The PI Network

The Pi-Network circuit, as depicted above, is found in both antenna tuners and as tank circuits in older tube style transmitters and linear amplifiers. The PI network is considered to be a low-pass filter and normally contains 3 tunable components. Generally the design calls for fairly large variable capacitor values and as such can be expensive to build. A concern for the Amateur Radio Operator is that the PI network, just like the T network, can produce resonance at more
than one set of component values. The difference between the settings results in higher or lower efficiency and thus power lost to heating.

Component values generally range from 10-500 pf for the variable caps and 2-20uh for the roller inductor or tapped coil.

The circuit depicted above is unbalanced in that the "hot" side is treated differently than the "shield" side of the tuner. Basically it's not symmetrical
and thus better suited for coax to coax antenna tuners.

Pros:

• Broad banded, able to match a wide range of impedances

• Can be modified to allow for the selection of lower value variable components by adding additional fixed components to the input side of the network

• Simple to implement

Cons:

• More knobs means more time required to find a good match

• More than one resonance point can be found using different component values

• If tuned to an incorrect resonance point, power loss due to heating of the roller inductor will result

• Power loss due to heating in this network is higher than that of the L network and significantly more than that of the symmetric PI network

• 500 pf capacitors can be large and expensive

The topology of this network must be changed to support the direction of match. Depending on whether the impedance is greater on the output or on the input side, the shunt capacitor must move from one side of the inductor to the other. There are many tuners that accomplish this task by using a set of rotary switches configured to move the capacitance from one side of the inductor to the other to further add additional capacitance to the network. Typical component values for the variable capacitors range from 20 - 500pf and up to 1000pf for fixed value capacitors. Roller inductor values range from 2uh to 20uh. In some matching conditions the larger variable components may be switched out altogether in favor of much smaller fixed value components. This is true for both the capacitors and the inductor.

The L network depicted above is not symmetric and thus unbalanced, meaning that it is best utilized in coax to coax antenna tuners. The L network topology can be changed to be fully symmetric for better support of coax to ladder/window line transmission line.

Pros:

• Very few variable components required

• Will resolve to only one point of resonance for a given frequency and load impedance

• Inexpensive to construct if the matching requirements for the feed line and load are not particularly wild

• Lower power losses due to heating than the standard PI-network illustrated above

• Simple to implement for either hi-Z or low-z, but not both

Cons:

• In order to service both hi-Z and low-Z impedance matches, the topology of the network must be changed.

• Considerably more expensive and complicated to implement if switch gear, power and relays are required to cover both high and low Z impedances.

The T network is a 3 component design that is found in the vast majority of 3 element tuners available today. Note that the T network is really just 2 L networks fixed back to back with a common inductor. This design allows for easier matching of the 50 ohm load on the input side as well as wide range matching on the load side. The network illustrated above is unbalanced and thus good for feeding a coax to coax implementation. This network has a higher loss factor than the L network described above.

Typical component values for the capacitors are 17-500pf and 2-20uh for the inductor

Note that this network topology can be implemented using a differential capacitor, meaning that the 2 capacitors are really part of the same component but with the rotors of the caps aligned 180 degrees out of phase with each other. This configuration effectively transfers capacitance from one side of the network to the other when attempting to match either high or low impedances. Additionally, by running this network using a differential capacitor, only one resonance solution will be found for any particular band.

Pros:

• Widely used by the Amateur Radio community

• Very simple construction

• Components easily acquired as new or at swap meets

• Can be implemented with a differential capacitor

Cons:

• Higher power loss than the L network

• Differential capacitors are scarce and expensive

• When used with 2 independent capacitors, multiple resonance points can be found for a specific frequency

The symmetric PI-Network is definitely the most complicated design, although the fully switchable L network can be just as difficult to implement. This network topology is rare in the Amateur Radio world, due to the cost of assembling one of these units. Four separate components are required, including 2 large roller inductors and 2 variable capacitors. Additionally, this is a fully balanced network, meaning that the components are mirrored from side to side. Due to symmetric construction of this network, the entire matching section must be floated above ground (disconnected from the case) and because of this stray capacitance can play a factor in tuner performance.

Note that in this and other symmetric designs, a current balun (1:1) is required to achieve equal currents on both sides of the network and must be installed BEFORE the tuner instead of AFTER the tuner. Because the balun is installed before the tuner, the number of toroids required to provide the correct amount of common mode current impedance protection is reduced as the toroid is not subject to the wild voltages that may be encountered on the feed line side of the network.

Component values for the capacitors range from 10-300pf and the roller inductors have a value of 2-20uh. The roller inductors in this network must be linked together through a belt system, which further complicates the implementation of the device and adds to the additional expense of the unit.

Additional capacitance may be required on the output side when attempting to create a match on 160 with an antenna that is "short" for that band - i.e., less than 1/2 wavelength long. This is usually accomplished by adding switched capacitance in shunt to the output.

Symmetric PI network tuners can be a bit difficult to tune, but the resonant match will generally be very good. The most important fact about the symmetric tuner is that it is incredibly efficient. Where other tuners may lose up to 10% of input power due to heating, the symmetric tuner loses a fraction of that with no heating problems.

Pros:

• Provides a very good match on all frequencies where a "long" antenna is utilized -i.e., 1/2 wavelength or better

• Extremely efficient (See the section below for proof)

• Provides excellent common mode current stopping power when used in the configuration illustrated above

• Specifically designed for balanced antenna systems

Cons:

• Difficult to implement

• Requires more components and thus is more expensive

• Specifically designed for balanced antenna systems

• Improper construction can lead to stray capacitance issues