Superbird problem gets mudd

ROB MATSON (ROBERT.D.MATSON@cpmx.saic.com)
11 Feb 1997 14:23:12 -0800

Hi Ron,

You asked, "Rob, can you determine the flash time for the main
surface?"  I assume you mean flashes from the "business side" -- yes, my
program was computing both.  It just wasn't differentiating between them.  Now
it plots blue tracks for one side, and red ones for the other.  Just one
problem -- there ARE no bright-side reflections right now (during nighttime)
for North America.  I checked forward to the middle of April, and still
nothing.  However, this all becomes moot because of my failure to address one
simple question:

Why is Rainer only seeing flashes at the rotation rate (1 every 23 and a half
seconds), when we were seeing them at the half-period?

I think I have an answer which explains this.  Unfortunately, it completely
changes the rotation axis calculation.

First, some assumptions.  (1) There are two solar arrays.  (2) The
two arrays are not necessarily parallel to one another, nor to the satellite's
axis of rotation.  (3) The two arrays are essentially identical in terms of
their dimensions and reflective properties. (4) The satellite's rotation
period is approximately 23 1/2 seconds.

Fact #1:  At least for western U.S. observers last Fall, 2 flashes were
visible per satellite rotation.  There are only two ways this could happen. 
(1)  At least one of the arrays is very nearly parallel to the axis of
rotation, or (2) each array is tilted by the same angle relative to the
rotation axis, but in opposite directions (one clockwise, the other
counterclockwise).  Solution #2 appears to be ruled out, because of ...

Fact #2:  Rainer Kracht did not observe 2 flashes per rotation on February 1,
1997.

For similar reasons, solution #1 can be refined to read, "One of the arrays is
very nearly parallel to the axis of rotation."  If both arrays were close to
parallel to the axis of rotation, Rainer would have seen flashes at the
half-period.

Conclusions:

1.  Bright reflections can be produced by either side of either solar panel.

2.  Western U.S. observers saw reflections from both sides of panel #1, which
is very nearly parallel to the satellite axis of rotation.

3.  Rainer Kracht observed reflections from ONE side of panel #2, which is
tilted more than panel #1 relative to the satellite rotation axis.  (It still
wouldn't have to be tilted much.)

These conclusions would help to explain the flash pattern you observed last
Fall.  If panel #1 is angled just a half-degree from parallel to the rotation
axis, the direction of the reflection would shift by 2 degrees from the first
flash to the next one 11.7 or so seconds later.  As the sun-satellite-observer
geometry changed, you would see a phasing effect.  The peak of the flashes
from one side of the array would not correspond to the peak from the other
side.  At the beginning the odd flashes would be brighter; by the end the even
ones would be.  But remember that in order for both odd and even flashes to be
visible at roughly the same time, the responsible array must be very close to
parallel to the rotation axis.  As you increase the angle between the plane of
the array and the rotation axis, the reflection tracks on the earth's surface
diverge, eventually reaching a point where one of the reflections (front-side
or back-side) never sweeps over your location.  In this case, you would only
see one flash per rotation.  This is what is going on with panel #2 at
Rainer's location.

Note that if "Rainer's solar panel" isn't tilted TOO much, he might still be
in position to observe the reflections from panel #1, or possibly even the
backside of his panel.  To check this he could try looking at Superbird over a
longer window.

Interestingly enough, I think YOU might have observed some dim flashes from
panel #2 last Fall.  Consider your entry:

[snip]

started observing by 850 PM MDT (250 UT 29 sep 96).

At about 9:13:30 PM (3:13:30 UT), I observed faint flashes
of about 9th magnitude.  I was unable to get a flash period
due to moonlight glare and the fact that the satellite was
only visible about 2 minutes.  Not very exciting visually.

Then about 9:24 PM MDT (3:24 UT), the flashes resumed.  The 
first period was 23.4 seconds.  Note that this was going from 
BRIGHT flash to Faint flash back to BRIGHT flash.  The same
period was observed on one additional attempt.

It exhibited a definite ramp up and down in brightness, eventually
becoming visible in binoculars.  Estimated max magnitude was 3-4.
It should have been visible to naked eye without glare from
nearly full moon.  

Last flash seen was around 9:37:15 PM MDT (3:37:15 UT 29 sep).

It may have been visible for a short period around 9:48 PM but that
was not a definite sighting.

[end snip]

My analysis of this observation is that at 3:13:30 UT you were actually
observing grazing flashes from one side of panel #2.  Then from 3:24 to
3:37:15 UT you observed flashes from both sides of panel #1.  Finally, at 3:48
you observed grazing flashes from the OTHER side of panel #2.  I don't think
it's an accident that the dim flashes you observed before and after the main
flash period were symmetrically 17 and a quarter minutes on either side of the
middle of the main flash window.  Such symmetry is required if the panels are
positioned as I've described.

Anyway, I've got some work ahead of me.  I can't use Rainer's observation
coupled with observations from last fall to come up with a spin axis -- his
observations are of a different panel than ours.  I should have enough
observations from last fall, however, to come up with a crude rotation axis.