The ``Interim'' Correlator

For spectral-line observing, the original post-upgrade ``interim'' correlation spectrometer provides four independent sub-correlators, each having eight chips with 1024 lags per chip. Each sub-correlator can be set up with its own independent bandwidth and configuration. Table 4 specifies the available configurations. The maximum bandwidth per sub-correlator is 50 MHz, with 8 other alternative bandwidths being available in decreasing octave steps. Popular configurations combine chips in groups of four to provide 9-level sampling (96% efficiency), interleaved operation (50-MHz bandwidth) or auto+cross correlation (all four Stokes parameters).

Table: ``Interim'' correlator configurations with all 4 digital filter boards

Config	Max Bw	Pol/Sbc$^{\ast}$	Boards	Lags/Sbc$^{\ast}$ & Resn -
	per Sbc$^{\ast}$ (MHz)		Used	(kms$^{-1}$ @ 1420 MHz)
9-level	25	1	4	2048 (2.6)
9-level	25	2	4	1024 (5.2)
3-level	25	2	4	2048 (2.6)
3-level interleaved	50	1	4	4096 (2.6)
3-level interleaved	50	2	4	2048 (5.2)
3-level Stokes	25	Full Stokes	4	2048 (2.6)

$^{\ast}$ Sbc = Subcorrelator

Notes:
1) The spectrometer has its available bandwidths defined by the following filters: a) Analog filter - 50 MHz, b) Digital filters - 25, 12.5, 6.25, 3.125, 1.563, 0.781, 0.391, 0.195 MHz
2) Double Nyquist sampling can be used with all configurations except interleaved, but decreases the maximum bandwidth by a factor of two.
3) 3-level, double Nyquist, 12.5-MHz bw and below will give 4 sub-bands with better resolution than the corresponding 9-level configuration.
4) 9-level operation achieves 96% of the signal-to-noise of analog correlation, whereas 3-level achieves 81%.
5) Different correlator boards can operate with different configurations to each other.
6) The fastest dump rate for spectral-line usage is about 10 Hz.
7) The number of lags used can be reduced by a factor of 2$^{n}$ down to 16 lags. The size of the output file will be proportional to the number of lags.