During the first week in January of 2010, my son made three additional recordings of VO-52 passes over his home in Silver Spring, MD. For all three he recorded the satellite beacon as before but on channel 2 of the recording. On channel 1 he recorded the audio from WWV, Fort Collins, CO, the National Institute of Standards time station. By comparing the location of the time markers from WWV to the beginning of the file, this provides timing accuracy of close to 10 msecs.
The improved timing accuracy is evident in the range comparison plots. In these plots I estimated the satellite’s range using the time and distance at closet approach, and the velocity derived from the Doppler shift. These estimates were calculated at the same time values that the PREDICT satellite tracking program reported for the pass. In this way I could directly compare the two predictions. The difference between the PREDICT and Doppler estimates is shown on each plot.
Maximum Elevation: 24 degrees 
Maximum Elevation: 41 degrees
Maximum elevation: 71 degrees
These three passes seem to show that the difference in predicted range varies directly with the pass elevation given by the PREDICT tracking software, ie high passes yield less accurate results than those down nearer the horizon.
Murmurs from the Earth... 