Background

Receivers B3 and W have an on-board interferometer allowing the image sideband to be rejected in "single-sideband" (SSB) mode or retained in "double-sideband" (DSB) operation. Note that the rejection is not perfect; in B3 it is at best about 13dB, and in W it has not been well determined. SSB operation results in better calibration in general. Receiver A3, however, does not have sideband rejection and thus is a DSB instrument by default. This has some special observational consequences, not only because the signals from both sidebands are observed simultaneously, but particularly for line calibration. The latter point is addressed here.

Consequences of DSB operation

Thus for any receiver operating in double-sideband mode, the strength of a given line observed in, say, the upper sideband (USB) is not likely to be exactly the same when observed in the lower sideband (LSB), and may be quite different. Observers should be aware of this effect and be sure to observe a source for which the same line has been observed as a "standard" wherever possible for comparison purposes. For A3 the presence of the pronounced "hump" in the receiver temperature between about 245 and 255 GHz already indicates that one needs to be especially careful in this region, and, because of the DSB performance of A3, be aware that having either sideband within the region is going to change the apparent line strength. Examples to illustrate this effectcan be had by looking at spectra of, say, NGC2071IR, in CO 2-1, CS 5-4, and HCN 3-2. In each case two spectra which place the line in the USB and LSB are shown superposed in green and red respectively. The CO spectrum is closely similar whether observed in LSB or USB, but CS should be observed in the USB, and HCN in the LSB, to avoid the calibration difficulties associated with the "hump" and the assumption of equal sideband gain. Further examples can be found in the "representative spectra" for A3. Similar databases exist for B3 and W.

Results using known lines

In this figure the integrated intensities of neighboring HC3N lines observed in CRL2688 are plotted as a function of the LO frequency. USB and LSB line intensities are shown in green and red respectively and compared with a scaled version of recent Trec values. The line intensities show a wild fluctuation across the A3 window, but it appears that virtually all this variation is due to the sideband ratio being non-unity. Data in the JCMT archive show that HC3N integrated intensities from CRL2688 remain similar into the 345-GHz window, and thus, dramatic changes in intrinsic line strengths across the A-band window are unlikely. Based on the sparse observations of isotopomers of HC3N it also appears that the transition is somewhat optically deep, limiting the variation of line brightness between transitions. Taking the average of pairs of neighboring transitions from the USB and LSB show that this is indeed the case (as indicated by the purple means symbols in the figure), and show that the sideband ratio in the region of the "hump" exhibits particularly strong variations.

The sideband question was also studied quite extensively during commissioning in December 1998, partly as an outcome of observations of SO and SO₂ at the time, and these early results are reported here. In this case there are more transitions within the band, but the line intensities require a more complicated excitation calculation.