Once the scientific goals of the observing program are in mind, you must make the following decisions:
Precise Observing Frequency: When choosing a precise frequency, consider both the absorption spectrum of the atmosphere and the possibility of contamination by galactic spectral line emission.
Observing Mode: Select the appropriate observing mode from those available, and decide upon the setup parameters. Observing modes supported are ON/OFF's, single beam and dual beam grid mapping, and utility programs such as five-point maps, sky tips, and focus checks.
Beam Throw: A particularly important setup parameter for beam-switched observations is the beam throw. The throw can be varied from 0 to +/-4.5'; (a throw larger than +/-3'; may be inefficient in terms of the blanking necessary)(Currently, the beam throw is fixed at +/- 2').
Observing Time Budget: As with all observations, you should carefully plan the observing program to include time not only for observations of the program sources, but for pointing, focusing, and calibration.
When beginning a continuum observing run, we recommend that you follow these steps:
We strongly recommend that you invest the time to perform these checks. Remember that the responsibility for the integrity of the data rests with the observer.
You should take some care in the precise choice of a continuum observing frequency. In addition to the scientific goals that determine the approximate frequency, you should also consider two other factors: what frequency gives the best atmospheric transmission, and will the observations be affected by contamination from spectral lines in the bandpass. In the 3 mm band, maximum transmission occurs near 90 GHz, while in the 1.3 mm band, maximum transmission occurs near 230 GHz. Figure 6.1 is a plot of the atmospheric transmission as a function of frequency for the bands covered by the 12m receivers.
If the observations are near molecular clouds, spectral line contamination can seriously affect continuum measurements. This is particularly true if the continuum bandpass contains strong molecular lines, such as those from CO, HCN, or HCO+, but even a series of weak lines, when integrated over the band, can be a problem. As receiver sensitivities have improved, observers have discovered that some molecular line sources contain what is nearly a continuum of spectral lines - the so-called "spectral line forest". We advise continuum observers to research the sources they wish to observe to find if any of these molecular line sources are either associated with the program source or along the line of sight. You should then consult a molecular line listing such as the F. J. Lovas Catalog (1986, J. Phys. Chem. Ref. Data, 15, 251) to pick the optimum observing frequency. For continuum measurements the 12m receivers operate in the double sideband mode with 600 MHz of bandwidth in each sideband and a sideband separation of 3.0 GHz.
You can perform several types of continuum observations at the 12m. The most common types are:
This section describes the basic considerations, terminology, and techniques for making these observations. Detailed descriptions
of the observing procedures are given in Section 6.4
In contrast to spectral line signals, continuum signals are fundamentally indistinguishable from other broadband noise sources, such as atmospheric emission, spillover radiation from the ground and surrounding buildings, and receiver gain fluctuations. Such noise sources (particularly atmospheric emission) are, in fact, usually stronger than the celestial signal. Furthermore, atmospheric emission may change over very short time scales (<1 second) and may vary significantly at sky positions separated by as little as a few arc minutes.
To distinguish the celestial signal from other broadband signals, the 12m systems provide various switching and subtraction techniques. Most of the techniques employ beam switching of one type or another. Beam switching is accomplished by a nutating subreflector (also known as a chopping secondary). The default subreflector switch rate is 4 Hz. The subreflector throw can be varied from 0'; to its physical limit of 4.5';. The two subreflector positions are termed the +BEAM and the -BEAM. With the analog phase- sensitive detectors, the source produces a positive deflection on the chart recorder when it is in the +BEAM and a negative deflection when in the -BEAM.
Two types of beam switching are employed: single beam and dual beam. For point source ON/OFF scans, a dual beam observation means that during the course of the scan, the telescope is positioned so that the source is alternately in the +BEAM and the -BEAM. For mapping observations, the dual beam mode means that source emission is expected in each of the subreflector beam positions and that the dual beam restoration algorithm (Emerson, Klein, & Haslam 1979, A&A, 76, 92) will be applied. A single beam observation, for either point source or mapping observations, means that source is observed in only one of the subreflector beam positions.
The other switching mode is total power position switching for which the subreflector is locked to a single position (the +BEAM, by convention). Observers occasionally use this mode to make accurate aperture efficiency measurements so as to avoid any efficiency losses incurred during subreflector switching. These observations usually use a planet whose strength is large compared to the emission differences of the atmosphere at the two sky positions of a position switched observation. In the rare circumstance in which a source is so extended that the dual beam restoration mapping technique is inapplicable, total power observations may be required. For most standard observing procedures, observers prefer beam switching over total power position switching, as it provides much better cancellation of the atmosphere.
In the following, we give two methods which can be used to measure the true subreflector beam throw.
Continuum observations at the 12m are usually calibrated by a direct conversion of the measured antenna temperature into flux density (Janskys). The scaling requires that the observer determine the atmospheric zenith optical depth, usually done with a tipping measurement. For a point source, the conversion is given by the standard equation
k is Boltzmann's constant (1.380662 x 10-23JK-1),
TA is the measured antenna temperature with no efficiency or atmospheric corrections applied,
T′A is the measured antenna temperature corrected for atmospheric attenuation,
T*R is the measured antenna temperature corrected for atmospheric attenuation, radiative loss, and rearward and forward scattering and spillover,
Ais the aperture efficiency,
Ap is the physical aperture (113.10 m2 for the 12m),
0 is the zenith optical depth, and
A is the number of airmasses.
The quantity = 24.4155 Jy K-1 for the 12m.
A convenient measure of sensitivity for continuum observations is the rms flux density per root integration time outside the Earth's atmosphere, S0. The sensitivity achieved in a given integration time t and under an atmosphere with zenith optical depth 0, is given by
S = S0t-1/2 exp( 0A) (6.2)
A table of S0 values for key frequencies is given in Table 6.1
Table 6.1:12m Receiver Continuum Sensitivities
The data from all standard continuum observations are processed through the digital backend (DBE). In 3.6.3 we gave a description of the hardware configuration of the DBE. It processes two independent channels and accumulates the signals from four switching phases, defined as:
where SIG and REF refer, respectively, to the signal and reference positions (+BEAM and -BEAM) of the nutating subreflector. CAL refers to the signal generated by the synchronous noise tube (which is available for 3 mm observations only). If the subreflector is fixed, as in a single-beam, total power measurement, the SIG and REF phases are the same (apart from random noise).
The blanking and delay parameters for the subreflector are set by the ARO staff regularly. The default switch period is 0.21 seconds (a 4 Hz switch rate), so that the control computer can acquire DBE data in synchronism with its basic timing cycle. The mechanical inertia of the subreflector imposes a limit of about 10 Hz on the maximum switch rate.
When the operator initiates a continuum scan, the control computer issues a signal to the DBE to begin a cycle of four phases. Each time a cycle is completed, the control computer reads the contents of the four phase registers of the DBE. Ideally, this occurs every 500 msec of elapsed time. The presence of error conditions (loss of phase lock or pointing out-of-tolerance) may cause the elapsed time to exceed the specified integration time. If the control computer detects an error condition during phases 1 or 2, it sends a restart signal to the subreflector and the phases are retaken. If the error condition occurs during phases 3 or 4, the computer rejects the entire 500 msec of data. The control system time counter counts down monotonically in 100 msec steps (the monitor updates only every second).
The data sent from the control computer and recorded in the sdd file contain the four phases, integrated over the number of dumps, for each sample or sky position. Thus, the raw data scan contains four times as many points as there were samples. If the sample number is denoted by Si and phases in that sample as i, the data are recorded in the order
No calibration factors have been applied to the data as it is recorded; the data are completely raw counts from the DBE. The raw data are converted into meaningful signals by the analysis program condar. The signals that can be computed from the phases are:
Switched Power: SP = (1 - 2 + 3 - 4) / 2
Total Power: TP = (1 + 2 + 3 + 4) / 4
Cal Signal: C = (1 + 2 - 3 - 4) / 2
Zero Level: Z = 1 - 2 - 3 + 4
Since the data stored in the analysis system are completely raw, there are several calibration parameters used by condar which must be set in order to get proper temperature and flux density scaling. Some of these can be set in the control system before you begin data acquisition:
This is the optical depth of the atmosphere at the zenith (in nepers).
The condar verbs that manipulate and calibrate DBE data are:
SWITCHED reorders the data array to contain only calibrated switched power data.
TOTALPWR reorders the data array to contain only calibrated total power data.
CALDBE reorders the data array to contain only calibrated calibration values.
ZERO calculates the calibrated rms of the "zero" values and stores the rms in the condar array vrms (1).
AVGcomputes the mean temperature and standard error of the mean of an ON/OFF sequence.
In general, you will not need to use these verbs explicitly as resident condar procedures (see below) are available for standard displays of switched and total power data.
The way in which the data are calibrated depends upon the type of calibration chosen when you acquired the data. For the three types of calibration available; vane, fixed, and noise tube, the data scaling factors are given by
where C is the average noise tube calibration signal from each sample, VANE is the signal measured from the vane, SKY is the signal measured from the sky, A is the airmass, and the zenith optical depth. Tc is the calibration scale factor defined above.
When using noise tube calibration, the verbs switched and totalpwr compute a system temperature and store the result in the header under the label tsys. The computation of tsys is by way of the equation
The verbs switched and totalpwr also calculate a signal-to-noise figure of merit called tpsn, which is stored in the adverb vrms(2). tpsn is the switched power signal divided by the total power signal and is given by
The basic condar data reduction commands are given below. #1 is either 1 or 2, for continuum channels 1 or 2. In the c#1 command below, #1 can be either 1, 2, or b for "both".
scan_number get table: produces a table listing of scan_number
scan_number s#1: displays a switched power ON/OFF sequence
scan_number m#1: displays a mapping row
scan_number f#1: displays a five-point map
scan_number sptip#1: displays vane-switched tipping scan
scan_number foca#1: displays a focus-check scan
scan_number c#1: stacks scans with specified weighting
The 12m has available several observing procedures for making ON/OFF measurements,
i.e., measurements of the difference between the output powers at a defined point (ON) and a nearby reference position (OFF). These procedures work with the subreflector nutating or not. The ON/OFF procedures are often used for measuring the flux density of weak point sources.
The standard point source ON/OFF observing procedure is called a sequence in the control system terminology. Using this procedure, the telescope moves between the ON and OFF source positions in the pattern OFF-ON-ON-OFF. A sequence is made up of one or more repeats of this basic cycle. This order of samples eliminates the effects of linear drifts in atmospheric noise or receiver gain on the measurements. You can specify integration time per position. (Procedures that are simply ONs, OFFs, or an ON-OFF pattern are described below.)
The standard set-up for an ON-OFF sequence in beam switched mode is double beam switching (DBS), i.e. with the ON and OFF positions separated by the subreflector throw. In this way the radio source to be studied can be cycled between the positive and negative beams by the movement of the telescope. Figure 6.2 displays the positional relationships between the OFF and ON samples. For the DBS mode, the output power difference between the ON and OFF phases is proportional to twice the source flux density. The advantage of the ON-OFF approach is that, to first order, it cancels imbalances between the two beams.
One can, of course, select the OFF position to be different from the position of the -BEAM (or the ON position different from the +BEAM). This is a single beam (SBS) observation. The most frequent use of the SBS mode is to determine the position of the
+BEAM and the -BEAM independently and thereby determine the subreflector throw and orientation.
To perform an ON-OFF sequence, you must give the operator the following setup parameters:
Figure 6.2: Position definitions for continuum sequences.
One option is available for mapping extended continuum sources: a grid-mapping procedure in which the telescope steps through a rectangular grid in the azimuth/elevation. For a detailed explanation regarding the pitfalls and perils of under sampling in mapping data, see 220.127.116.11.
In continuum grid mapping, a field is mapped in a rectangular grid in azimuth and elevation relative to the field center. This observing procedure has been developed for use with the dual-beam restoration algorithm of Emerson, Klein, & Haslam (1979 Astr. Ap.,76, 92). It will, however, work satisfactorily for either fixed beam (total-power) mapping or the mapping of fields that are smaller than the subreflector throw via the +BEAM. Note that continuum OTF mapping is functionally the same as continuum grid mapping for most applications. Given the fact that we currently have no analysis software for continuum grid mapping, but have analysis software for continuum OTF, we recommend that observers use continuum OTF mapping for all continuum mapping experiments.
The telescope builds a two-dimensional grid of observations by scanning rows at constant elevation relative to the source position. The telescope moves along a row in discrete steps, performing an integration at each position. The grid points are separated in azimuth by real angle. If a map has M columns and N rows the requested field center will lie at the central grid point INT[(M+1)/2], INT[(N+1)/2)] if N and M are odd numbers, where INT denotes an integer truncation. If N and M are even, the requested field center will fall at grid point [(M/2)+1, (N/2)+1]. A scan represents a row of the map in this observing mode. The mean sidereal time of each point is stored in the scan array. Analysis programs have been prepared to combine the scans into two-dimensional maps. You can process these maps using the dual beam restoration algorithm, then transform them into celestial coordinates and stack them. Unfortunately, we do not currently have an implementation of this analysis system at the telescope.
To observe such a map, proceed as follows:
To properly calibrate continuum data, you must correct for the attenuation of the signal as it traverses the atmosphere. We usually determine the atmospheric attenuation from the optical depth at the zenith, . The total power antenna temperature at a given airmass (elevation) is a function of and is given by the equation:
where Trx is the receiver noise temperature, l is the "warm spillover" efficiency (rear spillover, scattering, blockage, and ohmic loss efficiency), Tm is the mean atmospheric temperature, is the atmospheric optical depth at the zenith, A is the airmass, Tspill temperature of the warm spillover, and Tbg is the temperature of the cosmic background radiation. Note that Tm, Tspill, and Tbg are actually equivalent Rayleigh-Jeans temperatures of the point on the Planck blackbody curve corresponding to the same frequency. This correction factor is given by
where v is the observing frequency, T is the kinetic temperature, and h and k are the Planck and Boltzmann constants, respectively. For simplicity, we will retain the symbol "T" for temperatures, but in calculations "T" should be replaced by "J(v,T)".
The opacity of the atmosphere at the zenith, is usually determined by measuring the TA(sky) at several elevations and fitting the results with some form of Equation 6.9.
The mean atmospheric temperature and the spillover temperature are usually ~0.95 - 0.97 of the ambient temperature.
The basic observing procedure for measuring is called an sptip. This procedure steps the telescope through a series of elevation angles, moving from low to high elevation, and then back to low. The precise elevations, in steps of 0.3 airmasses, are listed in Table 6.2. In an sptip, the system makes total power observations of both the sky and the ambient temperature absorbing vane at each position in Table 6.2. Both sky and vane temperatures are recorded for analysis.
To perform an sptip observation, simply ask the operator to slew the telescope to an azimuth near your program source position and start the measurement. The condar data analysis package contains two methods of analyzing an sptip. If you wish on-line atmospheric corrections to be applied to continuum data displays, ask the operator to enter the new value of. Remember that the atmospheric absorption changes with time, and if the highest accuracy is required, a small correction should be applied off-line using all the opacity data for the run to interpolate the most-probable value for each scan.
Table 6.2: Sky Tip Antenna Positions
The sptip data reduction procedure uses the difference between vane and sky temperatures to derive . This is the simplest of all the tip analysis routines: it requires no input parameters, does not require accurate calibration of the temperature scale, nearly always produces at least a first order approximation to the optical depth, but makes a number of simplifying assumptions that may diminish its accuracy.
The antenna temperature of the sky is given in Equation 6.9, while the antenna temperature of the vane is given by
TA(vane) = Trx + Tvane (6.11)
where Tvane is the Rayleigh-Jeans-equivalent temperature of the vane (generally the ambient temperature). The difference between vane and sky temperatures is thus
The SPTIP analysis procedure makes the following assumptions:
Tvane = Tm = Tspill = Tamb,
hl = 1, and
Tbg = 0 (6.13)
where Tamb is the Rayleigh-Jeans-equivalent ambient temperature. Under these assumptions, one then finds that
ln(T-s) = ln(Tamb) - A. (6.14)
From the measured array of Tv-s as a function of A, the sptip procedure fits for by linear least squares. Among the assumptions 6.13, the assumptions that Jm = Jamb and that l = 1 are the most inaccurate. In addition, because the calibration scale is not always in calibrated in Kelvins, the term ln(Tamb) in Equation 6.14 is fit as a free parameter. This gives more freedom than it ought to have.
Experience with the sptip routine has shown that although it is the easiest of the available routines to use, it suffers from two deficiencies: (1) it tends to underestimate , especially for large values of , and (2) it often appears to give a misleadingly good fit, i.e. it may indicate that the atmosphere is more stable than it really is. Nevertheless, under moderately low 's (say <0.3), the sptip procedure usually gives an acceptably good measure of the atmospheric optical depth.
To use the sptip analysis procedure, in condar type:
Condar > scan_number sptip#1
where #1 is either 1 or 2 to choose channel 1 or channel 2. The data and fit are displayed in Figure 6.3
The stip reduction procedure fits Equation 6.12 to the data from an sptip observation by way of nonlinear least squares.
This procedure is the most direct of the two tipping analysis options: it makes no assumptions other than the accuracy of the basic model (Equation 6.9). To
give an accurate measure of
it does have several stringent requirements:
Figure 6.3: Sample SPTIP tipping scan.
The stip procedure can also be used to fit for l, a basic efficiency factor. Because it is a nonlinear (hence iterative) fitting routine, it may not converge if the data are of poor quality or, conversely, it may converge to a nonsensical result. Use stip with caution. To analyze an sptip scan with stip in condar:
Condar > scan_number stip
If you want to solve for hl, set etafree=1 in condar (etafree=0 by default). Figure 6.4 shows what the stip output looks like.
To accurately calibrate continuum data, you must set the temperature scale and/or flux density scale, correct for atmospheric attenuation, and correct for telescope systematics such as the gain-elevation effect. The method currently used to calibrate 12m continuum data is called "vane calibration".
Calibration of the antenna temperature scale for continuum data is accomplished using the vane calibration technique (see 5.6.1). In addition to a calibration of the antenna temperature scale, absolute calibration of continuum data also requires a measurement of the aperture efficiency or a factor for converting from antenna temperature to flux density. It may also be necessary to apply a correction for the gain-elevation effect if it is important at the particular observing frequency. To calibrate continuum data, follow the steps outlined below.
Many observers choose to skip the measurement of the temperature scale by hot/cold loads and calibrate their data strictly by comparisons with standard sources. In this relative calibration method, one merely observes a succession of standard sources, forms a calibration curve of flux density conversion factors, and through interpolation, corrects the data after the run is over.
Figure 6.4: Sample STIP tipping scan.
A hot/cold load calibration can be used to set the continuum temperature scale, to measure the receiver noise temperature, the noise tube temperature, the sky temperature, and the atmospheric optical depth. At present, we perform hot/cold loads by manually inserting hot and cold loads into the beam and measuring the deflections through the computer, on digital voltmeters, and sometimes on the analog chart recorders. The cold load usually consists of a styrofoam box holding a square of microwave absorber immersed in liquid nitrogen. The vaporization temperature of liquid nitrogen at the elevation of the 12 Meter Telescope is 80 K. For frequencies below about 230 GHz, we recommend that you use this temperature. For higher frequencies, the styrofoam may begin to absorb the radiation and will add to the temperature of the load. At these high frequencies, we recommend that you use another type of cold load known as a “dipper”. The dipper consists of a funnel-shaped box lined with absorber on the end of a wooden handle. The dipper is dipped into liquid nitrogen and held over the receiver feed during cold load cycles. A hot/cold measurement requires the active participation of the operator and at least one observer. To perform a hot/cold load calibration with a coherent receiver, proceed as follows:
To process the HOT/COLD measurement, in condar:
Condar > scan_number hc
The hc procedure will prompt you for the hot load temperature in Celsius and the cold load temperature in Kelvins. If the observation was made in double sideband mode, give the frequency of the local oscillator. The hot load temperature should be the temperature of the vane; the vane temperature is displayed on the Chopper Control" chassis in the control room. After entering these data, the crosshairs will appear on the graphics screen and you will be prompted to mark the zero offset, the cold load, and hot load samples. The procedure will then prompt you to enter any sky and noise tube samples that may exist. Ask the operator to update the value of Tc calculated by hc for the type of calibration, fixed or noise tube, that you are using.
If you wish to reduce the data by hand, you can get a table of the numbers in the scans by typing
In addition to the computer scan, you should reduce the data obtained from the voltmeters (or chart recorders) for comparison. The formulae for the quantities of interest are given below. The receiver noise temperature is a function of the “Y-factor”, defined by
where Vhot is the measured hot load voltage less the zero offset voltage, and Vcold is the measured cold load voltage less the zero offset voltage. The receiver noise temperature Trx is then given by
where Thot and Tcold are the equivalent Rayleigh-Jeans effective temperatures of the hot and cold loads, respectively, given by Equation 6.10. The Kelvins per volt, KPV, of the continuum system is given by
The Rayleigh-Jeans equivalent sky temperature is given by
Tsky = (Vsky - Vcold) KPV + Tcold (6.18)
where Tcold is the equivalent Rayleigh-Jeans temperature of the cold load given by Equation 6.10. If a noise tube was measured, its temperature is given by
Tnt = (Vnt - Vsky) KPV (6.19)
where Tnt is the Rayleigh-Jeans equivalent temperature of the noise tube and Vnt is the measured voltage of the noise tube. The optical depth of the atmosphere at the zenith is, from Equation 6.9
where A = 1 if the observation is at the zenith.
Equations 6.15 through 6.20 are used in the condar procedure hc and can be used to reduce the computer data manually if you so choose. If you are using a noise tube, use the value of Tnt given by Equation 6.19 for the calibration scale factor, Tc, entered into the control system. If you are not using a noise tube, the value of Tc to use is given by
where Thot and Tcold are the Rayleigh-Jeans equivalent hot and cold load temperatures, TA(hot) and TA(cold) are the apparent antenna temperatures of the hot and cold loads, respectively, and Tc(old value) is the old value of Tc that was in the computer during the hot/cold measurement. For each channel, there should be good consistency between the values computed via the computer and the voltmeter.
If you are using a noise tube to calibrate the data, the calibration scale should stay fairly accurate even if the gain of the receiver changes slightly. If you are not using a noise tube, the Tc's and noise temperatures are accurate only at the moment they are measured: receiver gain and tuning drifts will change these parameters. Depending on your choice of calibration methods, you may need to repeat the HOT/COLD measurements frequently.
Table 6.3: Planetary Flux Density Standards
For most continuum observations, the flux density scale is calibrated by observations of standard radio sources. In doing this, it should be remembered that, in addition to corrections for receiver and atmospheric effects, you should allow for the gain-elevation properties of the telescope if the observations cover a significant range of elevations. Current gain-elevation curves are given in The NRAO 12m Telescope Equipment and Calibration Status document.
At millimeter wavelengths, the flux densities of most extragalactic sources are variable and we recommend the use of the planets or compact HII regions for calibration. At least at 1mm, the brightest planets are usually significantly resolved. The peak flux densities of the planets should be computed using the planets utility program available on any of the mountain workstations. This program needs to know the effective observing frequency (GHz), the telescope HPBW (arcsec), the planetary unit semi-diameter in arcseconds (i.e. the semi-diameter of the planet as seen from a distance of 1 AU, available from the Astronomical Almanac or Table 6.3), the geocentric distance of the planet (AU), available from the Astronomical Almanac, and the brightness temperature of the planet at this frequency. The result is given in Jy/beam.
Table 6.3 gives the recommended brightness temperatures of the planets at 90, 150, and 227 GHz. Most of these are taken from the work of Ulich et al ., Griffin et al ., and Hildebrand et al .. The brightness temperature of Mars depends on solar distance. At 90 GHz, Ulich (1981, AJ, 86, 1619) suggests an effective temperature for Mars of
where R0 is the Mars-Sun distance in AU.
Some other radio sources are expected to be non-variable, and in the case of HII regions, unpolarized. Peak flux densities for sources measured with the 12 Meter Telescope that make suitable flux density calibrators at 90 GHz are given in Table 6.4.
Table 6.4: Flux Density Calibrators
A sample continuum status monitor display is shown in Figure 6.5. Each numbered box in Figure 6.5 indicates a section of the display which describes a particular set of attributes of a continuum measurement:
Box 1: Scan number, source name, and timing information.
Box 2: RA/Dec and lII/bII position information.
Box 3: Apex position information.
Box 4: Azimuth and elevation position and pointing offset information.
Box 5: Receiver calibration and tipper information.
Box 6: Subreflector beam and quadrant detector position information.
Box 7: Frequency, velocity, and calibration information.
Box 8: Telescope tracking and weather information.
Box 9: Current observation scan and integration time information.
Copyright Arizona Radio Observatory.