In selecting keywords for local DSN use, we try to avoid duplication by using local keywords used at other institutions or by existing software packages when no ambiguity can result:
Three strategies for recording data modifications were identified.
Depending on the type of correlator being used, the data acquisition software would initially set these keywords to either 'perform' or 'omit' as appropriate. During the processing steps, those with values of 'perform' would get changed into 'complete'.
At the completion of autocorrlator data processing, one has a spectrum similar to that obtained from other instruments, with the except that spectra obtained from autocorrelators are normalized, that is, the average value of the spectrum is 1. Multiplying this with a scaling constant (e.g. system temperature in K, spectral flux density in W/hz, etc.) yields a calibrated total power spectrum.
| Type | Purpose | Processing Required |
|---|---|---|
| raw | Diagnose equipment | none |
| total power | monitor receiver performance, RFI, etc. | normalize and scale to physical units |
| difference | measure weak signals | various, depending on observing mode and spectrometer capability |
The following figure shows all the possible ways in which spectra might be processed to obtain calibrated data.
The usual procedure is for the observatory staff to determine how the desired data type is obtained from each instrument, given its capabilities, the capabilities of the receiver to which it is connected, and the observing mode.
| The IRAF package uses a keyword DATA-TYP with values of 'BIAS', 'DOME FLAT', 'SKY FLAT', 'OBJECT', 'COMPARISON', 'DARK'. It is used at least by the University of Victoria and by NRO. Such a keyword is also needed for radio spectroscopy, and then a subset of these keyword values might serve: |
|
||||||||||||||
| The risk of this strategy is that if one used an IRAF-based package written elsewhere to process DSN data, this keyword and its values might trigger unwanted operations. SEST La Silla uses the keyword OBSTYPE for a similar purpose and, because it is a radio observatory, might be closer to our needs: |
|
| DATATYPE = | Meaning |
|---|---|
| sig | data with the putative signal present |
| ref | data with the signal absent or frequency-shifted |
| switched | sig-ref differenced in hardware, or at least not in the data acquisition program |
| amb-load | data taken on the ambient load |
| cal-source | data with a noise diode on |
| test | could be anything; look in the COMMENT fields |
| zero | data with the front-end disabled |
With this convention, the initial values of the reduction keywords will depend on observing mode and data type.
| Keyword | Source | Meaning | Example |
|---|---|---|---|
| ELAPTIME | UCO/Lick | The duration in seconds of the scan (starttime - stoptime) | For a single sweep of a (non-averaging) spectrum analyzer, ELAPTIME is the sweep time. |
| OBSTIME | SEST COMB |
The amount of time, in seconds, that each resolution element collected useful data. | For a multichannel spectrometer,
OBSTIME = ELAPTIME.
For a scanning spectrum analyzer, OBSTIME is the resolution bandwidth divided by the sweep rate. (The sweep rate is the span divided by the sweep time.) |
| N-SPEC | The number of spectra comprising this scan. | For a spectrometer with an inherently short integration time, OBSTIME may consist of many spectra. For an integrating spectrum analyzer, this is the number of sweeps comprising a displayed spectrum. | |
| EXPOSURE | SDF | Effective integration time so that the radiometer
equation  is satisfied. |
For a Dicke-switched spectrum, EXPOSURE = OBSTIME/4 |
| Keyword | Cal Scan | Ref Scan | Sig Scan | Notes |
|---|---|---|---|---|
| FRONTEND | Ka-core | Ka-core | Ka-core | The core Ka-band front end at DSS-13 has only a single horn |
| BACKEND | WBSA | WBSA | WBSA | The Wide Band Spectrum Analyzer is a digitial FT spectrometer which accumulates only total power spectra |
| OBSMODE | LINE+PSSW | LINE+PSSW | LINE+PSSW | |
| DATATYPE | cal | ref | sig | |
| TCAL-DIF | omit | omit | perform | process with the pink equation in the above figure |
| CALSCALE | omit | omit | perform | process with the brown equation in the above figure |
Frequency switching will be the same except for the OBSMODE keyword.
| Keyword | Ref Scan | Sig Scan | Notes |
|---|---|---|---|
| FRONTEND | Ka-core | Ka-core | |
| BACKEND | WBSA | WBSA | |
| OBSMODE | LINE+PSSW | LINE+PSSW | |
| DATATYPE | ref | sig | |
| TSYS-DIF | omit | perform | process with the blue equation in the above figure |
| CALSCALE | omit | perform | process with the brown equation in the above figure |
Frequency switching will be the same except for the OBSMODE keyword.
| Keyword | Ref Scan | Sig Scan |
|---|---|---|
| FRONTEND | Ka-core | Ka-core |
| BACKEND | SP500-512 | SP500-512 |
| AC-NORM | perform | perform |
| FOURIER | perform | perform |
| OBSMODE | LINE+PSSW | LINE+PSSW |
| DATATYPE | ref | sig |
| TSYS-DIF | omit | perform |
| CALSCALE | omit | perform |
Note that the order of the keywords TSYS-DIF and CALSCALE is significant. Performed in the order shown here, the average baseline value will be zero. If the order is reversed, the average baseline value will be the Tsys difference between the "sig" and "ref" scans.
| Keyword | Ref Scan | Switched Scan |
|---|---|---|
| FRONTEND | K | K |
| BACKEND | SpectraData | SpectraData |
| OBSMODE | LINE+BMSW | LINE+BMSW |
| DATATYPE | ref | switched |
| TSYS-RAT | omit | perform |
| CALSCALE | omit | perform |