How does a stub filter work page 5

Reflection Signal Measurement

8) The test set up is now changed as per Figure 17 with the Stub Filter replaced. The forward voltage is monitored by the oscilloscope channel connected to the 6 dB splitter and this channel is used for triggering. The Signal Generator is set to pulse at 300 Hz and the oscilloscope trigger adjusted to display the beginning of the burst of RF. This will probably mean performing single sweeps manually in order to obtain a waveform that starts at the zero crossing point of a rising signal voltage. 

Figure 17

[Set up for measuring Forward Voltage]

The waveform displayed on the oscilloscope shows the voltage building up as the reflected signal increases over each 180° phase change. Only the first 360° has major changes in the voltage levels and these should be similar to the calculated voltages.

Load Signal Measurement

9) The test set up is changed as per Figure 18 by connecting the coaxial cable output to the oscilloscope input, which is still set to 50 Ohms. 

Figure 18

[Set up for measuring Reflected Voltage]

The waveform displayed on the oscilloscope shows the voltage starts high and gradually reduces over each 180° phase change and these should be similar to the calculated voltages.


10) The calculations in this study have shown that the stub filter reflects the signal sourced from the generator back into the generator and phase cancels the signal going towards the load. The reflected signal towards the generator consists of partly the reflection caused by the mismatch at the T junction (which causes 180° phase reversal and is present at all frequencies) but  mainly by the reflection from the end of the stub filter and this reflection also has 180° phase reversal in relation to the point of the T junction at the filtered frequency. The fundamental behaviour of the circuit at the filtered frequency could therefore be regarded as similar to that of a short circuit placed at the T junction with a total reflection of power and a 180° phase reversal. The reality is that there is not a short circuit at the T junction and there are minor behaviors that show this. It is also apparent that there is an insertion loss at the T junction (0.51 dB) at all frequencies and so a three stub antenna tuner would have 1.53 dB insertion loss at the required transmit frequency, which with a 50 Watt transmitter would dissipate 15 Watts in the stubs.

 If you wish to study further and at higher frequencies, we do make a Stub Filter FLEXI-BOX which is not sophisticated technology but is  very easy to use in a laboratory environment. It is also very useful for making high dynamic range measurements using a spectrum analyser.

William J Highton  8/6/2011


Technical Article Index   < 4  5    Page Number