CUTLASS Pulse sequences


Background

The CUTLASS radars can operate a great variety of pulse sequences. The chosen pulse sequence will determine the alias points and resolution of the derived spectra, the duty cycle, the cross range noise and the actual parameters derived. Listed below are some of the actual pulse sequences used in the radar, the radops programs that use them, their vital statistics and what these actually mean for the data taken.

These pages are under development, so currently only a small subset of this information is actually documented.


normal scan

Normal scan is the radar program run for "standard" radar operations. Thus all of common time operations should run this pulse sequence. The current version of the pulse sequence was instigated at the beginning of August, 1995.

Picture of pulse sequence

The pulse length Tp is 300 microseconds, corresponding to 45 km range gates

The basic lag separation TL is 2400 microseconds

There are seven pulses in the sequence, and a total of 18 lags are calculated. Each lag is unique, although some lags are missing. The radar keeps as many of these pulse trains in the air at once as it can, with each pulse sequence lasting 67.2 milliseconds

The actual calculated lags are derived from the pulse pairs listed below:

 0 26 20  9 22 22 20 20 12  0 12  9  0  9 12 12  9  9
 0 27 22 12 26 27 26 27 20  9 22 20 12 22 26 27 26 27

The calculation of the lags is illustrated in the enclosed schematic.

A discussion on resolution

There are three distinct resolutions which can be identified in the operation of the radar and each resolution is associated with a different temporal parameter.
Pulse Length, Tp - Range resolution
Integration time, Ti - Temporal resolution
Lag separation, TL - Spectral resolution

Tp determines the range resolution. This is given explicitly by

where c is the speed of light.

Ti is the integration time of the radar receiver. This is the dwell time for the radar to look along one particular beam direction. N pulse sequences will be transmitted, received (scattering target permitting) and averaged during this time. In normal scan, it is possible to process about 10 pulse sequences every second.


Normal scan is operated with a 7 s integration time. A full scan of 16 beam directions takes 112 s to complete. In normal scan each new scan is begun on a 2 minute boundary. This timing restriction is to allow the synchronisation between different SuperDARN radars at the expense of approximately an 8 s inactive period at the end of every scan.


TL, the minimum lag separation, is the smallest sampling interval in constructing the backscatter autocorrelation function (ACF). If a power spectrum is created by an FFT of the ACF, the spectrum will have alias points at +/- (1/2TL), the Nyquist frequency. ACF and SPECTRUM

The smallest frequency change that can be directly resolved corresponds to the longest lag measured/calculated in the ACF, i.e. 1/(NmaxTL), Nmax=18 for normal scan. The actual frequency resolution obtained is far better than this since the mean Doppler frequency is obtained from a fit to all the points in the ACF (as is routinely performed by the radar FITACF task) or alternatively by calculating the spectral moments.

The Nyquist frequency and spectral resolution can be converted from frequency to Doppler velocity by the relation:

where ft is the radar frequency.

Other Pulse Schemes

High resolution modes

Many special programmes are run where the time and spatial resolution of the radar is increased. The pulse scheme is the same as Normal scan except that the integration time Ti and/or the pulse length Tp are changed.

Tp can be reduced to 100 microseconds giving a range resolution of 15 km. The choice of beam integration time is currently limited to a minimum of 1 s.

Further ideas (Just to stimulate thought really!)

The radar operating software allows the implementation of almost any pulse sequence you can think of a good reason to use. Although, some ideas may confuse bits of the software.

For example here are the lags that can be calculated from the three pulse scheme (0,2,3)

 0 3 2 3  
 0 2 0 0

0 1 2 3 ... lags

Why oh why would one do such a thing? The pulse scheme gives much reduced spectral information. However, one can still calculate a backscatter power and an estimate of the the Doppler velocity. The advantage is that one can fit more of such sequences into each second and so improve the SNR for small Ti.


These pages are administered by: WebMaster

Document last updated