CHAMP data processing and Precise Orbit Determination

Questions and Answers

The list below is intended to be of assistance in solving technical problems with CHAMP data processing. It has the nature of a Frequently Asked Questions summary, capturing the essence of a number of topics that were discussed during the CHAMP meeting at GFZ Potsdam on November 9th, 2001.  Any further relevant details that come up in the future may be added to this list in due time. The questions have been grouped more or less by subject, but in many cases there are overlaps between instrument handling, data processing and orbit computation. For this reason there is no explicit division in sections.

No. The receiver data is downloaded from on-board storage whenever possible, typically during three or four contact moments per day with the ground station. The receiver data is then recombined into daily files.
  The receiver has at present a limit of 8 simultaneous satellites, essentially because a larger number results in unnecessary processing complexity. This is a software limit, not a hardware limit.
  Nothing is wrong, other than an unintended duplication of some data records. You can either delete these records or process them the normal way.
  The GRACE data centre is under development and will be similar to the CHAMP ISDC structure.
  In short :
The pre-processed accelerometer data no longer has data spikes. The raw data contains spikes, some of which can be associated with electronic events on-board, others are still under investigation. This does not affect the L2 pre-processed data product anymore.

NOTE : a new release of all available accelerometer data up to day 251/2001 will be made available through ISDC, after some improvements in the pre-processing.
 
 

 Yes, for two reasons. First, the radial sensitivity is less due to the instrument design. Second, the internal control loop of the instrument has been identified to have an electronic problem, which may still affect output data after pre-processing.
  No. In principle, it is impossible to know the linear components of the attitude manoeuvre (due to thruster asymmetry) because the attitude pulses are too short to be properly absorbed in the 10-second accelerometer normal points. Any spikes that may or may not be thruster-related are removed in the pre-processing. Short gaps in the data that may result from this are bridged via interpolation during the pre-processing stage.
  The firing times and duration of the manoeuvres are given in the accelerometer L2 product, as well as the orientation of the rotation axis. You can either try to calibrate all manoeuvres independently, or you can make assumptions on systematic misfiring properties per thruster pair, and then estimate just a limited number of manoeuvre calibration parameters for your entire solution arc. The latter method is at present used by GFZ.
  These '999' indicate missing telemetry information. There may or may not have been an attitude manoeuvre at that point. In most cases, there is none.
  This is hard to say. The values for these parameters should be purely instrument-related but in practice many other effects seem to get absorbed in the accelerometer calibration parameters, depending on the processing method you use. In particular, if you have a poor a priori gravity model your dynamic orbit solution will never be very good, and the accelerometer parameters will absorb unwanted mismodelling effects.
  The altimeter mode has been a late addition, and is considered as experimental rather than as a normal CHAMP data product. The associated data is not distributed through ISDC in the same way as the other flight receiver data.
  A single event occurred associated with the along-track measurement direction, and the instrument was switched to the redundant ICU board which has been used ever since.
  The on-board data is time-tagged by means of the GPS clock. Any small differences with respect to true GPStime or UTC can have no significant influence on the accelerometer data output. The timing errors will typically be below microseconds level, during which the displacement of the satellites will be in the order of a millimeter. The acceleration of the satellite does not change significantly at this resolution of space and time.
  Under certain rare circumstances, the attitude sensors can get moon-blinded and produce false quaternions. These are used internally by the on-board attitude control system, which will try to flip the satellite. The spacecraft will recuperate by itself as soon as its sensors turn away from the moon and get back to nominal operations.
  Batch-request processing will be made available by GFZ to a wider range of users than until now - please keep an eye on the GFZ web page for announcements, or contact the ISDC support people directly with your request.
  GFZ have promised to be reasonable on this point. If you really need a larger volume of data, you can contact the ISDC support and request temporary or permanent modifications to your account limit, but obviously GFZ need to balance the number of requests of this kind against the available resources.
  The bias and scale factor parameter values are very stable, and for typical CHAMP arc lengths of 1 day or so they can normally be considered as a constant.