Micro electro mechanical gyroscopes are extending their scope toward demanding applications that require low noise, stable bias and robust operation. This work addresses the design and validation of a mode split amplitude modulation gyroscope and explores an alternative Force to Rebalance path. The objective is to approach navigation grade targets through architectural choices, electromechanical modeling and development of the signal chain. Starting from a platform of automotive origin, a coherent model is established for both the device and the electronic loops of drive, amplitude control and open-loop sense. Nonlinearities and parasitic couplings are analyzed, the dependence on modal decoupling is quantified, and design rules are derived for gain, bandwidth and noise. Once a dedicated board has been designed, a measurement campaign is carried out. In this configuration an Angle Random Walk of 0.037 °/ sqrt(h) and a Bias Instability of 0.155°/h are achieved without thermal control, consistent with the noise budget and with the bias instability model. Based on these results, the gyroscope was redesigned.Thus, Wafer-level measurements confirmed the presence of the expected vibration modes and demonstrated effective electrostatic tuning of the frequency split. However, an unexpected behavior of the quality factor was observed, prompting an in-depth analysis of the underlying physical phenomenon. In operation with a custom electronic board, the redesigned device exhibited a lower noise level compared to the previous version. Nevertheless, the presence of spurious modes prevented achieving the expected bias drift performance. To increase useful bandwidth and full scale range, a Force to Rebalance route is developed. A digital board samples the sense output, implements in logic a phase locked loop for demodulation and for remodulation, applies the required filtering and control blocks, then reconverts and reinjects the control action on the sense actuators through a DAC. The measurement campaign shows linear response up to 2000 dps with error below 0.03 %, in band white noise around 1.8 mdps / sqrt(Hz) from Allan analysis, and a Bias Instability near 0.2°/h at the optimal averaging time. The work provides a complete path from technology choice to experimental validation and identifies the key elements for a higher level of integration with improved margins on noise and stability. The thesis is organized in six chapters that cover introduction and state of the art, device and control loops, experimental characterization, modeling of the quality factor, the Force to Rebalance path, and on board electronics with test results.
I giroscopi microelettromeccanici stanno estendendo il loro campo di applicazione verso utilizzi più esigenti, che richiedono basso rumore, bias stabile e funzionamento robusto. Questo lavoro affronta la progettazione e la validazione di un giroscopio a modulazione di ampiezza con split modale ed esplora un percorso alternativo di Force to Rebalance. L’obiettivo è avvicinarsi a prestazioni di livello navigazionale attraverso scelte architetturali, modellazione elettromeccanica e sviluppo della catena di segnale. Partendo da una piattaforma di origine automotive, viene definito un modello coerente sia del dispositivo sia dei loop elettronici di drive, controllo dell’ampiezza e sense in open-loop. Vengono analizzate le non linearità e gli accoppiamenti parassiti, quantificata la dipendenza dal disaccoppiamento modale e derivate regole di progetto per guadagno, banda e rumore. Una volta progettata una scheda dedicata, viene condotta una campagna di misure. In questa configurazione si ottengono un Angle Random Walk di 0,037 °/sqrt(h) e una Bias Instability di 0,155°/h senza controllo termico, in accordo con il budget di rumore e con il modello di instabilità del bias. Sulla base di questi risultati, il giroscopio è stato riprogettato. Le misure a livello di wafer hanno confermato la presenza dei modi di vibrazione attesi e dimostrato un’efficace taratura elettrostatica della separazione in frequenza. Tuttavia, è stato osservato un comportamento inatteso del fattore di qualità, che ha richiesto un’analisi approfondita del fenomeno fisico sottostante. In funzionamento con una scheda elettronica custom, il dispositivo riprogettato ha mostrato un livello di rumore inferiore rispetto alla versione precedente. Ciononostante, la presenza di modi spurii ha impedito di raggiungere le prestazioni attese in termini di deriva del bias. Per aumentare la banda utile e il full scale range, viene sviluppato un percorso Force to Rebalance. Una scheda digitale campiona l’uscita di sense, implementa in logica un phase locked loop per la demodulazione e la rimodulazione, applica i blocchi di filtraggio e controllo richiesti, quindi riconverte e reinietta l’azione di controllo sugli attuatori di sense tramite un DAC. La campagna di misure mostra una risposta lineare fino a 2000 dps con errore inferiore allo 0,03%, rumore bianco in banda intorno a 1,8 mdps/sqrt(Hz) dall’analisi di Allan e una Bias Instability prossima a 0,2°/h al tempo di mediazione ottimale. Il lavoro fornisce un percorso completo dalla scelta tecnologica alla validazione sperimentale e individua gli elementi chiave per un livello superiore di integrazione, con margini migliorati in termini di rumore e stabilità. La tesi è organizzata in sei capitoli che coprono introduzione e stato dell’arte, dispositivo e loop di controllo, caratterizzazione sperimentale, modellazione del fattore di qualità, il percorso Force to Rebalance ed elettronica di bordo con risultati di test.
Design and validation of a mode split amplitude modulation MEMS gyroscope in open loop and force to rebalance for navigation grade performance
DE PACE, MARCO
2025/2026
Abstract
Micro electro mechanical gyroscopes are extending their scope toward demanding applications that require low noise, stable bias and robust operation. This work addresses the design and validation of a mode split amplitude modulation gyroscope and explores an alternative Force to Rebalance path. The objective is to approach navigation grade targets through architectural choices, electromechanical modeling and development of the signal chain. Starting from a platform of automotive origin, a coherent model is established for both the device and the electronic loops of drive, amplitude control and open-loop sense. Nonlinearities and parasitic couplings are analyzed, the dependence on modal decoupling is quantified, and design rules are derived for gain, bandwidth and noise. Once a dedicated board has been designed, a measurement campaign is carried out. In this configuration an Angle Random Walk of 0.037 °/ sqrt(h) and a Bias Instability of 0.155°/h are achieved without thermal control, consistent with the noise budget and with the bias instability model. Based on these results, the gyroscope was redesigned.Thus, Wafer-level measurements confirmed the presence of the expected vibration modes and demonstrated effective electrostatic tuning of the frequency split. However, an unexpected behavior of the quality factor was observed, prompting an in-depth analysis of the underlying physical phenomenon. In operation with a custom electronic board, the redesigned device exhibited a lower noise level compared to the previous version. Nevertheless, the presence of spurious modes prevented achieving the expected bias drift performance. To increase useful bandwidth and full scale range, a Force to Rebalance route is developed. A digital board samples the sense output, implements in logic a phase locked loop for demodulation and for remodulation, applies the required filtering and control blocks, then reconverts and reinjects the control action on the sense actuators through a DAC. The measurement campaign shows linear response up to 2000 dps with error below 0.03 %, in band white noise around 1.8 mdps / sqrt(Hz) from Allan analysis, and a Bias Instability near 0.2°/h at the optimal averaging time. The work provides a complete path from technology choice to experimental validation and identifies the key elements for a higher level of integration with improved margins on noise and stability. The thesis is organized in six chapters that cover introduction and state of the art, device and control loops, experimental characterization, modeling of the quality factor, the Force to Rebalance path, and on board electronics with test results.| File | Dimensione | Formato | |
|---|---|---|---|
|
2026_02_DePace_Thesis.pdf
non accessibile
Descrizione: Thesis
Dimensione
29.26 MB
Formato
Adobe PDF
|
29.26 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/249097