Advanced SDM control systems for MEMS capacitive sensing accelerometers
Advanced SDM control systems for MEMS capacitive sensing accelerometers
In this thesis a novel approach, applying multi stage noise shaping (MASH) closed-loop scheme to a capacitive bulk micromachined electromechanical sensing element, is presented. The MASH architecture is implemented by cascading a purely electron ΣΔΜ analogue-to-digital converter to the single-loop 2nd order electromechanical ΣΔ-interfaced accelerometer to provide higher-orders of modulation.
Using analytical derivations and Matlab and Simulink simulations, it has been shown that the addition of the second stage reduces the modulation noise inside the baseband by approximately 20dB. This is sufficient to ensure that modulation noise is no longer the restricting factor in characterising the performance of the accelerometer.
Both the single-loop accelerometer and the second electronic ΣΔΜ loop were implemented on PCB and prototype boards. The output of the two loops was recorded using a data acquisition card and then digitally recombined in Matlab. Improvements of up to 10dB in the modulation noise level were achieved.
Using linear approximation models and simulations, a study was made into the effects of variations in the sensor dynamics upon the performance of the MASH. Environmental changes were also included in the study, which showed that for variations of ±20% in the sensor dynamics the performance of the MASH system remains superior to the single-loop scheme.
Evidence from analytical models suggested that the addition of higher-order or further loops above the 2-1 architecture would lead to additional noise suppression within the baseband. However Matlab and Simulink simulations showed no additional performance enhancement were obtained for the 2-2 and the 2-1 architectures, beyond that of the 2-1 model. This discrepancy is caused by the low dc gain of the sensing element, causing correlation between the signal and quantisation noise in the first loop, which the linear model does not take into account.
University of Southampton
2005
Mokhtari, Mir
(2005)
Advanced SDM control systems for MEMS capacitive sensing accelerometers.
University of Southampton, Doctoral Thesis.
Record type:
Thesis
(Doctoral)
Abstract
In this thesis a novel approach, applying multi stage noise shaping (MASH) closed-loop scheme to a capacitive bulk micromachined electromechanical sensing element, is presented. The MASH architecture is implemented by cascading a purely electron ΣΔΜ analogue-to-digital converter to the single-loop 2nd order electromechanical ΣΔ-interfaced accelerometer to provide higher-orders of modulation.
Using analytical derivations and Matlab and Simulink simulations, it has been shown that the addition of the second stage reduces the modulation noise inside the baseband by approximately 20dB. This is sufficient to ensure that modulation noise is no longer the restricting factor in characterising the performance of the accelerometer.
Both the single-loop accelerometer and the second electronic ΣΔΜ loop were implemented on PCB and prototype boards. The output of the two loops was recorded using a data acquisition card and then digitally recombined in Matlab. Improvements of up to 10dB in the modulation noise level were achieved.
Using linear approximation models and simulations, a study was made into the effects of variations in the sensor dynamics upon the performance of the MASH. Environmental changes were also included in the study, which showed that for variations of ±20% in the sensor dynamics the performance of the MASH system remains superior to the single-loop scheme.
Evidence from analytical models suggested that the addition of higher-order or further loops above the 2-1 architecture would lead to additional noise suppression within the baseband. However Matlab and Simulink simulations showed no additional performance enhancement were obtained for the 2-2 and the 2-1 architectures, beyond that of the 2-1 model. This discrepancy is caused by the low dc gain of the sensing element, causing correlation between the signal and quantisation noise in the first loop, which the linear model does not take into account.
This record has no associated files available for download.
More information
Published date: 2005
Identifiers
Local EPrints ID: 465663
URI: http://eprints.soton.ac.uk/id/eprint/465663
PURE UUID: 61582f8a-a889-4f61-acbf-74cfa0c95c52
Catalogue record
Date deposited: 05 Jul 2022 02:25
Last modified: 05 Jul 2022 02:25
Export record
Contributors
Author:
Mir Mokhtari
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics