C* 2228: WOW!, OBS requested, data analysis

Walter Nissen (dk058@cleveland.Freenet.Edu)
Thu, 25 Sep 1997 19:01:06 -0400 (EDT)

This morning, Cosmos 2228 put on a display for the Ages, the epitome of 
irregular irregularity.  Fortunately, I was better prepared than ever 
before to record her gyrations, but what a wonderful thing it would have 
been if a high-speed photometer would have been aimed her way to provide a 
detailed light curve. 
 
The pass reached a more than adequate altitude of 88 degrees, in a 
reasonably dark sky.  The general magnitude was 5th or dimmer, but there 
were a few very bright Flashes of mag 0 or 1(?), real stunners.  There was 
an amazing variety of flashing behavior, very brief flashes, longer 
flashes, faint flashes, bright flashes, a period when she was varying so 
rapidly and so little that she became difficult to measure.  After 
culmination the brightness faded rapidly and she became increasing 
difficult to see, finally disappearing altogether at mag 6 or 7(?). 
 
After a careful analysis, there definitely seem to be different periods 
imbedded within the so-called "regular" behavior. 
 
Can it be that different parts of 2228 are rotating with respect to one 
another?  I would be particularly interested to have the opinions of Paul 
Maley, the Russians, etc., on this question. 
 
Bart, does this data suggest anything to you about the rotation axis?  It 
isn't a simple "synodic anomaly" as you have previously discussed. 
 
I want to thank those who have provided observations, and to assure you 
that they are much appreciated and will be mentioned in further analysis. 
Further observations are requested. 
 
2228 
1 22286U 92094A   97268.06711100 +.00000261 +00000-0 +36168-4 0 01733 
2 22286 082.5241 265.3634 0028412 062.8399 297.5702 14.73817619255389 
 
Data analysis: 
 
1000 data   0     f 
1010 data   3.03     c 
1020 data   8.08   a 
1030 data   9.35 
1040 data  10.48  f 
1050 data  13.24     c 
1060 data  18.45   a 
1070 data  20.72  F 
1080 data  28.81   a 
1090 data  31.16  F 
1100 data  34.32     c 
1110 data  39.21   a 
1120 data  40.23 
1130 data  41.59  f 
1140 data  49.80   a 
1150 data  52.13  f 
1160 data  60.23   a 
1170 data  62.47  F 
1180 data  65.32     c 
1190 data  70.83   a 
1200 data  72.88  F 
1210 data  81.37   a 
1220 data  83.37  F 
1230 data  91.88   a 
1240 data  93.75  F 
1250 data 104.21 
1260 data 112.84   a 
1270 data 114.87  F 
1280 data 133.59    b 
1290 data 143.29 
1300 data 144.13    b 
1310 data 154.35    b 
1320 data 164.49    b 
1330 data 165.30 
1340 data 169.57 
1350 data 170.14      d 
1360 data 174.78    b 
1370 data 180.63      d 
1380 data 185.03    b 
1390 data 195.18    b 
1400 data 200.42      d 
Times are in seconds measured from the first flash observed as C* 2228 
cleared the obscuring roof to my North, 0 s == 1997-09-25 10:21:10.87s. 
Some flashes are missing, generally very faint or possibly very flat; 
especially while I was rotating my body near Culmination (h = 88 degrees, 
t = 98 s). 
The flashes were dtm during, and also surrounding, the gap between 114.87 
and 133.59 s. 
The codes given identify arbitrary strings of roughly equally spaced 
events.  Except that F has its usual meaning, a bright Flash; more 
particularly, the Fs were identified by double clicks, the first for the 
timing, the second (not shown above, coming about .3 seconds later) to 
mark the event as a "main" or "dominant" flash.  The fs are other timings 
which appear consistent with the spacing of the Fs. 
 
I regressed the identifed strings with these results: 
string  intercept       period 
Ff        -.0746        10.4322 
c         2.9840        10.3974 
a         7.9018        10.4888 
b       133.7572        10.25 
d       170.31          10.065 
 
I don't know if the results of regression are suitable for PPAS, but if 
so, which of these seem reasonable?: 
 
PPAS(beginning): 
Walter I. Nissen, Jr., CDP, dk058@cleveland.freenet.edu, 55 Barrett RD #808, 
Berea, OH 44017-1657, USA, 440-243-4980, -81d 51.823', 41d 22.413', 256m, 7x35 
92- 94 A 97-09-25 10:21:10.9 WN  114.8  .2  11  10.432 irreg, FF, C* 2228  F 
92- 94 A 97-09-25 10:21:13.9 WN   62.4  .2   6  10.397 irreg, C* 2228      a 
92- 94 A 97-09-25 10:21:19.0 WN  104.9  .2  10  10.489 irreg, C* 2228      c 
92- 94 A 97-09-25 10:23:24.5 WN   61.5  .2   6  10.250 irreg, C* 2228      b 
92- 94 A 97-09-25 10:24: 1.0 WN   20.1  .4   2  10.065 irreg, C* 2228      d 
These are all results of linear regression.  The intervals have all been 
obtained by multiplying the derived period times the assumed number of 
periods.  The accuracy is arbitrary and is probably stated as worse than 
actual. 
 
Although these strings were recorded chronologically, and hence, all mixed 
up with each other, their internal consistency is far greater than would 
be expected if they merely represented coincidence arising from inaccurate 
measurement about a "real" period of, say, 10.4 s. 
 
Consider this fragment of a regression program (a real regression program 
would minimize the sum of the squares of ct) containing a subset of the Ff 
string: 
 
200 defint i-n 
220 defsng a-h,o-z 
300 let start = 20.72 
400 for i = 0 to 8 
410  read t 
420p let c = start + i * 10.436 
430  let ct = c - t 
460  print using "####.## ####.## ###.##"; t, c, ct 
500 next i 
700 end 
1220 data  20.72 
1230 data  31.16 
1240 data  41.59 
1250 data  52.13 
1260 data  62.47 
1270 data  72.88 
1280 data  83.37 
1290 data  93.75 
1300 data 104.21 
 
and its output: 
 
 20.72   20.72   0.00 
 31.16   31.16  -0.00 
 41.59   41.59   0.00 
 52.13   52.03  -0.10 
 62.47   62.46  -0.01 
 72.88   72.90   0.02 
 83.37   83.34  -0.03 
 93.75   93.77   0.02 
104.21  104.21  -0.00 
 
8 of these values lie astonishing close to a straight line.  It may well 
be that I was late recording one of the values; and also that the other 
3 values (of Ff) also are a couple of tenths (.2) of a second later; it 
may even be that the true period was closer to 10.4322; but even with all 
that, 8 of the 12 values I focused the greatest part of my attention on 
are very consistent internally. 
 
(In this regard, you may recall my earlier claim that I can synchronize my 
stopwatch to within about .02 s of WWV, or especially, 1-900-410-TIME). 
 
Can any of you, especially such experienced regressers as Mike McCants, 
Rainer Kracht, Bj"orn Gimle, and someone else whose identity I can't 
recall, make more sense of this data? 
 
I also want to set the record straight that Russell Eberst observed 
flashing behavior of C* 2228 before I reported my observations here. 
 
More later. 
 
Cheers. 
 
Walter Nissen                   dk058@cleveland.freenet.edu 
-81.8637, 41.3735, 256m elevation 
 
--- 
 
Astronomy is looking up!