GB2280037A - Minature silicon accelerometer and method - Google Patents

Abstract

A miniature accelerometer is fabricated using integrated circuit manufacturing and silicon micromachining techniques to form a closed loop, force balance sensor utilizing a silicon proofmass 16 formed from and connected to a layer of silicon 24 by a split flexure 22 etched therein. Sensor circuitry 14 detects AC signals coupled from the proofmass to a pair of electrodes 30, 32 formed on glass layers 26, 26 anodically bonded to the silicon layer. A DC restorative force is applied to the electrodes 30, 32 in response to the detected AC signals to balance acceleration forces applied to the proofmass. The sensor design is highly symmetrical. The proofmass receives applied voltages from DC bias source 35 and AC source 38, the resultant signals on electrodes 30, 32 being fed to amplifier 60 via a diode quad bridge. The amplifier output signal, at point 34, is of magnitude and sign corresponding to the sensed acceleration and is fed back to the electrodes 30, 32 (via an inverter 62 in the latter case). Conductive guard bands 64, 66 are provided on the glass layers 26, 28 around the electrodes. To prevent the proofmass touching one of the electrodes if the device is under acceleration at start-up, the guard electrodes have inward fingers 74 - 77 corresponding to cut-outs in the electrodes, and contact posts are provided on the proofmass as at 78, 79, fig. 3. <IMAGE>

Description

MINIATURE SILICON ACCELEROMETER AND METHOD This invention relates to miniature accelerometers for medium accuracy applications. There is a need for small and inexpensive accelerometers for use in tactical systems, such as the miniature accelerometers required in smart artillery and mortar operation where large stocks of rounds and accelerometers may be expended. A tactical accelerometer for such applications may be considered to be a medium quality instrument, but must have abdut five orders of magnitude dynamic range and be operable in a high vibration environment and over a wide temperature range.

The present invention provides a miniature accelerometer comprising: a pair of spaced electrode layers, each electrode layer comprising forcer electrode means and guard electrode means; a silicon proofmass; means for suspending said proofmass between said electrode layers; means responsive to differential capacitive coupling between said forcer electrode means and said proofmass for opposing acceleration forces applied thereto; wherein said guard electrode means prevents contact between said proofmass and said forcer electrode means.

In a further aspect, the invention provides a method of sensing acceleration using a proofmass and electrode layers forming an accelerometer, comprising the steps of providing a pair of electrode layers each said electrode layer comprising forcer electrode means and guard electrode means; suspending said proofmass in a balanced position between said electrode layers to define substantially equal coupling capacitances between said forcer electrode means and said proofmass; electrically sensing differential capacitive coupling between said forcer electrode means and said proofmass; generating a force on said proofmass which will oppose acceleration forces on said proofmass; and preventing contact between said forcer electrode means and said proofmass using said guard electrode means.

Emodiments of the present invention increase the dynamic range of a miniature accelerometer by a factor of 100 to a 1000 over the previous state-of-the-art micromachined accelerometers which have had a dynamic range of about 100.

Reference is hereby made to co-pending Divisional Application Serial No. 2251693 which claims a miniature accelerometer, comprising a pair of spaced electrode layers; a silicon proofmass; means for suspending said proofmass between said electrode layers; means for applying an AC signal to said proofmass; means for detecting AC signals capactively coupled to said electrode layers; and means responsive to the AC signals detected for applying acceleration opposing forces to the proofmass and generating an electrical signal which is a measure of acceleration.

FIGURE 1 is an illustration of a miniature accelerometer constructed according to one embodiment of the present invention including a partial cross sectional view of an accelerometer transducer shown in combination with a schematic diagram of sensor circuitry.

FIGURE 2 is an isometric view of the compontnts of thc acceleromter transduca of FIGURE 1 shown with the upper part opened, and with portions of thc lower part shown in cross section.

FIGURE 3 is a top plan view of the region of the proofmass in the region 3-3 of FIGURE2.

FIGURE 1 shows the miniature accelerometer 10 of one embodiment of the present invention including a transducer 12 and circuitry 14 and lia The transducer 12 of accelerometer 10 is shown in partial cross sectional view and comprises proofmass 16 shown in thc ncutral position, midway between upper and lower electrodes 30 and 32. Proofmass 16 is formed of silicon suspended by a weak flexure 22 integrally formed from silicon layer 24. Flexure 22 dcfincs a bend linc along the mid-plane of proofmass 16, indicated by rnid-planc axis 23 to mi"1misc vibration rectification. By arranging thc flexurc to occur symmetrically about the mid-plane of the proofmass, shaking of the accelerometer will not induce a rectification so that the instrument will have high lincarity.

Silicon layer 24 is anodically bonded between upper and lower glass layers 26 and 28 upon which have bccn deposited upper and lower metal electrodes 30 and 32, respectively, of aluminum or gold or other conduce In addition, the top electrode glass layer and the bottom electrode glass layer arc identical picses, fabricated from thc samc mask first applied to the top, and then to the bottom layer to keep thc symmetry exact The proofmass is very symmetric so biases arc not induccd in thc device This also contributes to a high valuc of dynamic range with high lincarity.

While electrodes 30 and 32 may conveniently be fabricated fmm evaporated aluminum, there is a certain moisturc sensitivity inherently present in using al lmsnum bc- cause it adsorbs significant surface moisturc which changes thc dielectric pmpcnics of the surface and adds capacitance which is indisdnguishablc from other capacitances being mc Accordingly, it is preferred to usc a metal which docs not adsorb moisturn, such as gold which may be conveniently sputtcr deposited. In order to keep thc proofmass surfacc stable, it is also preferably coated with gold Accelerometer 10 functions as a force balance instrument in which the force of acceleration upon proofmass 16 is balanced by appropriate electrical restoring forces applied by sensor circuitry 14 through upper and lower forcer electrodes 30 and 32.

The magnitude of the required restoring force is a function of the acceleration and may be measured at acceleration output 34 of sensor circuitry 14.

A DC bias is applied to proofmass 16 by a reference voltage source, such as stable semiconductorvoltagereference source 35, through decoupling orACisolatingresistor 36.

In parallel, an AC signal generated by signal source 38 is coupled to prooEmass 16 by coupling capacitor 40. Capacitors 42 and 44 couple the capacitances, formed between upper and lower electrodes 30 and 32 and proofmass 16, to differential bridgecir- cuit 46 which includes diodes 50,52,54 and 56, filter capacitor 58 and high impedance, high gain operational amplifier 60. The output of amplifier 60 is used to directly drive upper electrode 30 and, after inversion by inverter 62, to drive lower electrode 32.

Acceleration output 34 is applied via amplifier 60 and inverter 62 to upper and lower elects 30 and 32 to apply electrostatic restoring forces to proofmass 16 to resist forces applied thereto by acceleration. Such electrostatic forces would normally be nonlinear, because they are proportional to the square of the applied voltage. TheDC bias applied by stable voltage source 35, and the out-of-phase signals applied to upper and lower electrodes 30 and 32 from amplifier 60 and inverter 62, serve to remove this nonlincarity.

The forcer and proofmass voltage offset give a linear output with the electrostatic force feedback even though the basic forcer equations are nonlinear. The basic forcer equations are:

where Fu is the upper force, (Fl is the lower force), A is the forcer electrode area, Vu is the voltage between the upper forcer electrode and the prooftnass, Ve is the voltage be- tween the lower electrode and the proofmass, and d is the electrode/proofmass spacing.

By setting Vu = VOA and Vl = VO+ (2) where Ve is the voltage at 34 and A is a fixed voltage, e.g., lSV, an exact linearizatton of the force equation is obtained, as shown by: Ma = F,-FL= -2#0A#V0. (3) d where M is the mass of proofmass and a is the acceleration, so that the scale factor becomes: 2#0A# - (4) Md # is the voltage generated by source 35. Thus, the output voltage at 34 is exactly proportional to acceleration when proofmass 16 is centered between upper and lower electrodes 30 and 32. The voltage may then be used as a measurement of the accelera- tion.

Linearization may also be achieved by having proofmass 16 at zero bias and applying a fixed bias voltage in series with amplifier output to upper electrode 30 while applying the same fixed bias voltage, with the opposite polarity, in series with amplifier output to lower electrode 32.

Remaining system nonlinearities due, for cxamplc, to imperfect centering of the proofmass in the gap, may be reduced by bias offsets applied to amplifier 60 from a voltage bias source Vb through a large resistor ( > 250kSl) R3. The use of such bias offsets pcrmits the reduction of the effects of imperfect centering and vibration rectification essentially to zero.

In addition to upper and lower electrodes 30 and 32, upper and lower glass layers 26 and 28 may include upper and lower guard band electrodes 64 and 66. As shown in FIGURE 1, and discussed further below with regard to FIGURE 2, these guard band electrodes may be substantially thicker than upper and lower electrodes 30 and 32 to stand off proofmass 16 from upper and lower electrodes 30 and 31 This permits successful turn on of closed loop accelerometer 10 under acceleration without additional special circuitry.

The gaps on upper glass layer 26 between upper electrode 30 and upper guard band electrode 64 and, on lower glass layer 28 between lower electrode 32 and lower guard bandelectrode 66 may present special problems due to charging of the glass and leakage currents there across. When the potential on one such electrode is changed, the potential in the intervening insulator region or gap will also adjust from one value to another due to leakage currents on upper and lower glass layers 26 and 28. The potential distribution in these gaps changes as the insulation characteristics of these gaps change with humidity and other environmental conditions.

The potential in these gaps acts as an extension of the potential of upper or lower electrodes 30 and 32.

The electrostatic force effects of the potentials in these gaps is indistinguishable from the forces applied by upper and lower electrodes 30 and 32. The slowly changing potentials in the gaps therefore produce force transients at turn on and at acceleration changes.

By way of explnnstion consider the proofinass which sees only the forcer trans ducer electrode 30 and 32 of the upper and lower force However, the guard ring which is around the proofmass to prevent turn on problems is at a different potential from the electrodes 30 and 32. When the potential on electrodes 30 and 32 changes because of an acceleration change, the glass region between the electrodes 30 and 64 on the top or the glass region between electrode 32 and 66 on the bottom will have to change its potential as well. Glass between the main electrodes and the guard ring can be considered as a resistive material of very high resistance. If the potentials of electrodes 30 and 32 change, the region between these electrodes and their guard rings will change, but that change will be slow because of the high resistance of the glass.

The way to avoid having significant influence from the glass region between the electrodes is to make the main electrode 30 and electrode 32 much larger than the proofmass everywhere. However, the guard electrodes must be brought in at four small locations or fingers 74,75,76,77, to touch the proofmass (FIGURES 2 and 3), should the proofmass be forced against the upper or lower glass layers. The electrode shape is arranged then such that the guard ring extends over the proofmass region only in these fingers Otherwise, the glass region which is slow to charge is outside of the region where it can influence the proofmass. Nonetheless, the four fingers 74-77 do stick in and the effect of the glass in these regions around the four fingers is important In order to reduce the effect of this glass region, a V-groove 21 is placed in the proofmass underneath the glass region so that if the charge on the glass varies slowly, it will have a minimum effect in forcing the proofmass to move because of the increased distance from the glass to the proofmass. Each groove 21 defines a post, 78 and 79 being shown in FIGURE 3. The posts are there to contact the electrodes 64,66 while keeping the the electrodes 30,32 clear of the proofmass when the device is started up under accelera- tion. The gap between the proofinass and the glass is no longer 3 microns in this area but is much larger, closer to 20 or 30 microns.

In this way, the acceleration transient response times arc reduced or eliminatei As shown in greater detail in FIGURE 3, trench or V-groove 21 is etched in proofmass 16 directly beneath the gaps between upper and lower electrodes 30 and 32 and upper and lower guard band electrodes 64 and 66 on upper and lower glass layers 26 and 28, respectively. V-groove 21 reduces the effect of the potential in the gaps on proofmass 6 by providing a large space therebetween, such as 10 times the normal spacing between upper glass layer 26 and proofmass 16. Upper and lower electrodes 30 and 32 may extend beyond prooitsass 16 so that the effect of the fields from glass between upper electrode 30 and upper guard band electrode 64 and between lower electrode 32 and lower guard band electrode 66 are minimized The top and bottom glass surfaces are metalized at 26a, 28a and connected to ground to shield the transducer from outside electric fields and from the influence of outside conductors.

Alternatively, a slightly conducting surface coating (1010ohms per square) may be placed on upper and lower glass layers 26 and 28, particularly in the gaps. The potential effects of the gaps would not be eliminated, but the potential distribution in the gap would be almost instantaneously established and held constant, permitting satisfactory acceleration sensing without unwanted transients as long as the resistance of the surface coating is high enough to not interfere with the sensing.

Referring now to FIGURE 2, accelerometer transducer 12 is shown in a partially exploded isometric view in which upper glass layer 26 has been rotated into an open position and the remaining portion cut away for clarity. As noted above, upper electrode 30 has been fabricated on upper glass layer 26 by evaporation of aluminum or gold.

Surrounding upper electrode 30 is upper guard band electrode 64. In a similar manner, a portion of lower guard band electrode 66 on lower glass layer 28 is visible through the partially cutaway portion of silicon layer 24. The end of lower electrode 32 is also visible.

Metallic bond pad 81 on silicon layer 24, electrically contacts this layer. Bond wires such as 1 mil gold or aluminum wire connect the ends of clccaodcs 32,66,30, and 64 and bond pad 81 to electrical circuit 14 and 14a The bond wires are attached in one of several manners common to the integrated circuit industry.

;Uppcr.and lower guard band electrodes 64 and 66 may conveniently be fabricated with a double thickness layer of gold metalization, compared to the thickness of upper and lower electrodes 30 and 32. This additional thickness serves to stand proofatass 16 off from upper and lower electrodes 30 and 32 during turn on of accelerometer 10.

Upper and lower guard band electrodes 64 and 66 are maintained at the same potential as proofmass 16.

As may be further seen in FIGURE 2, upper and lower glass layers 26 and 28 may be designed to be identical to simplify fabrication. Wherever possible, transduccr 12 has been designed and constructed symmetrically to reduce mechanical stresses and drift. The material used for upper and lower glass layers 26 and 28, such as 7740 brand borosilicate Pyrex glass, should be matched as closcly as possible to that of proofmass 16 for thermal expansion. One can also use silicon with appropriate insulating layers for the top and bottom portions of the transduccr sandwich Grooves 68 and 70 are positioned on the upper surface of silicon layer 24 as shown to insure isolation of upper electrode 30 and upper guard band electrode 64 from contact with silicon layer 24 while similar insulating grooves, not visible in this view, are po sitioned on thG lowcr surface of silicon layer 24 to insulate lower guard band electrode 66 and lower electrode 32 from electrical contact with silicon layer 24.

Proofmass 16 is visible in the partial cutaway view, connected to silicon layer 24 by flexure 22. Flcxurc 22may be constructed from a single hinge conncction but, as shown, it is prefaablc to utilize a symmetrically balanced split flexure, such as that formed by hinges 25 and 27, to =inirnize prootnas offsets due to bending of accelerometer transducer 12. A minimum strength is required in flexure 22 to resist breakage due to forces in the plane of the silicon and to resist an electrostatic cocking instability.

The construction of accelerometer transducer 12, including proofmass 16 and flexure 22, may be accomplished with available integrated circuit and micromachining fabrication technologies. Various controlled etch techniques are available for forming proofmass 16 from surrounding silicon layer 24, leaving a limited physical connection therebetween to form flexure 22. Flexure 22 may be annealed at high temperature, such as 1000 C for several hours, to reduce any non-uniform stresses resulting during fabrication from high doping gradients andlor concentrations. Upper and lower glass layers 26 and 28 are anodically bonded to silicon layer 24 simultaneously to prevent warpage.

The upper and lower surfaces of proofmass 16 must be formed by etching away the surfaces of silicon layer 24 so that the appropriate gaps, on the order of about 2pm, will be created between proofmass 16 and upper and lower glass layers 26 and 28.

The outer surfaces of glass layers 26a and 28a are preferably metalized with a conductive coating so that during use, this outer metalization may be connected to a well defined potential, such as the system ground, and shield the internal structures from the effects of external electric fields which might otherwise cause errors The transducer can be mounted on a pad 29 of resilient material, as shown in FIG URE 2, to reduce stresses from a mounting surface which may have a different expansion characteristic and to reduce vibration transmission from such a mounting surface.

The pad 29 prevents warpage of the transducer being caused by differential thermal expansion with respect to the material of the package or by mechanical distortions of the package during use. If wallowed, such warpage would introduce small errors into the transducer so limiting its dynamic range. The resilient pad is preferably made electrically conductive by loading with an appropriate material, such as carbon-black or silver powder, permitting easy electrical contact to the lower metalized surface of layer 28.

Sensor circuitry 14 may conveniently be formed by conventional integrated circuit fabrication techniques as part of the same physical package as accelcrometer transduccr 12. Accelerometer transducer 12 and sensor circuitry 14 may be packaged within the same hermetically sealed flat pack or other convenient hybrid package.

During operation, if there is any imbalance in the capacitances between proofmass 16 and upper and lower electrodes 30 and 32 when proofmass 16 is centered, sensor circuitry 14 will balance the capacitances by applying electromotive forces to proofmass 16 to move it to an off centered position and hold it there. This off-center positioning will produce a second order nonlinearity unless compensated by the application of an appropriate bias offset to amplifier 60 or a compensation capacitance added between 16 and 30 or between 16 and 32.

With reference again to FIGURE 1, signal source 38 may be, for example, a 4 volt peak to peak AC source, such as a square wave source operating at say, IOMHz This AC signal, together with the appropriate DC bias from reference source 3 5, is applied to silicon layer 24 and therefore to proofmass 16. If proofmass 16 is in its neutral position, the AC signal is coupled equally to upper and lower electrodes 30 and 32. The capacitive pickup between proofmass 16 and upper and lower electrodes 30 and 32 is compared in differential bridge circuit 46.

It is important to note that capacitors 42 and 44 are large compared to the capacitance formed between upper and lower electrodes 30 and 32 and prooftnass 16 and thus have negligible effect on the neutral position of proofmass 16. In the neutral position, the values are balanced and no differential signal is applied to amplifier 60 so no differential DC signal is applied thereby to upper and lower electrodes 30 and 32 These electrodes pick up an rf signal namely a 10 megahertz signal from generator 38 coupled through the appropriate gaps from the proofmass to 30 and from the proofmass to 32.

To trace the signal flow, the AC signal from 38 is coupled to the proofmass where it is further coupled across the air gaps to 30 and 32. Following the path from 30, the signal sees on one hand, the resistor Rl, of relatively high impedance, between 10 and 30kilo ohms, isolating it from amplifier 60, and capacitor 42 which is a large capacitor allowing it to easily pass to the diode quad bridge. On the other hand, the dc feedback signal which comes from amplifier 60 goes along the path to resistor R1. Since there is very little dc current drawn, there is negligible de voltage drop across R1, from which the de signal is applied directly to electrode 30. Capacitor 42 serves to block the de signal from the diode quad bridge. Similarly the same arguments apply to the effect of resistor R2 and capacitor 44.

In operation, acceleration causes relative motion between proofmass 16 and upper and lower electrodes 30 and 32 as proofmass 16 attempts to pivot about flexure 22.

As proofmass 16 moves and approaches one electrode, such as lower electrode 32, the increased capacitive pickup of the AC signal from signal source 38 causes a voltage to be applied to amplifier 60 via differential bridge circuit 46. The resultant increased DC signal voltage applied to lower electrode 32 via inverter 62 and the decreased voltage applied to upper electrode 30 from amplifier 60 serves to apply an electrostatic force to proofmass 16 to resist the force of acceleration and restore proofmass 16 to its neutral position. The decreased signal applied to inverter 62 may be monitored at acceleration output 34 and is proportional to the acceleration force resisted For sufficiently high gain of amplifier 60, there is a negligible departure of proofmass 16 from its nominal position in the gap when acceleration is applied, thus assuring linearity. In order to maintain stability of the response of closed loop accelerometer 10 with high gain amplifier 60, itis necessary to provide a roll off of the high gain at high frequencies. One particularly advantageous mechanism for achieving high frequency gain roll off is to provide viscous damping by for example, the use of a gas or liquid in the gap between upper and lower electrodes 30 and 32 and proofmass 16 as shown in FIGURE 1.

Thus, there has been provided a novel miniature accelerometer design and method micromachined in silicon which has among other features a large dynamic range with an excellent ratio of full scale to accuracy. The design is highly symmetric. In particular, the hinge is in the center plane of the proofmass, and the structure is balanced to preventwarpage. This gives good bias stability to range ratio. The design is closed loop with an exceptionally high loop gain and a very weak hinge compared with the pendulosity and electrostatic forces. This also gives good bias stability to range ratio. The differential forcing with offset works in conjunction with the high loop gain to minimize non-linearity errors and vibro-pendulosity errors.

Secondary perturbations, which can give bias errors, have been eliminated in the present dcsign by a number of featurcs herein disclosed including 1) electrostatic shielding the outside of the chip 2) extending the electrodes to minimize fringing fields 3) choosing eleroode metals for surface stability at atomic levels i.e. lacking hygroscopic or oxide growth potential 4) etching grooves to reduce the effect of residual fringing 5) use of a resistive coating to stabilize charge distribution in electrode gaps 6) splitting the hinge to reduce the effect of structural warpage 7) use of a resilient mount to prevent structural warpage as a result of differential cxpansion or other external forces.

Additionally, the present invention simplifies the scrvo operation of the accelerom eter by applying the cxcitation to the proofmass, and by arranging the construction so that the pick off and forcer electrodes are now the same.

Claims (16)

1. CLAIMS 1. A miniature accelerometer, comprising: a pair of spaced electrode layers, each electrode layer comprising forcer electrode means and guard electrode means; a silicon proofmass; means for suspending said proofmass between said electrode layers; means responsive to differential capacitive coupling between said forcer electrode means and said proofmass for opposing acceleration forces applied thereto; wherein said guard electrode means prevents contact between said proofmass and said forcer electrode means.
2. 2. The miniature accelerometer claimed in Claim 1 wherein said means for opposing acceleration forces includes an amplifier responsive to differential capacitive coupling between said forcer electrode means and said proofmass for generating an electrical signal for opposing acceleration forces applied thereto.
3. 3. The miniature accelerometer claimed in Claim 1 or Claim 2, wherein said proofmass is a portion of a silicon layer sealed between said electrode layers;
4. 4. The miniature accelerometer claimed in Claim 3, wherein said silicon layer further comprises: .a flexure operating at the midplane of suspending said proofmass.
5. 5. The accelerometer as in Claim 4 wherein said flexure is formed in a split construction using symmetrically balanced hingers.
6. 6. The miniature accelerometer claimed in Claim 3, further comprising: a thin film of gas or liquid in the gap between said proofmass and said electrode layers to provide viscous damping therebetween.
7. 7. The miniature accelerometer claimed in any preceding claim, wherein said forcer electrode means are adapted to apply electrostatic forces to said proofmass in response to the means for opposing acceleration forces.
8. 8. The miniature acceleromater claimed in any preceding claim, wherein each said forcer electrode means comprises: a metalized layer for applying an electrostatic force to said proofmasst and each said guard electrode comprises a metalized layer substantially thicker than the metalized layer of said forcer electrode means.
9. 9. The miniature accelerometer claimed in Claim 8, further comprising: means for reducing transient electrostatic force changes resulting from differences in potential between the guard and forcer electrode metalizations.
10. 10. The miniature accelerometer claimed in Claim 9, wherein said means for reducing transient electrostatic force changes comprises: groove means in said proofmass opposite the portion of said electrode layer between said forcer and guard electrode metalizations for reducing the effects of electrostatic forces applied to said proofmass therefrom.
11. 11. The miniature accelerometer claimed in Claim 9, wherein said electrode layers are formed on respective glass layers and said means for reducing transient electrostatic force changes comprises: resistive coating means on a portion of said glass layer between said forcer and guard electrode metalizations for maintaining a well defined potential distribution therebetween.
12. 12. The accelerometer as in Claim 11, further including: conductor metalization layers deposited on top and bottom glass surfaces to shield the accelerometer from external electric fields.
13. 13. A method of sensing acceleration using a proofmass and electrode layers forming an accelerometer, comprising the steps of: providing a pair of electrode layers each said electrode layer comprising forcer electrode means and guard electrode means; suspending said proofmass in a balanced position between said electrode layers to define substantially equal coupling capacitances between said forcer electrode means and said proofmass; electrically sensing differential capacitive coupling between said forcer electrode means and said proofmass; generating a force on said proofmass which will oppose acceleration forces on said proofmass; and preventing contact between said forcer electrode means and said proofmass using said guard electrode means.
14. 14. The method of sensing acceleration claimed in Claim 13, further comprising: reducing transient electrostatic force changes resulting from differences in potential between said forcer and guard electrode means.
15. 15. The method of sensing acceleration claimed in Claim 14, wherein said electrode layers are formed on respective glass layers, and said transient reducing step comprises: increasing the distance between said proofmass and the portion of each said glass layer between said forcer and guard electrode means.
16. 16. The method of sensing acceleration claimed in Claim 14, wherein said electrode layers are formed on respective glass layers, and said transient reduction step comprises: applying a resistive coating to the portion of each said glass layer between said forcer and guard electrode means to maintain a constant potential distribution therebetween.