The free-fall mode experiment on LPF: first results

Transcription

1 The free-fall mode experiment on LPF: first results Roberta Giusteri on behalf of the LPF collaboration University of Trento and INFN-TIFPA LISA Symposium XI, 5-9 September

2 LPF and actuation control LPF as an accelerometer Down-scaled version of one LISA arm TM1 drag free (geodesic reference) TM2 continuously controlled along x 2 F V act δ(f) 2F δv act V act S Δg [m s -2 Hz -1/2 ] Actuation x/ϕ 1/2 S δvact /V act S 1/2 1/2 Δg 2 Δg DC S δvact /V act! 5 ppm/ Hz Actuation ϕ Actuation along x not present in LISA [M Armano et al. PRL (2016)] 1/11

3 Free Fall experiment Proposed to subtract the actuation noise on TM2 Actuation limited to short impulses Repeated stretches of parabolic flights Tfly = s Timp = 1 s Texp = s fexp 2.85 mhz Ø No dependence on the actuator calibration Ø Acceleration noise in actuation-free condition: characterization for LISA 2/11

4 FF mode vs standard science mode Predicted actuation noise in different authority configurations Nominal: F max x2 = 2200 pn N max ϕ2 = 10.4 pn m URLA: F max x2 = 50 pn S Δg [m s -2 Hz -1/2 ] FF in Nominal (ϕ) FF in URLA (ϕ) Science mode in URLA (x/ϕ) Science mode in Nominal (x/ϕ) N max ϕ2 = 1 pn m 3/11

5 FF mode vs standard science mode Predicted actuation noise in different authority configurations Nominal: F max x2 = 2200 pn N max ϕ2 = 10.4 pn m URLA: F max x2 = 50 pn S Δg [m s -2 Hz -1/2 ] FF in Nominal (ϕ) FF in URLA (ϕ) Science mode in URLA (x/ϕ) Science mode in Nominal (x/ϕ) N max ϕ2 = 1 pn m Δx ~ 35 nm Δg [m s -2 ] F cmd ~ 3 pn F kick ~ 1.5 nn Δg predicted ~ 600 pm/s 2 3/11

6 Experiment calibration Equation of motion to fit: Δ!!x = ω 2 2 Δx Δω 2 (x 1 x SC )+ Δg+!g 0 t Filtering and decimation: T samp ~ 17 s f samp ~ 57 mhz n fit = 14 x 10-7 ω 2 2 Time [s] values [m s -2 ] x values [s -2 ] x 10-6 Δω 2 x Parameter Preliminary results Mean value Error ω 2 2 [10-7 s -2 ] Δω 2 [10-7 s -2 ] Δg [pm s-2 ] !g 0 [pm s -2 d -1 ] [FF on June 10 th in URLA authority] values [s -2 ]!g 0 values [m s -2 d -1 ] Δg 4/11

7 LTPDA dev (R2014b), :30: UTC, LPF_DA_Module: 3224d1b, ltpda: f476c3b, iplot Data analysis challenge Kicks: high-noise configuration analysis of actuation-free motion à gaps in data Δg [m s -2 ] Analysis techniques: Time [s] Blackmann-Harris low pass filtering Windowing technique Constraint-Gaussian Gap Patching low pass filtering + downsampling and set kicks to zero masking the gaps with spectral windows filling the gaps with artificial data Δg [m s -2 ] Amplitude [] Amplitude [] Amplitude [N] # commanded force Origin: :34: Time [s] # Window HF Origin: :34: Time [s] # 10 4 [Poster: G. Russano LISA Pathfinder Free Fall Mode, A18] 5/11

8 BH low pass approach Δg estimation at 10Hz Δg(t) = Δ!!x + Δω 2 (x 1 x SC )+ω 2 2 Δx!g 0 t Δg [m s -2 ] BH low-pass filter n tot per T exp, n keep per T fly Set to zero n gap = n tot n keep Evaluate the spectrum PSD correction for n tot /n keep and BH transfer function T samp = T exp / n tot BH window, T win = T fly 2 margin (n keep 1) T samp filtering and downsampling Δg [m s -2 ] Time [s] Configuration: n tot 34 n keep 25 n gap 9 T samp [s] 10.3 f samp [mhz] 97 margin [s] 2 T win [s] 98 6/11

9 Calibration of BH technique Investigation into the potential BIAS introduced by the BH technique: 1. Science measurement (URLA, May 16-18) 2. Apply BH filtering method 3. Insert artificial gaps 4. Estimate the spectrum 5. Correction for filter transfer function S Δg Bias estimation necessary to correct the spectrum accurately. Bias estimation under consolidation Correction related to n tot /n keep (work in progress) S Δg 7/11

10 First results applying BH method FF Nominal x/ϕ authority on TM2, 1 day (F max = 2200 pn, N max = 10.4 pn m): Parameter Mean value Error ω 2 2 [10-7 s -2 ] Δω 2 [10-7 s -2 ] Δg [pm s-2 ] !g 0 [pm s -2 d -1 ] FF URLA x/ϕ authority on TM2, 1 day (F max = 50 pn, N max = 1 pn m): Parameter Mean value Error ω 2 2 [10-7 s -2 ] Δω 2 [10-7 s -2 ] Δg [pm s-2 ] S Δg Preliminary results!g 0 [pm s -2 d -1 ] /11

11 Comparison with standard mode Nominal authority: Science mode: May Free Fall: June 9 th S Δg [m s -2 Hz -1/2 ] Preliminary results URLA authority: S Δg Preliminary results Science mode: May Free Fall: June 10 th 9/11

12 Comparison among all FF runs URLA authority, BH technique applied Averaged spectral values in [0.1, 0.4] mhz band FF run S [fm s -2 Hz -1/2 ] June run (1 d) ± July run (1.5 d) ± August run (3 d) ± S Δg Preliminary results 10/11

13 Conclusions and perspectives Ø Free-fall mode control achieved and maintained Ø All analysis techniques have extracted spectra Ø Free Fall results consistent with actuation models predicted Ø Free Fall results consistent with long science measurements (URLA) Future work: q Analysis consolidation of the two last FF experiments q Correction for BIAS introduced by each method q New Free Fall experiment in same conditions of the first one (and longer) q New design Free Fall experiment with kicks also on ϕ 2 Perspectives: Free Fall implementation à closer to LISA performance Spectral estimation in presence of gaps 11/11

14 Thank you 14