[Isocops] IBEX Star Tracker Outage %valid

[Reno, Michelle]

Nevermind, I just got the thumbs up from the SCB team- the new %valid threshold for using the Quaternions+Static Z Rate is changing to 50%. We should be able to incorporate the changes into orbit 167.

________________________________

From: isocops-bounces at lists.sr.unh.edu on behalf of Reno, Michelle
Sent: Thu 5/3/2012 5:03 PM
To: isocops at lists.sr.unh.edu
Subject: [Isocops] IBEX Star Tracker Outage %valid


Hi guys,
 
Changing the Star Tracker outage threshold from 90% valid to ~50% valid will take more time than anticipated. It needs more analysis from the SC team. Stay tuned.
 
I wrote this up for our new SSMO rep today, and I figured it may be useful for anyone who wants to know more about the star tracker outage topic. Most of you probably already know this...
 
 
  The Attitude Control System (ACS) sends a spin pulse to the payload Combined Electronics Unit (CEU) once per spin, when a predefined meridian is crossed. The CEU uses the time between spin pulses to set the temporal size of the 6-degree bins the data is put into. 
 
  We need to 'de-spin' the data when there is a difference between the spin period as determined by the spin pulse and the actual spin period. This happens when there is a star tracker outage: the star tracker has a field of view fully or partially obscured by the Earth or Moon.
 
  The ACS telemetry that defines the level of obscuration is called '%valid' and is based on the percentage of the spin that contains reasonable quaternions. The ACS creates spin pulses based on these %valid readings. Above ~90% valid the spin pulses are based on the Kalman Filter output of processed quaternions only. Between ~20-90% valid the spin pulses are created based on a combination of the Kalman Filter and the 'Static Z Rate', where the Static Z Rate is more heavily weighted. Below ~20% only the Static Z Rate is used. The Static Z Rate is a constant stored in the ACS memory; it is the actual spin rate measured 15 minutes after the thrusters are disabled for inertial precession maneuvers. This constant is reset after each inertial maneuver because the spin rate changes slightly with each maneuver.
 
  The Static Z Rate is close to the actual spin rate, but there is always some error between the stored Static Z Rate and the actual spin rate. This spin period error causes the CEU to size the bins slightly incorrectly (the time allotted to each bin covers a portion of sky slightly smaller than 6 degrees). These errors themselves are small, but they are cumulative; the error from one spin adds to the error from the next. Over the course of a 4-day arc this can be seen in the data as the data slowly marching across the sky; aka bin 0 is ecliptic north normally, but after 1 day of outage it may be ecliptic north + 8 degrees, after 2 days of outage Bin 0 is ecliptic north + 16 degrees, and so on. In order to fix the maps the ISOC will use a bright signal (the moon in the Lo star sensor, the ribbon, etc) and apply a correction to the angles so the bright signal remains aligned with its actual position in the sky before and after the outage.

   When we have an outage that spans a standard apogee or perigee maneuver window, we perform a Sun Precession maneuver. Since the ACS is not getting valid quaternions, the Static Z Rate is not reset directly following the Sun Maneuver (it is reset manually after the outage is over). This introduces an additional error in the ACS reported spin period during the outage - there is a slight change to the spin rate due to the maneuver that is not captured because the ACS is still using the Static Z Rate set before the Sun Maneuver. The ISOC software can de-spin angular errors introduced by Sun Maneuvers.
 
   Another way to introduce error is by spin-down pulses during the Sun Maneuver. There is an allowable max spin rate for standard operations. After each maneuver (of any kind) the spin rate increases slightly and the spin rate slowly creeps towards this upper limit. When the spin rate goes above the max allowable spin rate, the ACS waits until the next time thrusters are enabled for a standard maneuver and then executes spin-down pulses until the spin rate is below the max threshold. This keeps us in a range of ~4.16-4.19 rpm. The effect of spin down pulses on the spin rate is much larger than the small change due to a standard maneuver.
 
   When spin down pulses occur during an inertial precession maneuver it does not affect the error in the ACS reported spin pulse because the post-spin down rate is measured and stored as the Static Z Rate. When a spin down pulse occurs during a Sun Maneuver, the error on the ACS reported spin rate is high because the Static Z Rate is not reset. The actual spin rate has changed non-negligibly since the last Static Z-Rate measurement was taken.  
 
   A washing machine arc is one which is very hard to de-spin because the difference between the ACS-provided spin pulse and the actual spin period is larger than our standard de-spinning tolerance. The data does not march across the sky in an ordered fashion but contains angular jumps. This seems to happen only when we have a spin down pulse while we are doing a Sun Maneuver. This does not happen frequently.

 
 
 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Chelle Reno
Austin Mission Consulting
(210) 478-7337 (c)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.sr.unh.edu/mailman/private/isocops/attachments/20120504/b4eb5bed/attachment.html