Is the stepped frequency bandwidth calculation in the phased array toolbox correct?

Phased Array System Toolbox Version 3.2 (R2016a) 10-Feb-2016

Example: hwav =

phased.SteppedFMWaveform with properties:

               SampleRate: 12000000
    DurationSpecification: 'Pulse width'
               PulseWidth: 1.3333e-06
                      PRF: 8000
    PRFSelectionInputPort: 0
            FrequencyStep: 7.5000e+03
                 NumSteps: 4
             OutputFormat: 'Pulses'
                NumPulses: 1

K>> bandwidth(hwav)

ans =

   3.0000e+04

K>> burst_bw_hz

burst_bw_hz =

      772500

K>> 1/hwav.PulseWidth

ans =

      750000

where burst_bw_hz=pulse_bw_hz+((nfreqs-1)*del_freq_hz);

How can the burst bandwidth be smaller than the pulse bandwidth?

Answer 1

[Honglei Chen]

Could you clarify the burst bandwidth computation? I would be interested if you could point me to a reference.

The stepped FM waveform is a way to achieve frequency sweeping using simple pulses. Thus the bandwidth of each pulse is independent to the bandwidth achieved by the entire waveform. Therefore, I'm not sure if there is a definite relation between the two.

Thanks.

[Thomas Medl]

The BW of the "entire waveform" or "burst" IS dependent upon the pulse BW and the frequency step between pulses within a burst. See NumSteps and FrequencyStep attributes of the WF above. More documentation on the WF can be found by typing "doc phased.SteppedFMWaveform" in Matlab if you have the toolbox.

[Honglei Chen]

I couldn't find the burst bandwidth computation in the following reference page. Would you mind pointing me to that?

https://www.mathworks.com/help/phased/ref/phased.steppedfmwaveform-class.html

The NumSteps property help did say that when there is no frequency step, the waveform becomes a simple rectangular waveform, so in that case the bandwidth is the reciprocal of the pulse width. But if you have multiple frequency steps and if you process the entire waveform coherently, then what matters is the bandwidth covered by the steps, not the bandwidth for each single pulse.