GMRT Observations

325 MHz observations for continuum imaging of a sample of nine candidate SNRs were done during the period of 1998-2000. The parameters of these observations are summarized in Table 4.1. Most of the objects of interest in these fields were extended, with emission at angular scales $\sim2-20$ arcmin requiring reliable observations at the smallest available baselines. Observations at 325 MHz are often affected by intermittent radio frequency interference (RFI) and, since RFI from nearby sources of emission remains partially correlated for the smallest baselines, data from the smallest baselines are also most severely affected by RFI. Fortunately, RFI is often narrow band and can be identified if the observing band is split into a number of narrow band frequency channels. The observations were therefore done with the full 16 MHz band split into 128 frequency channels of each having a width of $\sim125$ kHz.

The current single side-band GMRT correlator measures only the co-polar visibilities, i.e. only signals of the same polarization are correlated using the Indian mode of the VLBA Multiplier and Accumulator chips (MACs). All the 128 frequency channels of the co-polar visibilities corresponding to the right and left circular polarized signals were recorded for all available baselines. The data was kept in the multi-channel format throughout the imaging process to minimize the band-width smearing of sources away from the phase centre (see Section 4.2.2). After editing the RFI-affected or otherwise bad data, a typical bandwidth of about $5-6$ MHz was finally used for imaging giving a typical RMS noise of $\sim10$ mJy/beam.

The background sky temperature can vary by a factor of $2-3$ within the Galactic plane, resulting in a change in the total power output by similar factors. The correlator samplers are, however, optimized to work with an input signal of 0 dBm. To keep the sampler inputs at this level, automatic level controllers (ALCs) are introduced at the output of the baseband (BB) which effectively changes the gain to keep the output at 0 dBm. To keep the ALC operating point within its linear operating range, an attenuation of 16 db was typically used for the IF and BB signals.

Table 4.1: Parameters of observations with GMRT

Frequency of Observations (MHz)	325
RF Bandwidth (MHz)	16
Bandwidth used (MHz)	$5-6$
Integration time (sec)	16.9
Average time spent on the source (hr)	5
No. of Antennas used	$20-27$
Max. baseline (k$\lambda$)	$15-25$
Min. baseline ($\lambda$)	$60-100$
Max. spatial scale(arcmin)	$\sim30$
Average Antenna Sensitivity (K/Jy)	0.32
Primary beam (degree)	1.4
Synthesized resolution (arcsec)	$10-20$
RMS noise (mJy/beam)	$\sim10$

As mentioned earlier, the GMRT was in a state of being debugged during the period of the observations. Consequently, the number of antennas and the longest baseline available changed from observation to observation. As a result, the resolution changes from observation to observation. However, most of the Central Square antennas, plus some of the arm antennas were available for all observations giving angular resolution in the range $\approx 60-10$ arcsec. Some of the fields, which needed higher resolution were observed when the long baselines antennas were available. Since most of the target objects were extended, it was essential that the short antenna spacings were well sampled. Hence the three antennas, C05, C06 and C09 which provide the shortest spacings, were used for all observations giving a shortest baseline of $\sim 100\lambda$ (after flagging bad data) for which reliable data was measured. This corresponds to a largest angular scale of $\sim30$ arcmin. Most of the sources of interest were well within this limit in angular size and we believe that most of the emission up to angular scales of $\sim30$ arcmin is well represented in the GMRT 325-MHz images.