ALFA Motor RFI Measurement

The following describes a preliminary assessment of possible interference from the ALFA-motor control system.

The motor electronics are housed in a shielded box located on the rotary floor and next to the ALFA receiver.

The plots (see below) show the results for the X & Y channels of beam 0, based on the measurements made on the 11th June 2004, when the ALFA motor was "on" (but not running). (100 MHz BW centered at 1420 MHz, 1024 lags or channels, 0.1 second dumps [of course, in the pulsar mode], ~240 sec spans).

The door of the control box was "open" in one case, and "closed" in the other. For this to be a meaningful test, the ALFA feeds were uncovered (to allow sampling of the possible interference), but the array was deliberately not kept at the focus (to reduce possible confusion from sky contribution as well as the usually expected RFI at L-band). The telescope was not driven during this test.

The correlations between temporal fluctuations in intensities of each spectral channel pairs is displayed in the form of a map in the central panel. The left and bottom panels show average spectrum (black), auto/cross-correlation for pairs at the same frequency (green) and the efficiency (blue). The efficiency, used here as a sensitive detector of RFI/instabilities, is estimated as the ratio of signal-to-noise observed to that ideally expected. This nominal magnitude of this so-called efficiency is as expected for the 3-level quantization (i.e. about 0.81).

The interference in the "door open" state may be understood (pending detailed analysis) as occurring, in principle, at all the harmonics of about 13.33333 (i.e. 40/3).... e.g., the features at 1400, 1413.33.., and 1440 MHz where there is noticeable contribution on the average, correspond to 105th, 106th, and 108th harmonics respectively. The apparent width of these features at L-band is suspected to be a little more than the resolution in these spectra (~100 KHz; or ~20 km/s). ..... meaning that the width of the fundamental feature (i.e. at 13.333 MHz) may be about 1 KHz.... consistent with the generally expected stability of modern-day free-running oscillators/clocks.

Also, note that the efficiency drops significantly at the locations of the RFI harmonics within the observed band (while the presence of the RFI may not always be apparent in the average spectrum), suggesting that the RFI is variable on the time scales much shorter than the observed span. The apparent (-ve) correlation of some of the RFI features with other parts of the band (seen as the cross features in the correlation map), signals occasional gain-compression due to strong RFI.

It is reassuring that the "door closed" state makes a huge difference, providing significant and desirable isolation as from the RFI. However, this isolation from the motor control-system interference needs to be assessed further, e.g., over longer integration times and when the Motor is actually running.


The Plots

Door Closed, auto correlation, polarization X.
Door Closed, auto correlation, polarization Y.
Door Closed, cross correlation.
Door Open, auto correlation, polarization X.
Door Open, auto correlation, polarization X.
Door Open, cross correlation.