HOW TO EVALUATE ARGOS LOCATIONS
2. ESTIMATING THE PTT TIMING

The timing of PTT transmissions (period ON of the duty cycle) is programmed by manufacturer. The advertised expected deviation of +-5% from actual time is often exceeded and in some cases it can reach as much as +15%. If the deviation is constant, the timing will drift and the real transmission time will differ from the programmed duty cycle. Even very small difference can produce a considerable shift of the beginning (or end) of the transmission in aggregate and after few weeks. Determining the timing of transmission is very important before undertaking a PTT recovery trip. The Argos DIAG files (and specially PRV/A in DS format) provide data about the PTTs. From these data actual timing of transmission can be directly calculated. When this type of data is not available, the exact timing can be estimated by more complicated method described here.

My experience is solely with the PTT-100 (Microwave Telemetry, Columbia, MD, USA).  It provides in every fourth transmission the data on the number of hours passed since the first transmission of the current ON period (2nd number in the row). By examining the timing of the most recent cycles, the real time elapsed between the start of ON periods can be determined and accurate long term predictions can be made about when transmissions will occur.

PTTs made by other manufacturers may provide different data or the location of data on timing may be elsewhere in the transmitted information. For PTTs that do not provide these data, one must compare the timing of a larger number of sequential ON periods and determine the mean time between cycles. The simplest method is to use a graphical means.

Procedure:

1) Update regulary your database created for each PTT in a spreadsheet file. Keep individual values from each Argos record in one line. Record also LC Z.

2) Create one column for "date&time numeric value". From these figures subtract the number so that you get a meaningful value (e.g. day from the 1st of January or June etc.). In this "date&time" column below the Argos data add formulas that are calculating (for all studied periods) each beginning and end of the period ON from estimated parameters: ON, OFF, koef, shift and start). Each cell is calculated from the previous one in this way:
start+shift/24..(cell above)+koef*ON/24..(cell above)+koef*OFF/24..(cell above)+koef*ON/24 etc.

3) Test different values of first four parameters in order to find a combination that produces the theoretical distribution of ON in the past. ALL ARGOS REPORTED RECORDS MUST GRAPHICALLY FIT IN. One can also find the potential variation of these figures.

4) This formula series can be then projected to the future and the expected timing ON deducted from the graph or calculated by converting "date&time" values back to the date and time.

Following figures show the Quattro calculations ("date&time" in column D) and one of produced graphs (red - observed, blue - predicted period ON).

Next figure shows the timing of 2.5 years old solar PTT on the white stork recently flying in Sudan. Excel calculations ("date&time" in column E) are projected in the more complex graph (blue - observed, violet - predicted period ON).
In Excel one can also change parameters dragging them in the additional graph by mouse.

Back