The caution flag indicates PHA overflow and/or gain switching in progress. The LET E buffer indicator (bits 9 and 10) has the following states:

4-9	4-10		4-9	4-10
  0	   0	DUBL	  1	   1	WDSTP
  0	   1	TRPL	  1	   0	WDPEN

LB Triple
    &     LET B    
cmd state Double DUBL TRPL WDPEN WDPEN wL1 WDSTP
   F48      B48   4C2  5C6  BCA     FCA     9CE
   F08      B68
   F68

Rate Scalers

Eight rate accumulators (numbered A through H) are used. In two of the accumulators (F and H), the input signals are subcommutated sixteen times. (The same subcom sequence controls status readout.) Table 7 shows rate readout as a function of accumulator letter and subcom state.

The first five rates are rates of "events" (an event is a triggering of the HIC by an energetic particle or the PHA and TAG data generated by such a triggering) as defined below, the remainder are singles rates from the detectors (LE1 - LE5, LB1 - LB4) or slant discriminators (SLB in LET B, SB in LET E). LBTRP is the rate of LB1.LB2.LB3.LB4* coincidences. The requirements for the various types of events are discussed in the event section.

The spin of the spacecraft at 3 rpm nominal will allow calculation of anisotropies from the rate scalers after the fact. Note that the eight rate scalers are read every 2 seconds; ten times per spin. It will be necessary to understand the buffering delay to get the phase right (or perhaps vice-versa).

Rate Compression

For example, zero counts leaves the accumulator at all-one's. Zero shifts are required to up-shift a 1 into position 24. All 8 MSB's are 1's. The transmitted number has 5 leading 0 bits and 7 trailing 1 bits (octal 0177). One count leaves the accumulator at all-zero's. Thirty-one up-shifts are done searching for a 1 but none is found. This case is the exception to the MSB = 1 rule. The transmitted number has 5 leading 1's and 7 trailing 0's (octal 7600). Two counts leave the accumulator at 1. Twenty-three shifts are required. The transmitted number is octal 5600. For 7200 counts (the internal calibrator) the accumulator state will be binary 1110000011111 ; eleven shifts will be required and the italicized bits will form the mantissa. The result is 010111100000 in binary, 5E0 in hex, 2740 in octal.

To decompress rates, the 7-bit mantissa is picked up and put in position 17 through 23 of a computer word with at least 24 bits (numbered 1 to 24). Bit number 24 is set to 1. The word is then down-shifted the number of times indicated by the 5-bit exponent. The word is then incremented to compensate for the all-one's reset state of the accumulator. In pseudo-FORTRAN notation,

rate = (128+mantissa)*(2**16)
rate = rate/2**exponent
rate = rate + 1

If the resulting rate is greater than 256, a better estimate for many data processing applications is obtained by using 128.5 ± 0.5 instead of 128 in the formula above. For example, hex 5E0 decodes as 7185±16.

Table 8 specifies examples for the smaller numbers likely to be encounted in flight.

Command/Status Data

A Galileo command consists of two 8-bit bytes sent to the GAB as documented in the JPL IRD 512335. These 16 bits are decoded by the GAB. The first bit is spare, the second indicates "cal start" (BC28CAL), the third indicates "high voltage on" (BC28HVON), and the fourth indicates that the following 12 are a serial command. These 12 bits are sent on to the CRS electronic as was done on Voyager. As before, these 12 bits are interpreted as a four-bit column number or register address and eight bits of data for that column. Table 10 shows the interpretation of each bit. Recall that commandable functions are also shown in the rate definitions in Table 6 by brackets.

The JPL nomenclature indicated in parentheses above and in Table 9 must be used when speaking with them. The prefix BC is bus command; GEM/HIC is experiment number 28. When printing status data JPL/MTS uses hex. When specifying commands to be sent, they usually use binary for the eight bits of "data".

Analog Data

Analog data is readout by the spacecraft via a multiplexed line. In the telemetry we receive an eight-bit "data number" (dn). The multiplexor is stepped by the adapter board once each 7 minor frames, but this sequence is not synched to the spacecraft major frame (rim) structure. Note that two step signals are required to step the multiplexer; it may well take 14 minor frames to switch states. The reset signal cannot be sent. Synchronism must be achieved by inspecting the data. The data consists of power supply voltages and temperatures as listed in Table 11.

The temperature calibrations for the multiplexed analog data are roughly given by

°C = A₀ + A₁*(dn) + A₂*(dn)² + A₃*(dn)³

where

A₀ ~ 67.
A₁ ~ -1.
A₂ ~ 5. * ^-3
A₃ ~ -11. * 10^-6

There is also a separate temperature transducer on the telescope housing which is not multiplexed and which is readout by the spacecraft. The JPL acronym is TTEMP. Its calibration is given by

A₀ ~ -102.45
A₁ ~ +0.674666
A₂ ~ 90.524 * 10^-6
A₃ ~ 0.0

Packets

Do not confuse HIC instrument packets, described here, with CDS 02 telemetry packets, described in Galileo Project Doc. 625-205: 3-280, Phase 2 (available from JPL).
The GAB issues a fixed sequence of word gates to CRS to create a particular mixture of rates, status, and PHA's which are then sent on to the Galileo spacecraft by GAB or HIC. A HIC instrument packet consists of three minor frames. Each minor frame consists of eight 12-bit words. The third minor frame of the three in a packet contains subcommutated rate and status data, with a subcom depth of 16. Thus an instrument cycle consists of 16 packets, numbered 0 through 15 by the four MSB's in the status word. Figure 6 illustrates the packet format.

Error Protection Encoding

The CRC word consists of 8 bits of actual CRC (the CRC character) and four trailing bits of zero's. The CRC character is generated in a 8-bit shift register which applies the encoding polynomial x**8 + x**7 + x**6 + 1. On the ground, where the error rate is negligible, the CRC should be checked but no data correction is necessary.

The encoding circuit is illustrated in Figure 7. The following "subroutine" will perform the same encoding.

initialize 84-element array x() to zero
loop for n = 1 to 84
	input =  nth bit of 84-bit data stream
	x(0)  =  x(8) XOR input
	x(8)  =  x(7) XOR x(0)
	x(7)  =  x(6) XOR x(0)
	x(6)  =  x(5)
	x(5)  =  x(4)
	x(4)  =  x(3)
	x(3)  =  x(2)
	x(2)  =  x(1)
	x(1)  =  x(0)
end of loop
CRC = 128*x(8) + 64*x(7) + 32*x(6) + 16*x(5) + 8*x(4)
      + 4*x(3) + 2*x(2) + x(1)

Synchronism

There are some values which the 12-bit log compressed rate words never have (including F00 and 020) and others which are impossible in the Galileo application because the rapid readout (every 3 minor frames or every 2 seconds) means that we cannot accumulate more than about 100,000 counts in the accumulator. Thus if the first four bits of an rs word have value 1, 2, 3, 12, 13, or 14 the rs word is a status word and the index SCN must be 2 (possible values are 0, 1, 2) and mux state N is the value found. The following "subroutine" will determine all the HIC pointers after less than 30 minor frames. N is the value of the first four bits, and MUXN is the mux state,