Noise Characterization Using Transfer Functions: Examples from Spacecraft Doppler Tracking

1
Noise Characterization Using Transfer Functions
Examples from Spacecraft Doppler Tracking

• John Armstrong for the Cassini Radio Science GW
  Group
• R. Ambrosini, J.W. Armstrong,
• B. Bertotti, L. Iess, P. Tortora, H.D. Wahlquist

2
Noise Characterization Using Transfer Functions

• The Doppler technique its signal and noise
  transfer functions
• Examples of noise identification using transfer
  functions
• Frequency timing system
• Antenna mechanical noises
• Phase noise due to propagation through irregular
  media
• Model example for LISA identifying proof mass
  noises

3
DSS25 and Cassini
4
Cassini Radio System
5
Three-Pulse GW Response
6
(No Transcript)
7
Frequency/Timing Glitch
8
Antenna Mechanical Event
9
Plasma Events
10
(No Transcript)
11
Isolating Proof-Mass Noises with Transfer
Functions

• LISA unequal-arm geometry
• Nomenclature for proof mass disturbances vi and
  vi
• Idea isolate which proof mass glitched
  (Braginsky, 2001) -- or which proof mass is
  noisier -- through their transfer functions

12
Isolating Proof-Mass Noises with Transfer
Functions

• Temporal transfer fucntion of 6 PMs to the
  three unequal-arm Michelson combinations (X, Y,
  Z) symmetrical Sagnac (z)
• First generation TDI

13
GW Transfer Function to TDI X

• Source/detector geometry dependent
• 8-pulse response in general for first-gen TDI
  equal arm lengths assumed in this example so it
  degenerates to 7-pulse response ApJ 527, 814
  (1999)
• Will not be confused with PM glitches

14
Isolating Proof-Mass Noises with Transfer
Functions (Statistically)

• Noise-only spectra of first-generation TDI
  combinations X and Z (expressed here as spectra
  of fractional Doppler fluctuations, vs. phase)
• Proof mass 1 assumed 10X noisier than nominal
• PM1 does not enter in Z, so it is OK -- spectrum
  of X affected, however

CQG, 20, S283 (2003)
15
Concluding Comments

• In spacecraft Doppler tracking, at least, the
  data themselves are much more sensitive than any
  of the engineering housekeeping channels which
  monitor experiment health -- so you have to use
  the data themselves, in general, to understand
  the noises
• I would expect this to be true for LISA, too
• Transfer functions of the signal and noises to
  the TDI combinations offer a way to use the
  science data to understand the noises and
  instrument performance