Utterly cool.

We are seeing a sawtooth in the residuals @ ~6 minute intervals - is this due to attitude control, solution harmonics or what?

During the TCM-2 data arc, a couple passes of Doppler
residuals such as that shown in Figure 5 exhibited a
peculiar sawtooth pattem. Because it appeared that we
were having problems fitting the data without the gasleak
acceleration estimation, we became concerned that
this may have been more evidence of serious problems
in the SIC modeling, the SDST, DSN hardware or the
ODP. Several DSN and NAG experts helped analyze
these unusual patterns, but no definitive explanation
was found. The DSN tracking procedures for Odyssey
had been to follow the S/C’s downlink frequency within
a fairly tight bandwidth by periodically ramping the
uplink signal. It was believed that this ramping of the
signal could have contributed to this problem,
especially if the values of the ramp rates were being
truncated, but no evidence of this was found. In the case
that this data was incorrect, our procedures were to
remove the data, however, it was determined that the
data had little effect on the OD solutions, especially due
to the signatures’ high fi-equency nature.

Interesting item from 2001 Mars Odyssey orbit determination during interplanetary cruise:



The resolution of the graph makes it hard to read, but it look like the amplitude of the sawtooth is a bit larger than with MRO (~5x), and the period about 15 minutes.