HISCALE Level 0 processing is done by the procedure called ARCGEN (for Archive Generator), which uses the EDR, SEDR, and the auxiliary file IDF.DAT (for Instrument Description file). For those familiar with the JPL SPICE system, the IDF.DAT file plays the role of the I file in SPICE. ARCGEN and the command procedures that control its execution during production are designed to process HISCALE EDR for 24 hour periods synchronized on SCET days. The output file labeling scheme uses the day number in the file names. ARCGEN writes two different file types populated with .RAT records (standing for count rates) and .PHA (standing for pulse height analysis). The .PHA files have two record types, one with the digitized pulse heights and the other type with the MFSA outputs. MFSA stands for M and F spectrum accumulator, and these records report the readouts of HISCALE on-board accumulations of digitized pulse heights from the M, M', F, and F' detectors into 32 energy channels for each angular sector. The MFSA repeat cycle is 1024 seconds long.

There are several important programming strategies that were employed to simplify the HISCALE data at the ARCGEN step:

1. Each ARCGEN output record has an identically defined header block, regardless of record type or length. The record types are self-labeling within this universally defined record header.
2. The .RAT records are defined so that they can have the same structure for all time resolutions from the highest to lowest. Because the original file media for HISCALE outputs were to be 1/2" large reel 6250 bpi magnetic tapes, the .RAT records were defined to include the sectors rates for 10 spin groups. A spin group is two spins when Ulysses operates at the highest bit rate (1024 bps) and is 4, 8, or 16 spins as the telemetry rate is divided by successive factors of two. The principal reason that 10 spin groups were lumped together was to yield large records to minimize the tape overhead of inter record gaps and data overhead in the headers.
3. The .PHA and .MFSA records also have count rates included. These rates are time-synchronous with the PHA and MFSA cycles so that the pulse height analysis results can be reliably converted to absolute fluxes.
4. Because detector pitch angles with respect to the magnetic field are central to analysis and interpretation of charged particle observations, space is reserved in the HISCALE record headers for this information. Because of operational constraints on the availability of the magnetic field vectors (kindly furnished by the HED team), ARCGEN usually writes its output records with the pitch angle information vacant because of the unavailability of the magnetic data. Later, when magnetic field files are available, the procedure MAGMERGE computes this information and places it in the correct spin group of the archive records.
5. When .RAT data are averaged (by LANAVG) to longer intervals the averaged rates are all placed in the space reserved for spin group #1 and placed in files with extension .AVG. These .AVG records are created as an expedient intermediate step and are not intended for final archiving. A median filtering procedure is also available for the creation of averages of HISCALE data, provided that the averaging interval is long enough to include at least 10 repetitions of each data item. This median filtering procedure, called QCA (for Quality Controlled Averaging) writes records to a file labeled with extension .QCA. Because of the serious effects that telemetry noise can have on long HISCALE averages, it is strongly recommended that only QCA-based long averages be used for intervals of hourly or longer time resolution. The .AVG or .QCA records can be condensed by a 10:1 factor, if necessary, by a procedure called COMPRESS which places average rates in all 10 spin group slots of the record.

Next: Section 2.3.1 Programming Summary of Records and Variables