Figure 1 : Total ciruit (C4 is optional). The moveable part (rotor) of should be connected to ground.
Depending on the frequency range a flying realization in a metal box with SMA jacks is often well suited. For higher power levels N-type or PL-type jacks are recommanded, also air trimmers with large plate distances.
Design
The circuit has many similarities to the equivalent circuit of a transmission line and it is well-known that a l/4 already forms a good matching network. If you look at the element values of such a l/4 line at the frequency of interest, you will get pretty good starting values the circuit. For the 70cm band you get the following values (see figure 2) :
Figure 2 : Equivalent circuit of a 50W-l/4 line at 440MHz (calculation with LTRANS : L1=18.08nH, C1=C2=7.235pF)
The trim capacitors C1 and C2 should cover the calculated values (7,2pF) (end value about 3 times larger than calculated). L1 is choosen to be half the calculated value (here 8.2nH instead of 18nH). This is because the parallel capacitor C3 will increase the effective inductance (figure 3). C3 should have the same value as C1 and C2.
Figure 3 : The effective inductance of a parallel LC circuit (calculation with LCFILT : 8.2nH || 9pF results in 18,8nH at 440MHz)
C4 should be at least 5·C1. If you need a very low lower cut-off frequency and therefore a large capacitor (e.g. 470nF) you should use a parallel combination of capacitors (e.g. use a 470nF foil capacitor and a 470pF ceramic type). To modify the circuit for the 2m band or for FM radio you can use the same approach or you can scale all elements. For FM radio frequencies simply multiply all capacitance and inductor values by 440MHz:100MHz=4.4. Itīs also possible to cover the short wave band, but then you need quite large tuning capacitors which are not always available.
Tuning Procedure
Normally you would expect that a circuit with 3 tuning elements is not easy to use. But itīs not too difficults if you look at the function using the Smith chart. If all 3 trim capacitors are approximately in the center position, you will see that the circuit behaves nearly as a l/4 transmission line. 50W at the input results in an output impedance of 50W. This setup is a good starting point for all matching problems (figure 4).
Figure 4 : Basic circuit setup in the Smith chart (calculation with CSmith)
The main difference to an ideal transmission line is the low-pass character instead of the periodic frequency response (figure 5). The first harmonic will be reduced by about 14dB.
Figure 5 : Frequency response in basic setup (calculation with CSmith)
In the Smith chart you may modify the tuning elements. You will see that C3 has a strong influence on the real part of the impedance and that C1 and C2 have more influence on the imaginary part. In practice you will begin the tuning procedure with the trimmer with the strongest influence on matching. If you already know your circuit impedances, e.g. because you have measured them using a network analyser, you may first do the tuning procedure in CSmith and then make a pre adjust of the trim capacitors.
The bandwidth is quite large in figure 5, but if you need a match to more extreme impedances the bandwidth will decrease. Figure 6 shows the matching circuit for an 8W generator to 50W (VSWR 6.25:1). A small mismatch of 1.45:1 remains because the trim capacitors are not ideal. But the loss is still low (0.3dB instead of 3.3dB without matching network). The 1dB bandwidth is now only 200MHz and the harmonic rejection is increased.
Figure 6 : Tuned circuit for matching to 8W (CSmith)
For many impedances different tuning setups are possible. But this setups might differ in bandwidth and harmonic rejection. For PAīs also different efficiencies might be achieved, although the output power is nearly the same. If you choose large values of C3, you will get a steeper cut-off in the frequency domain (stop-band zero). Smaller values have the advantage of larger bandwidth and low sensitivity to component elements.
Application
RF matching networks are often needed in conjunction with RF amplifiers or antennas. The 1st application was the matching of a 70cm amateur walkie-talkie to the input external power amplifier. Without the matching network the output power was quite low and the RF VOX doesnīt work correctly.
Look up: In extreme positions the matching might be worse with the match box. This might result in PA dammage or oscillations. So always start from the center positon at make only small steps.
L1 = 8.2nH (air coil : about 3 turns, 0.5mm Cu wire, inner diameter 2.5mm)
C1=C2=C3=1..22pF
C4=470pF
Figure 7 : 70cm band matching network
