Live From Mars was active July 1996-December 1997.

My Journals


What Navigators Need to Know

by Pieter Kallemeyn

June 19, 1997

Right now we are preparing for our final maneuver, Trajectory Correction Maneuver #4, which will occur June 25. We must to be able to get the best current orbit model possible using the tracking data up to the last minute. What I'm doing right now is preparing an input file for the orbit determination program. This file describes the variation in the spin rate of the spacecraft since February. One of the things we have to be able to know and model in the program is is how the spacecraft is spinning because that affects how the tracking data looks.

The spacecraft controls how fast it spins. It receives ongoing instructions that say it is supposed to be within so many rpms. On June 8 at 23.46.46 UTC, Pathfinder was spinning at 1.957 rotations per minute. That's the value that the spacecraft thinks it is spinning at. It travels twice as fast as a second hand on a watch. This is on the list of things that navigators need to know!

Pathfinder is now within 100 km of its landing site. We did an orbit determination solution last week for the purpose of doing a preliminary design of TCM #4. The center of that solution, when mapped to the surface, is within our requirement ellipse, or footprint. The footprint measures 200 km from one edge to the other and 100 km from edge to edge in the other direction. What we want to do is put Pathfinder in the center of the ellipse as best as we know it. The chances of doing this are pretty good. Imagine there are two football-shaped ellipses: one is our requirement of what we want to be inside and the other is our current orbit knowledge. These two footballs are offset from each other by about 60 km off target. We want to be within 100 km in the worst case. We want to move those footballs so they're one on top of the other. This will be done with a course correction.

TCM #4 be the smallest maneuver we've done on Pathfinder--.028 meters per second (the velocity change). Once Pathfinder receives the command it takes only a few seconds for this change to take affect. We'll develop a sequence that will be clocked off at a given time and sent to the spacecraft hours before the event. Then as Pathfinder is racing toward Mars it will activate the command at the appropriate time.

Maneuvers are performed in one of two modes: turn and burn, or vector. In the turn and burn mode the spacecraft is turned in the direction we want the velocity change to occur. A pair of thrusters is fired for a predetermined period, which results in a velocity change. The vector mode is a maneuver that is performed in two parts: axial, which is along the direction of the spin axis, and lateral, which is roughly 90 degrees from that. The interesting thing about a lateral maneuver is since the spacecraft is spinning and the thrusters are mounted on the spacecraft, the thrusters have to be pulsed at given times. An axial maneuver is a continuous burn. If the lateral thrusters were continuously fired, they'd cancel each other out after one revolution. We wait until the thrusters are at just the right orientation in the spin and then we fire them for a few seconds and turn them off. The spacecraft spins another half a rev until the other thruster cluster comes into the same orientation and then the thrusters are fired. Basically, thrusters are fired every half rev.

The Deep Space Network (DSN) is measuring Pathfinder's doppler and range on a continuous basis. We receive our tracking data from the DSN in batches--one a day. The Radiometer Data Conditioning group looks at the data first, applying any adjustments that they think ought to be made. They do some editing, some error checking and data accountability. When they believe the tracking data are good enough, they package it into a particular binary format and send us an email message that the data are ready to be picked up. We capture it via an ftp transfer and then add it to the master tracking data file, which contains data we've received since launch.

The output arrives around 10 a.m. every day, afterwhich it takes us about 30 minutes to do one orbit determination run. This is the first step in doing navigation -- using the tracking data with a model on the ground of what the trajectory is like. We have a program that uses that trajectory model and the timing of the data points on the tracking file. From this we can come up with a predicted value for each one of those points. This is what we call the predicted tracking measurement.

Point for point, we compare the predicted measurement with the actual measurement taken at the DSN. If everything was done correctly, the error ought to be zero. But there's always a little bit of error due to a number of reasons: our trajectory isn't perfect and there are errors outside of navigation's control that we have to take into account. For example, the atmosphere distorts the signal so we have to calibrate for that. We also have to calibrate for Earth's rotation. We're also modeling the spin rate of the spacecraft. So we're getting data from a bunch of places, incorporating it into the model, and looking at the residuals (the observed value minus the computed value). It takes the difference of those two values and determines what the true updated trajectory ought to look like. The end of this is a trajectory file. From the trajectory file we know where we're going. First you have to figure out where you are and where you want to go and the TCM takes you from one to the other.