NATIONAL RADIO ASTRONOMY OBSERVATORY Green Bank, West Virginia 12 March 1996 MEMORANDUM To: Y. Murata and J. Ulvestad From: Larry D'Addario Subj: CALIBRATION AND FLAGGING DATA FROM THE NRAO OVLBI TRACKING STATION AT GREEN BANK Cc: Bob Brown, Jon Romney In this memo, I describe our plans for supplying flagging and calibration data from the NRAO tracking station at Green Bank. I believe that the approach we are taking should be convenient for both ISAS and JPL; indeed, I think that it will simplify the work of everybody concerned. >From study of the recent documents on pulse calibration data files [1], correlator log files [2], and flag definitions [3], it seems that the main change over the past year is that much of the information that was in a single file (the correlator input log) has been removed, with the intention of putting it into separate files. The specification for pulse calibration files [1] is very new, and was adopted with little time for comment. The removal of flag records from the correlator input log [2] has created a need for a new mechanism for transmitting that information, and the details of this are not yet specified. There is also a need to provide additional kinds of calibration data to the user, and the methods for doing this also remain unspecified. Most of the discussion about this sort of data (flagging and calibration) has concentrated on supplying it to the end user in a form that is convenient for reading into AIPS. The file designs have reflected this. It seems to us that the existing specifications [1-3] are quite suitable and acceptable from this point of view, but that they are not at all suitable for handling the other end of the data path, namely the recording and transmission of the original measurements. In two specific cases -- the tone extraction (pulse cal) data and the flagging data -- most of the information must be generated at the tracking stations. Since the NRAO tracking station must begin to implement some form of transmission of flag data immediately, and since the software for generating correlator log data is already partially complete, our station will supply flag records, tone extraction records, and receiver gain records as part of the correlator input log file. The formats for these record types are given in Appendix A of this memo. This is consistent with the latest version of the correlator input log specification [2], which permits additional record types as long as they have names distinct from the ones that it defines. The correlator input log files that we produce are available to all mission elements, which may use them for multiple purposes. The creation of end-user flag tables and calibration tables is the responsibility of mission elements much closer to the end user; this could be ISAS (for VSOP) or the correlator operators. The use of our files at the input of a correlator is unaffected as long as the additional record types are ignored. However, upon request, we will also create a reduced version of each such file that is identical to the full version except that the extra record types are omitted. I think that the recently-adopted specification for pulse calibration data [1] should be retained, but it should be interpreted as applying only to the data supplied to the end user and not to the data from tracking stations (in spite of its title). The log records specified here should be the primary means of transmitting the tracking stations' measurements. It seems to me that this makes the data transfer much simpler for ISAS, as well as for us. It will also simplify the work at JPL if they will supply DSN stations' data in the same way. This approach has these features: A) There are fewer files to deal with, and the format is very simple. B) The file syntax is much more likely to be stable. There will be no need to change the format if changes should occur in end-user requirements or in AIPS. C) No new syntax needs to be defined for the flag data. Under the plans suggested in [1-3], an attempt is made to force the tracking station files to look like end user files. Yet the files remain seriously incomplete; processing is still needed at ISAS (or somewhere) to produce complete files. In that case, the multiplicity of file types produces no advantage for anyone. We at the NRAO would like to know whether this plan is acceptable to ISAS, and we would like to have the comments of JPL. Please consider it carefully and let us know your opinions as soon as possible. REFERENCES [1] J. Ulvestad & J. Romney, "VSOP phase calibration format for SVLBI tracking stations," version 1.0, 9 Feb 96. [2] J. Romney, "Alternative log...for space VLBI observations," draft dated 12 Feb 96. [3] J. Ulvestad et al., "Flag dictionary for space radio telescope log files," ver 2.0, dated 23 Jan 96. APPENDIX A SUPPLEMENTARY RECORD TYPES IN THE CORRELATOR INPUT LOGS CREATED BY THE NRAO OVLBI EARTH STATION AT GREEN BANK 1. INTRODUCTION This document specifies how certain calibration information and error flags will be supplied by the NRAO OVLBI Earth Station to other mission elements. In summary, they will be supplied as part of the log file that will be used at the input to the correlator. Generally, all information of "log type" will be supplied this way, unless a provision has been made to supply it differently. "Log type" means information that is gathered in real time during an observation (tracking pass), and that must be recorded in chronological order with time tags. (Examples of log-type data for which other provisions exist are the time corrections file and the performance log file.) Three specific data types are addressed here: tone extraction data; square law detector data; and error flags. Fortunately, the draft specification for the Correlator Input Log [1] allows for its extensibility. In particular, any record types whose names are distinct from those in the specification may be included with the understanding that they will be ignored by the VLBA correlator. Of course, they need not be ignored by other users; this provides us with a reasonable and consistent mechanism for supplying the required data. 2. RECORD SPECIFICATIONS The notation of [1] is adopted for giving the syntax of each record type. 2.1 Tone Extractor record yydddhhmmssxx/TONE/,,, where is the channel designator, identical to the field in a "BBC" record that preceeds this "TONE" record in the log; is the baseband frequency of the tone whose complex amplitude has been measured, in MHz; is the measured amplitude, as a fraction of the total rms signal level in the channel; is the measured phase, in degrees, relative to the data clock. Notes: It is an error if the log contains no "BBC" record for the same channel since the beginning of the current tracking pass. The observing frequency of the measured tone is the sum of the fields of the BBC and TONE records, provided that the signs have been correctly given; both records must use identical units (MHz). The scaling may in practice be approximate (since an accurate value for the total rms signal may not be available), but the scaling of all values for the same channel during the same tracking pass must be identical. Since is a relative "voltage," the fractional power in the tone is ^2. As stated, the reference is the data clock; this means that a tone in phase with a harmonic of the 1 Hz reference should be reported as having zero phase. Again, this may in practice be achievable only approximately; but the relative phases of all measurements should be consistent. 2.2 Square Law Detector record yydddhhmmssxx/SQLD/,,,,,,..., where is a decimal number proportional to the total power measured in channel #i at the given time, and is a decimal number proportional to the switched power measured in the same channel at the same time. Here "total power" refers to the measured power when the calibration noise source is OFF, and "switched power" refers to the difference between the power with the noise source ON and the power with the noise source OFF. The absolute scale of these values is arbitrary, but the two numbers from one channel must have the same scale. As indicated by the word "proportional," any known zero offset must have been removed. is the power spectral density of the calibration noise source for channel #i, in Kelvins, at the given time, referred to a known reference plane in the receiving system's signal path. Channel numbers follow a fixed sequence defined for each telescope. Data may be omitted (e.g., for channels not currently in use) by leaving the corresponding field blank or empty, but all deliminters (commas) must be included. Notes: This record type is applicable only to telescopes having a switched calibration noise source that adds a known (preferably constant) amount of noise spectral density to the astronomical signal when on; and having a synchronously-switched square law detector on each signal channel immediately prior to digitization (typically at baseband). Both VSOP and Radioastron include these features, as do most ground radio telescopes used for VLBI. Corrections for known hardware errors or peculiarities in a particular telescope (e.g., temperature dependence or non-linearity of the detectors) should have been applied if possible. We give in each record even though it is normally constant in order to allow for its variation with temperature or other influences. Under these conditions, the best available estimate of the GAIN from the reference plane to the digitizer is / (in arbitrary units), and the best available estimate of the SYSTEM TEMPERATURE is */ (in Kelvins). 2.3 Flag record dddhhmmssxx/FLAG/,[,] where is an integer specifying a particular error type; is an integer from 0 to 3, indicating the present severity of the error, as follows: 0 no error - the specified error does not now exist; 1 warning - the specified error exists, but is not believed to be having a significant effect on the quality of the observation now in progress; 2 error - the specified error exists, and the present observation is probably adversely affected; 3 severe - the specified error exists, and the present observation is almost certain to be adversely affected. The optional is a string, in double quotes, of at most 64 characters, giving a brief description of the error; this is intended for human readers of the file, so no automatic procedures (software) should depend on its existence or contents. The values of and shall be taken from a list [2] maintained by VSOG/RSOG. This list shall contain a full explanation (not limited to 64 characters) of each error type. This record type shall be included in the file whenever the value of a given error CHANGES. In particular, a record is included with =0 if the condition previously had non-zero . At the beginning of a tracking pass (specified by the TSID record, see [1]), all errors are assumed to have =0. REFERENCE [1] J. Romney, "Alternative log input to the VLBA correlator for space VLBI operations." Draft specification dated 1996 February 12. [2] J. Ulvestad et al., "Flag dictionary for space radio telescope log files," ver 2.0, dated 23 Jan 96.