Disturbances in GRACE-FO LRI Ranging Data: Thruster Effects and Single Event Upsets

The range between the two spacecraft in the GRACE-FO mission is measured by the Laser Ranging Interferometer (LRI) and the Microwave Instrument (MWI). The measured signal has variations of gravitational and non-gravitational origins, and can also contain unphysical disturbances, which we regard as noise. Some sporadic and unphysical range errors in LRI are expected to arise from radiation-induced Single Event Upsets (SEUs) in the ranging processor. Another source of non-gravitational range variations are parasitic accelerations produced by the attitude control thrusters (ACTs), which are short-lived events that can dominate the LRI range spectrum for frequencies between 35 and 200 mHz.

SEUs are errors in the processed data due to a transient perturbation of electronic components in the processing chain by high-energy particles that exist in Earth’s magnetic field and relentlessly inundate all electronic devices on board the spacecraft. SEUs can manifest as temporary bit upsets in the registers of two low-pass FIR filters, propagating through a series of decimators and producing a distinct pattern in the LRI downlinked data. We investigate bitflips in any of the four downlinked channels on each spacecraft at a time.

Knowing the architecture of the LRI, we produce SEU templates of the internal filtering chain’s impulse response over the parameter space. The free parameters are the exact sub-sample timing, the register number within the 500- or 140-taps filter and the bit number of the 64-bit depth of each register. The output patterns are then compared to the observed glitches through a pattern matching algorithm to detect and characterize the SEUs. About 20 SEU events were successfully detected and modelled using this algorithm.

A much more frequent disturbance arises from the activation of ACTs which keep the satellites aligned. Even though ACTs should only induce rotational accelerations, slightly misaligned mounts and imperfections of the valves can cause a small coupling into linear accelerations. These linear accelerations are measured by the accelerometer in all three axes and by the LRI in the line-of-sight direction when the second time-derivative of the ranging data is formed. As long as both the accelerometer and LRI correctly measure these accelerations, they should not introduce errors when performing gravitational field recovery.

We present an approach to separate the effects of ACTs from the dominant gravitational signal in LRI ranging data and characterize and model the acceleration profiles in the line-of-sight direction of different ACTs to remove the signatures from LRI ranging data for diagnostic purposes. Our analysis excludes ACT-induced unphysical phase jumps in LRI data, which are handled separately.

Our ACT results help to understand the ACT response differences in the accelerometer and the LRI data products, while identifying and removing SEU models provides cleaner ranging datasets. Both aspects have relevance for the design of future missions and instruments and might, in principle, improve the recovered gravity fields in GRACE-FO.