Voyager analysis

From Tauwiki

Contents 1 Analysis of Voyager data 1.1 Median filtering of data 1.2 Data Format 1.3 Data Modes 2 Steps in running the Voyager Pipeline 2.1 raw_read 2.1.1 clean 3 Results 3.1 Diffuse spectra

[edit] Analysis of Voyager data

[edit] Median filtering of data

The first step is to median filter the data with a 3 point filter which should just remove spikes and then rewrite the data in the same format.

[edit] Data Format

The data are in 768 byte records with specific elements being integer or float. However, the format is machine specific so I have to flip the byte order on my Mac. In IDL, this corresponds to just plain byteswap for integers and byteswap,/lswap for float: the four bytes within a longword are changed from (B0, B1, B2, B3), to (B3, B2, B1, B0).

IDL procedure for byteswapping float

byteorder,variable (for integer)

byteorder,variable,/lswap (for float)

[edit] Data Modes

In the 768 byte record, the integration time is integers(3) which is bytes 6 and 7 but note that this has to be reversed. Depending on the mode, the integration time is:

Integration times for Voyager data

Data Mode	Integration Time
0	time = 1
1	.32
2	3.84
3	12
4	48
5	48
6	192
7	576
8	720
9	240
10	3.84

Column numbers in data

Data Type	Column Number	Variable
Integer	1	S/C number
Integer	4	Year
Integer	5	Day
Integer	6	Hours
Integer	7	Minutes
Integer	8	Seconds
Float	60 - 185	Data values
Integer	1	S/C number

The data values have to be multiplied by a scale factor which depends on the integration time: scale_factor = flts(22,rec) * time

[edit] Steps in running the Voyager Pipeline

[edit] raw_read

Reads disk file into an array.

[edit] clean

Called by raw_read.pro

1. Select all the records with a common integration time.
2. Calculate the total number of counts in the record.
3. Run a 3 point median filter on the totals.
4. Reject any record which is different by more than 5 * sqrt(number of counts) from the total.
5. Write the valid times out to a file.

Uses clean to remove bad spectra using a median filter. The total flux is calculated for each record and a 3 point median filter run. If the total is greater than 5 sigma times the sum, we reject it.

The output is a 256 element array with the following indices:

Index
6 - 131	Spectrum
138	S/C ID
139	Integration time
143	Mission launch time
144:148	Date
149	FDS count
150	Spectrum number
151:152	Sun RA/Dec
155	RA
156	Dec
173 - 175	FPN CAl Scat Sp Num
176	# of 1-2-1 smoothings
177:178	Pitch and Yaw

[edit] Results

[edit] Diffuse spectra

4044 of the 6406 targets are diffuse observations, if we set the limit on confidence at 20%. Checking the data shows that the variation with dw is flat implying that no star is present.