I need to measure the reflexion at critical positions along travelling wave tube delay line structures that are based on folded waveguide, comb line or coupled cavities. These structures are highly dispersive, that is the wave velocity is a strong function of frequency. Hence, the time domain [/u] response is distorted, especially as the far end of the structure. The displayed response is spread out in time and is reduced in amplitude.

With the old VNA 8510 one could use the "table delay" function to correct for the dispersion (enter measured values for the electrical delay for each frequency at the desired position).

With the newer VNAs this is no longer possible as only "coaxial" or "waveguide" delay functions are available.

To overcome this problem we read out the measured frequency data and process it externally.

The following functions have been sucessfuly implimated and give the same as the internal routines of the VNA:

Window the frequency domain response (Kaiser-Bessel)

Discrete inverse Fourier transform from frequency to time domain (DIFT)

Electrical delay (waveguide)

However, I have not yet found suitable gate function so that I can isolate response from critical positions along the delay line.

I would be very grateful if you could provide me with further information regarding the gate function used by Agilent.

Best regards

Robert A. Nunn

With the old VNA 8510 one could use the "table delay" function to correct for the dispersion (enter measured values for the electrical delay for each frequency at the desired position).

With the newer VNAs this is no longer possible as only "coaxial" or "waveguide" delay functions are available.

To overcome this problem we read out the measured frequency data and process it externally.

The following functions have been sucessfuly implimated and give the same as the internal routines of the VNA:

Window the frequency domain response (Kaiser-Bessel)

Discrete inverse Fourier transform from frequency to time domain (DIFT)

Electrical delay (waveguide)

However, I have not yet found suitable gate function so that I can isolate response from critical positions along the delay line.

I would be very grateful if you could provide me with further information regarding the gate function used by Agilent.

Best regards

Robert A. Nunn

In fact, you can use equation editor to get a phase vs freq trace on another trace, and then use that trace to multiply by the DUT trace prior to the time domain transform.

You can also write your own custom .dll in the equation editor to process the data in whatever way you like, and then use the time domain transform on the processed data.