next up previous contents
Next: 6.5 Appendix Up: 6. Cross Correlators Previous: 6.3 The Correlator in   Contents

Subsections

6.4 The correlator on Plateau de Bure

As an example of a cross-correlator used in mm-wave interferometry, I briefly introduce the correlator system on Plateau-de-Bure. Only a spectroscopic correlator is in use. Continuum bands are synthesized by channel averages covering the desired bandwidths. Aspects concerning concrete observing projects are addressed in Chapter 8 by R.Neri.


6.4.1 The third-generation correlator

The third-generation correlator for the Plateau de Bure interferometer will allow for more flexibility, due to the following improvements: These improvements are made possible by using new, more integrated technology at both analog and digital signal processing steps.


The cross-correlator comprises eight independent units. Each consists of three parts: an IF processor (frequency setting, low-pass filter selection, oscillator phase control - i.e. the analog functions), a digital part, controlled by a master processor (i.e. delay steps, clipping correction, FFT, small delay corrections, bandpass correction), and a satellite micro reading out and further processing the correlations. Each unit can be placed in the $ [100,1100]$ MHz IF band 6.4, in steps of 0.625MHz (by using a third frequency conversion). There are seven combinations of bandwidth and channel spacing. The channel spacing follows a power-of-two sequence. Three out of the seven modes are synthesized by the adjacent upper and lower sidebands of an image rejection mixer. These bandwidths show the Gibbs phenomenon right in the middle of the band (i.e. at the edges of the IRM sidebands). The central two channels are flagged by default, the observer should avoid to place the most important part of the line there. The highest possible spectral resolution (channel spacing 0.039 MHz) is produced by slowing down the clock rate from 40 to 20MHz.


The spectroscopic capabilities of the cross-correlator at Plateau de Bure are summarized in Table 6.5. Part of the flexibility is achieved by using the ``time-multiplexing'' technique. For example, a time-multiplexing factor four means that the data, arriving at a rate of $ 160\times 10^6$ samples/s, are alternately put into four shift-registers. The shift registers are read out at the clock frequency of 40MHz, thus creating four data streams taken at a rate that is lower by a factor of four (as compared to the sampling speed). Equivalently, a time-multiplex factor two means two data streams at a rate of 80MHz each.


For further technical specifications see the Correlator Web page.

Table 6.5: The Complex Cross Correlator on Plateau de Bure
Bandwidth Sub-band Clock Time Number Complex Channel Spectral
  of IRM$ ^{(1)}$ Rate Multiplex of Lags Channels Spacing Resolution $ [$MHz$ ]$
$ [$MHz$ ]$   $ [$MHz$ ]$ Factor $ (2)$   [MHz] (3) (4)
$ 2\times 160$MHz LSB + USB 80 4 $ 2\times    128$ $ 2 \times    64$ 2.500 3.018 3.975
$ 1\times 160$MHz LSB or USB 80 4 $ 1\times    256$ $ 1 \times 128$ 1.250 1.509 1.988
$ 2\times    80$MHz LSB + USB 80 2 $ 2\times    256$ $ 2 \times 128$ 0.625 0.754 0.994
$ 1\times    80$MHz LSB or USB 80 2 $ 1\times    512$ $ 1 \times 256$ 0.312 0.377 0.497
$ 2\times    40$MHz LSB + USB 80 1 $ 2\times    512$ $ 2 \times 256$ 0.156 0.189 0.248
$ 1\times    40$MHz LSB or USB 80 1 $ 1\times 1024$ $ 1 \times 512$ 0.078 0.094 0.124
$ 1\times    20$MHz LSB or USB 40 1 $ 1\times 1024$ $ 1 \times 512$ 0.039 0.047 0.062
 
Notes: (1) image rejection mixer (2) with negative & positive time lags (3) box-shaped time-lag window (4) Welch time-lag window


next up previous contents
Next: 6.5 Appendix Up: 6. Cross Correlators Previous: 6.3 The Correlator in   Contents
Anne Dutrey