> At 23:55 27/04/02, Kevin Fetter wrote: > You must have a stop watch that measures time in 1/1000 sec. > My watch only measures at 1/100 sec. As for my timing accuracy, > its 40/100 second or 0.4 seconds. That's 0.2 seconds to star > the watch and 0.2 seconds to stop the watch. How good can you > measure time to. > Tony answered: ... > on a known minute. Thus the consistency of the reaction time is important, > for the period measurement, not its absolute value. I entered the cycle > number and flash time into an Excel spreadsheet, and fitted a staight > line for the cycle number vs Time. The slope of this line is the > best estimate of the period. Whats more EXCEL gives me a formal standard > deviation of the slope ( and the period), and residual plots. > The SD for most days was .003 or .004 seconds. The difference of the observed > times and the fitted times was always equal to or better than 0.2 seconds, > with quite a few .05 secs or better. > This technique is of course quite time consuming and only applicable > to slow moving satellites like 88-18A. I use the program SYNODIC that I wrote, which plots Tony's "difference of the observed times and the fitted times". Then it's easy to spot the bad timings, and even more important, if there is a phase shift, ie a different surface starts showing its flashes during a long observed arc, or during an intentional gap to avoid filling your stopwatch memories. This can result in two parallell lines, with a shift of a half period, or some other more or less explainable fraction, and I can use "all" observations to find the best common period, which is NOT necessarily close to (last-first) flash time divided by an integer! There is "always" a synodic effect, ie the observed period (P) is longer or shorter than the physical rotation period (Pp), because the Sun-Sat-Obs angle is changing during the observation. In the case of geosynchronous satellites, observations on consecutive days will span an integral number of physical rotations. If P/2 > SD * 86400 / P you can get a VERY accurate flash period, BUT : In most (all?) geometries I can think of, there is always a daily phase shift, and P should be close to (the lap covering one day) divided by (an integer +-0.5), but Pp is lap/integer. For an even number of days P = lap / integer. (To determine if Pp is <P or >P you need very accurate absolute timings of the same pass from two different observers, or Bart de Pontieu's DRA method from very accurate lap times over a large range of Sun-Sat-Obs angles - and some luck) -- bjorn.gimle@tietotech.se (office) -- -- b_gimle@algonet.se (home) http://www.algonet.se/~b_gimle -- -- COSPAR 5919, MALMA, 59.2576 N, 18.6172 E, 23 m -- -- COSPAR 5918, HAMMARBY, 59.2985 N, 18.1045 E, 44 m -- ----------------------------------------------------------------- Unsubscribe from SeeSat-L by sending a message with 'unsubscribe' in the SUBJECT to SeeSat-L-request@lists.satellite.eu.org http://www.satellite.eu.org/seesat/seesatindex.html
This archive was generated by hypermail 2b29 : Sun Apr 28 2002 - 09:02:36 EDT