1. Field of the Invention

The present invention relates to isolator arrangements for capacitive pressure sensors and to improved diaphragm mounting structures.

2. Prior Art

U.S. Patent No. 3,618,390, owned by the same assignee as the present invention teaches the use of a sensing diaphragm which bottoms out on excessive pressure to protect the isolating diaphragms. This invention provided great impetus to capacitance pressure measurement techniques, as manifested by substantial commercial exploitation and success. The present invention as taught herein may be used in cooperation with the structure claimed in U.S. Patent No. 3,618,590.

SUMMARY OF THE INTEN ION The present invention includes the use of a capacitance type pressure sensor -having a diaphragm disposed in a central chamber and having isolators so that the process fluid or other pressure ^" is applied to the isolator and the pressure is communicated to the sensor by a substantially incompressible fluid via passageway means to the central chamber. The diaphragm is then urged by the incompressible fluid to move to a position which along with an ^' electrically conductive surface disposed on an internal surface of a portion of the central chamber forms a variable capacitor, which when driven by a suitable circuit produces an electrical signal responsive to pressure.

The invention envisions remote isolators, electrical iso¬ lation of the sensor from the isolators, and improved sensor mounting, material selection and configuration to reduce the effects of static line pressure and temperature. Such deleterious effects are substantially reduced, resulting in an improved capacitive signal which is representative of pressure.

BRIEF DESCRIPTION OF THE DRAWINGS ^* Fig. 1 is a sectional view of a pressure transducer made in accordance with the instant invention.

Fig. 2 is a sectional view of the sensor housing of another embodiment of the sensor housing of the pressure trans¬ ducer.

Fig. 3 is a graph showing the results of testing a transducer in accord with an embodiment of the invention showing per¬ cent error (in tenths (1/10) of one percent (1.) ) v. differ- ential pressure for five different calibrations having from 0 PSIG (pounds per square inch gage) static line pressure applied to both sides of the transducer to 2000 PSIG to both sides of such ^' trans¬ ducers.

Fig. 4 is a graphic representation of the results of testing a transducer in accord with a embodiment of the invention showing output deviation in percent v. differential line pressure to show temperature effect (uncompensated) for several cali¬ brations.

DESCRIPTION OF ^* THE PREFERRED EMBODIMENT A transducer which preferably is used for differential pressure, gauge pressure, flow, level or other such pressure measurement is shown generally at 10. The transducer includes a transducer housing or frame 12 which supports a sensor housing 14 and a pair of isolator housings 16a and 16b. It is envisioned that housings 14, 16a and 16b may be included in cr spaced from housing 12. The pressures to be sensed are represented by arrows 18 and 20 at the transducer input ports. Pressures 18 and 20 act on isolator diaphragms 22 and 24, respectively. Diaphragms 22 and 24 preferably are very flexible and are formed in a con- ventional manner. The corrugations 26 of diaphragms 22 and 24 ^• represent a preferred isolator diaphragm construction having a plurality of convolution ^' s, as desired. Chambers 27 and 29 are

OMFI

-j- defined by diaphragms 22 and 24 in cooperation with the respective housings 16a and 16b. Chambers 27 and 29 are coupled to passage¬ ways 28 and 30, which preferably are formed by stainless steel tubing, but may be formed from other suitable materials. Sensor housing 14 is preferably machined and formed from a metal such as stainless steel, preferably an austenitic stain¬ less steel such as 304 is used. Generally, housing 14 is formed from two portions 32 and 34 which preferably are substantially equal in size and when assembled are generally divided by a sensor diaphragm 36 which is held at its edges and will elastically deflect under differential pressures and which typically is subjected to a desired radial tension. A central cone shaped cavity 38 with bores 42 and 44 is formed in portion 32 and, similarly a central cone shaped cavity 46 with bores 48 and 50 is formed in portion 34. Conduits 52 and 54 are formed in portions 32 and ^' 34 to communicate with passage¬ ways 28 and 30, respectively. The interior openings of the conduits 52 and 54 form continuations of the passageways 28 and 30 to com¬ municate with a chamber 53 defined in portion 32 by diaphragm 36 and the central section of material 60a in portion 32 and a second ^" chamber 55 defined in portion 34 by diaphragm 36 and the central section of material 60b in portion 34. An electrical conductor 56 is inserted through bore 42 to chamber 38 and, similarly, a conductor 58 is inserte through bore 48 to chamber 46. Conductors 56 and 58 may be metal tubes to aid in filling the sensor chambers with a non-compressible fluid.

The electrically conductive portions of the housing 14 are electrically insulated from the metal tubes forming passage¬ ways 28 and 30 and from conductors 56 and 58. An insulative non- porous material 60a and 60b such as glass or ceramic, is filled into cavities 38 and 46 and bores 42 and 48 and is bonded to housing portions 32 and 34 along a surface forming an angle θ with respect to the plane formed by the joining of portions 32 and 34. The central section of material 60a and 60b and the central area of housing portions 32 and 34, as well as inner ends of

OMPI

conductors 56 and 58 are then contoured or recessed by grinding or machining, preferably to provide a suitable stop surface for sensor diaphragm 36 when the diaphragm 36 deflects under an over- range pressure applied to the isolator diaphragms. Conduits 52, 54 may be a single cylinder as shown , or a plurality of small cylinders in accord with the teaching of U.S. Patent No. 3,618,390 to provide diaphragm support- in an overpressure condition.

A suitable electrically conductive material is then de¬ posited in a layer on the inner surface of material 60a and 60b in each housing portion as at 61 and 63. The layers face opposite sides of the sensor diaphragm 36 and are electrically coupled to conductors 56 and 58, respectively. Sensor diaphragm 36 prefer¬ ably is formed from a suitable electrically conductive material and is fixed into position between housing portions 32 and 34 and layers 61 and 63 by a continuous bead weld 62 thus forming a common plate for material 61 and 63, hence forming two ca¬ pacitors C. and Z- . A suitable conductor 64 is then coupled to sensor housing 14 which is at the same electrical potential as diaphragm ^" 36. Sensor diaphragm 36 can also be formed from a non- conductive material and have a conductive portion disposed in or on the diaphragm to form such common plate for a variable sensor capacitor. A suitable conductor 64 is then coupled to such conductive portion. Bolts 70 may then be added to take up the pressure forces on sensor housing 14. A suitable, substantially incompressible fluid, such as silicone oil, is then filled into each side of the transducer assembly through conductors 56 and 58 to the sensor diaphragm chamber formed in housing portion 32 by diaphragm 36, and to isolating chamber 27, and similarly to the sensor chamber in housing portion 34, and isolating chamber 29. .ϊhen such spaces are filled, conductors 56 and 58 are pinched off at their outer ends and suitable leadwires are attached thereto.

The action of pressure on isolator diaphragms 22 and 24, the substantially incompressible fluid in chambers 27 and 29,

OMPI

passageways 28, and 30, and on sensing diaphragm 36 is fully explained, for example, in United States Patent 3,618,390. The invention of the sensing diaphragm 36 bottoming out in an overpressure condition as taught in U.S. Patent 3,618,390 or the isolating diaphragm 22 or 24 bottoming out in an overpressure condition may, as desired, be used with the present invention.

The physical location of isolator diaphragms 22 and 24 spaced from sensor diaphragm 36 is shown somewhat schematically, as the location of isolator diaphragms 22 and 24 is not critical, pro- viding that such diaphragms are located so as not to apply undesired mechanical stress, other than the pressure through the incom¬ pressible fluid, to sensor housing 14. While sensor housing 14 preferably is fixedly mounted in housing 12 it is not required that it be rigidly mounted as by welding. As shown, it is re- tained by flexible straps 71, which are formed from an electrical insulative material to electrically isolate the sensor housing 14 from transducer housing 12 and to support sensor housing 14.

With chambers 27, 29, passageways 28 and 30 (including openings in ^' conduits 52 and 54) and the chambers between layers 61 and 63 and diaphragm 56.filled with incompressible fluid, differ¬ entials between pressures represented by arrows 18 and 20 will cause diaphragm 36 to deflect proportional to pressure differ¬ ential and its capacitance relative to layers 61 and 63 changes. Another embodiment of the invention is shown in Fig. 2. In this embodiment the sensor housing 14A is somewhat wider than the embodiment of Fig. 1. While the numbering corresponds to Fig. 1, (with a capital letter forming a part of the alphanumeric designations thereof) it is observed that with the increased width of sensor housing 14Λ, bores 44A and 50A are somewhat deeper in Fig. 2 than bores 44 and 50, and material 60A and 60B, has been filled to include a portion of such bores. An angle θ is the

included angle from the plane of diaphragm 56A at its rest position to the conical surface forming the recess in the respective housing portion in which material 60A and 60B is filled. This angle determines the effective depth of the material 60A,60B (or 60a, 60b in the first form of the invention) which backs the capacitor plates 61A and 63A (or 61 and 63) . Although an angle 6 of approxi¬ mately 45 is preferred for the embodiments of Fig. 1 and 2, it has been found that angles from 25 to 70 have resulted in improved stability and thus improved performance over known constructions which comprise, for example, a non-compressive bond ^' between the insulating material and metal. The angle also can be measured with respect to _^. the central axis of the sensor housing which is perpendicular to the plane of diaphragm 36A (or 36) when it is at rest. One significant advantage of the present invention is improvement of the static pressure effect on the pressure span of the transducer. In prior art embodiments, the effect of static pressure on span error has been found to be approximately a one percent (1%) change in output across the instrument span per one thousand (1000) pounds per square inch (PSI) change of static pressure. In such known transducers, the pressure on the outside of the sensor housing, caused by the pressure being sensed acting on the isolating diaphragms, and pressure from the inside of the sensor chamber c I by the pressure being sensed on the incom- pressible fluid resulted in sensor housing deformation outwardly in a known manner, as per Poisson's ratio.

Further, in known methods of manufacturing such capaci¬ tive transducers, the insulating material, upon which the con¬ ductive material is deposited to form the second plate of each of the variable capacitors, has been relatively thin in comparison to the insulative material thickness in the central cavity of the

sensor housing as disclosed herein. When the insulat'ive material is thin or when the insulating material to metal interface is somewhat parallel to the rest axis of the diaphragm (perpendicular to plane of the diaphragm), the insulating material-metal inter- faces (bond) 65a, 65b, 65A, 65B are then subjected to a shear force which may cause the bond to weaken or fracture. When pressure is applied to a sensor having a fractured bond, such pressure causes the insulating material to move away from the diaphragm. The movement of the insulator material causes an undesirable change in capacitance not representative of the sensed pressure, which adds to the error effect caused by the static line pressure. When the sensor is formed in accordance with the present disclosure, bonds 65a, 65b, 65A, 65B are substantially in compression and, consequently, much less vulnerable to such fracture.

By removing the isolators from the side of the sensor housing, the capacitor plate spacing on both sides of the dia¬ phragm 36 increases with increasing static line pressure applied at 18, 20 due to slight outward motion of the sensor portions - with respect to the sensor diaphragm. This static line pressure increase also causes portions 32 and 34 to warp slightly about their respective neutral axes (shown in Figs. 2 and 5 at X-X) as the two housing portions tend to contract adjacent the diaphragm (as shown by arrows 70A in the Figs. 2 § SB). (The insulation material is not specifically shown in Figs. 5A or 5B since these Figures are illustrative only and apply to the configurations of Figs. 1 and 2.) Such warping is perhaps best explained by reference to Fig. 5A which shows portions 32 and 34 at rest and in Fig. 5B which shows an exaggerated wεrp condition (for emphasis) as caused by increasing static line pressure. As the static line pressure is increased the capacitance spacing (d) of Fig. 5A be¬ tween the diaphragm 36 and capacitor plates 61 and 63 increases to d' (as shown in Fig. 5B) and such spacing change is not repre¬ sentative of the applied differential pressure. In accordance

with the present invention, the change in capacitance caused by such warping is substantially compensated for by the decrease in diaphragm radial tension caused by the contraction adjacent the diaphragm. Radial tension or prestress applied to the diaphragm 5 at time of construction along with suitable dimensions and materi¬ als results in the elastic stiffness of the diaphragm decreasing with increasing static pressure. Preferably the diaphragm material is high strength steel having good elastic characteristics. The compensation advantage is present at all static line pressures, 10 but more fully realized at static line pressures above 500 psi.

The following equations further explain the static line pressure compensation according to a preferred embodiment of the instant invention having a first capacitor C. and a second capaci¬ tor Cr as described herein:

the output signal from the differential pressure capacitance cell. CH = the capacitance of the

20 greater of C, or C-. ^~

CL = the capacitance of the lesser of C, or C _? .

Xp = diaphragm deflection with differential 25 pressure.

Xo = capacitance spacing at zero (0) static line gauge pressure.

Xo' = capacitance spacing at 30 elevated static line pressure.

c = diaphragm stretch at time of construction (initial prεstretch). c ' = diaphragm stretch at elevated static line pressure.

Simplifying:

0 « *E_ Xoό

When the transducer is formed in accordance with the present in- 10 vention as the static line pressure increases, the capacitance spacing Xo increases to Xo' and the diaphragm stretch (6 ) de¬ creases to δ ' . By holding the product of Xo-c substantially equal to Xo'-δ ' hence, substantially equal to a constant, the diaphragm deflection (Xp) is responsive to the differential 5 pressure applied thereto and the output (0) is thus independent of static line pressure.

A transducer, made in accordance with the embodiments of Fig. 1 and Fig. 2, but not having an angle 6 between 25 and 70 rather having a cylindrical, metal-insulating material interface bond; 0 that is, the bond interface was first generally ^' perpendicular (θ = 90 ) from diaphragm 36, then generally parallel (9 = 0 ) from diaphragm 36 generally as shown in U.S. Patent No. 5,618,390, was tested under actual loading conditions. This early form of the invention did not then include the compressive S bond taught herein, but rather had the prior art shear bond. The improved bond is helpful as taught to avoid bond fracture and it is beli-ved from ar.alyzation and evaluation that such fracture did not occur and, therefore the nature of the bond did not affect test results. In the embodiment tested, the other principles, of the invention weτ& followed,_sι.ιch as separating the isolator 16a, 16fc from sensor housing 14 and compen¬ sating for sensor housing 14 warp with a suitable sensor diaphragm 36 prestress. Sensor diaphragm 36 was l.S mils thick and ap-

proximately 1.12 inch in diameter and had approximately 105,000 PSI in prcstress applied (though prestress from SO,-000 to 200,000 PSI may be acceptable) and was formed from NiSpan C; insulative material 60a, 60b, 60A, 60B was Owens 0120 glass; and, sensor housing 14 was NiSpan C material approximately 1.250 inches in diameter. Capacitance spacing (Xo) in the center was approximately .0075 inches. Isolators 16a, 16b were formed from stainless steel (304SST) and were approxi¬ mately 3 inches in diameter and were coupled to chambers 53, 55 by passageways 28,30 formed from 1/16 inch O.D. stainless steel tubing. The results of such testing are shown in Fig. 3. As shown, all test points deviations due to static line pressure effect from zero (0) PSIG to two thousand (2000) PSIG are less than .2% across the differential pressure span of 0 to 240 inches of water. The curves of Fig. 3 show a very small mechanical hysteresis. Such mechanical hysteresis is not uncommon and depends not only on the instantaneous value of stress as caused by differential pressure and static line pressure but also on the previous history of such stress.

Yet a further improvement is attained by the present invention, as the zero stability of the transducer, which in known transducers varies with both temperature and static pressure, is improved because the isolator housings are not in direct physical contact with the sensor housing. Only the tubes forming passageways 28 and 30 are in direct contact with sensor housing 14 and these tubes yield to accommodate .loads or changes due to temperature on the isolators without stressing the sensor housing 14.

, A test also was conducted to demonstrate the improved uncompensated temperature effect with respect to the stability of the output capacitance signal of the present invention of the embodiment described above; such results are shown in Fig. 4. The "uncompensated" effect is the error present before any compensation of an electrical signal is applied. Electrical signal compensation

OMP

is commonly used to reduce errors further, but ^' it is highly advan¬ tageous to provide a structure having a low uncompensa ed error. Each curve of Fig. 4 represents a separate calibration. Several such calibrations were accomplished, seven of which are shown on Fig. 4, one at 100°F, then again at 100°F, then at 200°F, then

100°F, then 0°F, then 100°F, again 200°F, and finally at 100 F. The curves show that the configuration resulted in excellent stability and in a very low thermal hysteresis as the capacitance deviation at 100 F for three calibrations at that temperature was less than +_ .18.. Thermal hysteresis refers to the difference in calibration results at a specific temperature after coming to that calibration temperature from higher and lower temperatures respectively.

Many embodiments formed of different materials and having

material 60 was alkali lead glass, specifically Corning 1990 glass and the sensor housing 14 was austenitic stainless steel.

One further advantage of the present invention is that since the isolator diaphragms are no longer an integral part -of the sensor housing 14, the size of the isolator diaphragms may be increased relative to the sensor housing. This increase in size is important in some instances to reduce the effects of temperature and other factors on overall transducer performance. Further, sensor housing 14 preferably is electrically isolated from transducer housing 12 resulting in a simplification of the transducer circuitry when electrical isolation is desired which is often the case for industrial pressure measurements.

While the invention has been described using a variable ca- pacitaπce sensor, those skilled in the art understand that a variable impedance, that is a variable impedance variable reactance sensor, can be used with the invention as described herein.

In summary, the several listed advantages as well as those apparent to those skilled in the art are realized from the improvements of the present invention.