridges across the two output ports.
The quarter-wave sections transform the impedance so all three ports stay matched. The bridging resistor is where the cleverness lives. When the outputs are balanced and in phase, no voltage difference appears across the resistor, so it draws no current and burns no power. The divider stays low-loss for the signal you want. When unbalanced or reflected energy shows up at the outputs, a voltage difference does appear across the resistor, and the resistor soaks up that energy. That selective behavior is the source of the isolation: lossless for the wanted signal, lossy only for the unwanted difference between ports.
Wilkinson laid this out in 1960 in the IEEE Transactions on Microwave Theory and Techniques. His original experimental model reached about 27 dB of isolation, an output VSWR near 1.6, and a usable band of roughly 20 percent. Modern stripline and microstrip fabrication push those numbers well past the early prototype, with tighter matching and wider bandwidth, but the operating principle has not changed.
Wilkinson vs. Resistive vs. Lumped Element
Wilkinson is one of three common ways to build a power divider, and the right pick depends on your frequency range, power level, and how much loss you can spend.
A resistive power divider uses only resistors, with no transmission line sections. It works from DC upward and covers very wide bandwidth in a simple package. The tradeoff is real: it carries an extra 3 dB of loss above the split, because the resistors burn half the input power, and it gives little to no isolation. That makes it a good fit for broadband test paths where flat response down to DC outweighs the loss penalty.
A Wilkinson divider built in stripline or microstrip keeps excess loss down to a few tenths of a dB and delivers the high isolation a resistive part cannot. The cost is that its quarter-wave lines tie its performance to a frequency band, and the design cannot reach true DC.
A lumped element divider swaps the quarter-wave lines for discrete inductors and capacitors. That shrinks the part and lets it reach down toward DC, which is useful below a few GHz and in high-power builds where the resistors need room to dissipate heat.
A quick way to narrow the choice: if you need DC coverage and very wide bandwidth and can accept the loss, go resistive. If you need low loss and high isolation in a defined band above roughly 1 GHz, go with a Wilkinson stripline part. If you need a low minimum frequency or high average power in a compact footprint, look at lumped element designs. You can compare the full range of RF power dividers and combiners by construction technology to match the topology to your application.
The Specs That Decide Performance
Once you settle on a topology, a handful of numbers tell you whether a specific part will work in your signal chain.
Insertion loss. The theoretical 3.01 dB drop on a two-way split is not really a loss; it is the power division itself. What you watch is the excess loss above that figure. A good Wilkinson stripline divider adds only 0.1 to 0.3 dB on top of the split. Budget that excess across every stage in your network, because it adds up fast in a multi-stage feed.
Port-to-port isolation. This measures how well one output port rejects a signal present on a neighboring output. Typical values run 15 to 25 dB. Weak isolation lets outputs cross-couple, which can destabilize an amplifier combining network and distort the amplitude and phase pattern in an antenna feed.
VSWR and return loss. These describe the same impedance mismatch in different units. A VSWR of 1.50:1 works out to about 14 dB of return loss. In a cascaded chain, specifying a tighter figure such as 1.35:1 at each junction keeps mismatch from compounding across the whole network.
Amplitude and phase balance. These track how closely the outputs match each other across the band. Tight balance, on the order of 0.1 dB and 1 degree, carries the most weight in array systems, where small per-element errors stack up across hundreds of radiators.
Power handling. The combining power limit is usually set by the thermal rating of the internal isolation resistor, not by the transmission lines. That limit drops when you combine non-coherent signals, so confirm the rating for your actual signal conditions.
Where Wilkinson Dividers Earn Their Place
The topology shows up wherever a system has to split or combine RF power without giving up isolation or match.
Phased array feed networks lean on it heavily. A corporate feed splits the transmit signal through several stages of two-way division before the power reaches each radiating element, and every stage needs equal amplitude and equal phase at its outputs. Balance errors in the feed map directly to higher sidelobes and beam pointing error in the far field, so the per-output balance of each divider becomes a system-level spec.
High-power amplifier combining is another core use, especially in satcom ground terminals. Combining several amplifier outputs into one transmit path lets a system reach high output power using many smaller, lower-cost amplifiers instead of a single large tube. Here the isolation resistor’s thermal rating sets the practical ceiling, since any power that does not combine cleanly turns into heat inside that resistor.
RF test benches use Wilkinson and resistive dividers for coherent signal injection into multiple devices under test, for vector network analyzer reference paths, and for building calibration standards. These are exactly the markets MCLI’s power dividers serve, from defense radar programs to research labs.
Specifying the Right Wilkinson Divider
Start with frequency range and power level, then let those two numbers steer the construction choice. For coverage above 1 GHz into the millimeter-wave range with low loss and high isolation, a stripline Wilkinson divider is the right call. For sub-3 GHz commercial work in a smaller footprint, a microstrip Wilkinson part fits. For a low minimum frequency or high average power, a lumped element design handles it better.
Match the connector to the band as well. Below 6 GHz, SMA and Type N connectors handle most applications, with Type N giving better power handling at the same frequency. Above 18 GHz, precision connectors such as 2.92mm, 2.4mm, or 1.85mm are needed to hold the rated VSWR and insertion loss. And check performance at the band edges and across temperature, not only at center frequency on the bench. A part that meets spec at room temperature on day one can drift out of range at cold start or after years in the field.
MCLI builds Wilkinson power dividers in stripline (PS Series) covering 0.1 to 70 GHz and microstrip (PM Series) covering 0.15 to 2.7 GHz, alongside resistive and lumped element lines that span DC to 70 GHz and power levels from 1 watt to 10,000 watts. Many models ship the same day from stock, with full insertion loss, isolation, and VSWR data published on every datasheet. If you are specifying a divider for a feed network or a combining stage and want to talk through the specs that matter for your design, MCLI’s engineering team can help you size the part to your application.
Frequently Asked Questions
What is the Wilkinson power divider concept?
The Wilkinson power divider is a passive three-port circuit that splits one input into two equal, in-phase outputs while keeping those outputs isolated from each other. Ernest Wilkinson introduced it in 1960 to match all ports and isolate the outputs at the same time, something a plain splitter cannot do. The same part also runs in reverse as a combiner.
How does a Wilkinson power divider work?
It uses two quarter-wavelength transmission lines, one from the input to each output, plus a resistor bridged across the two output ports. For balanced, in-phase signals the resistor sees no voltage and burns no power, so the split stays low-loss. Reflected or unbalanced energy does develop a voltage across the resistor and gets absorbed, and that is what creates the isolation.
What is the equation for the Wilkinson power divider?
For an equal two-way design in a system impedance Z0, each quarter-wave line has a characteristic impedance of the square root of two times Z0, about 70.7 ohms in a 50 ohm system, and the isolation resistor equals two times Z0, or 100 ohms. The general N-way power divider uses a line impedance of the square root of N times Z0, with isolation resistors referenced to Z0. Each line is one quarter wavelength long at the center frequency.
Is the Wilkinson power divider lossless?
For the signal it splits, it is close to lossless. With matched, in-phase outputs the isolation resistor burns no power, so the only intended drop is the split itself, about 3 dB for a two-way. It is not lossless for reflected or unbalanced energy, which the resistor absorbs by design, and real parts add small conductor and dielectric losses on top, usually a few tenths of a dB.
What is the purpose of a power divider?
A power divider distributes one RF signal into several outputs of controlled amplitude and phase, or run in reverse, works as a power combiner. That lets a system feed many antenna elements, drive multiple test channels, or combine amplifier outputs while holding the impedance match at every port. Output isolation keeps a fault or reflection on one branch from disturbing the others.
How do you design a power divider?
Design starts with the frequency band, the number of outputs, the power level, and the isolation and balance the system needs. Those targets set the topology, the line impedances and resistor values, and the construction method, such as stripline or microstrip. Most production designs are modeled in EM simulation and confirmed by measurement, which is why many engineers specify a catalog or custom part rather than build one from scratch.
Clean Splits, Clean Combines
The Wilkinson power divider has stayed the workhorse of RF signal splitting for one reason: it does the hard part well. It splits or combines power while keeping the output ports isolated and matched, at very little loss. Pick the construction to your band and power, hold the isolation and balance specs your system needs, and the divider quietly disappears into the design the way a good passive component should.