AU753408C

AU753408C - Gravity gradiometer accelerometers

Info

Publication number: AU753408C
Application number: AU76318/98A
Authority: AU
Inventors: Gregory John Adams; Graham Clifford Goodwin; James Beresford Lee; Graeme Joseph O'keefe; Robert John Turner
Original assignee: Broken Hill Pty Co Ltd
Current assignee: FUGRO FINANCE AG
Priority date: 1997-06-11
Filing date: 1998-06-11
Publication date: 2003-08-28
Anticipated expiration: 2018-06-11
Also published as: AU753408B2; AU7631898A

Description

GRAVITY GRADIOMETERACCELEROMETERS

Technical Field

This invention concerns improvements to gravity gradient instruments (GGI), and in particular to the accelerometers that are paired within these instruments.

Background Art

The GGI consists of two pairs of high quality, low noise, matched accelerometers mounted on a block. Each of the accelerometers has an internal feedback loop for proper operation, and an external feedback loop for trim adjustment of the accelerometer scale factor and alignment of the accelerometer sensitive axis.

The normal configuration has the accelerometers mounted in opposing pairs, and equally spaced around the circumference of a circle, with their sensitive axes tangential to the circle. In use the block is rotated about a spin axis which is perpendicular to the plane of the circle, and passes through the centre of the circle. The outputs of the accelerometers of each pair are differenced and the difference signals are then combined. The overall effect is that the large common mode accelerometer output signals cancel to a high degree of precision, so that the residual differences which constitute the gradient signal are observable.

The accelerometers must be matched in their pairs so that the current/acceleration transfer function is matched in amplitude and phase at all frequencies of interest, to an accuracy of 1 part in 10¹⁰. The mismatch in accelerometer pairs is a result of the difference of the internal feedback closed loop errors, and thus the mismatch is also inversely proportional to the open loop gain. The existing external feedback scale factor adjustment can degrade, by an order of magnitude, the high frequency (> 1Hz) lateral sensitivity for a 2% mismatch within the accelerometers. The influence of vertical acceleration on the accelerometers is an additional complicating factor.

Summary of the Invention The invention, as currently envisaged, is an accelerometer having a proof mass suspended by a spring within a magnetic field. An internal feedback loop provides a signal related to movement of the proof mass back through a reaction coil retaining the proof mass in the magnetic field, to maintain the proof mass stationary. An external feedback loop adjusts the accelerometer scale factor. Wherein, the internal feedback loop provides second order compensation to the proof mass and the spring stiffness.

The internal feedback path may include high gain to reduce errors in the accelerometer transfer function.

A compensator in the internal feedback loop may provide double pole and double zero compensation:

m_s² + k_

where: m₁ is the mass of the proof mass k₁ is the spring constant of the spring m_Q is a nominal proof mass and k₀ is a nominal spring constant.

The nominal proof mass and spring constant represent the accelerometer characteristics to which both accelerometers of a pair are to be matched in order that the two accelerometers are closely matched to each other. In a GGI the compensation provided by the internal feedback loop may correct for the mass ratio (or Fo frequency ratio) mismatch between two paired accelerometers by providing the s² loop gain term as a mass compensator. The compensation may also correct for the spring ratio by providing a loop gain term which lumps together variations in pick-off gain, integrator capacitors and spring constant, as a spring stiffness k compensator.

The components of the compensation network are typically resistors and capacitors. Variable components are introduced so that the break frequencies can be tuned over a +7-5% range to match the masses m and the spring stiffnesses k of the accelerometer pairs to better than 0.5%. The closed loop gain blocks are trimmed in pairs to match the time constants to within 0.5%.

In a further aspect the invention is a method of matching accelerometer pairs, comprising the steps of : Testing the pair of accelerometers in back to back fixtures on a horizontal shaker which is aligned to the same vertical angle as in the GGI.

Selecting the accelerometers on the basis of the best scale factor match at 0.5Hz and the lowest lateral sensitivity at 10Hz.

Applying horizontal excitation at 0.25Hz. 1Hz and 10Hz (or higher) and detecting the system response with synchronous demodulation at the frequencies. A scale factor loop may be driven from the 0.25Hz signal or it may be adjusted manually while the excitation is applied.

Adjusting second order mass compensation to minimise the in phase (I) and the quadrature (Q) components of the signal at 10Hz. Adjusting second order spring stiffness compensation to minimise the components at 1Hz.

Iteratively repeating the adjusting steps to achieve matches better than 0.5%.

The external feedback path may provide the output signal, demodulated at the spin frequency Ω by a demodulator, to correct for mismatches in the strength B of the magnets in accelerometers. This feedback loop may also compensate for some part of the mismatches in proof mass and spring stiffness.

Brief Description of the Drawings

An example of the invention will now be described with reference to figure 1 which is a schematic diagram of a modified accelerometer pair embodying the present invention.

Best Modes for Carrying out the Invention

Accelerometer 1 comprises a proof mass 2 suspended by a spring 3 in a magnetic field provided by magnet 4. The proof mass 2 is associated with a position sensor 5 which provides a feedback signal through feedback loop 6 to a reaction coil 7 that retains the proof mass 2. The current il flowing through coil 7 is taken as an output from output port 8.

Feedback loop 6 includes a position sensor amplifier 9, a compensation network 10 and another amplifier 11 arranged in series. In use an external acceleration causes a force to be applied to the proof mass 2 within accelerometer 1. Any tendency to movement of the proof mass 2 causes a current to flow in feedback loop 6 and through coil 7. The current flowing through retaining coil 7 in the magnetic field provided by magnet 4 causes an equal and opposite force to be applied to the proof mass 2 to compensate for the tendency to movement. The current il flowing through the coil 7 is therefore a measure of the acceleration applied to the proof mass. In a GGI the accelerometers are arranged in opposing pairs and the output il from accelerometer 1 is added to the output i2 from its paired accelerometer 12 and is combined with the output signals from the other pair of accelerometers in the GGI, in a summing amplifier 14, and the output is then filtered 15 to provide a signal to the instrument output 16. An external feedback path 17 provides the output signal, demodulated at the spin frequency Ω by demodulator 18, to be applied to a small electromagnet 19 associated with the magnet 4. This feedback signal corrects for mismatches between the strength B of the magnet 4 in accelerometer 1 and the strength of the magnet in accelerometer 12. This feedback loop will also compensate for some part of the mismatches in proof mass m and spring stiffness k; in particular the compensation matches the real component of the of mismatch at the spin frequency, but it cannot match at any other frequency or match the imaginary component of the mismatch. Below 2Hz the dominant parameters affecting mismatch are the spring constant k differences. Above 2Hz the dominant parameters are the proof mass m differences. At the preferred frequency of operation, below 1Hz, the external feedback loop can be thought of as providing dynamic correction for the magnetic field ratio B/k where B is the magnetic field strength and k is the spring stiffness. The internal feedback path 6 includes high gains from the two amplifiers 9 and 11 to reduce errors in the accelerometer transfer function. This error reduction, as a function of frequency, is inversely proportional to the loop gain at that frequency. It follows that the mismatch in accelerometer pairs is a result of the difference of the closed loop errors. The compensator 10 in the feedback loop 6 provides double pole and double zero compensation:

mas² + £o

where: m_t is the mass of the proof mass k-_t is the spring constant of the spring m₀ is a nominal proof mass and Ao is a nominal spring constant. This compensation corrects for the mass ratio (or Fo frequency ratio) by providing an s² loop gain term as a mass compensator. The compensation also corrects for the spring ratio by providing a loop gain term which lumps together variations in pick-off gain, integrator capacitors and spring constant, as a spring stiffness k compensator.

The components of the compensation network 10 are typically resistors and capacitors. Variable components are introduced so that the break frequencies can be tuned over a +/-5% range to match the masses m and the spring stiffnesses k of the accelerometer pairs to better than 0.5%. The closed loop gain blocks are trimmed in pairs to match the time constants to within 0.5%.

The accelerometer pairs are then tested in back to back fixtures on a horizontal shaker. The accelerometers are selected on the basis of the best scale factor match at 0.5Hz and the lowest lateral sensitivity at 10Hz. Horizontal excitation at 0.25Hz, 1Hz and 10Hz (or higher) is input and the system response is detected with synchronous demodulation at these frequencies. A scale factor loop may be driven from the 0.25Hz signal or it may be adjusted manually while the excitation is applied. The mass compensation is adjusted to minimise the in phase (I) and the quadrature (Q) components of the signal at 10Hz, while the spring stiffness compensation is adjusted to minimise the components at 1Hz. The compensations are adjusted iteratively to achieve matches better than 0.5% since there is some interaction between the parameters.

It should be appreciated that although the invention has been described with reference to a particular example it could be embodied in many other forms. For instance, alternative parameterisation is possible besides the magnetic field ratio B/k, mass ratio and spring ratio. However the compensation must provide second order correction. It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific embodiments without departing from the spirit or scope of the invention as broadly described. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.

Claims

1. An accelerometer comprising; a proof mass suspended by a spring within a magnetic field; an internal feedback loop provides a signal related to movement of the proof mass back through a reaction coil acting in the magnetic field, to maintain the proof mass stationary; an external feedback loop to adjust the accelerometer scale factor; wherein, the internal feedback loop provides second order compensation to the proof mass and the spring stiffness.

2. An accelerometer according to claim 1, wherein the internal feedback path includes high gain to reduce errors in the accelerometer transfer function.

3. An accelerometer according to claim 1, wherein there is a compensator in the internal feedback loop to provide double pole and double zero compensation as follows: giiS² + k

where: m-i is the mass of the proof mass k_x is the spring constant of the spring m₀ is a nominal proof mass and ko is a nominal spring constant and, the nominal proof mass and spring constant represent the accelerometer characteristics to which both accelerometers of a pair are to be matched in order that the two accelerometers are closely matched to each other.

4. In a gravity gradient instrument compensation is provided by the internal feedback loop to correct for the mass ratio (or Fo frequency ratio) mismatch between two paired accelerometers by providing a s² loop gain term as a mass compensator.

5. Compensation according to claim 4 wherein the compensation also corrects for the spring ratio by providing a loop gain term which lumps together variations in pick-off gain, integrator capacitors and spring constant, as a spring stiffness k compensator.

6. Compensation according to claim 5 wherein the components of the compensation network include variable resistors or capacitors so that the break frequencies can be tuned over a +/-5% range to match the masses m and the spring stiffnesses k of the accelerometer pairs to better than 0.5%.

7. Compensation according to claim 6 wherein the closed loop gain blocks are trimmed in pairs to match the time constants to within 0.5%.

8. A method of matching accelerometer pairs, comprising the steps of: testing the pair of accelerometers in back to back fixtures on a horizontal shaker which is aligned to the same vertical angle as in the GGI; selecting the accelerometers on the basis of the best scale factor match at 0.5Hz and the lowest lateral sensitivity at 10Hz; applying horizontal excitation at a range of frequencies above 0.25Hz, and detecting the system response with synchronous demodulation at those frequencies; adjusting second order mass compensation to minimise the in phase (I) and the quadrature (Q) components of the signal at 10Hz; adjusting second order spring stiffness compensation to minimise the components at 1Hz; and iteratively repeating the adjusting steps to achieve matches better than 0.5%.

9. Compensation according to claim 8 wherein a scale factor loop may be driven from the horizontal exit action or it may be adjusted manually while the excitation is applied.

10. Compensation according to claim 9 wherein the external feedback path provides the output signal, demodulated at the spin frequency ╬⌐ by a demodulator, to correct for mismatches in the strength B of the magnets in accelerometers.