I am new into the RF measurements. I was measuring the S11/S22 (reflection) parameters of an RF attenuator (with different attenuation levels, e.g., 10 dB) via a vector network analyzer from Rhode & Schwarz.

I measured the magnitude of S11/S22 of the attenuator, which was typically below -30 dB. But when I measured the phase response of S11, then my teacher said it is not required. So, my question is, why do we not need phase response of S11/S22? I understand that we seek linear phase response in the useful frequency range of S12/S21 parameters (i.e., constant Group delay).

Additional question: Do we not need it at all for other components as well, e.g., Directional coupler?

  • If not, then why?

  • If yes, then why?

  • 4
    $\begingroup$ I'm voting to close this question as off-topic because it's really not a signal processing, but a microwave engineering question. I recommend moving it to electronics.stackexchange.com . $\endgroup$ – Marcus Müller Oct 9 '18 at 22:58
  • $\begingroup$ I agree with Marcus. The quick answer is that the phase response is not required as we are confirming (by ensuring S11/S22 is below -30 dB) that a minimum amount of power is reflected (meaning the attenuator has a good match). -30 dB is 1/1000 of the power, and results in a reflected phasor of 0.032 relative to the incident signal. (So knowing the phase of the reflection will have little impact). It is sufficient in this case for our objectives to just confirm the magnitude is sufficiently small. If you move this to electronics.stackexchange.com I will answer in more detail there. $\endgroup$ – Dan Boschen Oct 10 '18 at 2:11
  • $\begingroup$ Thank you for your suggestion and some quick answer. I am not sure whether I have rights to move this question to other forum (perhaps I don't have enough points?)? Can anyone help me with this? $\endgroup$ – learning Oct 10 '18 at 9:12